First stab at #396

.gitignore

@@ -1,4 +1,3 @@
-tantivy.iml
 *.swp
 target
 target/debug

CHANGELOG.md

@@ -1,39 +1,10 @@
-Tantivy 0.9.0
-=====================
-- Removed most unsafe (@fulmicoton)
-- Indexer memory footprint improved. (VInt comp, inlining the first block. (@fulmicoton)
-- Stemming in other language possible (@pentlander)
-- Segments with no docs are deleted earlier (@barrotsteindev)
-
-Tantivy 0.8.1
-=====================
-Hotfix of #476.
-
-Merge was reflecting deletes before commit was passed. 
-Thanks @barrotsteindev  for reporting the bug.
-
-
-Tantivy 0.8.0
-=====================
-*No change in the index format*
-- API Breaking change in the collector API. (@jwolfe, @fulmicoton)
-- Multithreaded search (@jwolfe, @fulmicoton) 
-
-
-Tantivy 0.7.1
-=====================
-*No change in the index format*
-- Bugfix: NGramTokenizer panics on non ascii chars
-- Added a space usage API
 
 Tantivy 0.7
 =====================
 - Skip data for doc ids and positions (@fulmicoton),
   greatly improving performance
 - Tantivy error now rely on the failure crate (@drusellers)
-- Added support for `AND`, `OR`, `NOT` syntax in addition to the `+`,`-` syntax
-- Added a snippet generator with highlight (@vigneshsarma, @fulmicoton)
-- Added a `TopFieldCollector` (@pentlander)
+
 
 Tantivy 0.6.1
 =========================

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tantivy"
-version = "0.9.0-dev"
+version = "0.7.0-dev"
 authors = ["Paul Masurel <paul.masurel@gmail.com>"]
 license = "MIT"
 categories = ["database-implementations", "data-structures"]
@@ -12,9 +12,10 @@ readme = "README.md"
 keywords = ["search", "information", "retrieval"]
 
 [dependencies]
-base64 = "0.10.0"
+base64 = "0.9.1"
 byteorder = "1.0"
 lazy_static = "1"
+tinysegmenter = "0.1.0"
 regex = "1.0"
 fst = {version="0.3", default-features=false}
 fst-regex = { version="0.2" }
@@ -29,54 +30,42 @@ serde = "1.0"
 serde_derive = "1.0"
 serde_json = "1.0"
 num_cpus = "1.2"
-fs2={version="0.4", optional=true}
-itertools = "0.8"
+itertools = "0.7"
 levenshtein_automata = {version="0.1", features=["fst_automaton"]}
 bit-set = "0.5"
-uuid = { version = "0.7", features = ["v4", "serde"] }
-crossbeam = "0.5"
+uuid = { version = "0.6", features = ["v4", "serde"] }
+crossbeam = "0.4"
+crossbeam-channel = "0.2"
 futures = "0.1"
 futures-cpupool = "0.1"
 owning_ref = "0.4"
 stable_deref_trait = "1.0.0"
 rust-stemmers = "1"
-downcast-rs = { version="1.0" }
+downcast = { version="0.9" }
 matches = "0.1"
 bitpacking = "0.5"
-census = "0.2"
+census = "0.1"
 fnv = "1.0.6"
 owned-read = "0.4"
 failure = "0.1"
-htmlescape = "0.3.1"
-fail = "0.2"
-scoped-pool = "1.0"
-murmurhash32 = "0.2"
 
 [target.'cfg(windows)'.dependencies]
 winapi = "0.2"
 
 [dev-dependencies]
-rand = "0.6"
-maplit = "1"
+rand = "0.5"
 
 [profile.release]
 opt-level = 3
 debug = false
+lto = true
 debug-assertions = false
 
-[profile.test]
-debug-assertions = true
-overflow-checks = true
-
 [features]
-# by default no-fail is disabled. We manually enable it when running test.
-default = ["mmap", "no_fail"]
-mmap = ["fst/mmap", "atomicwrites", "fs2"]
+default = ["mmap"]
+mmap = ["fst/mmap", "atomicwrites"]
 lz4-compression = ["lz4"]
-no_fail = ["fail/no_fail"]
-unstable = [] # useful for benches.
 
 [badges]
 travis-ci = { repository = "tantivy-search/tantivy" }
 
-

README.md

@@ -4,7 +4,6 @@
 [![Join the chat at https://gitter.im/tantivy-search/tantivy](https://badges.gitter.im/tantivy-search/tantivy.svg)](https://gitter.im/tantivy-search/tantivy?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![Build status](https://ci.appveyor.com/api/projects/status/r7nb13kj23u8m9pj/branch/master?svg=true)](https://ci.appveyor.com/project/fulmicoton/tantivy/branch/master)
-[![Say Thanks!](https://img.shields.io/badge/Say%20Thanks-!-1EAEDB.svg)](https://saythanks.io/to/fulmicoton)
 
 ![Tantivy](https://tantivy-search.github.io/logo/tantivy-logo.png)
 
@@ -21,7 +20,7 @@
 
 **Tantivy** is a **full text search engine library** written in rust.
 
-It is closer to [Apache Lucene](https://lucene.apache.org/) than to [Elasticsearch](https://www.elastic.co/products/elasticsearch) and [Apache Solr](https://lucene.apache.org/solr/) in the sense it is not
+It is closer to Lucene than to Elastic Search and Solr in the sense it is not
 an off-the-shelf search engine server, but rather a crate that can be used
 to build such a search engine.
 
@@ -33,8 +32,8 @@ Tantivy is, in fact, strongly inspired by Lucene's design.
 - Fast (check out the :racehorse: :sparkles: [benchmark](https://tantivy-search.github.io/bench/) :sparkles: :racehorse:)
 - Tiny startup time (<10ms), perfect for command line tools
 - BM25 scoring (the same as lucene)
-- Natural query language `(michael AND jackson) OR "king of pop"`
-- Phrase queries search (`"michael jackson"`)
+- Basic query language (`+michael +jackson`)
+- Phrase queries search (\"michael jackson\"`)
 - Incremental indexing
 - Multithreaded indexing (indexing English Wikipedia takes < 3 minutes on my desktop)
 - Mmap directory
@@ -44,14 +43,12 @@ Tantivy is, in fact, strongly inspired by Lucene's design.
 - LZ4 compressed document store
 - Range queries
 - Faceted search
-- Configurable indexing (optional term frequency and position indexing)
+- Configurable indexing (optional term frequency and position indexing
 - Cheesy logo with a horse
 
 # Non-features
 
-- Distributed search is out of the scope of tantivy. That being said, tantivy is meant as a
-library upon which one could build a distributed search. Serializable/mergeable collector state for instance, 
-are within the scope of tantivy.
+- Distributed search and will not be in the scope of tantivy.
 
 
 # Supported OS and compiler
@@ -76,14 +73,10 @@ It will walk you through getting a wikipedia search engine up and running in a f
 Tantivy compiles on stable rust but requires `Rust >= 1.27`.
 To check out and run tests, you can simply run :
 
-    git clone https://github.com/tantivy-search/tantivy.git
+    git clone git@github.com:tantivy-search/tantivy.git
     cd tantivy
     cargo build
 
-## Running tests
-
-Some tests will not run with just `cargo test` because of `fail-rs`.
-To run the tests exhaustively, run `./run-tests.sh`. 
 
 # Contribute
 

appveyor.yml

@@ -18,5 +18,5 @@ install:
 build: false
 
 test_script:
-  - REM SET RUST_LOG=tantivy,test & cargo test --verbose --no-default-features --features mmap -- --test-threads 1
-  - REM SET RUST_BACKTRACE=1 & cargo build --examples
+  - REM SET RUST_LOG=tantivy,test & cargo test --verbose
+  - REM SET RUST_BACKTRACE=1 & cargo build --examples

ci/script.sh

@@ -11,11 +11,12 @@ main() {
     else
         echo "Build"
         cross build --target $TARGET
+        cross build --target $TARGET --release
         if [ ! -z $DISABLE_TESTS ]; then
             return
         fi
         echo "Test"
-        cross test --target $TARGET --no-default-features --features mmap -- --test-threads 1
+        cross test --target $TARGET
     fi
     for example in $(ls examples/*.rs)
     do

doc/src/SUMMARY.md

@@ -9,7 +9,6 @@
 - [Facetting](./facetting.md)
 - [Innerworkings](./innerworkings.md)
   - [Inverted index](./inverted_index.md)
-- [Best practise](./inverted_index.md)
 
 [Frequently Asked Questions](./faq.md)
 [Examples](./examples.md)

doc/src/avant-propos.md

@@ -2,8 +2,8 @@
 
 > Tantivy is a **search** engine **library** for Rust.
 
-If you are familiar with Lucene, it's an excellent approximation to consider tantivy as Lucene for rust. tantivy is heavily inspired by Lucene's design and
-they both have the same scope and targetted use cases.
+If you are familiar with Lucene, tantivy is heavily inspired by Lucene's design and
+they both have the same scope and targetted users.
 
 If you are not familiar with Lucene, let's break down our little tagline.
 
@@ -17,18 +17,15 @@ relevancy, collapsing, highlighting, spatial search.
   experience. But keep in mind this is just a toolbox.
   Which bring us to the second keyword...
 
-- **Library** means that you will have to write code. tantivy is not an *all-in-one* server solution like elastic search for instance.
-
-  Sometimes a functionality will not be available in tantivy because it is too
-  specific to your use case. By design, tantivy should make it possible to extend
-  the available set of features using the existing rock-solid datastructures.
-
-  Most frequently this will mean writing your own `Collector`, your own `Scorer` or your own
-  `TokenFilter`... Some of your requirements may also be related to
-  something closer to architecture or operations. For instance, you may
-  want to build a large corpus on Hadoop, fine-tune the merge policy to keep your
-  index sharded in a time-wise fashion, or you may want to convert and existing
-  index from a different format.
-
-  Tantivy exposes a lot of low level API to do all of these things.
+- **Library** means that you will have to write code. tantivy is not an *all-in-one* server solution.
   
+  Sometimes a functionality will not be available in tantivy because it is too specific to your use case. By design, tantivy should make it possible to extend
+  the available set of features using the existing rock-solid datastructures. 
+
+   Most frequently this will mean writing your own `Collector`, your own `Scorer` or your own 
+   `Tokenizer/TokenFilter`... But some of your requirement may also be related to 
+   architecture or operations. For instance, you may want to build a large corpus on Hadoop,
+   fine-tune the merge policy to keep your index sharded in a time-wise fashion, or you may want
+   to convert and existing index from a different format.
+   
+   Tantivy exposes its API to do all of these things.

doc/src/basis.md

@@ -2,76 +2,47 @@
 
 ## Straight from disk
 
-Tantivy accesses its data using an abstracting trait called `Directory`.
-In theory, one can come and override the data access logic. In practise, the
-trait somewhat assumes that your data can be mapped to memory, and tantivy
-seems deeply married to using `mmap` for its io [^1], and the only persisting
-directory shipped with tantivy is the `MmapDirectory`.
+By default, tantivy accesses its data using its `MMapDirectory`. 
+While this design has some downsides, this greatly simplifies the source code of tantivy, 
+and entirely delegates the caching to the OS.
 
-While this design has some downsides, this greatly simplifies the source code of
-tantivy. Caching is also entirely delegated to the OS.
+`tantivy` works entirely (or almost) by directly reading the datastructures as they are layed on disk.
+As a result, the act of opening an indexing does not involve loading different datastructures 
+from the disk into random access memory : starting a process, opening an index, and performing a query 
+can typically be done in a matter of milliseconds. 
 
-`tantivy` works entirely (or almost) by directly reading the datastructures as they are layed on disk. As a result, the act of opening an indexing does not involve loading different datastructures from the disk into random access memory : starting a process, opening an index, and performing your first query can typically be done in a matter of milliseconds.
-
-This is an interesting property for a command line search engine, or for some multi-tenant log search engine : spawning a new process for each new query can be a perfectly sensible solution in some use case.
+This is an interesting property for a command line search engine, or for some multi-tenant log search engine.
+Spawning a new process for each new query can be a perfectly sensible solution in some use case.
 
 In later chapters, we will discuss tantivy's inverted index data layout.
-One key take away is that to achieve great performance, search indexes are extremely compact.
+One key take away is that to achieve great performance, search indexes are extremely compact. 
 Of course this is crucial to reduce IO, and ensure that as much of our index can sit in RAM.
 
-Also, whenever possible its data is accessed sequentially. Of course, this is an amazing property when tantivy needs to access the data from your spinning hard disk, but this is also
-critical for performance, if your data is read from and an `SSD` or even already in your pagecache.
+Also, whenever possible the data is accessed sequentially. Of course, this is an amazing property when tantivy needs to access
+the data from your spinning hard disk, but this is also a great property when working with `SSD` or `RAM`, 
+as it makes our read patterns very predictable for the CPU.
 
 
 ## Segments, and the log method
 
-That kind of compact layout comes at one cost: it prevents our datastructures from being dynamic.
-In fact, the `Directory` trait does not even allow you to modify part of a file.
+That kind compact layout comes at one cost: it prevents our datastructures from being dynamic.
+In fact, a trait called `Directory` is in charge of abstracting all of tantivy's data access
+and its API does not even allow editing these file once they are written.
 
 To allow the addition / deletion of documents, and create the illusion that
-your index is dynamic (i.e.: adding and deleting documents), tantivy uses a common database trick sometimes referred to as the *log method*.
+your index is dynamic (i.e.: adding and deleting documents), tantivy uses a common database trick sometimes
+referred to as the *log method*.
 
-Let's forget about deletes for a moment.
-
-As you add documents, these documents are processed and stored in a dedicated datastructure, in a `RAM` buffer. This datastructure is not ready for search, but it is useful to receive your data and rearrange it very rapidly.
-
-As you add documents, this buffer will reach its capacity and tantivy will transparently stop adding document to it and start converting this datastructure to its final read-only format on disk. Once written, an brand empty buffer is available to resume adding documents.
+Let's forget about deletes for a moment. As you add documents, these documents are processed and stored in 
+a dedicated datastructure, in a `RAM` buffer. This datastructure is designed to be dynamic but 
+cannot be accessed for search. As you add documents, this buffer will reach its capacity and tantivy will
+transparently stop adding document to it and start converting this datastructure to its final
+read-only format on disk. Once written, an brand empty buffer is available to resume adding documents. 
 
 The resulting chunk of index obtained after this serialization is called a `Segment`.
 
-> A segment is a self-contained atomic piece of index. It is identified with a UUID, and all of its files are identified using the naming scheme : `<UUID>.*`.
-
-Which brings us to the nature of a tantivy `Index`.
-
-> A tantivy `Index` is a collection of `Segments`.
-
-Physically, this really just means and index is a bunch of segment files in a given `Directory`,
-linked together by a `meta.json` file. This transparency can become extremely handy
-to get tantivy to fit your use case:
-
-*Example 1* You could for instance use hadoop to build a very large search index in a timely manner, copy all of the resulting segment files in the same directory and edit the `meta.json` to get a functional index.[^2]
-
-*Example 2* You could also disable your merge policy and enforce daily segments. Removing data after one week can then be done very efficiently by just editing the `meta.json` and deleting the files associated to segment `D-7`.
+> A segment is a self-contained atomic piece of index. It is identified with a UUID, and all of its files
+are identified using the naming scheme : `<UUID>.*`.
 
 
-
-
-
-# Merging
-
-As you index more and more data, your index will accumulate more and more segments.
-Having a lot of small segments is not really optimal. There is a bit of redundancy in having
-all these term dictionary. Also when searching, we will need to do term lookups as many times as we have segments.  It can hurt search performance a bit.
-
-That's where merging or compacting comes into place. Tantivy will continuously consider merge
-opportunities and start merging segments in the background.
-
-
-# Indexing throughput, number of indexing threads
-
-
-
-
-[^1]: This may eventually change.
-
-[^2]: Be careful however. By default these files will not be considered as *managed* by tantivy. This means they will never be garbage collected by tantivy, regardless of whether they become obsolete or not.
+> A tantivy `Index` is a collection of `Segments`.

doc/src/examples.md

@@ -1,3 +1 @@
 # Examples
-
-- [Basic search](/examples/basic_search.html)

examples/basic_search.rs

@@ -10,17 +10,17 @@
 // - search for the best document matchings "sea whale"
 // - retrieve the best document original content.
 
+
 extern crate tempdir;
 
 // ---
 // Importing tantivy...
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopDocs;
+use tantivy::collector::TopCollector;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::Index;
-use tempdir::TempDir;
 
 fn main() -> tantivy::Result<()> {
     // Let's create a temporary directory for the
@@ -35,7 +35,7 @@ fn main() -> tantivy::Result<()> {
     // be indexed".
 
     // first we need to define a schema ...
-    let mut schema_builder = Schema::builder();
+    let mut schema_builder = SchemaBuilder::default();
 
     // Our first field is title.
     // We want full-text search for it, and we also want
@@ -106,37 +106,37 @@ fn main() -> tantivy::Result<()> {
     // For convenience, tantivy also comes with a macro to
     // reduce the boilerplate above.
     index_writer.add_document(doc!(
-    title => "Of Mice and Men",
-    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-            over the yellow sands in the sunlight before reaching the narrow pool. On one \
-            side of the river the golden foothill slopes curve up to the strong and rocky \
-            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-            fresh and green with every spring, carrying in their lower leaf junctures the \
-            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-            limbs and branches that arch over the pool"
+        title => "Of Mice and Men",
+        body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
+                bank and runs deep and green. The water is warm too, for it has slipped twinkling \
+                over the yellow sands in the sunlight before reaching the narrow pool. On one \
+                side of the river the golden foothill slopes curve up to the strong and rocky \
+                Gabilan Mountains, but on the valley side the water is lined with trees—willows \
+                fresh and green with every spring, carrying in their lower leaf junctures the \
+                debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
+                limbs and branches that arch over the pool"
     ));
 
     index_writer.add_document(doc!(
-    title => "Of Mice and Men",
-    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-            over the yellow sands in the sunlight before reaching the narrow pool. On one \
-            side of the river the golden foothill slopes curve up to the strong and rocky \
-            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-            fresh and green with every spring, carrying in their lower leaf junctures the \
-            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-            limbs and branches that arch over the pool"
+        title => "Of Mice and Men",
+        body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
+                bank and runs deep and green. The water is warm too, for it has slipped twinkling \
+                over the yellow sands in the sunlight before reaching the narrow pool. On one \
+                side of the river the golden foothill slopes curve up to the strong and rocky \
+                Gabilan Mountains, but on the valley side the water is lined with trees—willows \
+                fresh and green with every spring, carrying in their lower leaf junctures the \
+                debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
+                limbs and branches that arch over the pool"
     ));
 
     // Multivalued field just need to be repeated.
     index_writer.add_document(doc!(
-    title => "Frankenstein",
-    title => "The Modern Prometheus",
-    body => "You will rejoice to hear that no disaster has accompanied the commencement of an \
-             enterprise which you have regarded with such evil forebodings.  I arrived here \
-             yesterday, and my first task is to assure my dear sister of my welfare and \
-             increasing confidence in the success of my undertaking."
+       title => "Frankenstein",
+       title => "The Modern Prometheus",
+       body => "You will rejoice to hear that no disaster has accompanied the commencement of an \
+                enterprise which you have regarded with such evil forebodings.  I arrived here \
+                yesterday, and my first task is to assure my dear sister of my welfare and \
+                increasing confidence in the success of my undertaking."
     ));
 
     // This is an example, so we will only index 3 documents
@@ -213,10 +213,15 @@ fn main() -> tantivy::Result<()> {
     //
     // We are not interested in all of the documents but
     // only in the top 10. Keeping track of our top 10 best documents
-    // is the role of the TopDocs.
+    // is the role of the TopCollector.
+    let mut top_collector = TopCollector::with_limit(10);
 
     // We can now perform our query.
-    let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
+    searcher.search(&*query, &mut top_collector)?;
+
+    // Our top collector now contains the 10
+    // most relevant doc ids...
+    let doc_addresses = top_collector.docs();
 
     // The actual documents still need to be
     // retrieved from Tantivy's store.
@@ -225,10 +230,14 @@ fn main() -> tantivy::Result<()> {
     // the document returned will only contain
     // a title.
 
-    for (_score, doc_address) in top_docs {
-        let retrieved_doc = searcher.doc(doc_address)?;
+    for doc_address in doc_addresses {
+        let retrieved_doc = searcher.doc(&doc_address)?;
         println!("{}", schema.to_json(&retrieved_doc));
     }
 
+
     Ok(())
 }
+
+
+use tempdir::TempDir;

examples/custom_collector.rs

@@ -1,187 +0,0 @@
-// # Custom collector example
-//
-// This example shows how you can implement your own
-// collector. As an example, we will compute a collector
-// that computes the standard deviation of a given fast field.
-//
-// Of course, you can have a look at the tantivy's built-in collectors
-// such as the `CountCollector` for more examples.
-
-extern crate tempdir;
-
-// ---
-// Importing tantivy...
-#[macro_use]
-extern crate tantivy;
-use tantivy::collector::{Collector, SegmentCollector};
-use tantivy::fastfield::FastFieldReader;
-use tantivy::query::QueryParser;
-use tantivy::schema::Field;
-use tantivy::schema::{Schema, FAST, INT_INDEXED, TEXT};
-use tantivy::Index;
-use tantivy::SegmentReader;
-
-#[derive(Default)]
-struct Stats {
-    count: usize,
-    sum: f64,
-    squared_sum: f64,
-}
-
-impl Stats {
-    pub fn count(&self) -> usize {
-        self.count
-    }
-
-    pub fn mean(&self) -> f64 {
-        self.sum / (self.count as f64)
-    }
-
-    fn square_mean(&self) -> f64 {
-        self.squared_sum / (self.count as f64)
-    }
-
-    pub fn standard_deviation(&self) -> f64 {
-        let mean = self.mean();
-        (self.square_mean() - mean * mean).sqrt()
-    }
-
-    fn non_zero_count(self) -> Option<Stats> {
-        if self.count == 0 {
-            None
-        } else {
-            Some(self)
-        }
-    }
-}
-
-struct StatsCollector {
-    field: Field,
-}
-
-impl StatsCollector {
-    fn with_field(field: Field) -> StatsCollector {
-        StatsCollector { field }
-    }
-}
-
-impl Collector for StatsCollector {
-    // That's the type of our result.
-    // Our standard deviation will be a float.
-    type Fruit = Option<Stats>;
-
-    type Child = StatsSegmentCollector;
-
-    fn for_segment(
-        &self,
-        _segment_local_id: u32,
-        segment: &SegmentReader,
-    ) -> tantivy::Result<StatsSegmentCollector> {
-        let fast_field_reader = segment.fast_field_reader(self.field)?;
-        Ok(StatsSegmentCollector {
-            fast_field_reader,
-            stats: Stats::default(),
-        })
-    }
-
-    fn requires_scoring(&self) -> bool {
-        // this collector does not care about score.
-        false
-    }
-
-    fn merge_fruits(&self, segment_stats: Vec<Option<Stats>>) -> tantivy::Result<Option<Stats>> {
-        let mut stats = Stats::default();
-        for segment_stats_opt in segment_stats {
-            if let Some(segment_stats) = segment_stats_opt {
-                stats.count += segment_stats.count;
-                stats.sum += segment_stats.sum;
-                stats.squared_sum += segment_stats.squared_sum;
-            }
-        }
-        Ok(stats.non_zero_count())
-    }
-}
-
-struct StatsSegmentCollector {
-    fast_field_reader: FastFieldReader<u64>,
-    stats: Stats,
-}
-
-impl SegmentCollector for StatsSegmentCollector {
-    type Fruit = Option<Stats>;
-
-    fn collect(&mut self, doc: u32, _score: f32) {
-        let value = self.fast_field_reader.get(doc) as f64;
-        self.stats.count += 1;
-        self.stats.sum += value;
-        self.stats.squared_sum += value * value;
-    }
-
-    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
-        self.stats.non_zero_count()
-    }
-}
-
-fn main() -> tantivy::Result<()> {
-    // # Defining the schema
-    //
-    // The Tantivy index requires a very strict schema.
-    // The schema declares which fields are in the index,
-    // and for each field, its type and "the way it should
-    // be indexed".
-
-    // first we need to define a schema ...
-    let mut schema_builder = Schema::builder();
-
-    // We'll assume a fictional index containing
-    // products, and with a name, a description, and a price.
-    let product_name = schema_builder.add_text_field("name", TEXT);
-    let product_description = schema_builder.add_text_field("description", TEXT);
-    let price = schema_builder.add_u64_field("price", INT_INDEXED | FAST);
-    let schema = schema_builder.build();
-
-    // # Indexing documents
-    //
-    // Lets index a bunch of fake documents for the sake of
-    // this example.
-    let index = Index::create_in_ram(schema.clone());
-
-    let mut index_writer = index.writer(50_000_000)?;
-    index_writer.add_document(doc!(
-        product_name => "Super Broom 2000",
-        product_description => "While it is ok for short distance travel, this broom \
-        was designed quiditch. It will up your game.",
-        price => 30_200u64
-    ));
-    index_writer.add_document(doc!(
-        product_name => "Turbulobroom",
-        product_description => "You might have heard of this broom before : it is the sponsor of the Wales team.\
-            You'll enjoy its sharp turns, and rapid acceleration",
-        price => 29_240u64
-    ));
-    index_writer.add_document(doc!(
-        product_name => "Broomio",
-        product_description => "Great value for the price. This broom is a market favorite",
-        price => 21_240u64
-    ));
-    index_writer.add_document(doc!(
-        product_name => "Whack a Mole",
-        product_description => "Prime quality bat.",
-        price => 5_200u64
-    ));
-    index_writer.commit()?;
-    index.load_searchers()?;
-
-    let searcher = index.searcher();
-    let query_parser = QueryParser::for_index(&index, vec![product_name, product_description]);
-
-    // here we want to get a hit on the 'ken' in Frankenstein
-    let query = query_parser.parse_query("broom")?;
-    if let Some(stats) = searcher.search(&query, &StatsCollector::with_field(price))? {
-        println!("count: {}", stats.count());
-        println!("mean: {}", stats.mean());
-        println!("standard deviation: {}", stats.standard_deviation());
-    }
-
-    Ok(())
-}

examples/custom_tokenizer.rs

@@ -3,14 +3,16 @@
 // In this example, we'll see how to define a tokenizer pipeline
 // by aligning a bunch of `TokenFilter`.
 
+
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopDocs;
+use tantivy::collector::TopCollector;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::tokenizer::NgramTokenizer;
 use tantivy::Index;
 
+
 fn main() -> tantivy::Result<()> {
     // # Defining the schema
     //
@@ -20,7 +22,7 @@ fn main() -> tantivy::Result<()> {
     // be indexed".
 
     // first we need to define a schema ...
-    let mut schema_builder = Schema::builder();
+    let mut schema_builder = SchemaBuilder::default();
 
     // Our first field is title.
     // In this example we want to use NGram searching
@@ -68,12 +70,12 @@ fn main() -> tantivy::Result<()> {
     // heap for the indexer can increase its throughput.
     let mut index_writer = index.writer(50_000_000)?;
     index_writer.add_document(doc!(
-    title => "The Old Man and the Sea",
-    body => "He was an old man who fished alone in a skiff in the Gulf Stream and \
-     he had gone eighty-four days now without taking a fish."
+        title => "The Old Man and the Sea",
+        body => "He was an old man who fished alone in a skiff in the Gulf Stream and \
+         he had gone eighty-four days now without taking a fish."
     ));
     index_writer.add_document(doc!(
-    title => "Of Mice and Men",
+       title => "Of Mice and Men",
        body => r#"A few miles south of Soledad, the Salinas River drops in close to the hillside
                 bank and runs deep and green. The water is warm too, for it has slipped twinkling
                 over the yellow sands in the sunlight before reaching the narrow pool. On one
@@ -84,7 +86,7 @@ fn main() -> tantivy::Result<()> {
                 limbs and branches that arch over the pool"#
     ));
     index_writer.add_document(doc!(
-    title => "Frankenstein",
+        title => "Frankenstein",
         body => r#"You will rejoice to hear that no disaster has accompanied the commencement of an
                 enterprise which you have regarded with such evil forebodings.  I arrived here
                 yesterday, and my first task is to assure my dear sister of my welfare and
@@ -104,10 +106,12 @@ fn main() -> tantivy::Result<()> {
     // here we want to get a hit on the 'ken' in Frankenstein
     let query = query_parser.parse_query("ken")?;
 
-    let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
+    let mut top_collector = TopCollector::with_limit(10);
+    searcher.search(&*query, &mut top_collector)?;
 
-    for (_, doc_address) in top_docs {
-        let retrieved_doc = searcher.doc(doc_address)?;
+    let doc_addresses = top_collector.docs();
+    for doc_address in doc_addresses {
+        let retrieved_doc = searcher.doc(&doc_address)?;
         println!("{}", schema.to_json(&retrieved_doc));
     }
 

examples/deleting_updating_documents.rs

@@ -10,10 +10,11 @@
 // Importing tantivy...
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopDocs;
-use tantivy::query::TermQuery;
+use tantivy::collector::TopCollector;
 use tantivy::schema::*;
 use tantivy::Index;
+use tantivy::query::TermQuery;
+
 
 // A simple helper function to fetch a single document
 // given its id from our index.
@@ -27,10 +28,11 @@ fn extract_doc_given_isbn(index: &Index, isbn_term: &Term) -> tantivy::Result<Op
     // The second argument is here to tell we don't care about decoding positions,
     // or term frequencies.
     let term_query = TermQuery::new(isbn_term.clone(), IndexRecordOption::Basic);
-    let top_docs = searcher.search(&term_query, &TopDocs::with_limit(1))?;
+    let mut top_collector = TopCollector::with_limit(1);
+    searcher.search(&term_query, &mut top_collector)?;
 
-    if let Some((_score, doc_address)) = top_docs.first() {
-        let doc = searcher.doc(*doc_address)?;
+    if let Some(doc_address) =  top_collector.docs().first() {
+        let doc = searcher.doc(doc_address)?;
         Ok(Some(doc))
     } else {
         // no doc matching this ID.
@@ -39,11 +41,12 @@ fn extract_doc_given_isbn(index: &Index, isbn_term: &Term) -> tantivy::Result<Op
 }
 
 fn main() -> tantivy::Result<()> {
+
     // # Defining the schema
     //
     // Check out the *basic_search* example if this makes
     // small sense to you.
-    let mut schema_builder = Schema::builder();
+    let mut schema_builder = SchemaBuilder::default();
 
     // Tantivy does not really have a notion of primary id.
     // This may change in the future.
@@ -123,6 +126,7 @@ fn main() -> tantivy::Result<()> {
        isbn => "978-9176370711",
     ));
 
+
     // You are guaranteed that your clients will only observe your index in
     // the state it was in after a commit.
     // In this example, your search engine will at no point be missing the *Frankenstein* document.
@@ -139,4 +143,4 @@ fn main() -> tantivy::Result<()> {
     );
 
     Ok(())
-}
+}

examples/faceted_search.rs

@@ -22,59 +22,60 @@ use tantivy::schema::*;
 use tantivy::Index;
 
 fn main() -> tantivy::Result<()> {
-    // Let's create a temporary directory for the
-    // sake of this example
-    let index_path = TempDir::new("tantivy_facet_example_dir")?;
-    let mut schema_builder = Schema::builder();
+  // Let's create a temporary directory for the
+  // sake of this example
+  let index_path = TempDir::new("tantivy_facet_example_dir")?;
+  let mut schema_builder = SchemaBuilder::default();
 
-    schema_builder.add_text_field("name", TEXT | STORED);
+  schema_builder.add_text_field("name", TEXT | STORED);
 
-    // this is our faceted field
-    schema_builder.add_facet_field("tags");
+  // this is our faceted field
+  schema_builder.add_facet_field("tags");
 
-    let schema = schema_builder.build();
+  let schema = schema_builder.build();
 
-    let index = Index::create_in_dir(&index_path, schema.clone())?;
+  let index = Index::create_in_dir(&index_path, schema.clone())?;
 
-    let mut index_writer = index.writer(50_000_000)?;
+  let mut index_writer = index.writer(50_000_000)?;
 
-    let name = schema.get_field("name").unwrap();
-    let tags = schema.get_field("tags").unwrap();
+  let name = schema.get_field("name").unwrap();
+  let tags = schema.get_field("tags").unwrap();
 
-    // For convenience, tantivy also comes with a macro to
-    // reduce the boilerplate above.
-    index_writer.add_document(doc!(
+  // For convenience, tantivy also comes with a macro to
+  // reduce the boilerplate above.
+  index_writer.add_document(doc!(
         name => "the ditch",
         tags => Facet::from("/pools/north")
     ));
 
-    index_writer.add_document(doc!(
+  index_writer.add_document(doc!(
         name => "little stacey",
         tags => Facet::from("/pools/south")
     ));
 
-    index_writer.commit()?;
+  index_writer.commit()?;
 
-    index.load_searchers()?;
+  index.load_searchers()?;
 
-    let searcher = index.searcher();
+  let searcher = index.searcher();
 
-    let mut facet_collector = FacetCollector::for_field(tags);
-    facet_collector.add_facet("/pools");
+  let mut facet_collector = FacetCollector::for_field(tags);
+  facet_collector.add_facet("/pools");
 
-    let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
+  searcher.search(&AllQuery, &mut facet_collector).unwrap();
 
-    // This lists all of the facet counts
-    let facets: Vec<(&Facet, u64)> = facet_counts.get("/pools").collect();
-    assert_eq!(
-        facets,
-        vec![
-            (&Facet::from("/pools/north"), 1),
-            (&Facet::from("/pools/south"), 1),
-        ]
-    );
+  let counts = facet_collector.harvest();
+  // This lists all of the facet counts
+  let facets: Vec<(&Facet, u64)> = counts.get("/pools").collect();
+  assert_eq!(
+    facets,
+    vec![
+      (&Facet::from("/pools/north"), 1),
+      (&Facet::from("/pools/south"), 1)
+    ]
+  );
 
-    Ok(())
+  Ok(())
 }
 
 use tempdir::TempDir;

examples/iterating_docs_and_positions.rs

@@ -7,18 +7,21 @@
 // the list of documents containing a term, getting
 // its term frequency, and accessing its positions.
 
+
 // ---
 // Importing tantivy...
 #[macro_use]
 extern crate tantivy;
 use tantivy::schema::*;
 use tantivy::Index;
-use tantivy::{DocId, DocSet, Postings};
+use tantivy::{DocSet, DocId, Postings};
 
 fn main() -> tantivy::Result<()> {
+
+
     // We first create a schema for the sake of the
     // example. Check the `basic_search` example for more information.
-    let mut schema_builder = Schema::builder();
+    let mut schema_builder = SchemaBuilder::default();
 
     // For this example, we need to make sure to index positions for our title
     // field. `TEXT` precisely does this.
@@ -44,6 +47,7 @@ fn main() -> tantivy::Result<()> {
     // there is actually only one segment here, but let's iterate through the list
     // anyway)
     for segment_reader in searcher.segment_readers() {
+
         // A segment contains different data structure.
         // Inverted index stands for the combination of
         // - the term dictionary
@@ -54,18 +58,19 @@ fn main() -> tantivy::Result<()> {
         // Let's go through all docs containing the term `title:the` and access their position
         let term_the = Term::from_field_text(title, "the");
 
+
         // This segment posting object is like a cursor over the documents matching the term.
         // The `IndexRecordOption` arguments tells tantivy we will be interested in both term frequencies
         // and positions.
         //
         // If you don't need all this information, you may get better performance by decompressing less
         // information.
-        if let Some(mut segment_postings) =
-            inverted_index.read_postings(&term_the, IndexRecordOption::WithFreqsAndPositions)
-        {
+        if let Some(mut segment_postings) = inverted_index.read_postings(&term_the, IndexRecordOption::WithFreqsAndPositions) {
+
             // this buffer will be used to request for positions
             let mut positions: Vec<u32> = Vec::with_capacity(100);
             while segment_postings.advance() {
+
                 // the number of time the term appears in the document.
                 let doc_id: DocId = segment_postings.doc(); //< do not try to access this before calling advance once.
 
@@ -93,6 +98,7 @@ fn main() -> tantivy::Result<()> {
         }
     }
 
+
     // A `Term` is a text token associated with a field.
     // Let's go through all docs containing the term `title:the` and access their position
     let term_the = Term::from_field_text(title, "the");
@@ -105,6 +111,7 @@ fn main() -> tantivy::Result<()> {
     // Also, for some VERY specific high performance use case like an OLAP analysis of logs,
     // you can get better performance by accessing directly the blocks of doc ids.
     for segment_reader in searcher.segment_readers() {
+
         // A segment contains different data structure.
         // Inverted index stands for the combination of
         // - the term dictionary
@@ -117,9 +124,7 @@ fn main() -> tantivy::Result<()> {
         //
         // If you don't need all this information, you may get better performance by decompressing less
         // information.
-        if let Some(mut block_segment_postings) =
-            inverted_index.read_block_postings(&term_the, IndexRecordOption::Basic)
-        {
+        if let Some(mut block_segment_postings) = inverted_index.read_block_postings(&term_the, IndexRecordOption::Basic) {
             while block_segment_postings.advance() {
                 // Once again these docs MAY contains deleted documents as well.
                 let docs = block_segment_postings.docs();
@@ -131,3 +136,4 @@ fn main() -> tantivy::Result<()> {
 
     Ok(())
 }
+

examples/snippet.rs

@@ -1,87 +0,0 @@
-// # Snippet example
-//
-// This example shows how to return a representative snippet of
-// your hit result.
-// Snippet are an extracted of a target document, and returned in HTML format.
-// The keyword searched by the user are highlighted with a `<b>` tag.
-extern crate tempdir;
-
-// ---
-// Importing tantivy...
-#[macro_use]
-extern crate tantivy;
-use tantivy::collector::TopDocs;
-use tantivy::query::QueryParser;
-use tantivy::schema::*;
-use tantivy::Index;
-use tantivy::{Snippet, SnippetGenerator};
-use tempdir::TempDir;
-
-fn main() -> tantivy::Result<()> {
-    // Let's create a temporary directory for the
-    // sake of this example
-    let index_path = TempDir::new("tantivy_example_dir")?;
-
-    // # Defining the schema
-    let mut schema_builder = Schema::builder();
-    let title = schema_builder.add_text_field("title", TEXT | STORED);
-    let body = schema_builder.add_text_field("body", TEXT | STORED);
-    let schema = schema_builder.build();
-
-    // # Indexing documents
-    let index = Index::create_in_dir(&index_path, schema.clone())?;
-
-    let mut index_writer = index.writer(50_000_000)?;
-
-    // we'll only need one doc for this example.
-    index_writer.add_document(doc!(
-    title => "Of Mice and Men",
-    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-            over the yellow sands in the sunlight before reaching the narrow pool. On one \
-            side of the river the golden foothill slopes curve up to the strong and rocky \
-            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-            fresh and green with every spring, carrying in their lower leaf junctures the \
-            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-            limbs and branches that arch over the pool"
-    ));
-    // ...
-    index_writer.commit()?;
-
-    index.load_searchers()?;
-
-    let searcher = index.searcher();
-    let query_parser = QueryParser::for_index(&index, vec![title, body]);
-    let query = query_parser.parse_query("sycamore spring")?;
-
-    let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
-
-    let snippet_generator = SnippetGenerator::create(&searcher, &*query, body)?;
-
-    for (score, doc_address) in top_docs {
-        let doc = searcher.doc(doc_address)?;
-        let snippet = snippet_generator.snippet_from_doc(&doc);
-        println!("Document score {}:", score);
-        println!("title: {}", doc.get_first(title).unwrap().text().unwrap());
-        println!("snippet: {}", snippet.to_html());
-        println!("custom highlighting: {}", highlight(snippet));
-    }
-
-    Ok(())
-}
-
-fn highlight(snippet: Snippet) -> String {
-    let mut result = String::new();
-    let mut start_from = 0;
-
-    for (start, end) in snippet.highlighted().iter().map(|h| h.bounds()) {
-        result.push_str(&snippet.fragments()[start_from..start]);
-        result.push_str(" --> ");
-        result.push_str(&snippet.fragments()[start..end]);
-        result.push_str(" <-- ");
-        start_from = end;
-    }
-
-    result.push_str(&snippet.fragments()[start_from..]);
-    result
-}

examples/stop_words.rs

@@ -15,103 +15,115 @@ extern crate tempdir;
 // Importing tantivy...
 #[macro_use]
 extern crate tantivy;
-use tantivy::collector::TopDocs;
+use tantivy::collector::TopCollector;
 use tantivy::query::QueryParser;
 use tantivy::schema::*;
 use tantivy::tokenizer::*;
 use tantivy::Index;
 
 fn main() -> tantivy::Result<()> {
-    // this example assumes you understand the content in `basic_search`
-    let mut schema_builder = Schema::builder();
+  // this example assumes you understand the content in `basic_search`
+  let index_path = TempDir::new("tantivy_stopwords_example_dir")?;
+  let mut schema_builder = SchemaBuilder::default();
 
-    // This configures your custom options for how tantivy will
-    // store and process your content in the index; The key
-    // to note is that we are setting the tokenizer to `stoppy`
-    // which will be defined and registered below.
-    let text_field_indexing = TextFieldIndexing::default()
-        .set_tokenizer("stoppy")
-        .set_index_option(IndexRecordOption::WithFreqsAndPositions);
-    let text_options = TextOptions::default()
-        .set_indexing_options(text_field_indexing)
-        .set_stored();
+  // This configures your custom options for how tantivy will
+  // store and process your content in the index; The key
+  // to note is that we are setting the tokenizer to `stoppy`
+  // which will be defined and registered below.
+  let text_field_indexing = TextFieldIndexing::default()
+    .set_tokenizer("stoppy")
+    .set_index_option(IndexRecordOption::WithFreqsAndPositions);
+  let text_options = TextOptions::default()
+    .set_indexing_options(text_field_indexing)
+    .set_stored();
 
-    // Our first field is title.
-    schema_builder.add_text_field("title", text_options);
+  // Our first field is title.
+  schema_builder.add_text_field("title", text_options);
 
-    // Our second field is body.
-    let text_field_indexing = TextFieldIndexing::default()
-        .set_tokenizer("stoppy")
-        .set_index_option(IndexRecordOption::WithFreqsAndPositions);
-    let text_options = TextOptions::default()
-        .set_indexing_options(text_field_indexing)
-        .set_stored();
-    schema_builder.add_text_field("body", text_options);
+  // Our second field is body.
+  let text_field_indexing = TextFieldIndexing::default()
+    .set_tokenizer("stoppy")
+    .set_index_option(IndexRecordOption::WithFreqsAndPositions);
+  let text_options = TextOptions::default()
+    .set_indexing_options(text_field_indexing)
+    .set_stored();
+  schema_builder.add_text_field("body", text_options);
 
-    let schema = schema_builder.build();
+  let schema = schema_builder.build();
 
-    let index = Index::create_in_ram(schema.clone());
+  let index = Index::create_in_dir(&index_path, schema.clone())?;
 
-    // This tokenizer lowers all of the text (to help with stop word matching)
-    // then removes all instances of `the` and `and` from the corpus
-    let tokenizer = SimpleTokenizer
-        .filter(LowerCaser)
-        .filter(StopWordFilter::remove(vec![
-            "the".to_string(),
-            "and".to_string(),
-        ]));
+  // This tokenizer lowers all of the text (to help with stop word matching)
+  // then removes all instances of `the` and `and` from the corpus
+  let tokenizer = SimpleTokenizer
+    .filter(LowerCaser)
+    .filter(StopWordFilter::remove(vec![
+      "the".to_string(),
+      "and".to_string(),
+    ]));
 
-    index.tokenizers().register("stoppy", tokenizer);
+  index.tokenizers().register("stoppy", tokenizer);
 
-    let mut index_writer = index.writer(50_000_000)?;
+  let mut index_writer = index.writer(50_000_000)?;
 
-    let title = schema.get_field("title").unwrap();
-    let body = schema.get_field("body").unwrap();
+  let title = schema.get_field("title").unwrap();
+  let body = schema.get_field("body").unwrap();
 
-    index_writer.add_document(doc!(
+  index_writer.add_document(doc!(
     title => "The Old Man and the Sea",
     body => "He was an old man who fished alone in a skiff in the Gulf Stream and \
      he had gone eighty-four days now without taking a fish."
+  ));
+
+  index_writer.add_document(doc!(
+        title => "Of Mice and Men",
+        body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
+                bank and runs deep and green. The water is warm too, for it has slipped twinkling \
+                over the yellow sands in the sunlight before reaching the narrow pool. On one \
+                side of the river the golden foothill slopes curve up to the strong and rocky \
+                Gabilan Mountains, but on the valley side the water is lined with trees—willows \
+                fresh and green with every spring, carrying in their lower leaf junctures the \
+                debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
+                limbs and branches that arch over the pool"
     ));
 
-    index_writer.add_document(doc!(
-    title => "Of Mice and Men",
-    body => "A few miles south of Soledad, the Salinas River drops in close to the hillside \
-            bank and runs deep and green. The water is warm too, for it has slipped twinkling \
-            over the yellow sands in the sunlight before reaching the narrow pool. On one \
-            side of the river the golden foothill slopes curve up to the strong and rocky \
-            Gabilan Mountains, but on the valley side the water is lined with trees—willows \
-            fresh and green with every spring, carrying in their lower leaf junctures the \
-            debris of the winter’s flooding; and sycamores with mottled, white, recumbent \
-            limbs and branches that arch over the pool"
+  index_writer.add_document(doc!(
+       title => "Frankenstein",
+       body => "You will rejoice to hear that no disaster has accompanied the commencement of an \
+                enterprise which you have regarded with such evil forebodings.  I arrived here \
+                yesterday, and my first task is to assure my dear sister of my welfare and \
+                increasing confidence in the success of my undertaking."
     ));
 
-    index_writer.add_document(doc!(
-    title => "Frankenstein",
-    body => "You will rejoice to hear that no disaster has accompanied the commencement of an \
-             enterprise which you have regarded with such evil forebodings.  I arrived here \
-             yesterday, and my first task is to assure my dear sister of my welfare and \
-             increasing confidence in the success of my undertaking."
-    ));
+  index_writer.commit()?;
 
-    index_writer.commit()?;
+  index.load_searchers()?;
 
-    index.load_searchers()?;
+  let searcher = index.searcher();
 
-    let searcher = index.searcher();
+  let query_parser = QueryParser::for_index(&index, vec![title, body]);
 
-    let query_parser = QueryParser::for_index(&index, vec![title, body]);
+  // this will have NO hits because it was filtered out
+  // because the query is run through the analyzer you
+  // actually will get an error here because the query becomes
+  // empty
+  assert!(query_parser.parse_query("the").is_err());
 
-    // stop words are applied on the query as well.
-    // The following will be equivalent to `title:frankenstein`
-    let query = query_parser.parse_query("title:\"the Frankenstein\"")?;
-    let top_docs = searcher.search(&query, &TopDocs::with_limit(10))?;
+  // this will have hits
+  let query = query_parser.parse_query("is")?;
 
-    for (score, doc_address) in top_docs {
-        let retrieved_doc = searcher.doc(doc_address)?;
-        println!("\n==\nDocument score {}:", score);
-        println!("{}", schema.to_json(&retrieved_doc));
-    }
+  let mut top_collector = TopCollector::with_limit(10);
 
-    Ok(())
+  searcher.search(&*query, &mut top_collector)?;
+
+  let doc_addresses = top_collector.docs();
+
+  for doc_address in doc_addresses {
+    let retrieved_doc = searcher.doc(&doc_address)?;
+    println!("{}", schema.to_json(&retrieved_doc));
+  }
+
+  Ok(())
 }
+
+use tempdir::TempDir;

examples/working_with_json.rs

@@ -9,7 +9,7 @@ fn main() -> tantivy::Result<()> {
     // Check out the basic example if this is confusing to you.
     //
     // first we need to define a schema ...
-    let mut schema_builder = Schema::builder();
+    let mut schema_builder = SchemaBuilder::default();
     schema_builder.add_text_field("title", TEXT | STORED);
     schema_builder.add_text_field("body", TEXT);
     schema_builder.add_u64_field("year", INT_INDEXED);

run-tests.sh

@@ -1,2 +0,0 @@
-#!/bin/bash
-cargo test --no-default-features --features mmap -- --test-threads 1

src/collector/chained_collector.rs

@@ -0,0 +1,142 @@
+use collector::Collector;
+use DocId;
+use Result;
+use Score;
+use SegmentLocalId;
+use SegmentReader;
+
+/// Collector that does nothing.
+/// This is used in the chain Collector and will hopefully
+/// be optimized away by the compiler.
+pub struct DoNothingCollector;
+impl Collector for DoNothingCollector {
+    #[inline]
+    fn set_segment(&mut self, _: SegmentLocalId, _: &SegmentReader) -> Result<()> {
+        Ok(())
+    }
+    #[inline]
+    fn collect(&mut self, _doc: DocId, _score: Score) {}
+    #[inline]
+    fn requires_scoring(&self) -> bool {
+        false
+    }
+}
+
+/// Zero-cost abstraction used to collect on multiple collectors.
+/// This contraption is only usable if the type of your collectors
+/// are known at compile time.
+///
+/// ```rust
+/// #[macro_use]
+/// extern crate tantivy;
+/// use tantivy::schema::{SchemaBuilder, TEXT};
+/// use tantivy::{Index, Result};
+/// use tantivy::collector::{CountCollector, TopCollector, chain};
+/// use tantivy::query::QueryParser;
+///
+/// # fn main() { example().unwrap(); }
+/// fn example() -> Result<()> {
+///     let mut schema_builder = SchemaBuilder::new();
+///     let title = schema_builder.add_text_field("title", TEXT);
+///     let schema = schema_builder.build();
+///     let index = Index::create_in_ram(schema);
+///     {
+///         let mut index_writer = index.writer(3_000_000)?;
+///         index_writer.add_document(doc!(
+///             title => "The Name of the Wind",
+///         ));
+///         index_writer.add_document(doc!(
+///             title => "The Diary of Muadib",
+///         ));
+///         index_writer.add_document(doc!(
+///             title => "A Dairy Cow",
+///         ));
+///         index_writer.add_document(doc!(
+///             title => "The Diary of a Young Girl",
+///         ));
+///         index_writer.commit().unwrap();
+///     }
+///
+///     index.load_searchers()?;
+///     let searcher = index.searcher();
+///
+///     {
+///         let mut top_collector = TopCollector::with_limit(2);
+///         let mut count_collector = CountCollector::default();
+///         {
+///             let mut collectors = chain().push(&mut top_collector).push(&mut count_collector);
+///             let query_parser = QueryParser::for_index(&index, vec![title]);
+///             let query = query_parser.parse_query("diary")?;
+///             searcher.search(&*query, &mut collectors).unwrap();
+///         }
+///         assert_eq!(count_collector.count(), 2);
+///         assert!(top_collector.at_capacity());
+///     }
+///
+///     Ok(())
+/// }
+/// ```
+pub struct ChainedCollector<Left: Collector, Right: Collector> {
+    left: Left,
+    right: Right,
+}
+
+impl<Left: Collector, Right: Collector> ChainedCollector<Left, Right> {
+    /// Adds a collector
+    pub fn push<C: Collector>(self, new_collector: &mut C) -> ChainedCollector<Self, &mut C> {
+        ChainedCollector {
+            left: self,
+            right: new_collector,
+        }
+    }
+}
+
+impl<Left: Collector, Right: Collector> Collector for ChainedCollector<Left, Right> {
+    fn set_segment(
+        &mut self,
+        segment_local_id: SegmentLocalId,
+        segment: &SegmentReader,
+    ) -> Result<()> {
+        self.left.set_segment(segment_local_id, segment)?;
+        self.right.set_segment(segment_local_id, segment)?;
+        Ok(())
+    }
+
+    fn collect(&mut self, doc: DocId, score: Score) {
+        self.left.collect(doc, score);
+        self.right.collect(doc, score);
+    }
+
+    fn requires_scoring(&self) -> bool {
+        self.left.requires_scoring() || self.right.requires_scoring()
+    }
+}
+
+/// Creates a `ChainedCollector`
+pub fn chain() -> ChainedCollector<DoNothingCollector, DoNothingCollector> {
+    ChainedCollector {
+        left: DoNothingCollector,
+        right: DoNothingCollector,
+    }
+}
+
+#[cfg(test)]
+mod tests {
+
+    use super::*;
+    use collector::{Collector, CountCollector, TopCollector};
+
+    #[test]
+    fn test_chained_collector() {
+        let mut top_collector = TopCollector::with_limit(2);
+        let mut count_collector = CountCollector::default();
+        {
+            let mut collectors = chain().push(&mut top_collector).push(&mut count_collector);
+            collectors.collect(1, 0.2);
+            collectors.collect(2, 0.1);
+            collectors.collect(3, 0.5);
+        }
+        assert_eq!(count_collector.count(), 3);
+        assert!(top_collector.at_capacity());
+    }
+}

src/collector/count_collector.rs

@@ -1,5 +1,4 @@
 use super::Collector;
-use collector::SegmentCollector;
 use DocId;
 use Result;
 use Score;
@@ -12,14 +11,14 @@ use SegmentReader;
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{Schema, TEXT};
+/// use tantivy::schema::{SchemaBuilder, TEXT};
 /// use tantivy::{Index, Result};
-/// use tantivy::collector::Count;
+/// use tantivy::collector::CountCollector;
 /// use tantivy::query::QueryParser;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = Schema::builder();
+///     let mut schema_builder = SchemaBuilder::new();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
 ///     let index = Index::create_in_ram(schema);
@@ -44,86 +43,59 @@ use SegmentReader;
 ///     let searcher = index.searcher();
 ///
 ///     {
+///	        let mut count_collector = CountCollector::default();
 ///         let query_parser = QueryParser::for_index(&index, vec![title]);
 ///         let query = query_parser.parse_query("diary")?;
-///         let count = searcher.search(&query, &Count).unwrap();
+///         searcher.search(&*query, &mut count_collector).unwrap();
 ///
-///         assert_eq!(count, 2);
+///         assert_eq!(count_collector.count(), 2);
 ///     }
 ///
 ///     Ok(())
 /// }
 /// ```
-pub struct Count;
-
-impl Collector for Count {
-    type Fruit = usize;
-
-    type Child = SegmentCountCollector;
-
-    fn for_segment(&self, _: SegmentLocalId, _: &SegmentReader) -> Result<SegmentCountCollector> {
-        Ok(SegmentCountCollector::default())
-    }
-
-    fn requires_scoring(&self) -> bool {
-        false
-    }
-
-    fn merge_fruits(&self, segment_counts: Vec<usize>) -> Result<usize> {
-        Ok(segment_counts.into_iter().sum())
-    }
-}
-
 #[derive(Default)]
-pub struct SegmentCountCollector {
+pub struct CountCollector {
     count: usize,
 }
 
-impl SegmentCollector for SegmentCountCollector {
-    type Fruit = usize;
+impl CountCollector {
+    /// Returns the count of documents that were
+    /// collected.
+    pub fn count(&self) -> usize {
+        self.count
+    }
+}
+
+impl Collector for CountCollector {
+    fn set_segment(&mut self, _: SegmentLocalId, _: &SegmentReader) -> Result<()> {
+        Ok(())
+    }
 
     fn collect(&mut self, _: DocId, _: Score) {
         self.count += 1;
     }
 
-    fn harvest(self) -> usize {
-        self.count
+    fn requires_scoring(&self) -> bool {
+        false
     }
 }
 
 #[cfg(test)]
 mod tests {
-    use super::{Count, SegmentCountCollector};
-    use collector::Collector;
-    use collector::SegmentCollector;
+
+    use collector::{Collector, CountCollector};
 
     #[test]
-    fn test_count_collect_does_not_requires_scoring() {
-        assert!(!Count.requires_scoring());
-    }
-
-    #[test]
-    fn test_segment_count_collector() {
-        {
-            let count_collector = SegmentCountCollector::default();
-            assert_eq!(count_collector.harvest(), 0);
-        }
-        {
-            let mut count_collector = SegmentCountCollector::default();
-            count_collector.collect(0u32, 1f32);
-            assert_eq!(count_collector.harvest(), 1);
-        }
-        {
-            let mut count_collector = SegmentCountCollector::default();
-            count_collector.collect(0u32, 1f32);
-            assert_eq!(count_collector.harvest(), 1);
-        }
-        {
-            let mut count_collector = SegmentCountCollector::default();
-            count_collector.collect(0u32, 1f32);
-            count_collector.collect(1u32, 1f32);
-            assert_eq!(count_collector.harvest(), 2);
-        }
+    fn test_count_collector() {
+        let mut count_collector = CountCollector::default();
+        assert_eq!(count_collector.count(), 0);
+        count_collector.collect(0u32, 1f32);
+        assert_eq!(count_collector.count(), 1);
+        assert_eq!(count_collector.count(), 1);
+        count_collector.collect(1u32, 1f32);
+        assert_eq!(count_collector.count(), 2);
+        assert!(!count_collector.requires_scoring());
     }
 
 }

src/collector/facet_collector.rs

@@ -1,17 +1,20 @@
 use collector::Collector;
-use collector::SegmentCollector;
 use docset::SkipResult;
 use fastfield::FacetReader;
 use schema::Facet;
 use schema::Field;
-use std::cmp::Ordering;
+use std::cell::UnsafeCell;
 use std::collections::btree_map;
 use std::collections::BTreeMap;
 use std::collections::BTreeSet;
 use std::collections::BinaryHeap;
 use std::collections::Bound;
 use std::iter::Peekable;
+use std::mem;
 use std::{u64, usize};
+use termdict::TermMerger;
+
+use std::cmp::Ordering;
 use DocId;
 use Result;
 use Score;
@@ -43,6 +46,12 @@ impl<'a> Ord for Hit<'a> {
     }
 }
 
+struct SegmentFacetCounter {
+    pub facet_reader: FacetReader,
+    pub facet_ords: Vec<u64>,
+    pub facet_counts: Vec<u64>,
+}
+
 fn facet_depth(facet_bytes: &[u8]) -> usize {
     if facet_bytes.is_empty() {
         0
@@ -82,14 +91,14 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{Facet, Schema, TEXT};
+/// use tantivy::schema::{Facet, SchemaBuilder, TEXT};
 /// use tantivy::{Index, Result};
 /// use tantivy::collector::FacetCollector;
 /// use tantivy::query::AllQuery;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = Schema::builder();
+///     let mut schema_builder = SchemaBuilder::new();
 ///
 ///     // Facet have their own specific type.
 ///     // It is not a bad practise to put all of your
@@ -132,10 +141,13 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 ///			let mut facet_collector = FacetCollector::for_field(facet);
 ///         facet_collector.add_facet("/lang");
 ///         facet_collector.add_facet("/category");
-///         let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
+///         searcher.search(&AllQuery, &mut facet_collector).unwrap();
+///
+///         // this object contains count aggregate for all of the facets.
+///         let counts = facet_collector.harvest();
 ///
 ///         // This lists all of the facet counts
-///         let facets: Vec<(&Facet, u64)> = facet_counts
+///         let facets: Vec<(&Facet, u64)> = counts
 ///             .get("/category")
 ///             .collect();
 ///         assert_eq!(facets, vec![
@@ -147,10 +159,13 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 ///     {
 ///			let mut facet_collector = FacetCollector::for_field(facet);
 ///         facet_collector.add_facet("/category/fiction");
-///         let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
+///         searcher.search(&AllQuery, &mut facet_collector).unwrap();
+///
+///         // this object contains count aggregate for all of the facets.
+///         let counts = facet_collector.harvest();
 ///
 ///         // This lists all of the facet counts
-///         let facets: Vec<(&Facet, u64)> = facet_counts
+///         let facets: Vec<(&Facet, u64)> = counts
 ///             .get("/category/fiction")
 ///             .collect();
 ///         assert_eq!(facets, vec![
@@ -163,10 +178,13 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 ///    {
 ///			let mut facet_collector = FacetCollector::for_field(facet);
 ///         facet_collector.add_facet("/category/fiction");
-///         let facet_counts = searcher.search(&AllQuery, &facet_collector).unwrap();
+///         searcher.search(&AllQuery, &mut facet_collector).unwrap();
+///
+///         // this object contains count aggregate for all of the facets.
+///         let counts = facet_collector.harvest();
 ///
 ///         // This lists all of the facet counts
-///         let facets: Vec<(&Facet, u64)> = facet_counts.top_k("/category/fiction", 1);
+///         let facets: Vec<(&Facet, u64)> = counts.top_k("/category/fiction", 1);
 ///         assert_eq!(facets, vec![
 ///             (&Facet::from("/category/fiction/fantasy"), 2)
 ///         ]);
@@ -176,19 +194,19 @@ fn facet_depth(facet_bytes: &[u8]) -> usize {
 /// }
 /// ```
 pub struct FacetCollector {
+    facet_ords: Vec<u64>,
     field: Field,
-    facets: BTreeSet<Facet>,
-}
+    ff_reader: Option<UnsafeCell<FacetReader>>,
+    segment_counters: Vec<SegmentFacetCounter>,
 
-pub struct FacetSegmentCollector {
-    reader: FacetReader,
-    facet_ords_buf: Vec<u64>,
     // facet_ord -> collapse facet_id
-    collapse_mapping: Vec<usize>,
+    current_segment_collapse_mapping: Vec<usize>,
     // collapse facet_id -> count
-    counts: Vec<u64>,
+    current_segment_counts: Vec<u64>,
     // collapse facet_id -> facet_ord
-    collapse_facet_ords: Vec<u64>,
+    current_collapse_facet_ords: Vec<u64>,
+
+    facets: BTreeSet<Facet>,
 }
 
 fn skip<'a, I: Iterator<Item = &'a Facet>>(
@@ -197,7 +215,7 @@ fn skip<'a, I: Iterator<Item = &'a Facet>>(
 ) -> SkipResult {
     loop {
         match collapse_it.peek() {
-            Some(facet_bytes) => match facet_bytes.encoded_str().as_bytes().cmp(target) {
+            Some(facet_bytes) => match facet_bytes.encoded_bytes().cmp(target) {
                 Ordering::Less => {}
                 Ordering::Greater => {
                     return SkipResult::OverStep;
@@ -222,8 +240,15 @@ impl FacetCollector {
     /// is of the proper type.
     pub fn for_field(field: Field) -> FacetCollector {
         FacetCollector {
+            facet_ords: Vec::with_capacity(255),
+            segment_counters: Vec::new(),
             field,
-            facets: BTreeSet::default(),
+            ff_reader: None,
+            facets: BTreeSet::new(),
+
+            current_segment_collapse_mapping: Vec::new(),
+            current_collapse_facet_ords: Vec::new(),
+            current_segment_counts: Vec::new(),
         }
     }
 
@@ -253,100 +278,141 @@ impl FacetCollector {
         }
         self.facets.insert(facet);
     }
-}
-
-impl Collector for FacetCollector {
-    type Fruit = FacetCounts;
-
-    type Child = FacetSegmentCollector;
-
-    fn for_segment(
-        &self,
-        _: SegmentLocalId,
-        reader: &SegmentReader,
-    ) -> Result<FacetSegmentCollector> {
-        let facet_reader = reader.facet_reader(self.field)?;
-
-        let mut collapse_mapping = Vec::new();
-        let mut counts = Vec::new();
-        let mut collapse_facet_ords = Vec::new();
 
+    fn set_collapse_mapping(&mut self, facet_reader: &FacetReader) {
+        self.current_segment_collapse_mapping.clear();
+        self.current_collapse_facet_ords.clear();
+        self.current_segment_counts.clear();
         let mut collapse_facet_it = self.facets.iter().peekable();
-        collapse_facet_ords.push(0);
-        {
-            let mut facet_streamer = facet_reader.facet_dict().range().into_stream();
-            if facet_streamer.advance() {
-                'outer: loop {
-                    // at the begining of this loop, facet_streamer
-                    // is positionned on a term that has not been processed yet.
-                    let skip_result = skip(facet_streamer.key(), &mut collapse_facet_it);
-                    match skip_result {
-                        SkipResult::Reached => {
-                            // we reach a facet we decided to collapse.
-                            let collapse_depth = facet_depth(facet_streamer.key());
-                            let mut collapsed_id = 0;
-                            collapse_mapping.push(0);
-                            while facet_streamer.advance() {
-                                let depth = facet_depth(facet_streamer.key());
-                                if depth <= collapse_depth {
-                                    continue 'outer;
-                                }
-                                if depth == collapse_depth + 1 {
-                                    collapsed_id = collapse_facet_ords.len();
-                                    collapse_facet_ords.push(facet_streamer.term_ord());
-                                    collapse_mapping.push(collapsed_id);
-                                } else {
-                                    collapse_mapping.push(collapsed_id);
-                                }
-                            }
-                            break;
+        self.current_collapse_facet_ords.push(0);
+        let mut facet_streamer = facet_reader.facet_dict().range().into_stream();
+        if !facet_streamer.advance() {
+            return;
+        }
+        'outer: loop {
+            // at the begining of this loop, facet_streamer
+            // is positionned on a term that has not been processed yet.
+            let skip_result = skip(facet_streamer.key(), &mut collapse_facet_it);
+            match skip_result {
+                SkipResult::Reached => {
+                    // we reach a facet we decided to collapse.
+                    let collapse_depth = facet_depth(facet_streamer.key());
+                    let mut collapsed_id = 0;
+                    self.current_segment_collapse_mapping.push(0);
+                    while facet_streamer.advance() {
+                        let depth = facet_depth(facet_streamer.key());
+                        if depth <= collapse_depth {
+                            continue 'outer;
                         }
-                        SkipResult::End | SkipResult::OverStep => {
-                            collapse_mapping.push(0);
-                            if !facet_streamer.advance() {
-                                break;
-                            }
+                        if depth == collapse_depth + 1 {
+                            collapsed_id = self.current_collapse_facet_ords.len();
+                            self.current_collapse_facet_ords
+                                .push(facet_streamer.term_ord());
+                            self.current_segment_collapse_mapping.push(collapsed_id);
+                        } else {
+                            self.current_segment_collapse_mapping.push(collapsed_id);
                         }
                     }
+                    break;
+                }
+                SkipResult::End | SkipResult::OverStep => {
+                    self.current_segment_collapse_mapping.push(0);
+                    if !facet_streamer.advance() {
+                        break;
+                    }
                 }
             }
         }
-
-        counts.resize(collapse_facet_ords.len(), 0);
-
-        Ok(FacetSegmentCollector {
-            reader: facet_reader,
-            facet_ords_buf: Vec::with_capacity(255),
-            collapse_mapping,
-            counts,
-            collapse_facet_ords,
-        })
     }
 
-    fn requires_scoring(&self) -> bool {
-        false
+    fn finalize_segment(&mut self) {
+        if self.ff_reader.is_some() {
+            self.segment_counters.push(SegmentFacetCounter {
+                facet_reader: self.ff_reader.take().unwrap().into_inner(),
+                facet_ords: mem::replace(&mut self.current_collapse_facet_ords, Vec::new()),
+                facet_counts: mem::replace(&mut self.current_segment_counts, Vec::new()),
+            });
+        }
     }
 
-    fn merge_fruits(&self, segments_facet_counts: Vec<FacetCounts>) -> Result<FacetCounts> {
-        let mut facet_counts: BTreeMap<Facet, u64> = BTreeMap::new();
-        for segment_facet_counts in segments_facet_counts {
-            for (facet, count) in segment_facet_counts.facet_counts {
-                *(facet_counts.entry(facet).or_insert(0)) += count;
+    /// Returns the results of the collection.
+    ///
+    /// This method does not just return the counters,
+    /// it also translates the facet ordinals of the last segment.
+    pub fn harvest(mut self) -> FacetCounts {
+        self.finalize_segment();
+
+        let collapsed_facet_ords: Vec<&[u64]> = self.segment_counters
+            .iter()
+            .map(|segment_counter| &segment_counter.facet_ords[..])
+            .collect();
+        let collapsed_facet_counts: Vec<&[u64]> = self.segment_counters
+            .iter()
+            .map(|segment_counter| &segment_counter.facet_counts[..])
+            .collect();
+
+        let facet_streams = self.segment_counters
+            .iter()
+            .map(|seg_counts| seg_counts.facet_reader.facet_dict().range().into_stream())
+            .collect::<Vec<_>>();
+
+        let mut facet_merger = TermMerger::new(facet_streams);
+        let mut facet_counts = BTreeMap::new();
+
+        while facet_merger.advance() {
+            let count = facet_merger
+                .current_kvs()
+                .iter()
+                .map(|it| {
+                    let seg_ord = it.segment_ord;
+                    let term_ord = it.streamer.term_ord();
+                    collapsed_facet_ords[seg_ord]
+                        .binary_search(&term_ord)
+                        .map(|collapsed_term_id| {
+                            if collapsed_term_id == 0 {
+                                0
+                            } else {
+                                collapsed_facet_counts[seg_ord][collapsed_term_id]
+                            }
+                        })
+                        .unwrap_or(0)
+                })
+                .sum();
+            if count > 0u64 {
+                let bytes: Vec<u8> = facet_merger.key().to_owned();
+                // may create an corrupted facet if the term dicitonary is corrupted
+                let facet = unsafe { Facet::from_encoded(bytes) };
+                facet_counts.insert(facet, count);
             }
         }
-        Ok(FacetCounts { facet_counts })
+        FacetCounts { facet_counts }
     }
 }
 
-impl SegmentCollector for FacetSegmentCollector {
-    type Fruit = FacetCounts;
+impl Collector for FacetCollector {
+    fn set_segment(&mut self, _: SegmentLocalId, reader: &SegmentReader) -> Result<()> {
+        self.finalize_segment();
+        let facet_reader = reader.facet_reader(self.field)?;
+        self.set_collapse_mapping(&facet_reader);
+        self.current_segment_counts
+            .resize(self.current_collapse_facet_ords.len(), 0);
+        self.ff_reader = Some(UnsafeCell::new(facet_reader));
+        Ok(())
+    }
 
     fn collect(&mut self, doc: DocId, _: Score) {
-        self.reader.facet_ords(doc, &mut self.facet_ords_buf);
+        let facet_reader: &mut FacetReader = unsafe {
+            &mut *self.ff_reader
+                .as_ref()
+                .expect("collect() was called before set_segment. This should never happen.")
+                .get()
+        };
+        facet_reader.facet_ords(doc, &mut self.facet_ords);
         let mut previous_collapsed_ord: usize = usize::MAX;
-        for &facet_ord in &self.facet_ords_buf {
-            let collapsed_ord = self.collapse_mapping[facet_ord as usize];
-            self.counts[collapsed_ord] += if collapsed_ord == previous_collapsed_ord {
+        for &facet_ord in &self.facet_ords {
+            let collapsed_ord = self.current_segment_collapse_mapping[facet_ord as usize];
+            self.current_segment_counts[collapsed_ord] += if collapsed_ord == previous_collapsed_ord
+            {
                 0
             } else {
                 1
@@ -355,24 +421,8 @@ impl SegmentCollector for FacetSegmentCollector {
         }
     }
 
-    /// Returns the results of the collection.
-    ///
-    /// This method does not just return the counters,
-    /// it also translates the facet ordinals of the last segment.
-    fn harvest(self) -> FacetCounts {
-        let mut facet_counts = BTreeMap::new();
-        let facet_dict = self.reader.facet_dict();
-        for (collapsed_facet_ord, count) in self.counts.iter().cloned().enumerate() {
-            if count == 0 {
-                continue;
-            }
-            let mut facet = vec![];
-            let facet_ord = self.collapse_facet_ords[collapsed_facet_ord];
-            facet_dict.ord_to_term(facet_ord as u64, &mut facet);
-            // TODO
-            facet_counts.insert(Facet::from_encoded(facet).unwrap(), count);
-        }
-        FacetCounts { facet_counts }
+    fn requires_scoring(&self) -> bool {
+        false
     }
 }
 
@@ -404,9 +454,9 @@ impl FacetCounts {
         let right_bound = if facet.is_root() {
             Bound::Unbounded
         } else {
-            let mut facet_after_bytes: String = facet.encoded_str().to_owned();
-            facet_after_bytes.push('\u{1}');
-            let facet_after = Facet::from_encoded_string(facet_after_bytes);
+            let mut facet_after_bytes: Vec<u8> = facet.encoded_bytes().to_owned();
+            facet_after_bytes.push(1u8);
+            let facet_after = unsafe { Facet::from_encoded(facet_after_bytes) }; // ok logic
             Bound::Excluded(facet_after)
         };
         let underlying: btree_map::Range<_, _> = self.facet_counts.range((left_bound, right_bound));
@@ -426,8 +476,9 @@ impl FacetCounts {
             heap.push(Hit { count, facet });
         }
 
-        let mut lowest_count: u64 = heap.peek().map(|hit| hit.count).unwrap_or(u64::MIN); //< the `unwrap_or` case may be triggered but the value
-                                                                                          // is never used in that case.
+        let mut lowest_count: u64 = heap.peek().map(|hit| hit.count)
+            .unwrap_or(u64::MIN); //< the `unwrap_or` case may be triggered but the value
+                                  // is never used in that case.
 
         for (facet, count) in it {
             if count > lowest_count {
@@ -453,14 +504,14 @@ mod tests {
     use core::Index;
     use query::AllQuery;
     use rand::distributions::Uniform;
-    use rand::prelude::SliceRandom;
     use rand::{thread_rng, Rng};
-    use schema::{Document, Facet, Field, Schema};
+    use schema::Field;
+    use schema::{Document, Facet, SchemaBuilder};
     use std::iter;
 
     #[test]
     fn test_facet_collector_drilldown() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let facet_field = schema_builder.add_facet_field("facet");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -485,10 +536,12 @@ mod tests {
         index_writer.commit().unwrap();
         index.load_searchers().unwrap();
         let searcher = index.searcher();
+
         let mut facet_collector = FacetCollector::for_field(facet_field);
         facet_collector.add_facet(Facet::from("/top1"));
-        let counts = searcher.search(&AllQuery, &facet_collector).unwrap();
+        searcher.search(&AllQuery, &mut facet_collector).unwrap();
 
+        let counts: FacetCounts = facet_collector.harvest();
         {
             let facets: Vec<(String, u64)> = counts
                 .get("/top1")
@@ -501,17 +554,18 @@ mod tests {
                     ("/top1/mid1", 50),
                     ("/top1/mid2", 50),
                     ("/top1/mid3", 50),
-                ]
-                .iter()
-                .map(|&(facet_str, count)| (String::from(facet_str), count))
-                .collect::<Vec<_>>()
+                ].iter()
+                    .map(|&(facet_str, count)| (String::from(facet_str), count))
+                    .collect::<Vec<_>>()
             );
         }
     }
 
     #[test]
-    #[should_panic(expected = "Tried to add a facet which is a descendant of \
-                               an already added facet.")]
+    #[should_panic(
+        expected = "Tried to add a facet which is a descendant of \
+                    an already added facet."
+    )]
     fn test_misused_facet_collector() {
         let mut facet_collector = FacetCollector::for_field(Field(0));
         facet_collector.add_facet(Facet::from("/country"));
@@ -520,7 +574,7 @@ mod tests {
 
     #[test]
     fn test_doc_unsorted_multifacet() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let facet_field = schema_builder.add_facet_field("facets");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -537,7 +591,8 @@ mod tests {
         assert_eq!(searcher.num_docs(), 1);
         let mut facet_collector = FacetCollector::for_field(facet_field);
         facet_collector.add_facet("/subjects");
-        let counts = searcher.search(&AllQuery, &facet_collector).unwrap();
+        searcher.search(&AllQuery, &mut facet_collector).unwrap();
+        let counts = facet_collector.harvest();
         let facets: Vec<(&Facet, u64)> = counts.get("/subjects").collect();
         assert_eq!(facets[0].1, 1);
     }
@@ -551,7 +606,7 @@ mod tests {
 
     #[test]
     fn test_facet_collector_topk() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let facet_field = schema_builder.add_facet_field("facet");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -564,15 +619,9 @@ mod tests {
                 let doc = doc!(facet_field => facet);
                 iter::repeat(doc).take(count)
             })
-            .map(|mut doc| {
-                doc.add_facet(
-                    facet_field,
-                    &format!("/facet/{}", thread_rng().sample(&uniform)),
-                );
-                doc
-            })
+            .map(|mut doc| { doc.add_facet(facet_field, &format!("/facet/{}", thread_rng().sample(&uniform) )); doc})
             .collect();
-        docs[..].shuffle(&mut thread_rng());
+        thread_rng().shuffle(&mut docs[..]);
 
         let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
         for doc in docs {
@@ -585,8 +634,9 @@ mod tests {
 
         let mut facet_collector = FacetCollector::for_field(facet_field);
         facet_collector.add_facet("/facet");
-        let counts: FacetCounts = searcher.search(&AllQuery, &facet_collector).unwrap();
+        searcher.search(&AllQuery, &mut facet_collector).unwrap();
 
+        let counts: FacetCounts = facet_collector.harvest();
         {
             let facets: Vec<(&Facet, u64)> = counts.top_k("/facet", 3);
             assert_eq!(
@@ -609,13 +659,13 @@ mod bench {
     use query::AllQuery;
     use rand::{thread_rng, Rng};
     use schema::Facet;
-    use schema::Schema;
+    use schema::SchemaBuilder;
     use test::Bencher;
     use Index;
 
     #[bench]
     fn bench_facet_collector(b: &mut Bencher) {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let facet_field = schema_builder.add_facet_field("facet");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -639,8 +689,8 @@ mod bench {
 
         b.iter(|| {
             let searcher = index.searcher();
-            let facet_collector = FacetCollector::for_field(facet_field);
-            searcher.search(&AllQuery, &facet_collector).unwrap();
+            let mut facet_collector = FacetCollector::for_field(facet_field);
+            searcher.search(&AllQuery, &mut facet_collector).unwrap();
         });
     }
 }

src/collector/int_facet_collector.rs

@@ -79,7 +79,7 @@ mod tests {
     // make sure we have facet counters correctly filled
     fn test_facet_collector_results() {
 
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::new();
         let num_field_i64 = schema_builder.add_i64_field("num_i64", FAST);
         let num_field_u64 = schema_builder.add_u64_field("num_u64", FAST);
         let text_field = schema_builder.add_text_field("text", STRING);
@@ -88,7 +88,7 @@ mod tests {
         let index = Index::create_in_ram(schema.clone());
 
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             {
                 for i in 0u64..10u64 {
                     index_writer.add_document(doc!(

src/collector/mod.rs

@@ -1,91 +1,7 @@
 /*!
-
-# Collectors
-
-Collectors define the information you want to extract from the documents matching the queries.
-In tantivy jargon, we call this information your search "fruit".
-
-Your fruit could for instance be :
-- [the count of matching documents](./struct.Count.html)
-- [the top 10 documents, by relevancy or by a fast field](./struct.TopDocs.html)
-- [facet counts](./struct.FacetCollector.html)
-
-At one point in your code, you will trigger the actual search operation by calling
-[the `search(...)` method of your `Searcher` object](../struct.Searcher.html#method.search).
-This call will look like this.
-
-```verbatim
-let fruit = searcher.search(&query, &collector)?;
-```
-
-Here the type of fruit is actually determined as an associated type of the collector (`Collector::Fruit`).
-
-
-# Combining several collectors
-
-A rich search experience often requires to run several collectors on your search query.
-For instance,
-- selecting the top-K products matching your query
-- counting the matching documents
-- computing several facets
-- computing statistics about the matching product prices
-
-A simple and efficient way to do that is to pass your collectors as one tuple.
-The resulting `Fruit` will then be a typed tuple with each collector's original fruits
-in their respective position.
-
-```rust
-# extern crate tantivy;
-# use tantivy::schema::*;
-# use tantivy::*;
-# use tantivy::query::*;
-use tantivy::collector::{Count, TopDocs};
-#
-# fn main() -> tantivy::Result<()> {
-# let mut schema_builder = Schema::builder();
-#     let title = schema_builder.add_text_field("title", TEXT);
-#     let schema = schema_builder.build();
-#     let index = Index::create_in_ram(schema);
-#     let mut index_writer = index.writer(3_000_000)?;
-#       index_writer.add_document(doc!(
-#       title => "The Name of the Wind",
-#      ));
-#     index_writer.add_document(doc!(
-#        title => "The Diary of Muadib",
-#     ));
-#     index_writer.commit().unwrap();
-#     index.load_searchers()?;
-#     let searcher = index.searcher();
-#     let query_parser = QueryParser::for_index(&index, vec![title]);
-#     let query = query_parser.parse_query("diary")?;
-let (doc_count, top_docs): (usize, Vec<(Score, DocAddress)>) =
-    searcher.search(&query, &(Count, TopDocs::with_limit(2)))?;
-#     Ok(())
-# }
-```
-
-The `Collector` trait is implemented for up to 4 collectors.
-If you have more than 4 collectors, you can either group them into
-tuples of tuples `(a,(b,(c,d)))`, or rely on `MultiCollector`'s.
-
-# Combining several collectors dynamically
-
-Combining collectors into a tuple is a zero-cost abstraction: everything
-happens as if you had manually implemented a single collector
-combining all of our features.
-
-Unfortunately it requires you to know at compile time your collector types.
-If on the other hand, the collectors depend on some query parameter,
-you can rely on `MultiCollector`'s.
-
-
-# Implementing your own collectors.
-
-See the `custom_collector` example.
-
+Defines how the documents matching a search query should be processed.
 */
 
-use downcast_rs;
 use DocId;
 use Result;
 use Score;
@@ -93,272 +9,238 @@ use SegmentLocalId;
 use SegmentReader;
 
 mod count_collector;
-pub use self::count_collector::Count;
+pub use self::count_collector::CountCollector;
 
 mod multi_collector;
 pub use self::multi_collector::MultiCollector;
 
 mod top_collector;
-
-mod top_score_collector;
-pub use self::top_score_collector::TopDocs;
-
-mod top_field_collector;
-pub use self::top_field_collector::TopDocsByField;
+pub use self::top_collector::TopCollector;
 
 mod facet_collector;
 pub use self::facet_collector::FacetCollector;
 
-/// `Fruit` is the type for the result of our collection.
-/// e.g. `usize` for the `Count` collector.
-pub trait Fruit: Send + downcast_rs::Downcast {}
-
-impl<T> Fruit for T where T: Send + downcast_rs::Downcast {}
+mod chained_collector;
+pub use self::chained_collector::{chain, ChainedCollector};
 
 /// Collectors are in charge of collecting and retaining relevant
 /// information from the document found and scored by the query.
 ///
+///
 /// For instance,
 ///
 /// - keeping track of the top 10 best documents
 /// - computing a breakdown over a fast field
 /// - computing the number of documents matching the query
 ///
-/// Our search index is in fact a collection of segments, so
-/// a `Collector` trait is actually more of a factory to instance
-/// `SegmentCollector`s for each segments.
+/// Queries are in charge of pushing the `DocSet` to the collector.
 ///
-/// The collection logic itself is in the `SegmentCollector`.
+/// As they work on multiple segments, they first inform
+/// the collector of a change in a segment and then
+/// call the `collect` method to push the document to the collector.
+///
+/// Temporally, our collector will receive calls
+/// - `.set_segment(0, segment_reader_0)`
+/// - `.collect(doc0_of_segment_0)`
+/// - `.collect(...)`
+/// - `.collect(last_doc_of_segment_0)`
+/// - `.set_segment(1, segment_reader_1)`
+/// - `.collect(doc0_of_segment_1)`
+/// - `.collect(...)`
+/// - `.collect(last_doc_of_segment_1)`
+/// - `...`
+/// - `.collect(last_doc_of_last_segment)`
 ///
 /// Segments are not guaranteed to be visited in any specific order.
-pub trait Collector: Sync {
-    /// `Fruit` is the type for the result of our collection.
-    /// e.g. `usize` for the `Count` collector.
-    type Fruit: Fruit;
-
-    /// Type of the `SegmentCollector` associated to this collector.
-    type Child: SegmentCollector<Fruit = Self::Fruit>;
-
+pub trait Collector {
     /// `set_segment` is called before beginning to enumerate
     /// on this segment.
-    fn for_segment(
-        &self,
+    fn set_segment(
+        &mut self,
         segment_local_id: SegmentLocalId,
         segment: &SegmentReader,
-    ) -> Result<Self::Child>;
-
-    /// Returns true iff the collector requires to compute scores for documents.
-    fn requires_scoring(&self) -> bool;
-
-    /// Combines the fruit associated to the collection of each segments
-    /// into one fruit.
-    fn merge_fruits(&self, segment_fruits: Vec<Self::Fruit>) -> Result<Self::Fruit>;
-}
-
-/// The `SegmentCollector` is the trait in charge of defining the
-/// collect operation at the scale of the segment.
-///
-/// `.collect(doc, score)` will be called for every documents
-/// matching the query.
-pub trait SegmentCollector: 'static {
-    /// `Fruit` is the type for the result of our collection.
-    /// e.g. `usize` for the `Count` collector.
-    type Fruit: Fruit;
-
+    ) -> Result<()>;
     /// The query pushes the scored document to the collector via this method.
     fn collect(&mut self, doc: DocId, score: Score);
 
-    /// Extract the fruit of the collection from the `SegmentCollector`.
-    fn harvest(self) -> Self::Fruit;
+    /// Returns true iff the collector requires to compute scores for documents.
+    fn requires_scoring(&self) -> bool;
 }
 
-// -----------------------------------------------
-// Tuple implementations.
-
-impl<Left, Right> Collector for (Left, Right)
-where
-    Left: Collector,
-    Right: Collector,
-{
-    type Fruit = (Left::Fruit, Right::Fruit);
-    type Child = (Left::Child, Right::Child);
-
-    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
-        let left = self.0.for_segment(segment_local_id, segment)?;
-        let right = self.1.for_segment(segment_local_id, segment)?;
-        Ok((left, right))
+impl<'a, C: Collector> Collector for &'a mut C {
+    fn set_segment(
+        &mut self,
+        segment_local_id: SegmentLocalId,
+        segment: &SegmentReader,
+    ) -> Result<()> {
+        (*self).set_segment(segment_local_id, segment)
+    }
+    /// The query pushes the scored document to the collector via this method.
+    fn collect(&mut self, doc: DocId, score: Score) {
+        C::collect(self, doc, score)
     }
 
     fn requires_scoring(&self) -> bool {
-        self.0.requires_scoring() || self.1.requires_scoring()
-    }
-
-    fn merge_fruits(
-        &self,
-        children: Vec<(Left::Fruit, Right::Fruit)>,
-    ) -> Result<(Left::Fruit, Right::Fruit)> {
-        let mut left_fruits = vec![];
-        let mut right_fruits = vec![];
-        for (left_fruit, right_fruit) in children {
-            left_fruits.push(left_fruit);
-            right_fruits.push(right_fruit);
-        }
-        Ok((
-            self.0.merge_fruits(left_fruits)?,
-            self.1.merge_fruits(right_fruits)?,
-        ))
+        C::requires_scoring(self)
     }
 }
 
-impl<Left, Right> SegmentCollector for (Left, Right)
-where
-    Left: SegmentCollector,
-    Right: SegmentCollector,
-{
-    type Fruit = (Left::Fruit, Right::Fruit);
-
-    fn collect(&mut self, doc: DocId, score: Score) {
-        self.0.collect(doc, score);
-        self.1.collect(doc, score);
-    }
-
-    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
-        (self.0.harvest(), self.1.harvest())
-    }
-}
-
-// 3-Tuple
-
-impl<One, Two, Three> Collector for (One, Two, Three)
-where
-    One: Collector,
-    Two: Collector,
-    Three: Collector,
-{
-    type Fruit = (One::Fruit, Two::Fruit, Three::Fruit);
-    type Child = (One::Child, Two::Child, Three::Child);
-
-    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
-        let one = self.0.for_segment(segment_local_id, segment)?;
-        let two = self.1.for_segment(segment_local_id, segment)?;
-        let three = self.2.for_segment(segment_local_id, segment)?;
-        Ok((one, two, three))
-    }
-
-    fn requires_scoring(&self) -> bool {
-        self.0.requires_scoring() || self.1.requires_scoring() || self.2.requires_scoring()
-    }
-
-    fn merge_fruits(&self, children: Vec<Self::Fruit>) -> Result<Self::Fruit> {
-        let mut one_fruits = vec![];
-        let mut two_fruits = vec![];
-        let mut three_fruits = vec![];
-        for (one_fruit, two_fruit, three_fruit) in children {
-            one_fruits.push(one_fruit);
-            two_fruits.push(two_fruit);
-            three_fruits.push(three_fruit);
-        }
-        Ok((
-            self.0.merge_fruits(one_fruits)?,
-            self.1.merge_fruits(two_fruits)?,
-            self.2.merge_fruits(three_fruits)?,
-        ))
-    }
-}
-
-impl<One, Two, Three> SegmentCollector for (One, Two, Three)
-where
-    One: SegmentCollector,
-    Two: SegmentCollector,
-    Three: SegmentCollector,
-{
-    type Fruit = (One::Fruit, Two::Fruit, Three::Fruit);
-
-    fn collect(&mut self, doc: DocId, score: Score) {
-        self.0.collect(doc, score);
-        self.1.collect(doc, score);
-        self.2.collect(doc, score);
-    }
-
-    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
-        (self.0.harvest(), self.1.harvest(), self.2.harvest())
-    }
-}
-
-// 4-Tuple
-
-impl<One, Two, Three, Four> Collector for (One, Two, Three, Four)
-where
-    One: Collector,
-    Two: Collector,
-    Three: Collector,
-    Four: Collector,
-{
-    type Fruit = (One::Fruit, Two::Fruit, Three::Fruit, Four::Fruit);
-    type Child = (One::Child, Two::Child, Three::Child, Four::Child);
-
-    fn for_segment(&self, segment_local_id: u32, segment: &SegmentReader) -> Result<Self::Child> {
-        let one = self.0.for_segment(segment_local_id, segment)?;
-        let two = self.1.for_segment(segment_local_id, segment)?;
-        let three = self.2.for_segment(segment_local_id, segment)?;
-        let four = self.3.for_segment(segment_local_id, segment)?;
-        Ok((one, two, three, four))
-    }
-
-    fn requires_scoring(&self) -> bool {
-        self.0.requires_scoring()
-            || self.1.requires_scoring()
-            || self.2.requires_scoring()
-            || self.3.requires_scoring()
-    }
-
-    fn merge_fruits(&self, children: Vec<Self::Fruit>) -> Result<Self::Fruit> {
-        let mut one_fruits = vec![];
-        let mut two_fruits = vec![];
-        let mut three_fruits = vec![];
-        let mut four_fruits = vec![];
-        for (one_fruit, two_fruit, three_fruit, four_fruit) in children {
-            one_fruits.push(one_fruit);
-            two_fruits.push(two_fruit);
-            three_fruits.push(three_fruit);
-            four_fruits.push(four_fruit);
-        }
-        Ok((
-            self.0.merge_fruits(one_fruits)?,
-            self.1.merge_fruits(two_fruits)?,
-            self.2.merge_fruits(three_fruits)?,
-            self.3.merge_fruits(four_fruits)?,
-        ))
-    }
-}
-
-impl<One, Two, Three, Four> SegmentCollector for (One, Two, Three, Four)
-where
-    One: SegmentCollector,
-    Two: SegmentCollector,
-    Three: SegmentCollector,
-    Four: SegmentCollector,
-{
-    type Fruit = (One::Fruit, Two::Fruit, Three::Fruit, Four::Fruit);
-
-    fn collect(&mut self, doc: DocId, score: Score) {
-        self.0.collect(doc, score);
-        self.1.collect(doc, score);
-        self.2.collect(doc, score);
-        self.3.collect(doc, score);
-    }
-
-    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
-        (
-            self.0.harvest(),
-            self.1.harvest(),
-            self.2.harvest(),
-            self.3.harvest(),
-        )
-    }
-}
-
-impl_downcast!(Fruit);
-
 #[cfg(test)]
-pub mod tests;
+pub mod tests {
+
+    use super::*;
+    use core::SegmentReader;
+    use fastfield::BytesFastFieldReader;
+    use fastfield::FastFieldReader;
+    use schema::Field;
+    use DocId;
+    use Score;
+    use SegmentLocalId;
+
+    /// Stores all of the doc ids.
+    /// This collector is only used for tests.
+    /// It is unusable in practise, as it does not store
+    /// the segment ordinals
+    pub struct TestCollector {
+        offset: DocId,
+        segment_max_doc: DocId,
+        docs: Vec<DocId>,
+        scores: Vec<Score>,
+    }
+
+    impl TestCollector {
+        /// Return the exhalist of documents.
+        pub fn docs(self) -> Vec<DocId> {
+            self.docs
+        }
+
+        pub fn scores(self) -> Vec<Score> {
+            self.scores
+        }
+    }
+
+    impl Default for TestCollector {
+        fn default() -> TestCollector {
+            TestCollector {
+                offset: 0,
+                segment_max_doc: 0,
+                docs: Vec::new(),
+                scores: Vec::new(),
+            }
+        }
+    }
+
+    impl Collector for TestCollector {
+        fn set_segment(&mut self, _: SegmentLocalId, reader: &SegmentReader) -> Result<()> {
+            self.offset += self.segment_max_doc;
+            self.segment_max_doc = reader.max_doc();
+            Ok(())
+        }
+
+        fn collect(&mut self, doc: DocId, score: Score) {
+            self.docs.push(doc + self.offset);
+            self.scores.push(score);
+        }
+
+        fn requires_scoring(&self) -> bool {
+            true
+        }
+    }
+
+    /// Collects in order all of the fast fields for all of the
+    /// doc in the `DocSet`
+    ///
+    /// This collector is mainly useful for tests.
+    pub struct FastFieldTestCollector {
+        vals: Vec<u64>,
+        field: Field,
+        ff_reader: Option<FastFieldReader<u64>>,
+    }
+
+    impl FastFieldTestCollector {
+        pub fn for_field(field: Field) -> FastFieldTestCollector {
+            FastFieldTestCollector {
+                vals: Vec::new(),
+                field,
+                ff_reader: None,
+            }
+        }
+
+        pub fn vals(self) -> Vec<u64> {
+            self.vals
+        }
+    }
+
+    impl Collector for FastFieldTestCollector {
+        fn set_segment(&mut self, _: SegmentLocalId, reader: &SegmentReader) -> Result<()> {
+            self.ff_reader = Some(reader.fast_field_reader(self.field)?);
+            Ok(())
+        }
+
+        fn collect(&mut self, doc: DocId, _score: Score) {
+            let val = self.ff_reader.as_ref().unwrap().get(doc);
+            self.vals.push(val);
+        }
+        fn requires_scoring(&self) -> bool {
+            false
+        }
+    }
+
+    /// Collects in order all of the fast field bytes for all of the
+    /// docs in the `DocSet`
+    ///
+    /// This collector is mainly useful for tests.
+    pub struct BytesFastFieldTestCollector {
+        vals: Vec<u8>,
+        field: Field,
+        ff_reader: Option<BytesFastFieldReader>,
+    }
+
+    impl BytesFastFieldTestCollector {
+        pub fn for_field(field: Field) -> BytesFastFieldTestCollector {
+            BytesFastFieldTestCollector {
+                vals: Vec::new(),
+                field,
+                ff_reader: None,
+            }
+        }
+
+        pub fn vals(self) -> Vec<u8> {
+            self.vals
+        }
+    }
+
+    impl Collector for BytesFastFieldTestCollector {
+        fn set_segment(&mut self, _segment_local_id: u32, segment: &SegmentReader) -> Result<()> {
+            self.ff_reader = Some(segment.bytes_fast_field_reader(self.field)?);
+            Ok(())
+        }
+
+        fn collect(&mut self, doc: u32, _score: f32) {
+            let val = self.ff_reader.as_ref().unwrap().get_val(doc);
+            self.vals.extend(val);
+        }
+
+        fn requires_scoring(&self) -> bool {
+            false
+        }
+    }
+}
+
+#[cfg(all(test, feature = "unstable"))]
+mod bench {
+    use collector::{Collector, CountCollector};
+    use test::Bencher;
+
+    #[bench]
+    fn build_collector(b: &mut Bencher) {
+        b.iter(|| {
+            let mut count_collector = CountCollector::default();
+            let docs: Vec<u32> = (0..1_000_000).collect();
+            for doc in docs {
+                count_collector.collect(doc, 1f32);
+            }
+            count_collector.count()
+        });
+    }
+}

src/collector/multi_collector.rs

@@ -1,115 +1,26 @@
 use super::Collector;
-use super::SegmentCollector;
-use collector::Fruit;
-use std::marker::PhantomData;
 use DocId;
 use Result;
 use Score;
 use SegmentLocalId;
 use SegmentReader;
-use TantivyError;
-
-pub struct MultiFruit {
-    sub_fruits: Vec<Option<Box<Fruit>>>,
-}
-
-pub struct CollectorWrapper<TCollector: Collector>(TCollector);
-
-impl<TCollector: Collector> Collector for CollectorWrapper<TCollector> {
-    type Fruit = Box<Fruit>;
-    type Child = Box<BoxableSegmentCollector>;
-
-    fn for_segment(
-        &self,
-        segment_local_id: u32,
-        reader: &SegmentReader,
-    ) -> Result<Box<BoxableSegmentCollector>> {
-        let child = self.0.for_segment(segment_local_id, reader)?;
-        Ok(Box::new(SegmentCollectorWrapper(child)))
-    }
-
-    fn requires_scoring(&self) -> bool {
-        self.0.requires_scoring()
-    }
-
-    fn merge_fruits(&self, children: Vec<<Self as Collector>::Fruit>) -> Result<Box<Fruit>> {
-        let typed_fruit: Vec<TCollector::Fruit> = children
-            .into_iter()
-            .map(|untyped_fruit| {
-                untyped_fruit.downcast::<TCollector::Fruit>()
-                    .map(|boxed_but_typed| *boxed_but_typed)
-                    .map_err(|_| {
-                        TantivyError::InvalidArgument("Failed to cast child fruit.".to_string())
-                    })
-            })
-            .collect::<Result<_>>()?;
-        let merged_fruit = self.0.merge_fruits(typed_fruit)?;
-        Ok(Box::new(merged_fruit))
-    }
-}
-
-impl SegmentCollector for Box<BoxableSegmentCollector> {
-    type Fruit = Box<Fruit>;
-
-    fn collect(&mut self, doc: u32, score: f32) {
-        self.as_mut().collect(doc, score);
-    }
-
-    fn harvest(self) -> Box<Fruit> {
-        BoxableSegmentCollector::harvest_from_box(self)
-    }
-}
-
-pub trait BoxableSegmentCollector {
-    fn collect(&mut self, doc: u32, score: f32);
-    fn harvest_from_box(self: Box<Self>) -> Box<Fruit>;
-}
-
-pub struct SegmentCollectorWrapper<TSegmentCollector: SegmentCollector>(TSegmentCollector);
-
-impl<TSegmentCollector: SegmentCollector> BoxableSegmentCollector
-    for SegmentCollectorWrapper<TSegmentCollector>
-{
-    fn collect(&mut self, doc: u32, score: f32) {
-        self.0.collect(doc, score);
-    }
-
-    fn harvest_from_box(self: Box<Self>) -> Box<Fruit> {
-        Box::new(self.0.harvest())
-    }
-}
-
-pub struct FruitHandle<TFruit: Fruit> {
-    pos: usize,
-    _phantom: PhantomData<TFruit>,
-}
-
-impl<TFruit: Fruit> FruitHandle<TFruit> {
-    pub fn extract(self, fruits: &mut MultiFruit) -> TFruit {
-        let boxed_fruit = fruits.sub_fruits[self.pos].take().expect("");
-        *boxed_fruit.downcast::<TFruit>().map_err(|_| ()).expect("Failed to downcast collector fruit.")
-    }
-}
 
 /// Multicollector makes it possible to collect on more than one collector.
 /// It should only be used for use cases where the Collector types is unknown
 /// at compile time.
-///
-/// If the type of the collectors is known, you can just group yours collectors
-/// in a tuple. See the
-/// [Combining several collectors section of the collector documentation](./index.html#combining-several-collectors).
+/// If the type of the collectors is known, you should prefer to use `ChainedCollector`.
 ///
 /// ```rust
 /// #[macro_use]
 /// extern crate tantivy;
-/// use tantivy::schema::{Schema, TEXT};
+/// use tantivy::schema::{SchemaBuilder, TEXT};
 /// use tantivy::{Index, Result};
-/// use tantivy::collector::{Count, TopDocs, MultiCollector};
+/// use tantivy::collector::{CountCollector, TopCollector, MultiCollector};
 /// use tantivy::query::QueryParser;
 ///
 /// # fn main() { example().unwrap(); }
 /// fn example() -> Result<()> {
-///     let mut schema_builder = Schema::builder();
+///     let mut schema_builder = SchemaBuilder::new();
 ///     let title = schema_builder.add_text_field("title", TEXT);
 ///     let schema = schema_builder.build();
 ///     let index = Index::create_in_ram(schema);
@@ -133,115 +44,55 @@ impl<TFruit: Fruit> FruitHandle<TFruit> {
 ///     index.load_searchers()?;
 ///     let searcher = index.searcher();
 ///
-///     let mut collectors = MultiCollector::new();
-///     let top_docs_handle = collectors.add_collector(TopDocs::with_limit(2));
-///     let count_handle = collectors.add_collector(Count);
-///     let query_parser = QueryParser::for_index(&index, vec![title]);
-///     let query = query_parser.parse_query("diary")?;
-///     let mut multi_fruit = searcher.search(&query, &collectors)?;
-///
-///     let count = count_handle.extract(&mut multi_fruit);
-///     let top_docs = top_docs_handle.extract(&mut multi_fruit);
-///
-///     # assert_eq!(count, 2);
-///     # assert_eq!(top_docs.len(), 2);
+///     {
+///         let mut top_collector = TopCollector::with_limit(2);
+///         let mut count_collector = CountCollector::default();
+///         {
+///             let mut collectors =
+///                 MultiCollector::from(vec![&mut top_collector, &mut count_collector]);
+///             let query_parser = QueryParser::for_index(&index, vec![title]);
+///             let query = query_parser.parse_query("diary")?;
+///             searcher.search(&*query, &mut collectors).unwrap();
+///         }
+///         assert_eq!(count_collector.count(), 2);
+///         assert!(top_collector.at_capacity());
+///     }
 ///
 ///     Ok(())
 /// }
 /// ```
-#[allow(clippy::type_complexity)]
-#[derive(Default)]
 pub struct MultiCollector<'a> {
-    collector_wrappers:
-        Vec<Box<Collector<Child = Box<BoxableSegmentCollector>, Fruit = Box<Fruit>> + 'a>>,
+    collectors: Vec<&'a mut Collector>,
 }
 
 impl<'a> MultiCollector<'a> {
-    /// Create a new `MultiCollector`
-    pub fn new() -> Self {
-        Default::default()
-    }
-
-    /// Add a new collector to our `MultiCollector`.
-    pub fn add_collector<'b: 'a, TCollector: Collector + 'b>(
-        &mut self,
-        collector: TCollector,
-    ) -> FruitHandle<TCollector::Fruit> {
-        let pos = self.collector_wrappers.len();
-        self.collector_wrappers
-            .push(Box::new(CollectorWrapper(collector)));
-        FruitHandle {
-            pos,
-            _phantom: PhantomData,
-        }
+    /// Constructor
+    pub fn from(collectors: Vec<&'a mut Collector>) -> MultiCollector {
+        MultiCollector { collectors }
     }
 }
 
 impl<'a> Collector for MultiCollector<'a> {
-    type Fruit = MultiFruit;
-    type Child = MultiCollectorChild;
-
-    fn for_segment(
-        &self,
+    fn set_segment(
+        &mut self,
         segment_local_id: SegmentLocalId,
         segment: &SegmentReader,
-    ) -> Result<MultiCollectorChild> {
-        let children = self
-            .collector_wrappers
-            .iter()
-            .map(|collector_wrapper| collector_wrapper.for_segment(segment_local_id, segment))
-            .collect::<Result<Vec<_>>>()?;
-        Ok(MultiCollectorChild { children })
-    }
-
-    fn requires_scoring(&self) -> bool {
-        self.collector_wrappers.iter().any(|c| c.requires_scoring())
-    }
-
-    fn merge_fruits(&self, segments_multifruits: Vec<MultiFruit>) -> Result<MultiFruit> {
-        let mut segment_fruits_list: Vec<Vec<Box<Fruit>>> = (0..self.collector_wrappers.len())
-            .map(|_| Vec::with_capacity(segments_multifruits.len()))
-            .collect::<Vec<_>>();
-        for segment_multifruit in segments_multifruits {
-            for (idx, segment_fruit_opt) in segment_multifruit.sub_fruits.into_iter().enumerate() {
-                if let Some(segment_fruit) = segment_fruit_opt {
-                    segment_fruits_list[idx].push(segment_fruit);
-                }
-            }
+    ) -> Result<()> {
+        for collector in &mut self.collectors {
+            collector.set_segment(segment_local_id, segment)?;
         }
-        let sub_fruits = self
-            .collector_wrappers
-            .iter()
-            .zip(segment_fruits_list)
-            .map(|(child_collector, segment_fruits)| {
-                Ok(Some(child_collector.merge_fruits(segment_fruits)?))
-            })
-            .collect::<Result<_>>()?;
-        Ok(MultiFruit { sub_fruits })
+        Ok(())
     }
-}
-
-pub struct MultiCollectorChild {
-    children: Vec<Box<BoxableSegmentCollector>>,
-}
-
-impl SegmentCollector for MultiCollectorChild {
-    type Fruit = MultiFruit;
 
     fn collect(&mut self, doc: DocId, score: Score) {
-        for child in &mut self.children {
-            child.collect(doc, score);
+        for collector in &mut self.collectors {
+            collector.collect(doc, score);
         }
     }
-
-    fn harvest(self) -> MultiFruit {
-        MultiFruit {
-            sub_fruits: self
-                .children
-                .into_iter()
-                .map(|child| Some(child.harvest()))
-                .collect(),
-        }
+    fn requires_scoring(&self) -> bool {
+        self.collectors
+            .iter()
+            .any(|collector| collector.requires_scoring())
     }
 }
 
@@ -249,42 +100,20 @@ impl SegmentCollector for MultiCollectorChild {
 mod tests {
 
     use super::*;
-    use collector::{Count, TopDocs};
-    use query::TermQuery;
-    use schema::IndexRecordOption;
-    use schema::{Schema, TEXT};
-    use Index;
-    use Term;
+    use collector::{Collector, CountCollector, TopCollector};
 
     #[test]
     fn test_multi_collector() {
-        let mut schema_builder = Schema::builder();
-        let text = schema_builder.add_text_field("text", TEXT);
-        let schema = schema_builder.build();
-
-        let index = Index::create_in_ram(schema);
+        let mut top_collector = TopCollector::with_limit(2);
+        let mut count_collector = CountCollector::default();
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-            index_writer.add_document(doc!(text=>"abc"));
-            index_writer.add_document(doc!(text=>"abc abc abc"));
-            index_writer.add_document(doc!(text=>"abc abc"));
-            index_writer.commit().unwrap();
-            index_writer.add_document(doc!(text=>""));
-            index_writer.add_document(doc!(text=>"abc abc abc abc"));
-            index_writer.add_document(doc!(text=>"abc"));
-            index_writer.commit().unwrap();
+            let mut collectors =
+                MultiCollector::from(vec![&mut top_collector, &mut count_collector]);
+            collectors.collect(1, 0.2);
+            collectors.collect(2, 0.1);
+            collectors.collect(3, 0.5);
         }
-        index.load_searchers().unwrap();
-        let searcher = index.searcher();
-        let term = Term::from_field_text(text, "abc");
-        let query = TermQuery::new(term, IndexRecordOption::Basic);
-
-        let mut collectors = MultiCollector::new();
-        let topdocs_handler = collectors.add_collector(TopDocs::with_limit(2));
-        let count_handler = collectors.add_collector(Count);
-        let mut multifruits = searcher.search(&query, &mut collectors).unwrap();
-
-        assert_eq!(count_handler.extract(&mut multifruits), 5);
-        assert_eq!(topdocs_handler.extract(&mut multifruits).len(), 2);
+        assert_eq!(count_collector.count(), 3);
+        assert!(top_collector.at_capacity());
     }
 }

src/collector/tests.rs

@@ -1,201 +0,0 @@
-use super::*;
-use core::SegmentReader;
-use fastfield::BytesFastFieldReader;
-use fastfield::FastFieldReader;
-use schema::Field;
-use DocAddress;
-use DocId;
-use Score;
-use SegmentLocalId;
-
-/// Stores all of the doc ids.
-/// This collector is only used for tests.
-/// It is unusable in pr
-///
-/// actise, as it does not store
-/// the segment ordinals
-pub struct TestCollector;
-
-pub struct TestSegmentCollector {
-    segment_id: SegmentLocalId,
-    fruit: TestFruit,
-}
-
-#[derive(Default)]
-pub struct TestFruit {
-    docs: Vec<DocAddress>,
-    scores: Vec<Score>,
-}
-
-impl TestFruit {
-    /// Return the list of matching documents exhaustively.
-    pub fn docs(&self) -> &[DocAddress] {
-        &self.docs[..]
-    }
-
-    pub fn scores(&self) -> &[Score] {
-        &self.scores[..]
-    }
-}
-
-impl Collector for TestCollector {
-    type Fruit = TestFruit;
-    type Child = TestSegmentCollector;
-
-    fn for_segment(
-        &self,
-        segment_id: SegmentLocalId,
-        _reader: &SegmentReader,
-    ) -> Result<TestSegmentCollector> {
-        Ok(TestSegmentCollector {
-            segment_id,
-            fruit: TestFruit::default(),
-        })
-    }
-
-    fn requires_scoring(&self) -> bool {
-        true
-    }
-
-    fn merge_fruits(&self, mut children: Vec<TestFruit>) -> Result<TestFruit> {
-        children.sort_by_key(|fruit| {
-            if fruit.docs().is_empty() {
-                0
-            } else {
-                fruit.docs()[0].segment_ord()
-            }
-        });
-        let mut docs = vec![];
-        let mut scores = vec![];
-        for child in children {
-            docs.extend(child.docs());
-            scores.extend(child.scores);
-        }
-        Ok(TestFruit { docs, scores })
-    }
-}
-
-impl SegmentCollector for TestSegmentCollector {
-    type Fruit = TestFruit;
-
-    fn collect(&mut self, doc: DocId, score: Score) {
-        self.fruit.docs.push(DocAddress(self.segment_id, doc));
-        self.fruit.scores.push(score);
-    }
-
-    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
-        self.fruit
-    }
-}
-
-/// Collects in order all of the fast fields for all of the
-/// doc in the `DocSet`
-///
-/// This collector is mainly useful for tests.
-pub struct FastFieldTestCollector {
-    field: Field,
-}
-
-pub struct FastFieldSegmentCollector {
-    vals: Vec<u64>,
-    reader: FastFieldReader<u64>,
-}
-
-impl FastFieldTestCollector {
-    pub fn for_field(field: Field) -> FastFieldTestCollector {
-        FastFieldTestCollector { field }
-    }
-}
-
-impl Collector for FastFieldTestCollector {
-    type Fruit = Vec<u64>;
-    type Child = FastFieldSegmentCollector;
-
-    fn for_segment(
-        &self,
-        _: SegmentLocalId,
-        reader: &SegmentReader,
-    ) -> Result<FastFieldSegmentCollector> {
-        Ok(FastFieldSegmentCollector {
-            vals: Vec::new(),
-            reader: reader.fast_field_reader(self.field)?,
-        })
-    }
-
-    fn requires_scoring(&self) -> bool {
-        false
-    }
-
-    fn merge_fruits(&self, children: Vec<Vec<u64>>) -> Result<Vec<u64>> {
-        Ok(children.into_iter().flat_map(|v| v.into_iter()).collect())
-    }
-}
-
-impl SegmentCollector for FastFieldSegmentCollector {
-    type Fruit = Vec<u64>;
-
-    fn collect(&mut self, doc: DocId, _score: Score) {
-        let val = self.reader.get(doc);
-        self.vals.push(val);
-    }
-
-    fn harvest(self) -> Vec<u64> {
-        self.vals
-    }
-}
-
-/// Collects in order all of the fast field bytes for all of the
-/// docs in the `DocSet`
-///
-/// This collector is mainly useful for tests.
-pub struct BytesFastFieldTestCollector {
-    field: Field,
-}
-
-pub struct BytesFastFieldSegmentCollector {
-    vals: Vec<u8>,
-    reader: BytesFastFieldReader,
-}
-
-impl BytesFastFieldTestCollector {
-    pub fn for_field(field: Field) -> BytesFastFieldTestCollector {
-        BytesFastFieldTestCollector { field }
-    }
-}
-
-impl Collector for BytesFastFieldTestCollector {
-    type Fruit = Vec<u8>;
-    type Child = BytesFastFieldSegmentCollector;
-
-    fn for_segment(
-        &self,
-        _segment_local_id: u32,
-        segment: &SegmentReader,
-    ) -> Result<BytesFastFieldSegmentCollector> {
-        Ok(BytesFastFieldSegmentCollector {
-            vals: Vec::new(),
-            reader: segment.bytes_fast_field_reader(self.field)?,
-        })
-    }
-
-    fn requires_scoring(&self) -> bool {
-        false
-    }
-
-    fn merge_fruits(&self, children: Vec<Vec<u8>>) -> Result<Vec<u8>> {
-        Ok(children.into_iter().flat_map(|c| c.into_iter()).collect())
-    }
-}
-
-impl SegmentCollector for BytesFastFieldSegmentCollector {
-    type Fruit = Vec<u8>;
-
-    fn collect(&mut self, doc: u32, _score: f32) {
-        let data = self.reader.get_val(doc);
-        self.vals.extend(data);
-    }
-
-    fn harvest(self) -> <Self as SegmentCollector>::Fruit {
-        self.vals
-    }
-}

src/collector/top_collector.rs

@@ -1,224 +1,244 @@
-use serde::export::PhantomData;
+use super::Collector;
 use std::cmp::Ordering;
 use std::collections::BinaryHeap;
 use DocAddress;
 use DocId;
 use Result;
+use Score;
 use SegmentLocalId;
 use SegmentReader;
 
-/// Contains a feature (field, score, etc.) of a document along with the document address.
-///
-/// It has a custom implementation of `PartialOrd` that reverses the order. This is because the
-/// default Rust heap is a max heap, whereas a min heap is needed.
-///
-/// WARNING: equality is not what you would expect here.
-/// Two elements are equal if their feature is equal, and regardless of whether `doc`
-/// is equal. This should be perfectly fine for this usage, but let's make sure this
-/// struct is never public.
-struct ComparableDoc<T, D> {
-    feature: T,
-    doc: D,
+// Rust heap is a max-heap and we need a min heap.
+#[derive(Clone, Copy)]
+struct GlobalScoredDoc {
+    score: Score,
+    doc_address: DocAddress,
 }
 
-impl<T: PartialOrd, D> PartialOrd for ComparableDoc<T, D> {
-    fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
+impl PartialOrd for GlobalScoredDoc {
+    fn partial_cmp(&self, other: &GlobalScoredDoc) -> Option<Ordering> {
         Some(self.cmp(other))
     }
 }
 
-impl<T: PartialOrd, D> Ord for ComparableDoc<T, D> {
+impl Ord for GlobalScoredDoc {
     #[inline]
-    fn cmp(&self, other: &Self) -> Ordering {
+    fn cmp(&self, other: &GlobalScoredDoc) -> Ordering {
         other
-            .feature
-            .partial_cmp(&self.feature)
-            .unwrap_or_else(|| Ordering::Equal)
+            .score
+            .partial_cmp(&self.score)
+            .unwrap_or_else(|| other.doc_address.cmp(&self.doc_address))
     }
 }
 
-impl<T: PartialOrd, D> PartialEq for ComparableDoc<T, D> {
-    fn eq(&self, other: &Self) -> bool {
+impl PartialEq for GlobalScoredDoc {
+    fn eq(&self, other: &GlobalScoredDoc) -> bool {
         self.cmp(other) == Ordering::Equal
     }
 }
 
-impl<T: PartialOrd, D> Eq for ComparableDoc<T, D> {}
+impl Eq for GlobalScoredDoc {}
 
-pub(crate) struct TopCollector<T> {
+/// The Top Collector keeps track of the K documents
+/// with the best scores.
+///
+/// The implementation is based on a `BinaryHeap`.
+/// The theorical complexity for collecting the top `K` out of `n` documents
+/// is `O(n log K)`.
+///
+/// ```rust
+/// #[macro_use]
+/// extern crate tantivy;
+/// use tantivy::schema::{SchemaBuilder, TEXT};
+/// use tantivy::{Index, Result, DocId, Score};
+/// use tantivy::collector::TopCollector;
+/// use tantivy::query::QueryParser;
+///
+/// # fn main() { example().unwrap(); }
+/// fn example() -> Result<()> {
+///     let mut schema_builder = SchemaBuilder::new();
+///     let title = schema_builder.add_text_field("title", TEXT);
+///     let schema = schema_builder.build();
+///     let index = Index::create_in_ram(schema);
+///     {
+///         let mut index_writer = index.writer_with_num_threads(1, 3_000_000)?;
+///         index_writer.add_document(doc!(
+///             title => "The Name of the Wind",
+///         ));
+///         index_writer.add_document(doc!(
+///             title => "The Diary of Muadib",
+///         ));
+///         index_writer.add_document(doc!(
+///             title => "A Dairy Cow",
+///         ));
+///         index_writer.add_document(doc!(
+///             title => "The Diary of a Young Girl",
+///         ));
+///         index_writer.commit().unwrap();
+///     }
+///
+///     index.load_searchers()?;
+///     let searcher = index.searcher();
+///
+///     {
+///	        let mut top_collector = TopCollector::with_limit(2);
+///         let query_parser = QueryParser::for_index(&index, vec![title]);
+///         let query = query_parser.parse_query("diary")?;
+///         searcher.search(&*query, &mut top_collector).unwrap();
+///
+///         let score_docs: Vec<(Score, DocId)> = top_collector
+///           .score_docs()
+///           .into_iter()
+///           .map(|(score, doc_address)| (score, doc_address.doc()))
+///           .collect();
+///
+///         assert_eq!(score_docs, vec![(0.7261542, 1), (0.6099695, 3)]);
+///     }
+///
+///     Ok(())
+/// }
+/// ```
+pub struct TopCollector {
     limit: usize,
-    _marker: PhantomData<T>,
+    heap: BinaryHeap<GlobalScoredDoc>,
+    segment_id: u32,
 }
 
-impl<T> TopCollector<T>
-where
-    T: PartialOrd + Clone,
-{
+impl TopCollector {
     /// Creates a top collector, with a number of documents equal to "limit".
     ///
     /// # Panics
     /// The method panics if limit is 0
-    pub fn with_limit(limit: usize) -> TopCollector<T> {
+    pub fn with_limit(limit: usize) -> TopCollector {
         if limit < 1 {
             panic!("Limit must be strictly greater than 0.");
         }
         TopCollector {
-            limit,
-            _marker: PhantomData,
-        }
-    }
-
-    pub fn limit(&self) -> usize {
-        self.limit
-    }
-
-    pub fn merge_fruits(
-        &self,
-        children: Vec<Vec<(T, DocAddress)>>,
-    ) -> Result<Vec<(T, DocAddress)>> {
-        if self.limit == 0 {
-            return Ok(Vec::new());
-        }
-        let mut top_collector = BinaryHeap::new();
-        for child_fruit in children {
-            for (feature, doc) in child_fruit {
-                if top_collector.len() < self.limit {
-                    top_collector.push(ComparableDoc { feature, doc });
-                } else if let Some(mut head) = top_collector.peek_mut() {
-                    if head.feature < feature {
-                        *head = ComparableDoc { feature, doc };
-                    }
-                }
-            }
-        }
-        Ok(top_collector
-            .into_sorted_vec()
-            .into_iter()
-            .map(|cdoc| (cdoc.feature, cdoc.doc))
-            .collect())
-    }
-
-    pub(crate) fn for_segment(
-        &self,
-        segment_id: SegmentLocalId,
-        _: &SegmentReader,
-    ) -> Result<TopSegmentCollector<T>> {
-        Ok(TopSegmentCollector::new(segment_id, self.limit))
-    }
-}
-
-/// The Top Collector keeps track of the K documents
-/// sorted by type `T`.
-///
-/// The implementation is based on a `BinaryHeap`.
-/// The theorical complexity for collecting the top `K` out of `n` documents
-/// is `O(n log K)`.
-pub(crate) struct TopSegmentCollector<T> {
-    limit: usize,
-    heap: BinaryHeap<ComparableDoc<T, DocId>>,
-    segment_id: u32,
-}
-
-impl<T: PartialOrd> TopSegmentCollector<T> {
-    fn new(segment_id: SegmentLocalId, limit: usize) -> TopSegmentCollector<T> {
-        TopSegmentCollector {
             limit,
             heap: BinaryHeap::with_capacity(limit),
-            segment_id,
+            segment_id: 0,
         }
     }
-}
 
-impl<T: PartialOrd + Clone> TopSegmentCollector<T> {
-    pub fn harvest(self) -> Vec<(T, DocAddress)> {
-        let segment_id = self.segment_id;
-        self.heap
-            .into_sorted_vec()
+    /// Returns K best documents sorted in decreasing order.
+    ///
+    /// Calling this method triggers the sort.
+    /// The result of the sort is not cached.
+    pub fn docs(&self) -> Vec<DocAddress> {
+        self.score_docs()
             .into_iter()
-            .map(|comparable_doc| {
-                (
-                    comparable_doc.feature,
-                    DocAddress(segment_id, comparable_doc.doc),
-                )
-            })
+            .map(|score_doc| score_doc.1)
+            .collect()
+    }
+
+    /// Returns K best ScoredDocument sorted in decreasing order.
+    ///
+    /// Calling this method triggers the sort.
+    /// The result of the sort is not cached.
+    pub fn score_docs(&self) -> Vec<(Score, DocAddress)> {
+        let mut scored_docs: Vec<GlobalScoredDoc> = self.heap.iter().cloned().collect();
+        scored_docs.sort();
+        scored_docs
+            .into_iter()
+            .map(|GlobalScoredDoc { score, doc_address }| (score, doc_address))
             .collect()
     }
 
     /// Return true iff at least K documents have gone through
     /// the collector.
-    #[inline(always)]
-    pub(crate) fn at_capacity(&self) -> bool {
+    #[inline]
+    pub fn at_capacity(&self) -> bool {
         self.heap.len() >= self.limit
     }
+}
 
-    /// Collects a document scored by the given feature
-    ///
-    /// It collects documents until it has reached the max capacity. Once it reaches capacity, it
-    /// will compare the lowest scoring item with the given one and keep whichever is greater.
-    #[inline(always)]
-    pub fn collect(&mut self, doc: DocId, feature: T) {
+impl Collector for TopCollector {
+    fn set_segment(&mut self, segment_id: SegmentLocalId, _: &SegmentReader) -> Result<()> {
+        self.segment_id = segment_id;
+        Ok(())
+    }
+
+    fn collect(&mut self, doc: DocId, score: Score) {
         if self.at_capacity() {
             // It's ok to unwrap as long as a limit of 0 is forbidden.
-            if let Some(limit_feature) = self.heap.peek().map(|head| head.feature.clone()) {
-                if limit_feature < feature {
-                    if let Some(mut head) = self.heap.peek_mut() {
-                        head.feature = feature;
-                        head.doc = doc;
-                    }
-                }
+            let limit_doc: GlobalScoredDoc = *self.heap
+                .peek()
+                .expect("Top collector with size 0 is forbidden");
+            if limit_doc.score < score {
+                let mut mut_head = self.heap
+                    .peek_mut()
+                    .expect("Top collector with size 0 is forbidden");
+                mut_head.score = score;
+                mut_head.doc_address = DocAddress(self.segment_id, doc);
             }
         } else {
-            // we have not reached capacity yet, so we can just push the
-            // element.
-            self.heap.push(ComparableDoc { feature, doc });
+            let wrapped_doc = GlobalScoredDoc {
+                score,
+                doc_address: DocAddress(self.segment_id, doc),
+            };
+            self.heap.push(wrapped_doc);
         }
     }
+
+    fn requires_scoring(&self) -> bool {
+        true
+    }
 }
 
 #[cfg(test)]
 mod tests {
-    use super::{TopCollector, TopSegmentCollector};
-    use DocAddress;
+
+    use super::*;
+    use collector::Collector;
+    use DocId;
     use Score;
 
     #[test]
     fn test_top_collector_not_at_capacity() {
-        let mut top_collector = TopSegmentCollector::new(0, 4);
+        let mut top_collector = TopCollector::with_limit(4);
         top_collector.collect(1, 0.8);
         top_collector.collect(3, 0.2);
         top_collector.collect(5, 0.3);
-        assert_eq!(
-            top_collector.harvest(),
-            vec![
-                (0.8, DocAddress(0, 1)),
-                (0.3, DocAddress(0, 5)),
-                (0.2, DocAddress(0, 3))
-            ]
-        );
+        assert!(!top_collector.at_capacity());
+        let score_docs: Vec<(Score, DocId)> = top_collector
+            .score_docs()
+            .into_iter()
+            .map(|(score, doc_address)| (score, doc_address.doc()))
+            .collect();
+        assert_eq!(score_docs, vec![(0.8, 1), (0.3, 5), (0.2, 3)]);
     }
 
     #[test]
     fn test_top_collector_at_capacity() {
-        let mut top_collector = TopSegmentCollector::new(0, 4);
+        let mut top_collector = TopCollector::with_limit(4);
         top_collector.collect(1, 0.8);
         top_collector.collect(3, 0.2);
         top_collector.collect(5, 0.3);
         top_collector.collect(7, 0.9);
         top_collector.collect(9, -0.2);
-        assert_eq!(
-            top_collector.harvest(),
-            vec![
-                (0.9, DocAddress(0, 7)),
-                (0.8, DocAddress(0, 1)),
-                (0.3, DocAddress(0, 5)),
-                (0.2, DocAddress(0, 3))
-            ]
-        );
+        assert!(top_collector.at_capacity());
+        {
+            let score_docs: Vec<(Score, DocId)> = top_collector
+                .score_docs()
+                .into_iter()
+                .map(|(score, doc_address)| (score, doc_address.doc()))
+                .collect();
+            assert_eq!(score_docs, vec![(0.9, 7), (0.8, 1), (0.3, 5), (0.2, 3)]);
+        }
+        {
+            let docs: Vec<DocId> = top_collector
+                .docs()
+                .into_iter()
+                .map(|doc_address| doc_address.doc())
+                .collect();
+            assert_eq!(docs, vec![7, 1, 5, 3]);
+        }
     }
 
     #[test]
     #[should_panic]
     fn test_top_0() {
-        let _collector: TopCollector<Score> = TopCollector::with_limit(0);
+        TopCollector::with_limit(0);
     }
+
 }

src/collector/top_field_collector.rs

@@ -1,250 +0,0 @@
-use super::Collector;
-use collector::top_collector::TopCollector;
-use collector::top_collector::TopSegmentCollector;
-use collector::SegmentCollector;
-use fastfield::FastFieldReader;
-use fastfield::FastValue;
-use schema::Field;
-use DocAddress;
-use Result;
-use SegmentLocalId;
-use SegmentReader;
-
-/// The Top Field Collector keeps track of the K documents
-/// sorted by a fast field in the index
-///
-/// The implementation is based on a `BinaryHeap`.
-/// The theorical complexity for collecting the top `K` out of `n` documents
-/// is `O(n log K)`.
-///
-/// ```rust
-/// #[macro_use]
-/// extern crate tantivy;
-/// # use tantivy::schema::{Schema, Field, FAST, TEXT};
-/// # use tantivy::{Index, Result, DocAddress};
-/// # use tantivy::query::{Query, QueryParser};
-/// use tantivy::collector::TopDocs;
-///
-/// # fn main() {
-/// #   let mut schema_builder = Schema::builder();
-/// #   let title = schema_builder.add_text_field("title", TEXT);
-/// #   let rating = schema_builder.add_u64_field("rating", FAST);
-/// #   let schema = schema_builder.build();
-/// #   let index = Index::create_in_ram(schema);
-/// #   let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-/// #   index_writer.add_document(doc!(
-/// #       title => "The Name of the Wind",
-/// #       rating => 92u64,
-/// #   ));
-/// #   index_writer.add_document(doc!(title => "The Diary of Muadib", rating => 97u64));
-/// #   index_writer.add_document(doc!(title => "A Dairy Cow", rating => 63u64));
-/// #   index_writer.add_document(doc!(title => "The Diary of a Young Girl", rating => 80u64));
-/// #   index_writer.commit().unwrap();
-/// #   index.load_searchers().unwrap();
-///	#   let query = QueryParser::for_index(&index, vec![title]).parse_query("diary").unwrap();
-/// #   let top_docs = docs_sorted_by_rating(&index, &query, rating).unwrap();
-/// #   assert_eq!(top_docs,
-/// #            vec![(97u64, DocAddress(0u32, 1)),
-/// #                 (80u64, DocAddress(0u32, 3))]);
-/// # }
-/// #
-/// /// Searches the document matching the given query, and
-/// /// collects the top 10 documents, order by the `field`
-/// /// given in argument.
-/// ///
-/// /// `field` is required to be a FAST field.
-/// fn docs_sorted_by_rating(index: &Index, query: &Query, sort_by_field: Field)
-///     -> Result<Vec<(u64, DocAddress)>> {
-///
-///     // This is where we build our collector!
-///     let top_docs_by_rating = TopDocs::with_limit(2).order_by_field(sort_by_field);
-///
-///     // ... and here is our documents. Not this is a simple vec.
-///     // The `u64` in the pair is the value of our fast field for each documents.
-///     index.searcher()
-///          .search(query, &top_docs_by_rating)
-/// }
-/// ```
-pub struct TopDocsByField<T> {
-    collector: TopCollector<T>,
-    field: Field,
-}
-
-impl<T: FastValue + PartialOrd + Clone> TopDocsByField<T> {
-    /// Creates a top field collector, with a number of documents equal to "limit".
-    ///
-    /// The given field name must be a fast field, otherwise the collector have an error while
-    /// collecting results.
-    ///
-    /// # Panics
-    /// The method panics if limit is 0
-    pub(crate) fn new(field: Field, limit: usize) -> TopDocsByField<T> {
-        TopDocsByField {
-            collector: TopCollector::with_limit(limit),
-            field,
-        }
-    }
-}
-
-impl<T: FastValue + PartialOrd + Send + Sync + 'static> Collector for TopDocsByField<T> {
-    type Fruit = Vec<(T, DocAddress)>;
-
-    type Child = TopFieldSegmentCollector<T>;
-
-    fn for_segment(
-        &self,
-        segment_local_id: SegmentLocalId,
-        reader: &SegmentReader,
-    ) -> Result<TopFieldSegmentCollector<T>> {
-        let collector = self.collector.for_segment(segment_local_id, reader)?;
-        let reader = reader.fast_field_reader(self.field)?;
-        Ok(TopFieldSegmentCollector { collector, reader })
-    }
-
-    fn requires_scoring(&self) -> bool {
-        false
-    }
-
-    fn merge_fruits(
-        &self,
-        segment_fruits: Vec<Vec<(T, DocAddress)>>,
-    ) -> Result<Vec<(T, DocAddress)>> {
-        self.collector.merge_fruits(segment_fruits)
-    }
-}
-
-pub struct TopFieldSegmentCollector<T: FastValue + PartialOrd> {
-    collector: TopSegmentCollector<T>,
-    reader: FastFieldReader<T>,
-}
-
-impl<T: FastValue + PartialOrd + Send + Sync + 'static> SegmentCollector
-    for TopFieldSegmentCollector<T>
-{
-    type Fruit = Vec<(T, DocAddress)>;
-
-    fn collect(&mut self, doc: u32, _score: f32) {
-        let field_value = self.reader.get(doc);
-        self.collector.collect(doc, field_value);
-    }
-
-    fn harvest(self) -> Vec<(T, DocAddress)> {
-        self.collector.harvest()
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::TopDocsByField;
-    use collector::Collector;
-    use collector::TopDocs;
-    use query::Query;
-    use query::QueryParser;
-    use schema::Field;
-    use schema::IntOptions;
-    use schema::{Schema, FAST, TEXT};
-    use DocAddress;
-    use Index;
-    use IndexWriter;
-    use TantivyError;
-
-    const TITLE: &str = "title";
-    const SIZE: &str = "size";
-
-    #[test]
-    fn test_top_collector_not_at_capacity() {
-        let mut schema_builder = Schema::builder();
-        let title = schema_builder.add_text_field(TITLE, TEXT);
-        let size = schema_builder.add_u64_field(SIZE, FAST);
-        let schema = schema_builder.build();
-        let (index, query) = index("beer", title, schema, |index_writer| {
-            index_writer.add_document(doc!(
-                title => "bottle of beer",
-                size => 12u64,
-            ));
-            index_writer.add_document(doc!(
-                title => "growler of beer",
-                size => 64u64,
-            ));
-            index_writer.add_document(doc!(
-                title => "pint of beer",
-                size => 16u64,
-            ));
-        });
-        let searcher = index.searcher();
-
-        let top_collector = TopDocs::with_limit(4).order_by_field(size);
-        let top_docs: Vec<(u64, DocAddress)> = searcher.search(&query, &top_collector).unwrap();
-        assert_eq!(
-            top_docs,
-            vec![
-                (64, DocAddress(0, 1)),
-                (16, DocAddress(0, 2)),
-                (12, DocAddress(0, 0))
-            ]
-        );
-    }
-
-    #[test]
-    #[should_panic]
-    fn test_field_does_not_exist() {
-        let mut schema_builder = Schema::builder();
-        let title = schema_builder.add_text_field(TITLE, TEXT);
-        let size = schema_builder.add_u64_field(SIZE, FAST);
-        let schema = schema_builder.build();
-        let (index, _) = index("beer", title, schema, |index_writer| {
-            index_writer.add_document(doc!(
-                title => "bottle of beer",
-                size => 12u64,
-            ));
-        });
-        let searcher = index.searcher();
-        let top_collector: TopDocsByField<u64> = TopDocs::with_limit(4).order_by_field(Field(2));
-        let segment_reader = searcher.segment_reader(0u32);
-        top_collector
-            .for_segment(0, segment_reader)
-            .expect("should panic");
-    }
-
-    #[test]
-    fn test_field_not_fast_field() {
-        let mut schema_builder = Schema::builder();
-        let title = schema_builder.add_text_field(TITLE, TEXT);
-        let size = schema_builder.add_u64_field(SIZE, IntOptions::default());
-        let schema = schema_builder.build();
-        let (index, _) = index("beer", title, schema, |index_writer| {
-            index_writer.add_document(doc!(
-                title => "bottle of beer",
-                size => 12u64,
-            ));
-        });
-        let searcher = index.searcher();
-        let segment = searcher.segment_reader(0);
-        let top_collector: TopDocsByField<u64> = TopDocs::with_limit(4).order_by_field(size);
-        assert_matches!(
-            top_collector
-                .for_segment(0, segment)
-                .map(|_| ())
-                .unwrap_err(),
-            TantivyError::FastFieldError(_)
-        );
-    }
-
-    fn index(
-        query: &str,
-        query_field: Field,
-        schema: Schema,
-        mut doc_adder: impl FnMut(&mut IndexWriter) -> (),
-    ) -> (Index, Box<Query>) {
-        let index = Index::create_in_ram(schema);
-
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-        doc_adder(&mut index_writer);
-        index_writer.commit().unwrap();
-        index.load_searchers().unwrap();
-
-        let query_parser = QueryParser::for_index(&index, vec![query_field]);
-        let query = query_parser.parse_query(query).unwrap();
-        (index, query)
-    }
-}

src/collector/top_score_collector.rs

@@ -1,200 +0,0 @@
-use super::Collector;
-use collector::top_collector::TopCollector;
-use collector::top_collector::TopSegmentCollector;
-use collector::SegmentCollector;
-use collector::TopDocsByField;
-use fastfield::FastValue;
-use schema::Field;
-use DocAddress;
-use DocId;
-use Result;
-use Score;
-use SegmentLocalId;
-use SegmentReader;
-
-/// The Top Score Collector keeps track of the K documents
-/// sorted by their score.
-///
-/// The implementation is based on a `BinaryHeap`.
-/// The theorical complexity for collecting the top `K` out of `n` documents
-/// is `O(n log K)`.
-///
-/// ```rust
-/// #[macro_use]
-/// extern crate tantivy;
-/// use tantivy::DocAddress;
-/// use tantivy::schema::{Schema, TEXT};
-/// use tantivy::{Index, Result};
-/// use tantivy::collector::TopDocs;
-/// use tantivy::query::QueryParser;
-///
-/// # fn main() { example().unwrap(); }
-/// fn example() -> Result<()> {
-///     let mut schema_builder = Schema::builder();
-///     let title = schema_builder.add_text_field("title", TEXT);
-///     let schema = schema_builder.build();
-///     let index = Index::create_in_ram(schema);
-///     {
-///         let mut index_writer = index.writer_with_num_threads(1, 3_000_000)?;
-///         index_writer.add_document(doc!(
-///             title => "The Name of the Wind",
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "The Diary of Muadib",
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "A Dairy Cow",
-///         ));
-///         index_writer.add_document(doc!(
-///             title => "The Diary of a Young Girl",
-///         ));
-///         index_writer.commit().unwrap();
-///     }
-///
-///     index.load_searchers()?;
-///     let searcher = index.searcher();
-///
-///     let query_parser = QueryParser::for_index(&index, vec![title]);
-///     let query = query_parser.parse_query("diary")?;
-///     let top_docs = searcher.search(&query, &TopDocs::with_limit(2))?;
-///
-///     assert_eq!(&top_docs[0], &(0.7261542, DocAddress(0, 1)));
-///     assert_eq!(&top_docs[1], &(0.6099695, DocAddress(0, 3)));
-///
-///     Ok(())
-/// }
-/// ```
-pub struct TopDocs(TopCollector<Score>);
-
-impl TopDocs {
-    /// Creates a top score collector, with a number of documents equal to "limit".
-    ///
-    /// # Panics
-    /// The method panics if limit is 0
-    pub fn with_limit(limit: usize) -> TopDocs {
-        TopDocs(TopCollector::with_limit(limit))
-    }
-
-    /// Set top-K to rank documents by a given fast field.
-    ///
-    /// (By default, `TopDocs` collects the top-K documents sorted by
-    /// the similarity score.)
-    pub fn order_by_field<T: PartialOrd + FastValue + Clone>(
-        self,
-        field: Field,
-    ) -> TopDocsByField<T> {
-        TopDocsByField::new(field, self.0.limit())
-    }
-}
-
-impl Collector for TopDocs {
-    type Fruit = Vec<(Score, DocAddress)>;
-
-    type Child = TopScoreSegmentCollector;
-
-    fn for_segment(
-        &self,
-        segment_local_id: SegmentLocalId,
-        reader: &SegmentReader,
-    ) -> Result<Self::Child> {
-        let collector = self.0.for_segment(segment_local_id, reader)?;
-        Ok(TopScoreSegmentCollector(collector))
-    }
-
-    fn requires_scoring(&self) -> bool {
-        true
-    }
-
-    fn merge_fruits(&self, child_fruits: Vec<Vec<(Score, DocAddress)>>) -> Result<Self::Fruit> {
-        self.0.merge_fruits(child_fruits)
-    }
-}
-
-/// Segment Collector associated to `TopDocs`.
-pub struct TopScoreSegmentCollector(TopSegmentCollector<Score>);
-
-impl SegmentCollector for TopScoreSegmentCollector {
-    type Fruit = Vec<(Score, DocAddress)>;
-
-    fn collect(&mut self, doc: DocId, score: Score) {
-        self.0.collect(doc, score)
-    }
-
-    fn harvest(self) -> Vec<(Score, DocAddress)> {
-        self.0.harvest()
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::TopDocs;
-    use query::QueryParser;
-    use schema::Schema;
-    use schema::TEXT;
-    use DocAddress;
-    use Index;
-    use Score;
-
-    fn make_index() -> Index {
-        let mut schema_builder = Schema::builder();
-        let text_field = schema_builder.add_text_field("text", TEXT);
-        let schema = schema_builder.build();
-        let index = Index::create_in_ram(schema);
-        {
-            // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-            index_writer.add_document(doc!(text_field=>"Hello happy tax payer."));
-            index_writer.add_document(doc!(text_field=>"Droopy says hello happy tax payer"));
-            index_writer.add_document(doc!(text_field=>"I like Droopy"));
-            assert!(index_writer.commit().is_ok());
-        }
-        index.load_searchers().unwrap();
-        index
-    }
-
-    #[test]
-    fn test_top_collector_not_at_capacity() {
-        let index = make_index();
-        let field = index.schema().get_field("text").unwrap();
-        let query_parser = QueryParser::for_index(&index, vec![field]);
-        let text_query = query_parser.parse_query("droopy tax").unwrap();
-        let score_docs: Vec<(Score, DocAddress)> = index
-            .searcher()
-            .search(&text_query, &TopDocs::with_limit(4))
-            .unwrap();
-        assert_eq!(
-            score_docs,
-            vec![
-                (0.81221175, DocAddress(0u32, 1)),
-                (0.5376842, DocAddress(0u32, 2)),
-                (0.48527452, DocAddress(0, 0))
-            ]
-        );
-    }
-
-    #[test]
-    fn test_top_collector_at_capacity() {
-        let index = make_index();
-        let field = index.schema().get_field("text").unwrap();
-        let query_parser = QueryParser::for_index(&index, vec![field]);
-        let text_query = query_parser.parse_query("droopy tax").unwrap();
-        let score_docs: Vec<(Score, DocAddress)> = index
-            .searcher()
-            .search(&text_query, &TopDocs::with_limit(2))
-            .unwrap();
-        assert_eq!(
-            score_docs,
-            vec![
-                (0.81221175, DocAddress(0u32, 1)),
-                (0.5376842, DocAddress(0u32, 2)),
-            ]
-        );
-    }
-
-    #[test]
-    #[should_panic]
-    fn test_top_0() {
-        TopDocs::with_limit(0);
-    }
-
-}

src/common/bitpacker.rs

@@ -1,6 +1,9 @@
-use byteorder::{ByteOrder, LittleEndian, WriteBytesExt};
+use common::serialize::BinarySerializable;
 use std::io;
+use std::io::Write;
+use std::mem;
 use std::ops::Deref;
+use std::ptr;
 
 pub(crate) struct BitPacker {
     mini_buffer: u64,
@@ -15,7 +18,7 @@ impl BitPacker {
         }
     }
 
-    pub fn write<TWrite: io::Write>(
+    pub fn write<TWrite: Write>(
         &mut self,
         val: u64,
         num_bits: u8,
@@ -25,14 +28,14 @@ impl BitPacker {
         let num_bits = num_bits as usize;
         if self.mini_buffer_written + num_bits > 64 {
             self.mini_buffer |= val_u64.wrapping_shl(self.mini_buffer_written as u32);
-            output.write_u64::<LittleEndian>(self.mini_buffer)?;
+            self.mini_buffer.serialize(output)?;
             self.mini_buffer = val_u64.wrapping_shr((64 - self.mini_buffer_written) as u32);
             self.mini_buffer_written = self.mini_buffer_written + num_bits - 64;
         } else {
             self.mini_buffer |= val_u64 << self.mini_buffer_written;
             self.mini_buffer_written += num_bits;
             if self.mini_buffer_written == 64 {
-                output.write_u64::<LittleEndian>(self.mini_buffer)?;
+                self.mini_buffer.serialize(output)?;
                 self.mini_buffer_written = 0;
                 self.mini_buffer = 0u64;
             }
@@ -40,18 +43,17 @@ impl BitPacker {
         Ok(())
     }
 
-    pub fn flush<TWrite: io::Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
+    pub fn flush<TWrite: Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
         if self.mini_buffer_written > 0 {
             let num_bytes = (self.mini_buffer_written + 7) / 8;
-            let mut arr: [u8; 8] = [0u8; 8];
-            LittleEndian::write_u64(&mut arr, self.mini_buffer);
+            let arr: [u8; 8] = unsafe { mem::transmute::<u64, [u8; 8]>(self.mini_buffer.to_le()) };
             output.write_all(&arr[..num_bytes])?;
             self.mini_buffer_written = 0;
         }
         Ok(())
     }
 
-    pub fn close<TWrite: io::Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
+    pub fn close<TWrite: Write>(&mut self, output: &mut TWrite) -> io::Result<()> {
         self.flush(output)?;
         // Padding the write file to simplify reads.
         output.write_all(&[0u8; 7])?;
@@ -100,10 +102,37 @@ where
             addr + 8 <= data.len(),
             "The fast field field should have been padded with 7 bytes."
         );
-        let val_unshifted_unmasked: u64 = LittleEndian::read_u64(&data[addr..]);
+        let val_unshifted_unmasked: u64 =
+            u64::from_le(unsafe { ptr::read_unaligned(data[addr..].as_ptr() as *const u64) });
         let val_shifted = (val_unshifted_unmasked >> bit_shift) as u64;
         val_shifted & mask
     }
+
+    /// Reads a range of values from the fast field.
+    ///
+    /// The range of values read is from
+    /// `[start..start + output.len()[`
+    pub fn get_range(&self, start: u32, output: &mut [u64]) {
+        if self.num_bits == 0 {
+            for val in output.iter_mut() {
+                *val = 0u64;
+            }
+        } else {
+            let data: &[u8] = &*self.data;
+            let num_bits = self.num_bits;
+            let mask = self.mask;
+            let mut addr_in_bits = (start as usize) * num_bits;
+            for output_val in output.iter_mut() {
+                let addr = addr_in_bits >> 3;
+                let bit_shift = addr_in_bits & 7;
+                let val_unshifted_unmasked: u64 =
+                    unsafe { ptr::read_unaligned(data[addr..].as_ptr() as *const u64) };
+                let val_shifted = (val_unshifted_unmasked >> bit_shift) as u64;
+                *output_val = val_shifted & mask;
+                addr_in_bits += num_bits;
+            }
+        }
+    }
 }
 
 #[cfg(test)]
@@ -141,4 +170,17 @@ mod test {
         test_bitpacker_util(6, 14);
         test_bitpacker_util(1000, 14);
     }
+
+    #[test]
+    fn test_bitpacker_range() {
+        let (bitunpacker, vals) = create_fastfield_bitpacker(100_000, 12);
+        let buffer_len = 100;
+        let mut buffer = vec![0u64; buffer_len];
+        for start in vec![0, 10, 20, 100, 1_000] {
+            bitunpacker.get_range(start as u32, &mut buffer[..]);
+            for i in 0..buffer_len {
+                assert_eq!(buffer[i], vals[start + i]);
+            }
+        }
+    }
 }

src/common/bitset.rs

@@ -34,17 +34,17 @@ impl TinySet {
     }
 
     /// Returns the complement of the set in `[0, 64[`.
-    fn complement(self) -> TinySet {
+    fn complement(&self) -> TinySet {
         TinySet(!self.0)
     }
 
     /// Returns true iff the `TinySet` contains the element `el`.
-    pub fn contains(self, el: u32) -> bool {
+    pub fn contains(&self, el: u32) -> bool {
         !self.intersect(TinySet::singleton(el)).is_empty()
     }
 
     /// Returns the intersection of `self` and `other`
-    pub fn intersect(self, other: TinySet) -> TinySet {
+    pub fn intersect(&self, other: TinySet) -> TinySet {
         TinySet(self.0 & other.0)
     }
 
@@ -77,7 +77,7 @@ impl TinySet {
 
     /// Returns true iff the `TinySet` is empty.
     #[inline(always)]
-    pub fn is_empty(self) -> bool {
+    pub fn is_empty(&self) -> bool {
         self.0 == 0u64
     }
 
@@ -114,7 +114,7 @@ impl TinySet {
         self.0 = 0u64;
     }
 
-    pub fn len(self) -> u32 {
+    pub fn len(&self) -> u32 {
         self.0.count_ones()
     }
 }
@@ -266,14 +266,14 @@ mod tests {
 
     #[test]
     fn test_bitset_large() {
-        let arr = generate_nonunique_unsorted(100_000, 5_000);
+        let arr = generate_nonunique_unsorted(1_000_000, 50_000);
         let mut btreeset: BTreeSet<u32> = BTreeSet::new();
-        let mut bitset = BitSet::with_max_value(100_000);
+        let mut bitset = BitSet::with_max_value(1_000_000);
         for el in arr {
             btreeset.insert(el);
             bitset.insert(el);
         }
-        for i in 0..100_000 {
+        for i in 0..1_000_000 {
             assert_eq!(btreeset.contains(&i), bitset.contains(i));
         }
         assert_eq!(btreeset.len(), bitset.len());

src/common/composite_file.rs

@@ -4,8 +4,6 @@ use common::VInt;
 use directory::ReadOnlySource;
 use directory::WritePtr;
 use schema::Field;
-use space_usage::FieldUsage;
-use space_usage::PerFieldSpaceUsage;
 use std::collections::HashMap;
 use std::io::Write;
 use std::io::{self, Read};
@@ -74,8 +72,7 @@ impl<W: Write> CompositeWrite<W> {
         let footer_offset = self.write.written_bytes();
         VInt(self.offsets.len() as u64).serialize(&mut self.write)?;
 
-        let mut offset_fields: Vec<_> = self
-            .offsets
+        let mut offset_fields: Vec<_> = self.offsets
             .iter()
             .map(|(file_addr, offset)| (*offset, *file_addr))
             .collect();
@@ -168,17 +165,6 @@ impl CompositeFile {
             .get(&FileAddr { field, idx })
             .map(|&(from, to)| self.data.slice(from, to))
     }
-
-    pub fn space_usage(&self) -> PerFieldSpaceUsage {
-        let mut fields = HashMap::new();
-        for (&field_addr, &(start, end)) in self.offsets_index.iter() {
-            fields
-                .entry(field_addr.field)
-                .or_insert_with(|| FieldUsage::empty(field_addr.field))
-                .add_field_idx(field_addr.idx, end - start);
-        }
-        PerFieldSpaceUsage::new(fields)
-    }
 }
 
 #[cfg(test)]

src/common/mod.rs

@@ -10,7 +10,7 @@ pub(crate) use self::bitset::TinySet;
 pub(crate) use self::composite_file::{CompositeFile, CompositeWrite};
 pub use self::counting_writer::CountingWriter;
 pub use self::serialize::{BinarySerializable, FixedSize};
-pub use self::vint::{read_u32_vint, serialize_vint_u32, write_u32_vint, VInt};
+pub use self::vint::VInt;
 pub use byteorder::LittleEndian as Endianness;
 
 use std::io;

src/common/vint.rs

@@ -1,5 +1,4 @@
 use super::BinarySerializable;
-use byteorder::{ByteOrder, LittleEndian};
 use std::io;
 use std::io::Read;
 use std::io::Write;
@@ -10,101 +9,9 @@ pub struct VInt(pub u64);
 
 const STOP_BIT: u8 = 128;
 
-pub fn serialize_vint_u32(val: u32) -> (u64, usize) {
-    const START_2: u64 = 1 << 7;
-    const START_3: u64 = 1 << 14;
-    const START_4: u64 = 1 << 21;
-    const START_5: u64 = 1 << 28;
-
-    const STOP_1: u64 = START_2 - 1;
-    const STOP_2: u64 = START_3 - 1;
-    const STOP_3: u64 = START_4 - 1;
-    const STOP_4: u64 = START_5 - 1;
-
-    const MASK_1: u64 = 127;
-    const MASK_2: u64 = MASK_1 << 7;
-    const MASK_3: u64 = MASK_2 << 7;
-    const MASK_4: u64 = MASK_3 << 7;
-    const MASK_5: u64 = MASK_4 << 7;
-
-    let val = u64::from(val);
-    const STOP_BIT: u64 = 128u64;
-    match val {
-        0...STOP_1 => (val | STOP_BIT, 1),
-        START_2...STOP_2 => (
-            (val & MASK_1) | ((val & MASK_2) << 1) | (STOP_BIT << (8)),
-            2,
-        ),
-        START_3...STOP_3 => (
-            (val & MASK_1) | ((val & MASK_2) << 1) | ((val & MASK_3) << 2) | (STOP_BIT << (8 * 2)),
-            3,
-        ),
-        START_4...STOP_4 => (
-            (val & MASK_1)
-                | ((val & MASK_2) << 1)
-                | ((val & MASK_3) << 2)
-                | ((val & MASK_4) << 3)
-                | (STOP_BIT << (8 * 3)),
-            4,
-        ),
-        _ => (
-            (val & MASK_1)
-                | ((val & MASK_2) << 1)
-                | ((val & MASK_3) << 2)
-                | ((val & MASK_4) << 3)
-                | ((val & MASK_5) << 4)
-                | (STOP_BIT << (8 * 4)),
-            5,
-        ),
-    }
-}
-
-/// Returns the number of bytes covered by a
-/// serialized vint `u32`.
-///
-/// Expects a buffer data that starts
-/// by the serialized `vint`, scans at most 5 bytes ahead until
-/// it finds the vint final byte.
-///
-/// # May Panic
-/// If the payload does not start by a valid `vint`
-fn vint_len(data: &[u8]) -> usize {
-    for (i, &val) in data.iter().enumerate().take(5) {
-        if val >= STOP_BIT {
-            return i + 1;
-        }
-    }
-    panic!("Corrupted data. Invalid VInt 32");
-}
-
-/// Reads a vint `u32` from a buffer, and
-/// consumes its payload data.
-///
-/// # Panics
-///
-/// If the buffer does not start by a valid
-/// vint payload
-pub fn read_u32_vint(data: &mut &[u8]) -> u32 {
-    let vlen = vint_len(*data);
-    let mut result = 0u32;
-    let mut shift = 0u64;
-    for &b in &data[..vlen] {
-        result |= u32::from(b & 127u8) << shift;
-        shift += 7;
-    }
-    *data = &data[vlen..];
-    result
-}
-
-/// Write a `u32` as a vint payload.
-pub fn write_u32_vint<W: io::Write>(val: u32, writer: &mut W) -> io::Result<()> {
-    let (val, num_bytes) = serialize_vint_u32(val);
-    let mut buffer = [0u8; 8];
-    LittleEndian::write_u64(&mut buffer, val);
-    writer.write_all(&buffer[..num_bytes])
-}
-
 impl VInt {
+
+
     pub fn val(&self) -> u64 {
         self.0
     }
@@ -113,13 +20,14 @@ impl VInt {
         VInt::deserialize(reader).map(|vint| vint.0)
     }
 
-    pub fn serialize_into_vec(&self, output: &mut Vec<u8>) {
+    pub fn serialize_into_vec(&self, output: &mut Vec<u8>){
         let mut buffer = [0u8; 10];
         let num_bytes = self.serialize_into(&mut buffer);
         output.extend(&buffer[0..num_bytes]);
     }
 
-    pub fn serialize_into(&self, buffer: &mut [u8; 10]) -> usize {
+    fn serialize_into(&self, buffer: &mut [u8; 10]) -> usize {
+
         let mut remaining = self.0;
         for (i, b) in buffer.iter_mut().enumerate() {
             let next_byte: u8 = (remaining % 128u64) as u8;
@@ -159,19 +67,18 @@ impl BinarySerializable for VInt {
                     return Err(io::Error::new(
                         io::ErrorKind::InvalidData,
                         "Reach end of buffer while reading VInt",
-                    ));
+                    ))
                 }
             }
         }
     }
 }
 
+
 #[cfg(test)]
 mod tests {
 
-    use super::serialize_vint_u32;
     use super::VInt;
-    use byteorder::{ByteOrder, LittleEndian};
     use common::BinarySerializable;
 
     fn aux_test_vint(val: u64) {
@@ -182,10 +89,10 @@ mod tests {
         }
         assert!(num_bytes > 0);
         if num_bytes < 10 {
-            assert!(1u64 << (7 * num_bytes) > val);
+            assert!(1u64 << (7*num_bytes) > val);
         }
         if num_bytes > 1 {
-            assert!(1u64 << (7 * (num_bytes - 1)) <= val);
+            assert!(1u64 << (7*(num_bytes-1)) <= val);
         }
         let serdeser_val = VInt::deserialize(&mut &v[..]).unwrap();
         assert_eq!(val, serdeser_val.0);
@@ -198,35 +105,11 @@ mod tests {
         aux_test_vint(5);
         aux_test_vint(u64::max_value());
         for i in 1..9 {
-            let power_of_128 = 1u64 << (7 * i);
+            let power_of_128 = 1u64 << (7*i);
             aux_test_vint(power_of_128 - 1u64);
-            aux_test_vint(power_of_128);
+            aux_test_vint(power_of_128 );
             aux_test_vint(power_of_128 + 1u64);
         }
         aux_test_vint(10);
     }
-
-    fn aux_test_serialize_vint_u32(val: u32) {
-        let mut buffer = [0u8; 10];
-        let mut buffer2 = [0u8; 10];
-        let len_vint = VInt(val as u64).serialize_into(&mut buffer);
-        let (vint, len) = serialize_vint_u32(val);
-        assert_eq!(len, len_vint, "len wrong for val {}", val);
-        LittleEndian::write_u64(&mut buffer2, vint);
-        assert_eq!(&buffer[..len], &buffer2[..len], "array wrong for {}", val);
-    }
-
-    #[test]
-    fn test_vint_u32() {
-        aux_test_serialize_vint_u32(0);
-        aux_test_serialize_vint_u32(1);
-        aux_test_serialize_vint_u32(5);
-        for i in 1..3 {
-            let power_of_128 = 1u32 << (7 * i);
-            aux_test_serialize_vint_u32(power_of_128 - 1u32);
-            aux_test_serialize_vint_u32(power_of_128);
-            aux_test_serialize_vint_u32(power_of_128 + 1u32);
-        }
-        aux_test_serialize_vint_u32(u32::max_value());
-    }
-}
+}

src/core/executor.rs

@@ -1,137 +0,0 @@
-use crossbeam::channel;
-use scoped_pool::{Pool, ThreadConfig};
-use Result;
-
-/// Search executor whether search request are single thread or multithread.
-///
-/// We don't expose Rayon thread pool directly here for several reasons.
-///
-/// First dependency hell. It is not a good idea to expose the
-/// API of a dependency, knowing it might conflict with a different version
-/// used by the client. Second, we may stop using rayon in the future.
-pub enum Executor {
-    SingleThread,
-    ThreadPool(Pool),
-}
-
-impl Executor {
-    /// Creates an Executor that performs all task in the caller thread.
-    pub fn single_thread() -> Executor {
-        Executor::SingleThread
-    }
-
-    // Creates an Executor that dispatches the tasks in a thread pool.
-    pub fn multi_thread(num_threads: usize, prefix: &'static str) -> Executor {
-        let thread_config = ThreadConfig::new().prefix(prefix);
-        let pool = Pool::with_thread_config(num_threads, thread_config);
-        Executor::ThreadPool(pool)
-    }
-
-    // Perform a map in the thread pool.
-    //
-    // Regardless of the executor (`SingleThread` or `ThreadPool`), panics in the task
-    // will propagate to the caller.
-    pub fn map<
-        A: Send,
-        R: Send,
-        AIterator: Iterator<Item = A>,
-        F: Sized + Sync + Fn(A) -> Result<R>,
-    >(
-        &self,
-        f: F,
-        args: AIterator,
-    ) -> Result<Vec<R>> {
-        match self {
-            Executor::SingleThread => args.map(f).collect::<Result<_>>(),
-            Executor::ThreadPool(pool) => {
-                let args_with_indices: Vec<(usize, A)> = args.enumerate().collect();
-                let num_fruits = args_with_indices.len();
-                let fruit_receiver = {
-                    let (fruit_sender, fruit_receiver) = channel::unbounded();
-                    pool.scoped(|scope| {
-                        for arg_with_idx in args_with_indices {
-                            scope.execute(|| {
-                                let (idx, arg) = arg_with_idx;
-                                let fruit = f(arg);
-                                if let Err(err) = fruit_sender.send((idx, fruit)) {
-                                    error!("Failed to send search task. It probably means all search threads have panicked. {:?}", err);
-                                }
-                            });
-                        }
-                    });
-                    fruit_receiver
-                    // This ends the scope of fruit_sender.
-                    // This is important as it makes it possible for the fruit_receiver iteration to
-                    // terminate.
-                };
-                // This is lame, but safe.
-                let mut results_with_position = Vec::with_capacity(num_fruits);
-                for (pos, fruit_res) in fruit_receiver {
-                    let fruit = fruit_res?;
-                    results_with_position.push((pos, fruit));
-                }
-                results_with_position.sort_by_key(|(pos, _)| *pos);
-                assert_eq!(results_with_position.len(), num_fruits);
-                Ok(results_with_position
-                    .into_iter()
-                    .map(|(_, fruit)| fruit)
-                    .collect::<Vec<_>>())
-            }
-        }
-    }
-}
-
-#[cfg(test)]
-mod tests {
-
-    use super::Executor;
-
-    #[test]
-    #[should_panic(expected = "panic should propagate")]
-    fn test_panic_propagates_single_thread() {
-        let _result: Vec<usize> = Executor::single_thread()
-            .map(
-                |_| {
-                    panic!("panic should propagate");
-                },
-                vec![0].into_iter(),
-            )
-            .unwrap();
-    }
-
-    #[test]
-    #[should_panic] //< unfortunately the panic message is not propagated
-    fn test_panic_propagates_multi_thread() {
-        let _result: Vec<usize> = Executor::multi_thread(1, "search-test")
-            .map(
-                |_| {
-                    panic!("panic should propagate");
-                },
-                vec![0].into_iter(),
-            )
-            .unwrap();
-    }
-
-    #[test]
-    fn test_map_singlethread() {
-        let result: Vec<usize> = Executor::single_thread()
-            .map(|i| Ok(i * 2), 0..1_000)
-            .unwrap();
-        assert_eq!(result.len(), 1_000);
-        for i in 0..1_000 {
-            assert_eq!(result[i], i * 2);
-        }
-    }
-
-}
-
-#[test]
-fn test_map_multithread() {
-    let result: Vec<usize> = Executor::multi_thread(3, "search-test")
-        .map(|i| Ok(i * 2), 0..10)
-        .unwrap();
-    assert_eq!(result.len(), 10);
-    for i in 0..10 {
-        assert_eq!(result[i], i * 2);
-    }
-}

src/core/index.rs

@@ -1,51 +1,40 @@
+use core::SegmentId;
+use error::TantivyError;
+use schema::Schema;
+use serde_json;
+use std::borrow::BorrowMut;
+use std::fmt;
+use std::sync::atomic::{AtomicUsize, Ordering};
+use std::sync::Arc;
+use Result;
+
 use super::pool::LeasedItem;
 use super::pool::Pool;
 use super::segment::create_segment;
 use super::segment::Segment;
 use core::searcher::Searcher;
-use core::Executor;
 use core::IndexMeta;
-use core::SegmentId;
 use core::SegmentMeta;
 use core::SegmentReader;
 use core::META_FILEPATH;
-use directory::ManagedDirectory;
 #[cfg(feature = "mmap")]
 use directory::MmapDirectory;
-use directory::INDEX_WRITER_LOCK;
-use directory::META_LOCK;
 use directory::{Directory, RAMDirectory};
-use error::DataCorruption;
-use error::TantivyError;
+use directory::{DirectoryClone, ManagedDirectory};
 use indexer::index_writer::open_index_writer;
 use indexer::index_writer::HEAP_SIZE_MIN;
 use indexer::segment_updater::save_new_metas;
+use indexer::DirectoryLock;
 use num_cpus;
-use schema::Field;
-use schema::FieldType;
-use schema::Schema;
-use serde_json;
-use std::borrow::BorrowMut;
-use std::fmt;
 use std::path::Path;
-use std::sync::atomic::{AtomicUsize, Ordering};
-use std::sync::Arc;
-use tokenizer::BoxedTokenizer;
 use tokenizer::TokenizerManager;
 use IndexWriter;
-use Result;
 
 fn load_metas(directory: &Directory) -> Result<IndexMeta> {
     let meta_data = directory.atomic_read(&META_FILEPATH)?;
     let meta_string = String::from_utf8_lossy(&meta_data);
     serde_json::from_str(&meta_string)
-        .map_err(|e| {
-            DataCorruption::new(
-                META_FILEPATH.clone(),
-                format!("Meta file cannot be deserialized. {:?}.", e),
-            )
-        })
-        .map_err(From::from)
+        .map_err(|_| TantivyError::CorruptedFile(META_FILEPATH.clone()))
 }
 
 /// Search Index
@@ -54,39 +43,10 @@ pub struct Index {
     schema: Schema,
     num_searchers: Arc<AtomicUsize>,
     searcher_pool: Arc<Pool<Searcher>>,
-    executor: Arc<Executor>,
     tokenizers: TokenizerManager,
 }
 
 impl Index {
-    /// Examines the director to see if it contains an index
-    pub fn exists<Dir: Directory>(dir: &Dir) -> bool {
-        dir.exists(&META_FILEPATH)
-    }
-
-    /// Accessor to the search executor.
-    ///
-    /// This pool is used by default when calling `searcher.search(...)`
-    /// to perform search on the individual segments.
-    ///
-    /// By default the executor is single thread, and simply runs in the calling thread.
-    pub fn search_executor(&self) -> &Executor {
-        self.executor.as_ref()
-    }
-
-    /// Replace the default single thread search executor pool
-    /// by a thread pool with a given number of threads.
-    pub fn set_multithread_executor(&mut self, num_threads: usize) {
-        self.executor = Arc::new(Executor::multi_thread(num_threads, "thrd-tantivy-search-"));
-    }
-
-    /// Replace the default single thread search executor pool
-    /// by a thread pool with a given number of threads.
-    pub fn set_default_multithread_executor(&mut self) {
-        let default_num_threads = num_cpus::get();
-        self.set_multithread_executor(default_num_threads);
-    }
-
     /// Creates a new index using the `RAMDirectory`.
     ///
     /// The index will be allocated in anonymous memory.
@@ -103,30 +63,9 @@ impl Index {
     #[cfg(feature = "mmap")]
     pub fn create_in_dir<P: AsRef<Path>>(directory_path: P, schema: Schema) -> Result<Index> {
         let mmap_directory = MmapDirectory::open(directory_path)?;
-        if Index::exists(&mmap_directory) {
-            return Err(TantivyError::IndexAlreadyExists);
-        }
-
         Index::create(mmap_directory, schema)
     }
 
-    /// Opens or creates a new index in the provided directory
-    #[cfg(feature = "mmap")]
-    pub fn open_or_create<Dir: Directory>(dir: Dir, schema: Schema) -> Result<Index> {
-        if Index::exists(&dir) {
-            let index = Index::open(dir)?;
-            if index.schema() == schema {
-                Ok(index)
-            } else {
-                Err(TantivyError::SchemaError(
-                    "An index exists but the schema does not match.".to_string(),
-                ))
-            }
-        } else {
-            Index::create(dir, schema)
-        }
-    }
-
     /// Creates a new index in a temp directory.
     ///
     /// The index will use the `MMapDirectory` in a newly created directory.
@@ -143,15 +82,13 @@ impl Index {
 
     /// Creates a new index given an implementation of the trait `Directory`
     pub fn create<Dir: Directory>(dir: Dir, schema: Schema) -> Result<Index> {
-        let directory = ManagedDirectory::wrap(dir)?;
+        let directory = ManagedDirectory::new(dir)?;
         Index::from_directory(directory, schema)
     }
 
     /// Create a new index from a directory.
-    ///
-    /// This will overwrite existing meta.json
     fn from_directory(mut directory: ManagedDirectory, schema: Schema) -> Result<Index> {
-        save_new_metas(schema.clone(), directory.borrow_mut())?;
+        save_new_metas(schema.clone(), 0, directory.borrow_mut())?;
         let metas = IndexMeta::with_schema(schema);
         Index::create_from_metas(directory, &metas)
     }
@@ -166,7 +103,6 @@ impl Index {
             num_searchers: Arc::new(AtomicUsize::new(n_cpus)),
             searcher_pool: Arc::new(Pool::new()),
             tokenizers: TokenizerManager::default(),
-            executor: Arc::new(Executor::single_thread()),
         };
         index.load_searchers()?;
         Ok(index)
@@ -177,27 +113,6 @@ impl Index {
         &self.tokenizers
     }
 
-    /// Helper to access the tokenizer associated to a specific field.
-    pub fn tokenizer_for_field(&self, field: Field) -> Result<Box<BoxedTokenizer>> {
-        let field_entry = self.schema.get_field_entry(field);
-        let field_type = field_entry.field_type();
-        let tokenizer_manager: &TokenizerManager = self.tokenizers();
-        let tokenizer_name_opt: Option<Box<BoxedTokenizer>> = match field_type {
-            FieldType::Str(text_options) => text_options
-                .get_indexing_options()
-                .map(|text_indexing_options| text_indexing_options.tokenizer().to_string())
-                .and_then(|tokenizer_name| tokenizer_manager.get(&tokenizer_name)),
-            _ => None,
-        };
-        match tokenizer_name_opt {
-            Some(tokenizer) => Ok(tokenizer),
-            None => Err(TantivyError::SchemaError(format!(
-                "{:?} is not a text field.",
-                field_entry.name()
-            ))),
-        }
-    }
-
     /// Opens a new directory from an index path.
     #[cfg(feature = "mmap")]
     pub fn open_in_dir<P: AsRef<Path>>(directory_path: P) -> Result<Index> {
@@ -207,7 +122,7 @@ impl Index {
 
     /// Open the index using the provided directory
     pub fn open<D: Directory>(directory: D) -> Result<Index> {
-        let directory = ManagedDirectory::wrap(directory)?;
+        let directory = ManagedDirectory::new(directory)?;
         let metas = load_metas(&directory)?;
         Index::create_from_metas(directory, &metas)
     }
@@ -233,8 +148,7 @@ impl Index {
     /// Each thread will receive a budget of  `overall_heap_size_in_bytes / num_threads`.
     ///
     /// # Errors
-    /// If the lockfile already exists, returns `Error::DirectoryLockBusy` or an `Error::IOError`.
-    ///
+    /// If the lockfile already exists, returns `Error::FileAlreadyExists`.
     /// # Panics
     /// If the heap size per thread is too small, panics.
     pub fn writer_with_num_threads(
@@ -242,21 +156,7 @@ impl Index {
         num_threads: usize,
         overall_heap_size_in_bytes: usize,
     ) -> Result<IndexWriter> {
-        let directory_lock = self
-            .directory
-            .acquire_lock(&INDEX_WRITER_LOCK)
-            .map_err(|err| {
-                TantivyError::LockFailure(
-                    err,
-                    Some(
-                        "Failed to acquire index lock. If you are using\
-                         a regular directory, this means there is already an \
-                         `IndexWriter` working on this `Directory`, in this process \
-                         or in a different process."
-                            .to_string(),
-                    ),
-                )
-            })?;
+        let directory_lock = DirectoryLock::lock(self.directory().box_clone())?;
         let heap_size_in_bytes_per_thread = overall_heap_size_in_bytes / num_threads;
         open_index_writer(
             self,
@@ -294,8 +194,7 @@ impl Index {
 
     /// Returns the list of segments that are searchable
     pub fn searchable_segments(&self) -> Result<Vec<Segment>> {
-        Ok(self
-            .searchable_segment_metas()?
+        Ok(self.searchable_segment_metas()?
             .into_iter()
             .map(|segment_meta| self.segment(segment_meta))
             .collect())
@@ -330,8 +229,7 @@ impl Index {
 
     /// Returns the list of segment ids that are searchable.
     pub fn searchable_segment_ids(&self) -> Result<Vec<SegmentId>> {
-        Ok(self
-            .searchable_segment_metas()?
+        Ok(self.searchable_segment_metas()?
             .iter()
             .map(|segment_meta| segment_meta.id())
             .collect())
@@ -344,18 +242,13 @@ impl Index {
         self.num_searchers.store(num_searchers, Ordering::Release);
     }
 
-    /// Update searchers so that they reflect the state of the last
-    /// `.commit()`.
+    /// Creates a new generation of searchers after
+
+    /// a change of the set of searchable indexes.
     ///
-    /// If indexing happens in the same process as searching,
-    /// you most likely want to call `.load_searchers()` right after each
-    /// successful call to `.commit()`.
-    ///
-    /// If indexing and searching happen in different processes, the way to
-    /// get the freshest `index` at all time, is to watch `meta.json` and
-    /// call `load_searchers` whenever a changes happen.
+    /// This needs to be called when a new segment has been
+    /// published or after a merge.
     pub fn load_searchers(&self) -> Result<()> {
-        let _meta_lock = self.directory().acquire_lock(&META_LOCK)?;
         let searchable_segments = self.searchable_segments()?;
         let segment_readers: Vec<SegmentReader> = searchable_segments
             .iter()
@@ -364,7 +257,7 @@ impl Index {
         let schema = self.schema();
         let num_searchers: usize = self.num_searchers.load(Ordering::Acquire);
         let searchers = (0..num_searchers)
-            .map(|_| Searcher::new(schema.clone(), self.clone(), segment_readers.clone()))
+            .map(|_| Searcher::new(schema.clone(), segment_readers.clone()))
             .collect();
         self.searcher_pool.publish_new_generation(searchers);
         Ok(())
@@ -399,79 +292,6 @@ impl Clone for Index {
             num_searchers: Arc::clone(&self.num_searchers),
             searcher_pool: Arc::clone(&self.searcher_pool),
             tokenizers: self.tokenizers.clone(),
-            executor: self.executor.clone(),
         }
     }
 }
-
-#[cfg(test)]
-mod tests {
-    use directory::RAMDirectory;
-    use schema::{Schema, INT_INDEXED, TEXT};
-    use Index;
-
-    #[test]
-    fn test_indexer_for_field() {
-        let mut schema_builder = Schema::builder();
-        let num_likes_field = schema_builder.add_u64_field("num_likes", INT_INDEXED);
-        let body_field = schema_builder.add_text_field("body", TEXT);
-        let schema = schema_builder.build();
-        let index = Index::create_in_ram(schema);
-        assert!(index.tokenizer_for_field(body_field).is_ok());
-        assert_eq!(
-            format!("{:?}", index.tokenizer_for_field(num_likes_field).err()),
-            "Some(SchemaError(\"\\\"num_likes\\\" is not a text field.\"))"
-        );
-    }
-
-    #[test]
-    fn test_index_exists() {
-        let directory = RAMDirectory::create();
-        assert!(!Index::exists(&directory));
-        assert!(Index::create(directory.clone(), throw_away_schema()).is_ok());
-        assert!(Index::exists(&directory));
-    }
-
-    #[test]
-    fn open_or_create_should_create() {
-        let directory = RAMDirectory::create();
-        assert!(!Index::exists(&directory));
-        assert!(Index::open_or_create(directory.clone(), throw_away_schema()).is_ok());
-        assert!(Index::exists(&directory));
-    }
-
-    #[test]
-    fn open_or_create_should_open() {
-        let directory = RAMDirectory::create();
-        assert!(Index::create(directory.clone(), throw_away_schema()).is_ok());
-        assert!(Index::exists(&directory));
-        assert!(Index::open_or_create(directory, throw_away_schema()).is_ok());
-    }
-
-    #[test]
-    fn create_should_wipeoff_existing() {
-        let directory = RAMDirectory::create();
-        assert!(Index::create(directory.clone(), throw_away_schema()).is_ok());
-        assert!(Index::exists(&directory));
-        assert!(Index::create(directory.clone(), Schema::builder().build()).is_ok());
-    }
-
-    #[test]
-    fn open_or_create_exists_but_schema_does_not_match() {
-        let directory = RAMDirectory::create();
-        assert!(Index::create(directory.clone(), throw_away_schema()).is_ok());
-        assert!(Index::exists(&directory));
-        assert!(Index::open_or_create(directory.clone(), throw_away_schema()).is_ok());
-        let err = Index::open_or_create(directory, Schema::builder().build());
-        assert_eq!(
-            format!("{:?}", err.unwrap_err()),
-            "SchemaError(\"An index exists but the schema does not match.\")"
-        );
-    }
-
-    fn throw_away_schema() -> Schema {
-        let mut schema_builder = Schema::builder();
-        let _ = schema_builder.add_u64_field("num_likes", INT_INDEXED);
-        schema_builder.build()
-    }
-}

src/core/index_meta.rs

@@ -46,19 +46,19 @@ impl fmt::Debug for IndexMeta {
 mod tests {
 
     use super::IndexMeta;
-    use schema::{Schema, TEXT};
+    use schema::{SchemaBuilder, TEXT};
     use serde_json;
 
     #[test]
     fn test_serialize_metas() {
         let schema = {
-            let mut schema_builder = Schema::builder();
+            let mut schema_builder = SchemaBuilder::new();
             schema_builder.add_text_field("text", TEXT);
             schema_builder.build()
         };
         let index_metas = IndexMeta {
             segments: Vec::new(),
-            schema,
+            schema: schema,
             opstamp: 0u64,
             payload: None,
         };

src/core/inverted_index_reader.rs

@@ -1,13 +1,13 @@
 use common::BinarySerializable;
 use directory::ReadOnlySource;
-use owned_read::OwnedRead;
-use positions::PositionReader;
 use postings::TermInfo;
 use postings::{BlockSegmentPostings, SegmentPostings};
 use schema::FieldType;
 use schema::IndexRecordOption;
 use schema::Term;
 use termdict::TermDictionary;
+use owned_read::OwnedRead;
+use positions::PositionReader;
 
 /// The inverted index reader is in charge of accessing
 /// the inverted index associated to a specific field.
@@ -32,7 +32,6 @@ pub struct InvertedIndexReader {
 }
 
 impl InvertedIndexReader {
-    #[cfg_attr(feature = "cargo-clippy", allow(clippy::needless_pass_by_value))] // for symetry
     pub(crate) fn new(
         termdict: TermDictionary,
         postings_source: ReadOnlySource,
@@ -55,12 +54,12 @@ impl InvertedIndexReader {
 
     /// Creates an empty `InvertedIndexReader` object, which
     /// contains no terms at all.
-    pub fn empty(field_type: &FieldType) -> InvertedIndexReader {
+    pub fn empty(field_type: FieldType) -> InvertedIndexReader {
         let record_option = field_type
             .get_index_record_option()
             .unwrap_or(IndexRecordOption::Basic);
         InvertedIndexReader {
-            termdict: TermDictionary::empty(&field_type),
+            termdict: TermDictionary::empty(field_type),
             postings_source: ReadOnlySource::empty(),
             positions_source: ReadOnlySource::empty(),
             positions_idx_source: ReadOnlySource::empty(),
@@ -101,6 +100,7 @@ impl InvertedIndexReader {
         block_postings.reset(term_info.doc_freq, postings_reader);
     }
 
+
     /// Returns a block postings given a `Term`.
     /// This method is for an advanced usage only.
     ///
@@ -111,7 +111,7 @@ impl InvertedIndexReader {
         option: IndexRecordOption,
     ) -> Option<BlockSegmentPostings> {
         self.get_term_info(term)
-            .map(move |term_info| self.read_block_postings_from_terminfo(&term_info, option))
+            .map(move|term_info| self.read_block_postings_from_terminfo(&term_info, option))
     }
 
     /// Returns a block postings given a `term_info`.
@@ -147,8 +147,7 @@ impl InvertedIndexReader {
             if option.has_positions() {
                 let position_reader = self.positions_source.clone();
                 let skip_reader = self.positions_idx_source.clone();
-                let position_reader =
-                    PositionReader::new(position_reader, skip_reader, term_info.positions_idx);
+                let position_reader = PositionReader::new(position_reader, skip_reader, term_info.positions_idx);
                 Some(position_reader)
             } else {
                 None

src/core/mod.rs

@@ -1,4 +1,3 @@
-mod executor;
 pub mod index;
 mod index_meta;
 mod inverted_index_reader;
@@ -10,7 +9,6 @@ mod segment_id;
 mod segment_meta;
 mod segment_reader;
 
-pub use self::executor::Executor;
 pub use self::index::Index;
 pub use self::index_meta::IndexMeta;
 pub use self::inverted_index_reader::InvertedIndexReader;
@@ -35,4 +33,10 @@ lazy_static! {
     /// Removing this file is safe, but will prevent the garbage collection of all of the file that
     /// are currently in the directory
     pub static ref MANAGED_FILEPATH: PathBuf = PathBuf::from(".managed.json");
+
+    /// Only one process should be able to write tantivy's index at a time.
+    /// This file, when present, is in charge of preventing other processes to open an IndexWriter.
+    ///
+    /// If the process is killed and this file remains, it is safe to remove it manually.
+    pub static ref LOCKFILE_FILEPATH: PathBuf = PathBuf::from(".tantivy-indexer.lock");
 }

src/core/pool.rs

@@ -87,8 +87,7 @@ impl<T> Deref for LeasedItem<T> {
     type Target = T;
 
     fn deref(&self) -> &T {
-        &self
-            .gen_item
+        &self.gen_item
             .as_ref()
             .expect("Unwrapping a leased item should never fail")
             .item // unwrap is safe here
@@ -97,8 +96,7 @@ impl<T> Deref for LeasedItem<T> {
 
 impl<T> DerefMut for LeasedItem<T> {
     fn deref_mut(&mut self) -> &mut T {
-        &mut self
-            .gen_item
+        &mut self.gen_item
             .as_mut()
             .expect("Unwrapping a mut leased item should never fail")
             .item // unwrap is safe here

src/core/searcher.rs

@@ -1,43 +1,16 @@
 use collector::Collector;
-use collector::SegmentCollector;
-use core::Executor;
 use core::InvertedIndexReader;
 use core::SegmentReader;
 use query::Query;
-use query::Scorer;
-use query::Weight;
 use schema::Document;
 use schema::Schema;
 use schema::{Field, Term};
-use space_usage::SearcherSpaceUsage;
 use std::fmt;
 use std::sync::Arc;
-use store::StoreReader;
 use termdict::TermMerger;
 use DocAddress;
-use Index;
 use Result;
 
-fn collect_segment<C: Collector>(
-    collector: &C,
-    weight: &Weight,
-    segment_ord: u32,
-    segment_reader: &SegmentReader,
-) -> Result<C::Fruit> {
-    let mut scorer = weight.scorer(segment_reader)?;
-    let mut segment_collector = collector.for_segment(segment_ord as u32, segment_reader)?;
-    if let Some(delete_bitset) = segment_reader.delete_bitset() {
-        scorer.for_each(&mut |doc, score| {
-            if !delete_bitset.is_deleted(doc) {
-                segment_collector.collect(doc, score);
-            }
-        });
-    } else {
-        scorer.for_each(&mut |doc, score| segment_collector.collect(doc, score));
-    }
-    Ok(segment_collector.harvest())
-}
-
 /// Holds a list of `SegmentReader`s ready for search.
 ///
 /// It guarantees that the `Segment` will not be removed before
@@ -45,43 +18,25 @@ fn collect_segment<C: Collector>(
 ///
 pub struct Searcher {
     schema: Schema,
-    index: Index,
     segment_readers: Vec<SegmentReader>,
-    store_readers: Vec<StoreReader>,
 }
 
 impl Searcher {
     /// Creates a new `Searcher`
-    pub(crate) fn new(
-        schema: Schema,
-        index: Index,
-        segment_readers: Vec<SegmentReader>,
-    ) -> Searcher {
-        let store_readers = segment_readers
-            .iter()
-            .map(|segment_reader| segment_reader.get_store_reader())
-            .collect();
+    pub(crate) fn new(schema: Schema, segment_readers: Vec<SegmentReader>) -> Searcher {
         Searcher {
             schema,
-            index,
             segment_readers,
-            store_readers,
         }
     }
-
-    /// Returns the `Index` associated to the `Searcher`
-    pub fn index(&self) -> &Index {
-        &self.index
-    }
-
     /// Fetches a document from tantivy's store given a `DocAddress`.
     ///
     /// The searcher uses the segment ordinal to route the
     /// the request to the right `Segment`.
-    pub fn doc(&self, doc_address: DocAddress) -> Result<Document> {
-        let DocAddress(segment_local_id, doc_id) = doc_address;
-        let store_reader = &self.store_readers[segment_local_id as usize];
-        store_reader.get(doc_id)
+    pub fn doc(&self, doc_address: &DocAddress) -> Result<Document> {
+        let DocAddress(segment_local_id, doc_id) = *doc_address;
+        let segment_reader = &self.segment_readers[segment_local_id as usize];
+        segment_reader.doc(doc_id)
     }
 
     /// Access the schema associated to the index of this searcher.
@@ -93,7 +48,7 @@ impl Searcher {
     pub fn num_docs(&self) -> u64 {
         self.segment_readers
             .iter()
-            .map(|segment_reader| u64::from(segment_reader.num_docs()))
+            .map(|segment_reader| segment_reader.num_docs() as u64)
             .sum::<u64>()
     }
 
@@ -102,9 +57,7 @@ impl Searcher {
     pub fn doc_freq(&self, term: &Term) -> u64 {
         self.segment_readers
             .iter()
-            .map(|segment_reader| {
-                u64::from(segment_reader.inverted_index(term.field()).doc_freq(term))
-            })
+            .map(|segment_reader| segment_reader.inverted_index(term.field()).doc_freq(term) as u64)
             .sum::<u64>()
     }
 
@@ -118,78 +71,19 @@ impl Searcher {
         &self.segment_readers[segment_ord as usize]
     }
 
-    /// Runs a query on the segment readers wrapped by the searcher.
-    ///
-    /// Search works as follows :
-    ///
-    ///  First the weight object associated to the query is created.
-    ///
-    ///  Then, the query loops over the segments and for each segment :
-    ///  - setup the collector and informs it that the segment being processed has changed.
-    ///  - creates a SegmentCollector for collecting documents associated to the segment
-    ///  - creates a `Scorer` object associated for this segment
-    ///  - iterate through the matched documents and push them to the segment collector.
-    ///
-    ///  Finally, the Collector merges each of the child collectors into itself for result usability
-    ///  by the caller.
-    pub fn search<C: Collector>(&self, query: &Query, collector: &C) -> Result<C::Fruit> {
-        let executor = self.index.search_executor();
-        self.search_with_executor(query, collector, executor)
-    }
-
-    /// Same as [`search(...)`](#method.search) but multithreaded.
-    ///
-    /// The current implementation is rather naive :
-    /// multithreading is by splitting search into as many task
-    /// as there are segments.
-    ///
-    /// It is powerless at making search faster if your index consists in
-    /// one large segment.
-    ///
-    /// Also, keep in my multithreading a single query on several
-    /// threads will not improve your throughput. It can actually
-    /// hurt it. It will however, decrease the average response time.
-    pub fn search_with_executor<C: Collector>(
-        &self,
-        query: &Query,
-        collector: &C,
-        executor: &Executor,
-    ) -> Result<C::Fruit> {
-        let scoring_enabled = collector.requires_scoring();
-        let weight = query.weight(self, scoring_enabled)?;
-        let segment_readers = self.segment_readers();
-        let fruits = executor.map(
-            |(segment_ord, segment_reader)| {
-                collect_segment(
-                    collector,
-                    weight.as_ref(),
-                    segment_ord as u32,
-                    segment_reader,
-                )
-            },
-            segment_readers.iter().enumerate(),
-        )?;
-        collector.merge_fruits(fruits)
+    /// Runs a query on the segment readers wrapped by the searcher
+    pub fn search<C: Collector>(&self, query: &Query, collector: &mut C) -> Result<()> {
+        query.search(self, collector)
     }
 
     /// Return the field searcher associated to a `Field`.
     pub fn field(&self, field: Field) -> FieldSearcher {
-        let inv_index_readers = self
-            .segment_readers
+        let inv_index_readers = self.segment_readers
             .iter()
             .map(|segment_reader| segment_reader.inverted_index(field))
             .collect::<Vec<_>>();
         FieldSearcher::new(inv_index_readers)
     }
-
-    /// Summarize total space usage of this searcher.
-    pub fn space_usage(&self) -> SearcherSpaceUsage {
-        let mut space_usage = SearcherSpaceUsage::new();
-        for segment_reader in self.segment_readers.iter() {
-            space_usage.add_segment(segment_reader.space_usage());
-        }
-        space_usage
-    }
 }
 
 pub struct FieldSearcher {
@@ -204,8 +98,7 @@ impl FieldSearcher {
     /// Returns a Stream over all of the sorted unique terms of
     /// for the given field.
     pub fn terms(&self) -> TermMerger {
-        let term_streamers: Vec<_> = self
-            .inv_index_readers
+        let term_streamers: Vec<_> = self.inv_index_readers
             .iter()
             .map(|inverted_index| inverted_index.terms().stream())
             .collect();
@@ -215,8 +108,7 @@ impl FieldSearcher {
 
 impl fmt::Debug for Searcher {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        let segment_ids = self
-            .segment_readers
+        let segment_ids = self.segment_readers
             .iter()
             .map(|segment_reader| segment_reader.segment_id())
             .collect::<Vec<_>>();

src/core/segment_component.rs

@@ -41,6 +41,6 @@ impl SegmentComponent {
             SegmentComponent::STORE,
             SegmentComponent::DELETE,
         ];
-        SEGMENT_COMPONENTS.iter()
+        SEGMENT_COMPONENTS.into_iter()
     }
 }

src/core/segment_id.rs

@@ -52,12 +52,12 @@ impl SegmentId {
     /// Picking the first 8 chars is ok to identify
     /// segments in a display message.
     pub fn short_uuid_string(&self) -> String {
-        (&self.0.to_simple_ref().to_string()[..8]).to_string()
+        (&self.0.simple().to_string()[..8]).to_string()
     }
 
     /// Returns a segment uuid string.
     pub fn uuid_string(&self) -> String {
-        self.0.to_simple_ref().to_string()
+        self.0.simple().to_string()
     }
 }
 

src/core/segment_meta.rs

@@ -50,7 +50,7 @@ impl<'a> serde::Deserialize<'a> for SegmentMeta {
     {
         let inner = InnerSegmentMeta::deserialize(deserializer)?;
         let tracked = INVENTORY.track(inner);
-        Ok(SegmentMeta { tracked })
+        Ok(SegmentMeta { tracked: tracked })
     }
 }
 

src/core/segment_reader.rs

@@ -4,7 +4,7 @@ use core::InvertedIndexReader;
 use core::Segment;
 use core::SegmentComponent;
 use core::SegmentId;
-use directory::ReadOnlySource;
+use core::SegmentMeta;
 use error::TantivyError;
 use fastfield::DeleteBitSet;
 use fastfield::FacetReader;
@@ -13,10 +13,10 @@ use fastfield::{self, FastFieldNotAvailableError};
 use fastfield::{BytesFastFieldReader, FastValue, MultiValueIntFastFieldReader};
 use fieldnorm::FieldNormReader;
 use schema::Cardinality;
+use schema::Document;
 use schema::Field;
 use schema::FieldType;
 use schema::Schema;
-use space_usage::SegmentSpaceUsage;
 use std::collections::HashMap;
 use std::fmt;
 use std::sync::Arc;
@@ -44,8 +44,7 @@ pub struct SegmentReader {
     inv_idx_reader_cache: Arc<RwLock<HashMap<Field, Arc<InvertedIndexReader>>>>,
 
     segment_id: SegmentId,
-    max_doc: DocId,
-    num_docs: DocId,
+    segment_meta: SegmentMeta,
 
     termdict_composite: CompositeFile,
     postings_composite: CompositeFile,
@@ -54,7 +53,7 @@ pub struct SegmentReader {
     fast_fields_composite: CompositeFile,
     fieldnorms_composite: CompositeFile,
 
-    store_source: ReadOnlySource,
+    store_reader: StoreReader,
     delete_bitset_opt: Option<DeleteBitSet>,
     schema: Schema,
 }
@@ -65,7 +64,7 @@ impl SegmentReader {
     /// Today, `tantivy` does not handle deletes, so it happens
     /// to also be the number of documents in the index.
     pub fn max_doc(&self) -> DocId {
-        self.max_doc
+        self.segment_meta.max_doc()
     }
 
     /// Returns the number of documents.
@@ -74,7 +73,7 @@ impl SegmentReader {
     /// Today, `tantivy` does not handle deletes so max doc and
     /// num_docs are the same.
     pub fn num_docs(&self) -> DocId {
-        self.num_docs
+        self.segment_meta.num_docs()
     }
 
     /// Returns the schema of the index this segment belongs to.
@@ -154,17 +153,15 @@ impl SegmentReader {
     /// Accessor to the `BytesFastFieldReader` associated to a given `Field`.
     pub fn bytes_fast_field_reader(&self, field: Field) -> fastfield::Result<BytesFastFieldReader> {
         let field_entry = self.schema.get_field_entry(field);
-        match *field_entry.field_type() {
-            FieldType::Bytes => {}
+        match field_entry.field_type() {
+            &FieldType::Bytes => {}
             _ => return Err(FastFieldNotAvailableError::new(field_entry)),
         }
-        let idx_reader = self
-            .fast_fields_composite
+        let idx_reader = self.fast_fields_composite
             .open_read_with_idx(field, 0)
             .ok_or_else(|| FastFieldNotAvailableError::new(field_entry))
             .map(FastFieldReader::open)?;
-        let values = self
-            .fast_fields_composite
+        let values = self.fast_fields_composite
             .open_read_with_idx(field, 1)
             .ok_or_else(|| FastFieldNotAvailableError::new(field_entry))?;
         Ok(BytesFastFieldReader::open(idx_reader, values))
@@ -178,7 +175,7 @@ impl SegmentReader {
                 "The field {:?} is not a \
                  hierarchical facet.",
                 field_entry
-            )));
+            )).into());
         }
         let term_ords_reader = self.multi_fast_field_reader(field)?;
         let termdict_source = self.termdict_composite.open_read(field).ok_or_else(|| {
@@ -189,7 +186,7 @@ impl SegmentReader {
                 field_entry.name()
             ))
         })?;
-        let termdict = TermDictionary::from_source(&termdict_source);
+        let termdict = TermDictionary::from_source(termdict_source);
         let facet_reader = FacetReader::new(term_ords_reader, termdict);
         Ok(facet_reader)
     }
@@ -197,7 +194,8 @@ impl SegmentReader {
     /// Accessor to the segment's `Field norms`'s reader.
     ///
     /// Field norms are the length (in tokens) of the fields.
-    /// It is used in the computation of the [TfIdf](https://fulmicoton.gitbooks.io/tantivy-doc/content/tfidf.html).
+    /// It is used in the computation of the [TfIdf]
+    /// (https://fulmicoton.gitbooks.io/tantivy-doc/content/tfidf.html).
     ///
     /// They are simply stored as a fast field, serialized in
     /// the `.fieldnorm` file of the segment.
@@ -215,8 +213,8 @@ impl SegmentReader {
     }
 
     /// Accessor to the segment's `StoreReader`.
-    pub fn get_store_reader(&self) -> StoreReader {
-        StoreReader::from_source(self.store_source.clone())
+    pub fn get_store_reader(&self) -> &StoreReader {
+        &self.store_reader
     }
 
     /// Open a new segment for reading.
@@ -225,8 +223,7 @@ impl SegmentReader {
         let termdict_composite = CompositeFile::open(&termdict_source)?;
 
         let store_source = segment.open_read(SegmentComponent::STORE)?;
-
-        fail_point!("SegmentReader::open#middle");
+        let store_reader = StoreReader::from_source(store_source);
 
         let postings_source = segment.open_read(SegmentComponent::POSTINGS)?;
         let postings_composite = CompositeFile::open(&postings_source)?;
@@ -263,14 +260,13 @@ impl SegmentReader {
         let schema = segment.schema();
         Ok(SegmentReader {
             inv_idx_reader_cache: Arc::new(RwLock::new(HashMap::new())),
-            max_doc: segment.meta().max_doc(),
-            num_docs: segment.meta().num_docs(),
+            segment_meta: segment.meta().clone(),
             termdict_composite,
             postings_composite,
             fast_fields_composite,
             fieldnorms_composite,
             segment_id: segment.id(),
-            store_source,
+            store_reader,
             delete_bitset_opt,
             positions_composite,
             positions_idx_composite,
@@ -286,8 +282,7 @@ impl SegmentReader {
     /// term dictionary associated to a specific field,
     /// and opening the posting list associated to any term.
     pub fn inverted_index(&self, field: Field) -> Arc<InvertedIndexReader> {
-        if let Some(inv_idx_reader) = self
-            .inv_idx_reader_cache
+        if let Some(inv_idx_reader) = self.inv_idx_reader_cache
             .read()
             .expect("Lock poisoned. This should never happen")
             .get(&field)
@@ -311,28 +306,25 @@ impl SegmentReader {
             // As a result, no data is associated to the inverted index.
             //
             // Returns an empty inverted index.
-            return Arc::new(InvertedIndexReader::empty(field_type));
+            return Arc::new(InvertedIndexReader::empty(field_type.clone()));
         }
 
         let postings_source = postings_source_opt.unwrap();
 
-        let termdict_source = self
-            .termdict_composite
+        let termdict_source = self.termdict_composite
             .open_read(field)
             .expect("Failed to open field term dictionary in composite file. Is the field indexed");
 
-        let positions_source = self
-            .positions_composite
+        let positions_source = self.positions_composite
             .open_read(field)
             .expect("Index corrupted. Failed to open field positions in composite file.");
 
-        let positions_idx_source = self
-            .positions_idx_composite
+        let positions_idx_source = self.positions_idx_composite
             .open_read(field)
             .expect("Index corrupted. Failed to open field positions in composite file.");
 
         let inv_idx_reader = Arc::new(InvertedIndexReader::new(
-            TermDictionary::from_source(&termdict_source),
+            TermDictionary::from_source(termdict_source),
             postings_source,
             positions_source,
             positions_idx_source,
@@ -349,6 +341,14 @@ impl SegmentReader {
         inv_idx_reader
     }
 
+    /// Returns the document (or to be accurate, its stored field)
+    /// bearing the given doc id.
+    /// This method is slow and should seldom be called from
+    /// within a collector.
+    pub fn doc(&self, doc_id: DocId) -> Result<Document> {
+        self.store_reader.get(doc_id)
+    }
+
     /// Returns the segment id
     pub fn segment_id(&self) -> SegmentId {
         self.segment_id
@@ -372,24 +372,6 @@ impl SegmentReader {
     pub fn doc_ids_alive(&self) -> SegmentReaderAliveDocsIterator {
         SegmentReaderAliveDocsIterator::new(&self)
     }
-
-    /// Summarize total space usage of this segment.
-    pub fn space_usage(&self) -> SegmentSpaceUsage {
-        SegmentSpaceUsage::new(
-            self.num_docs(),
-            self.termdict_composite.space_usage(),
-            self.postings_composite.space_usage(),
-            self.positions_composite.space_usage(),
-            self.positions_idx_composite.space_usage(),
-            self.fast_fields_composite.space_usage(),
-            self.fieldnorms_composite.space_usage(),
-            self.get_store_reader().space_usage(),
-            self.delete_bitset_opt
-                .as_ref()
-                .map(|x| x.space_usage())
-                .unwrap_or(0),
-        )
-    }
 }
 
 impl fmt::Debug for SegmentReader {
@@ -409,7 +391,7 @@ pub struct SegmentReaderAliveDocsIterator<'a> {
 impl<'a> SegmentReaderAliveDocsIterator<'a> {
     pub fn new(reader: &'a SegmentReader) -> SegmentReaderAliveDocsIterator<'a> {
         SegmentReaderAliveDocsIterator {
-            reader,
+            reader: reader,
             max_doc: reader.max_doc(),
             current: 0,
         }
@@ -447,12 +429,12 @@ impl<'a> Iterator for SegmentReaderAliveDocsIterator<'a> {
 #[cfg(test)]
 mod test {
     use core::Index;
-    use schema::{Schema, Term, STORED, TEXT};
+    use schema::{SchemaBuilder, Term, STORED, TEXT};
     use DocId;
 
     #[test]
     fn test_alive_docs_iterator() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         schema_builder.add_text_field("name", TEXT | STORED);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema.clone());

src/directory/directory.rs

@@ -1,102 +1,11 @@
-use directory::directory_lock::Lock;
-use directory::error::LockError;
 use directory::error::{DeleteError, OpenReadError, OpenWriteError};
 use directory::{ReadOnlySource, WritePtr};
 use std::fmt;
 use std::io;
-use std::io::Write;
 use std::marker::Send;
 use std::marker::Sync;
 use std::path::Path;
-use std::path::PathBuf;
 use std::result;
-use std::thread;
-use std::time::Duration;
-
-/// Retry the logic of acquiring locks is pretty simple.
-/// We just retry `n` times after a given `duratio`, both
-/// depending on the type of lock.
-struct RetryPolicy {
-    num_retries: usize,
-    wait_in_ms: u64,
-}
-
-impl RetryPolicy {
-    fn no_retry() -> RetryPolicy {
-        RetryPolicy {
-            num_retries: 0,
-            wait_in_ms: 0,
-        }
-    }
-
-    fn wait_and_retry(&mut self) -> bool {
-        if self.num_retries == 0 {
-            false
-        } else {
-            self.num_retries -= 1;
-            let wait_duration = Duration::from_millis(self.wait_in_ms);
-            thread::sleep(wait_duration);
-            true
-        }
-    }
-}
-
-/// The `DirectoryLock` is an object that represents a file lock.
-/// See  [`LockType`](struct.LockType.html)
-///
-/// It is transparently associated to a lock file, that gets deleted
-/// on `Drop.` The lock is released automatically on `Drop`.
-pub struct DirectoryLock(Box<Drop + Send + 'static>);
-
-struct DirectoryLockGuard {
-    directory: Box<Directory>,
-    path: PathBuf,
-}
-
-impl<T: Drop + Send + 'static> From<Box<T>> for DirectoryLock {
-    fn from(underlying: Box<T>) -> Self {
-        DirectoryLock(underlying)
-    }
-}
-
-impl Drop for DirectoryLockGuard {
-    fn drop(&mut self) {
-        if let Err(e) = self.directory.delete(&*self.path) {
-            error!("Failed to remove the lock file. {:?}", e);
-        }
-    }
-}
-
-enum TryAcquireLockError {
-    FileExists,
-    IOError(io::Error),
-}
-
-fn try_acquire_lock(
-    filepath: &Path,
-    directory: &mut Directory,
-) -> Result<DirectoryLock, TryAcquireLockError> {
-    let mut write = directory.open_write(filepath).map_err(|e| match e {
-        OpenWriteError::FileAlreadyExists(_) => TryAcquireLockError::FileExists,
-        OpenWriteError::IOError(io_error) => TryAcquireLockError::IOError(io_error.into()),
-    })?;
-    write.flush().map_err(TryAcquireLockError::IOError)?;
-    Ok(DirectoryLock::from(Box::new(DirectoryLockGuard {
-        directory: directory.box_clone(),
-        path: filepath.to_owned(),
-    })))
-}
-
-fn retry_policy(is_blocking: bool) -> RetryPolicy {
-    if is_blocking {
-        RetryPolicy {
-            num_retries: 100,
-            wait_in_ms: 100,
-        }
-    } else {
-        RetryPolicy::no_retry()
-    }
-}
 
 /// Write-once read many (WORM) abstraction for where
 /// tantivy's data should be stored.
@@ -164,42 +73,19 @@ pub trait Directory: DirectoryClone + fmt::Debug + Send + Sync + 'static {
     ///
     /// The file may or may not previously exist.
     fn atomic_write(&mut self, path: &Path, data: &[u8]) -> io::Result<()>;
-
-    /// Acquire a lock in the given directory.
-    ///
-    /// The method is blocking or not depending on the `Lock` object.
-    fn acquire_lock(&self, lock: &Lock) -> Result<DirectoryLock, LockError> {
-        let mut box_directory = self.box_clone();
-        let mut retry_policy = retry_policy(lock.is_blocking);
-        loop {
-            match try_acquire_lock(&lock.filepath, &mut *box_directory) {
-                Ok(result) => {
-                    return Ok(result);
-                }
-                Err(TryAcquireLockError::FileExists) => {
-                    if !retry_policy.wait_and_retry() {
-                        return Err(LockError::LockBusy);
-                    }
-                }
-                Err(TryAcquireLockError::IOError(io_error)) => {
-                    return Err(LockError::IOError(io_error));
-                }
-            }
-        }
-    }
 }
 
 /// DirectoryClone
 pub trait DirectoryClone {
-    /// Clones the directory and boxes the clone
-    fn box_clone(&self) -> Box<Directory>;
+  /// Clones the directory and boxes the clone
+  fn box_clone(&self) -> Box<Directory>;
 }
 
 impl<T> DirectoryClone for T
 where
-    T: 'static + Directory + Clone,
+  T: 'static + Directory + Clone,
 {
-    fn box_clone(&self) -> Box<Directory> {
-        Box::new(self.clone())
-    }
+  fn box_clone(&self) -> Box<Directory> {
+    Box::new(self.clone())
+  }
 }

src/directory/directory_lock.rs

@@ -1,56 +0,0 @@
-use std::path::PathBuf;
-
-/// A directory lock.
-///
-/// A lock is associated to a specific path and some
-/// [`LockParams`](./enum.LockParams.html).
-/// Tantivy itself uses only two locks but client application
-/// can use the directory facility to define their own locks.
-/// - [INDEX_WRITER_LOCK](./struct.INDEX_WRITER_LOCK.html)
-/// - [META_LOCK](./struct.META_LOCK.html)
-///
-/// Check out these locks documentation for more information.
-///
-#[derive(Debug)]
-pub struct Lock {
-    /// The lock needs to be associated with its own file `path`.
-    /// Depending on the platform, the lock might rely on the creation
-    /// and deletion of this filepath.
-    pub filepath: PathBuf,
-    /// `lock_params` describes whether acquiring the lock is meant
-    /// to be a blocking operation or a non-blocking.
-    ///
-    /// Acquiring a blocking lock blocks until the lock is
-    /// available.
-    /// Acquiring a blocking lock returns rapidly, either successfully
-    /// or with an error signifying that someone is already holding
-    /// the lock.
-    pub is_blocking: bool,
-}
-
-lazy_static! {
-     /// Only one process should be able to write tantivy's index at a time.
-    /// This lock file, when present, is in charge of preventing other processes to open an IndexWriter.
-    ///
-    /// If the process is killed and this file remains, it is safe to remove it manually.
-    ///
-    /// Failing to acquire this lock usually means a misuse of tantivy's API,
-    /// (creating more than one instance of the `IndexWriter`), are a spurious
-    /// lock file remaining after a crash. In the latter case, removing the file after
-    /// checking no process running tantivy is running is safe.
-    pub static ref INDEX_WRITER_LOCK: Lock = Lock {
-        filepath: PathBuf::from(".tantivy-writer.lock"),
-        is_blocking: false
-    };
-    /// The meta lock file is here to protect the segment files being opened by
-    /// `.load_searchers()` from being garbage collected.
-    /// It makes it possible for another process to safely consume
-    /// our index in-writing. Ideally, we may have prefered `RWLock` semantics
-    /// here, but it is difficult to achieve on Windows.
-    ///
-    /// Opening segment readers is a very fast process.
-    pub static ref META_LOCK: Lock = Lock {
-        filepath: PathBuf::from(".tantivy-meta.lock"),
-        is_blocking: true
-    };
-}

src/directory/error.rs

@@ -3,22 +3,6 @@ use std::fmt;
 use std::io;
 use std::path::PathBuf;
 
-/// Error while trying to acquire a directory lock.
-#[derive(Debug, Fail)]
-pub enum LockError {
-    /// Failed to acquired a lock as it is already hold by another
-    /// client.
-    /// - In the context of a blocking lock, this means the lock was not released within some `timeout` period.
-    /// - In the context of a non-blocking lock, this means the lock was busy at the moment of the call.
-    #[fail(
-        display = "Could not acquire lock as it is already held, possibly by a different process."
-    )]
-    LockBusy,
-    /// Trying to acquire a lock failed with an `IOError`
-    #[fail(display = "Failed to acquire the lock due to an io:Error.")]
-    IOError(io::Error),
-}
-
 /// General IO error with an optional path to the offending file.
 #[derive(Debug)]
 pub struct IOError {
@@ -26,12 +10,6 @@ pub struct IOError {
     err: io::Error,
 }
 
-impl Into<io::Error> for IOError {
-    fn into(self) -> io::Error {
-        self.err
-    }
-}
-
 impl fmt::Display for IOError {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         match self.path {

src/directory/managed_directory.rs

@@ -1,10 +1,7 @@
 use core::MANAGED_FILEPATH;
-use directory::error::{DeleteError, IOError, LockError, OpenReadError, OpenWriteError};
-use directory::DirectoryLock;
-use directory::Lock;
-use directory::META_LOCK;
+use directory::error::{DeleteError, IOError, OpenReadError, OpenWriteError};
 use directory::{ReadOnlySource, WritePtr};
-use error::DataCorruption;
+use error::TantivyError;
 use serde_json;
 use std::collections::HashSet;
 use std::io;
@@ -16,17 +13,6 @@ use std::sync::{Arc, RwLock};
 use Directory;
 use Result;
 
-/// Returns true iff the file is "managed".
-/// Non-managed file are not subject to garbage collection.
-///
-/// Filenames that starts by a "." -typically locks-
-/// are not managed.
-fn is_managed(path: &Path) -> bool {
-    path.to_str()
-        .map(|p_str| !p_str.starts_with('.'))
-        .unwrap_or(true)
-}
-
 /// Wrapper of directories that keeps track of files created by Tantivy.
 ///
 /// A managed directory is just a wrapper of a directory
@@ -54,24 +40,19 @@ fn save_managed_paths(
     wlock: &RwLockWriteGuard<MetaInformation>,
 ) -> io::Result<()> {
     let mut w = serde_json::to_vec(&wlock.managed_paths)?;
-    writeln!(&mut w)?;
+    write!(&mut w, "\n")?;
     directory.atomic_write(&MANAGED_FILEPATH, &w[..])?;
     Ok(())
 }
 
 impl ManagedDirectory {
     /// Wraps a directory as managed directory.
-    pub fn wrap<Dir: Directory>(directory: Dir) -> Result<ManagedDirectory> {
+    pub fn new<Dir: Directory>(directory: Dir) -> Result<ManagedDirectory> {
         match directory.atomic_read(&MANAGED_FILEPATH) {
             Ok(data) => {
                 let managed_files_json = String::from_utf8_lossy(&data);
                 let managed_files: HashSet<PathBuf> = serde_json::from_str(&managed_files_json)
-                    .map_err(|e| {
-                        DataCorruption::new(
-                            MANAGED_FILEPATH.clone(),
-                            format!("Managed file cannot be deserialized: {:?}. ", e),
-                        )
-                    })?;
+                    .map_err(|_| TantivyError::CorruptedFile(MANAGED_FILEPATH.clone()))?;
                 Ok(ManagedDirectory {
                     directory: Box::new(directory),
                     meta_informations: Arc::new(RwLock::new(MetaInformation {
@@ -94,9 +75,6 @@ impl ManagedDirectory {
     ///
     /// * `living_files` - List of files that are still used by the index.
     ///
-    /// The use a callback ensures that the list of living_files is computed
-    /// while we hold the lock on meta.
-    ///
     /// This method does not panick nor returns errors.
     /// If a file cannot be deleted (for permission reasons for instance)
     /// an error is simply logged, and the file remains in the list of managed
@@ -104,35 +82,25 @@ impl ManagedDirectory {
     pub fn garbage_collect<L: FnOnce() -> HashSet<PathBuf>>(&mut self, get_living_files: L) {
         info!("Garbage collect");
         let mut files_to_delete = vec![];
-
-        // It is crucial to get the living files after acquiring the
-        // read lock of meta informations. That way, we
-        // avoid the following scenario.
-        //
-        // 1) we get the list of living files.
-        // 2) someone creates a new file.
-        // 3) we start garbage collection and remove this file
-        // even though it is a living file.
-        //
-        // releasing the lock as .delete() will use it too.
         {
-            let meta_informations_rlock = self
-                .meta_informations
+            // releasing the lock as .delete() will use it too.
+            let meta_informations_rlock = self.meta_informations
                 .read()
                 .expect("Managed directory rlock poisoned in garbage collect.");
 
-            // The point of this second "file" lock is to enforce the following scenario
-            // 1) process B tries to load a new set of searcher.
-            // The list of segments is loaded
-            // 2) writer change meta.json (for instance after a merge or a commit)
-            // 3) gc kicks in.
-            // 4) gc removes a file that was useful for process B, before process B opened it.
-            if let Ok(_meta_lock) = self.acquire_lock(&META_LOCK) {
-                let living_files = get_living_files();
-                for managed_path in &meta_informations_rlock.managed_paths {
-                    if !living_files.contains(managed_path) {
-                        files_to_delete.push(managed_path.clone());
-                    }
+            // It is crucial to get the living files after acquiring the
+            // read lock of meta informations. That way, we
+            // avoid the following scenario.
+            //
+            // 1) we get the list of living files.
+            // 2) someone creates a new file.
+            // 3) we start garbage collection and remove this file
+            // even though it is a living file.
+            let living_files = get_living_files();
+
+            for managed_path in &meta_informations_rlock.managed_paths {
+                if !living_files.contains(managed_path) {
+                    files_to_delete.push(managed_path.clone());
                 }
             }
         }
@@ -166,8 +134,7 @@ impl ManagedDirectory {
         if !deleted_files.is_empty() {
             // update the list of managed files by removing
             // the file that were removed.
-            let mut meta_informations_wlock = self
-                .meta_informations
+            let mut meta_informations_wlock = self.meta_informations
                 .write()
                 .expect("Managed directory wlock poisoned (2).");
             {
@@ -189,17 +156,8 @@ impl ManagedDirectory {
     /// registering the filepath and creating the file
     /// will not lead to garbage files that will
     /// never get removed.
-    ///
-    /// File starting by "." are reserved to locks.
-    /// They are not managed and cannot be subjected
-    /// to garbage collection.
     fn register_file_as_managed(&mut self, filepath: &Path) -> io::Result<()> {
-        // Files starting by "." (e.g. lock files) are not managed.
-        if !is_managed(filepath) {
-            return Ok(());
-        }
-        let mut meta_wlock = self
-            .meta_informations
+        let mut meta_wlock = self.meta_informations
             .write()
             .expect("Managed file lock poisoned");
         let has_changed = meta_wlock.managed_paths.insert(filepath.to_owned());
@@ -237,10 +195,6 @@ impl Directory for ManagedDirectory {
     fn exists(&self, path: &Path) -> bool {
         self.directory.exists(path)
     }
-
-    fn acquire_lock(&self, lock: &Lock) -> result::Result<DirectoryLock, LockError> {
-        self.directory.acquire_lock(lock)
-    }
 }
 
 impl Clone for ManagedDirectory {
@@ -274,7 +228,7 @@ mod tests {
         let tempdir_path = PathBuf::from(tempdir.path());
         {
             let mmap_directory = MmapDirectory::open(&tempdir_path).unwrap();
-            let mut managed_directory = ManagedDirectory::wrap(mmap_directory).unwrap();
+            let mut managed_directory = ManagedDirectory::new(mmap_directory).unwrap();
             {
                 let mut write_file = managed_directory.open_write(*TEST_PATH1).unwrap();
                 write_file.flush().unwrap();
@@ -300,7 +254,7 @@ mod tests {
         }
         {
             let mmap_directory = MmapDirectory::open(&tempdir_path).unwrap();
-            let mut managed_directory = ManagedDirectory::wrap(mmap_directory).unwrap();
+            let mut managed_directory = ManagedDirectory::new(mmap_directory).unwrap();
             {
                 assert!(managed_directory.exists(*TEST_PATH1));
                 assert!(!managed_directory.exists(*TEST_PATH2));
@@ -324,7 +278,7 @@ mod tests {
         let living_files = HashSet::new();
 
         let mmap_directory = MmapDirectory::open(&tempdir_path).unwrap();
-        let mut managed_directory = ManagedDirectory::wrap(mmap_directory).unwrap();
+        let mut managed_directory = ManagedDirectory::new(mmap_directory).unwrap();
         managed_directory
             .atomic_write(*TEST_PATH1, &vec![0u8, 1u8])
             .unwrap();

src/directory/mmap_directory.rs

@@ -1,14 +1,8 @@
-extern crate fs2;
-
-use self::fs2::FileExt;
 use atomicwrites;
 use common::make_io_err;
-use directory::error::LockError;
 use directory::error::{DeleteError, IOError, OpenDirectoryError, OpenReadError, OpenWriteError};
 use directory::shared_vec_slice::SharedVecSlice;
 use directory::Directory;
-use directory::DirectoryLock;
-use directory::Lock;
 use directory::ReadOnlySource;
 use directory::WritePtr;
 use fst::raw::MmapReadOnly;
@@ -38,8 +32,7 @@ fn open_mmap(full_path: &Path) -> result::Result<Option<MmapReadOnly>, OpenReadE
         }
     })?;
 
-    let meta_data = file
-        .metadata()
+    let meta_data = file.metadata()
         .map_err(|e| IOError::with_path(full_path.to_owned(), e))?;
     if meta_data.len() == 0 {
         // if the file size is 0, it will not be possible
@@ -121,14 +114,6 @@ impl MmapCache {
 ///
 /// The Mmap object are cached to limit the
 /// system calls.
-///
-/// In the `MmapDirectory`, locks are implemented using the `fs2` crate definition of locks.
-///
-/// On MacOS & linux, it relies on `flock` (aka `BSD Lock`). These locks solve most of the
-/// problems related to POSIX Locks, but may their contract may not be respected on `NFS`
-/// depending on the implementation.
-///
-/// On Windows the semantics are again different.
 #[derive(Clone)]
 pub struct MmapDirectory {
     root_path: PathBuf,
@@ -227,21 +212,6 @@ impl MmapDirectory {
     }
 }
 
-/// We rely on fs2 for file locking. On Windows & MacOS this
-/// uses BSD locks (`flock`). The lock is actually released when
-/// the `File` object is dropped and its associated file descriptor
-/// is closed.
-struct ReleaseLockFile {
-    _file: File,
-    path: PathBuf,
-}
-
-impl Drop for ReleaseLockFile {
-    fn drop(&mut self) {
-        debug!("Releasing lock {:?}", self.path);
-    }
-}
-
 /// This Write wraps a File, but has the specificity of
 /// call `sync_all` on flush.
 struct SafeFileWriter(File);
@@ -339,8 +309,7 @@ impl Directory for MmapDirectory {
         // when the last reference is gone.
         mmap_cache.cache.remove(&full_path);
         match fs::remove_file(&full_path) {
-            Ok(_) => self
-                .sync_directory()
+            Ok(_) => self.sync_directory()
                 .map_err(|e| IOError::with_path(path.to_owned(), e).into()),
             Err(e) => {
                 if e.kind() == io::ErrorKind::NotFound {
@@ -383,26 +352,6 @@ impl Directory for MmapDirectory {
         meta_file.write(|f| f.write_all(data))?;
         Ok(())
     }
-
-    fn acquire_lock(&self, lock: &Lock) -> Result<DirectoryLock, LockError> {
-        let full_path = self.resolve_path(&lock.filepath);
-        // We make sure that the file exists.
-        let file: File = OpenOptions::new()
-            .write(true)
-            .create(true) //< if the file does not exist yet, create it.
-            .open(&full_path)
-            .map_err(LockError::IOError)?;
-        if lock.is_blocking {
-            file.lock_exclusive().map_err(LockError::IOError)?;
-        } else {
-            file.try_lock_exclusive().map_err(|_| LockError::LockBusy)?
-        }
-        // dropping the file handle will release the lock.
-        Ok(DirectoryLock::from(Box::new(ReleaseLockFile {
-            path: lock.filepath.clone(),
-            _file: file,
-        })))
-    }
 }
 
 #[cfg(test)]
@@ -413,11 +362,6 @@ mod tests {
 
     use super::*;
 
-    #[test]
-    fn test_open_non_existant_path() {
-        assert!(MmapDirectory::open(PathBuf::from("./nowhere")).is_err());
-    }
-
     #[test]
     fn test_open_empty() {
         // empty file is actually an edge case because those

src/directory/mod.rs

@@ -8,7 +8,6 @@ WORM directory abstraction.
 mod mmap_directory;
 
 mod directory;
-mod directory_lock;
 mod managed_directory;
 mod ram_directory;
 mod read_only_source;
@@ -17,12 +16,11 @@ mod shared_vec_slice;
 /// Errors specific to the directory module.
 pub mod error;
 
-pub use self::directory::DirectoryLock;
+use std::io::{BufWriter, Seek, Write};
+
 pub use self::directory::{Directory, DirectoryClone};
-pub use self::directory_lock::{Lock, INDEX_WRITER_LOCK, META_LOCK};
 pub use self::ram_directory::RAMDirectory;
 pub use self::read_only_source::ReadOnlySource;
-use std::io::{BufWriter, Seek, Write};
 
 #[cfg(feature = "mmap")]
 pub use self::mmap_directory::MmapDirectory;
@@ -40,4 +38,128 @@ impl<T: Seek + Write> SeekableWrite for T {}
 pub type WritePtr = BufWriter<Box<SeekableWrite>>;
 
 #[cfg(test)]
-mod tests;
+mod tests {
+
+    use super::*;
+    use std::io::{Seek, SeekFrom, Write};
+    use std::path::Path;
+
+    lazy_static! {
+        static ref TEST_PATH: &'static Path = Path::new("some_path_for_test");
+    }
+
+    #[test]
+    fn test_ram_directory() {
+        let mut ram_directory = RAMDirectory::create();
+        test_directory(&mut ram_directory);
+    }
+
+    #[test]
+    #[cfg(feature = "mmap")]
+    fn test_mmap_directory() {
+        let mut mmap_directory = MmapDirectory::create_from_tempdir().unwrap();
+        test_directory(&mut mmap_directory);
+    }
+
+    #[test]
+    #[should_panic]
+    fn ram_directory_panics_if_flush_forgotten() {
+        let mut ram_directory = RAMDirectory::create();
+        let mut write_file = ram_directory.open_write(*TEST_PATH).unwrap();
+        assert!(write_file.write_all(&[4]).is_ok());
+    }
+
+    fn test_simple(directory: &mut Directory) {
+        {
+            {
+                let mut write_file = directory.open_write(*TEST_PATH).unwrap();
+                assert!(directory.exists(*TEST_PATH));
+                write_file.write_all(&[4]).unwrap();
+                write_file.write_all(&[3]).unwrap();
+                write_file.write_all(&[7, 3, 5]).unwrap();
+                write_file.flush().unwrap();
+            }
+            let read_file = directory.open_read(*TEST_PATH).unwrap();
+            let data: &[u8] = &*read_file;
+            assert_eq!(data, &[4u8, 3u8, 7u8, 3u8, 5u8]);
+        }
+
+        assert!(directory.delete(*TEST_PATH).is_ok());
+        assert!(!directory.exists(*TEST_PATH));
+    }
+
+    fn test_seek(directory: &mut Directory) {
+        {
+            {
+                let mut write_file = directory.open_write(*TEST_PATH).unwrap();
+                write_file.write_all(&[4, 3, 7, 3, 5]).unwrap();
+                write_file.seek(SeekFrom::Start(0)).unwrap();
+                write_file.write_all(&[3, 1]).unwrap();
+                write_file.flush().unwrap();
+            }
+            let read_file = directory.open_read(*TEST_PATH).unwrap();
+            let data: &[u8] = &*read_file;
+            assert_eq!(data, &[3u8, 1u8, 7u8, 3u8, 5u8]);
+        }
+
+        assert!(directory.delete(*TEST_PATH).is_ok());
+    }
+
+    fn test_rewrite_forbidden(directory: &mut Directory) {
+        {
+            directory.open_write(*TEST_PATH).unwrap();
+            assert!(directory.exists(*TEST_PATH));
+        }
+        {
+            assert!(directory.open_write(*TEST_PATH).is_err());
+        }
+        assert!(directory.delete(*TEST_PATH).is_ok());
+    }
+
+    fn test_write_create_the_file(directory: &mut Directory) {
+        {
+            assert!(directory.open_read(*TEST_PATH).is_err());
+            let _w = directory.open_write(*TEST_PATH).unwrap();
+            assert!(directory.exists(*TEST_PATH));
+            assert!(directory.open_read(*TEST_PATH).is_ok());
+            assert!(directory.delete(*TEST_PATH).is_ok());
+        }
+    }
+
+    fn test_directory_delete(directory: &mut Directory) {
+        assert!(directory.open_read(*TEST_PATH).is_err());
+        let mut write_file = directory.open_write(*TEST_PATH).unwrap();
+        write_file.write_all(&[1, 2, 3, 4]).unwrap();
+        write_file.flush().unwrap();
+        {
+            let read_handle = directory.open_read(*TEST_PATH).unwrap();
+            {
+                assert_eq!(&*read_handle, &[1u8, 2u8, 3u8, 4u8]);
+
+                // Mapped files can't be deleted on Windows
+                if !cfg!(windows) {
+                    assert!(directory.delete(*TEST_PATH).is_ok());
+                    assert_eq!(&*read_handle, &[1u8, 2u8, 3u8, 4u8]);
+                }
+
+                assert!(directory.delete(Path::new("SomeOtherPath")).is_err());
+            }
+        }
+
+        if cfg!(windows) {
+            assert!(directory.delete(*TEST_PATH).is_ok());
+        }
+
+        assert!(directory.open_read(*TEST_PATH).is_err());
+        assert!(directory.delete(*TEST_PATH).is_err());
+    }
+
+    fn test_directory(directory: &mut Directory) {
+        test_simple(directory);
+        test_seek(directory);
+        test_rewrite_forbidden(directory);
+        test_write_create_the_file(directory);
+        test_directory_delete(directory);
+    }
+
+}

src/directory/ram_directory.rs

@@ -170,10 +170,10 @@ impl Directory for RAMDirectory {
         let path_buf = PathBuf::from(path);
         let vec_writer = VecWriter::new(path_buf.clone(), self.fs.clone());
 
-        let exists = self
-            .fs
+        let exists = self.fs
             .write(path_buf.clone(), &Vec::new())
             .map_err(|err| IOError::with_path(path.to_owned(), err))?;
+
         // force the creation of the file to mimic the MMap directory.
         if exists {
             Err(OpenWriteError::FileAlreadyExists(path_buf))
@@ -196,10 +196,6 @@ impl Directory for RAMDirectory {
     }
 
     fn atomic_write(&mut self, path: &Path, data: &[u8]) -> io::Result<()> {
-        fail_point!("RAMDirectory::atomic_write", |msg| Err(io::Error::new(
-            io::ErrorKind::Other,
-            msg.unwrap_or("Undefined".to_string())
-        )));
         let path_buf = PathBuf::from(path);
         let mut vec_writer = VecWriter::new(path_buf.clone(), self.fs.clone());
         self.fs.write(path_buf, &Vec::new())?;

src/directory/read_only_source.rs

@@ -5,6 +5,7 @@ use fst::raw::MmapReadOnly;
 use stable_deref_trait::{CloneStableDeref, StableDeref};
 use std::ops::Deref;
 
+
 /// Read object that represents files in tantivy.
 ///
 /// These read objects are only in charge to deliver

src/directory/tests.rs

@@ -1,182 +0,0 @@
-use super::*;
-use std::io::{Seek, SeekFrom, Write};
-use std::path::{Path, PathBuf};
-use std::time;
-
-lazy_static! {
-    static ref TEST_PATH: &'static Path = Path::new("some_path_for_test");
-}
-
-#[test]
-fn test_ram_directory() {
-    let mut ram_directory = RAMDirectory::create();
-    test_directory(&mut ram_directory);
-}
-
-#[test]
-#[cfg(feature = "mmap")]
-fn test_mmap_directory() {
-    let mut mmap_directory = MmapDirectory::create_from_tempdir().unwrap();
-    test_directory(&mut mmap_directory);
-}
-
-#[test]
-#[should_panic]
-fn ram_directory_panics_if_flush_forgotten() {
-    let mut ram_directory = RAMDirectory::create();
-    let mut write_file = ram_directory.open_write(*TEST_PATH).unwrap();
-    assert!(write_file.write_all(&[4]).is_ok());
-}
-
-fn test_simple(directory: &mut Directory) {
-    {
-        {
-            let mut write_file = directory.open_write(*TEST_PATH).unwrap();
-            assert!(directory.exists(*TEST_PATH));
-            write_file.write_all(&[4]).unwrap();
-            write_file.write_all(&[3]).unwrap();
-            write_file.write_all(&[7, 3, 5]).unwrap();
-            write_file.flush().unwrap();
-        }
-        let read_file = directory.open_read(*TEST_PATH).unwrap();
-        let data: &[u8] = &*read_file;
-        assert_eq!(data, &[4u8, 3u8, 7u8, 3u8, 5u8]);
-    }
-
-    assert!(directory.delete(*TEST_PATH).is_ok());
-    assert!(!directory.exists(*TEST_PATH));
-}
-
-fn test_seek(directory: &mut Directory) {
-    {
-        {
-            let mut write_file = directory.open_write(*TEST_PATH).unwrap();
-            write_file.write_all(&[4, 3, 7, 3, 5]).unwrap();
-            write_file.seek(SeekFrom::Start(0)).unwrap();
-            write_file.write_all(&[3, 1]).unwrap();
-            write_file.flush().unwrap();
-        }
-        let read_file = directory.open_read(*TEST_PATH).unwrap();
-        let data: &[u8] = &*read_file;
-        assert_eq!(data, &[3u8, 1u8, 7u8, 3u8, 5u8]);
-    }
-
-    assert!(directory.delete(*TEST_PATH).is_ok());
-}
-
-fn test_rewrite_forbidden(directory: &mut Directory) {
-    {
-        directory.open_write(*TEST_PATH).unwrap();
-        assert!(directory.exists(*TEST_PATH));
-    }
-    {
-        assert!(directory.open_write(*TEST_PATH).is_err());
-    }
-    assert!(directory.delete(*TEST_PATH).is_ok());
-}
-
-fn test_write_create_the_file(directory: &mut Directory) {
-    {
-        assert!(directory.open_read(*TEST_PATH).is_err());
-        let _w = directory.open_write(*TEST_PATH).unwrap();
-        assert!(directory.exists(*TEST_PATH));
-        assert!(directory.open_read(*TEST_PATH).is_ok());
-        assert!(directory.delete(*TEST_PATH).is_ok());
-    }
-}
-
-fn test_directory_delete(directory: &mut Directory) {
-    assert!(directory.open_read(*TEST_PATH).is_err());
-    let mut write_file = directory.open_write(*TEST_PATH).unwrap();
-    write_file.write_all(&[1, 2, 3, 4]).unwrap();
-    write_file.flush().unwrap();
-    {
-        let read_handle = directory.open_read(*TEST_PATH).unwrap();
-        {
-            assert_eq!(&*read_handle, &[1u8, 2u8, 3u8, 4u8]);
-
-            // Mapped files can't be deleted on Windows
-            if !cfg!(windows) {
-                assert!(directory.delete(*TEST_PATH).is_ok());
-                assert_eq!(&*read_handle, &[1u8, 2u8, 3u8, 4u8]);
-            }
-
-            assert!(directory.delete(Path::new("SomeOtherPath")).is_err());
-        }
-    }
-
-    if cfg!(windows) {
-        assert!(directory.delete(*TEST_PATH).is_ok());
-    }
-
-    assert!(directory.open_read(*TEST_PATH).is_err());
-    assert!(directory.delete(*TEST_PATH).is_err());
-}
-
-fn test_directory(directory: &mut Directory) {
-    test_simple(directory);
-    test_seek(directory);
-    test_rewrite_forbidden(directory);
-    test_write_create_the_file(directory);
-    test_directory_delete(directory);
-    test_lock_non_blocking(directory);
-    test_lock_blocking(directory);
-}
-
-fn test_lock_non_blocking(directory: &mut Directory) {
-    {
-        let lock_a_res = directory.acquire_lock(&Lock {
-            filepath: PathBuf::from("a.lock"),
-            is_blocking: false,
-        });
-        assert!(lock_a_res.is_ok());
-        let lock_b_res = directory.acquire_lock(&Lock {
-            filepath: PathBuf::from("b.lock"),
-            is_blocking: false,
-        });
-        assert!(lock_b_res.is_ok());
-        let lock_a_res2 = directory.acquire_lock(&Lock {
-            filepath: PathBuf::from("a.lock"),
-            is_blocking: false,
-        });
-        assert!(lock_a_res2.is_err());
-    }
-    let lock_a_res = directory.acquire_lock(&Lock {
-        filepath: PathBuf::from("a.lock"),
-        is_blocking: false,
-    });
-    assert!(lock_a_res.is_ok());
-}
-
-fn test_lock_blocking(directory: &mut Directory) {
-    let lock_a_res = directory.acquire_lock(&Lock {
-        filepath: PathBuf::from("a.lock"),
-        is_blocking: true,
-    });
-    assert!(lock_a_res.is_ok());
-    std::thread::spawn(move || {
-        //< lock_a_res is sent to the thread.
-        std::thread::sleep(time::Duration::from_millis(10));
-        // explicitely droping lock_a_res. It would have been sufficient to just force it
-        // to be part of the move, but the intent seems clearer that way.
-        drop(lock_a_res);
-    });
-    {
-        // A non-blocking call should fail, as the thread is running and holding the lock.
-        let lock_a_res = directory.acquire_lock(&Lock {
-            filepath: PathBuf::from("a.lock"),
-            is_blocking: false,
-        });
-        assert!(lock_a_res.is_err());
-    }
-    {
-        // the blocking call should wait for at least 10ms.
-        let start = time::Instant::now();
-        let lock_a_res = directory.acquire_lock(&Lock {
-            filepath: PathBuf::from("a.lock"),
-            is_blocking: true,
-        });
-        assert!(lock_a_res.is_ok());
-        assert!(start.elapsed().subsec_millis() >= 10);
-    }
-}

src/error.rs

@@ -2,143 +2,92 @@
 
 use std::io;
 
-use directory::error::LockError;
 use directory::error::{IOError, OpenDirectoryError, OpenReadError, OpenWriteError};
 use fastfield::FastFieldNotAvailableError;
 use query;
 use schema;
 use serde_json;
-use std::fmt;
 use std::path::PathBuf;
 use std::sync::PoisonError;
 
-pub struct DataCorruption {
-    filepath: Option<PathBuf>,
-    comment: String,
-}
-
-impl DataCorruption {
-    pub fn new(filepath: PathBuf, comment: String) -> DataCorruption {
-        DataCorruption {
-            filepath: Some(filepath),
-            comment,
-        }
-    }
-
-    pub fn comment_only(comment: String) -> DataCorruption {
-        DataCorruption {
-            filepath: None,
-            comment,
-        }
-    }
-}
-
-impl fmt::Debug for DataCorruption {
-    fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> {
-        write!(f, "Data corruption: ")?;
-        if let Some(ref filepath) = &self.filepath {
-            write!(f, "(in file `{:?}`)", filepath)?;
-        }
-        write!(f, ": {}.", self.comment)?;
-        Ok(())
-    }
-}
-
 /// The library's failure based error enum
 #[derive(Debug, Fail)]
 pub enum TantivyError {
     /// Path does not exist.
-    #[fail(display = "Path does not exist: '{:?}'", _0)]
+    #[fail(display = "path does not exist: '{:?}'", _0)]
     PathDoesNotExist(PathBuf),
     /// File already exists, this is a problem when we try to write into a new file.
-    #[fail(display = "File already exists: '{:?}'", _0)]
+    #[fail(display = "file already exists: '{:?}'", _0)]
     FileAlreadyExists(PathBuf),
-    /// Index already exists in this directory
-    #[fail(display = "Index already exists")]
-    IndexAlreadyExists,
-    /// Failed to acquire file lock
-    #[fail(display = "Failed to acquire Lockfile: {:?}. {:?}", _0, _1)]
-    LockFailure(LockError, Option<String>),
     /// IO Error.
-    #[fail(display = "An IO error occurred: '{}'", _0)]
+    #[fail(display = "an IO error occurred: '{}'", _0)]
     IOError(#[cause] IOError),
-    /// Data corruption.
-    #[fail(display = "{:?}", _0)]
-    DataCorruption(DataCorruption),
+    /// The data within is corrupted.
+    ///
+    /// For instance, it contains invalid JSON.
+    #[fail(display = "file contains corrupted data: '{:?}'", _0)]
+    CorruptedFile(PathBuf),
     /// A thread holding the locked panicked and poisoned the lock.
-    #[fail(display = "A thread holding the locked panicked and poisoned the lock")]
+    #[fail(display = "a thread holding the locked panicked and poisoned the lock")]
     Poisoned,
     /// Invalid argument was passed by the user.
-    #[fail(display = "An invalid argument was passed: '{}'", _0)]
+    #[fail(display = "an invalid argument was passed: '{}'", _0)]
     InvalidArgument(String),
     /// An Error happened in one of the thread.
-    #[fail(display = "An error occurred in a thread: '{}'", _0)]
+    #[fail(display = "an error occurred in a thread: '{}'", _0)]
     ErrorInThread(String),
     /// An Error appeared related to the schema.
     #[fail(display = "Schema error: '{}'", _0)]
     SchemaError(String),
     /// Tried to access a fastfield reader for a field not configured accordingly.
-    #[fail(display = "Fast field not available: '{:?}'", _0)]
+    #[fail(display = "fast field not available: '{:?}'", _0)]
     FastFieldError(#[cause] FastFieldNotAvailableError),
-    /// System error. (e.g.: We failed spawning a new thread)
-    #[fail(display = "System error.'{}'", _0)]
-    SystemError(String),
-}
-
-impl From<DataCorruption> for TantivyError {
-    fn from(data_corruption: DataCorruption) -> TantivyError {
-        TantivyError::DataCorruption(data_corruption)
-    }
 }
 
 impl From<FastFieldNotAvailableError> for TantivyError {
     fn from(fastfield_error: FastFieldNotAvailableError) -> TantivyError {
-        TantivyError::FastFieldError(fastfield_error)
-    }
-}
-
-impl From<LockError> for TantivyError {
-    fn from(lock_error: LockError) -> TantivyError {
-        TantivyError::LockFailure(lock_error, None)
+        TantivyError::FastFieldError(fastfield_error).into()
     }
 }
 
 impl From<IOError> for TantivyError {
     fn from(io_error: IOError) -> TantivyError {
-        TantivyError::IOError(io_error)
+        TantivyError::IOError(io_error).into()
     }
 }
 
 impl From<io::Error> for TantivyError {
     fn from(io_error: io::Error) -> TantivyError {
-        TantivyError::IOError(io_error.into())
+        TantivyError::IOError(io_error.into()).into()
     }
 }
 
 impl From<query::QueryParserError> for TantivyError {
     fn from(parsing_error: query::QueryParserError) -> TantivyError {
-        TantivyError::InvalidArgument(format!("Query is invalid. {:?}", parsing_error))
+        TantivyError::InvalidArgument(format!("Query is invalid. {:?}", parsing_error)).into()
     }
 }
 
 impl<Guard> From<PoisonError<Guard>> for TantivyError {
     fn from(_: PoisonError<Guard>) -> TantivyError {
-        TantivyError::Poisoned
+        TantivyError::Poisoned.into()
     }
 }
 
 impl From<OpenReadError> for TantivyError {
     fn from(error: OpenReadError) -> TantivyError {
         match error {
-            OpenReadError::FileDoesNotExist(filepath) => TantivyError::PathDoesNotExist(filepath),
-            OpenReadError::IOError(io_error) => TantivyError::IOError(io_error),
+            OpenReadError::FileDoesNotExist(filepath) => {
+                TantivyError::PathDoesNotExist(filepath).into()
+            }
+            OpenReadError::IOError(io_error) => TantivyError::IOError(io_error).into(),
         }
     }
 }
 
 impl From<schema::DocParsingError> for TantivyError {
     fn from(error: schema::DocParsingError) -> TantivyError {
-        TantivyError::InvalidArgument(format!("Failed to parse document {:?}", error))
+        TantivyError::InvalidArgument(format!("Failed to parse document {:?}", error)).into()
     }
 }
 
@@ -149,7 +98,7 @@ impl From<OpenWriteError> for TantivyError {
                 TantivyError::FileAlreadyExists(filepath)
             }
             OpenWriteError::IOError(io_error) => TantivyError::IOError(io_error),
-        }
+        }.into()
     }
 }
 
@@ -157,11 +106,11 @@ impl From<OpenDirectoryError> for TantivyError {
     fn from(error: OpenDirectoryError) -> TantivyError {
         match error {
             OpenDirectoryError::DoesNotExist(directory_path) => {
-                TantivyError::PathDoesNotExist(directory_path)
-            }
-            OpenDirectoryError::NotADirectory(directory_path) => {
-                TantivyError::InvalidArgument(format!("{:?} is not a directory", directory_path))
+                TantivyError::PathDoesNotExist(directory_path).into()
             }
+            OpenDirectoryError::NotADirectory(directory_path) => TantivyError::InvalidArgument(
+                format!("{:?} is not a directory", directory_path),
+            ).into(),
         }
     }
 }
@@ -169,6 +118,6 @@ impl From<OpenDirectoryError> for TantivyError {
 impl From<serde_json::Error> for TantivyError {
     fn from(error: serde_json::Error) -> TantivyError {
         let io_err = io::Error::from(error);
-        TantivyError::IOError(io_err.into())
+        TantivyError::IOError(io_err.into()).into()
     }
 }

src/fastfield/bytes/mod.rs

@@ -6,12 +6,12 @@ pub use self::writer::BytesFastFieldWriter;
 
 #[cfg(test)]
 mod tests {
-    use schema::Schema;
+    use schema::SchemaBuilder;
     use Index;
 
     #[test]
     fn test_bytes() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let field = schema_builder.add_bytes_field("bytesfield");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);

src/fastfield/bytes/writer.rs

@@ -51,7 +51,7 @@ impl BytesFastFieldWriter {
         self.next_doc();
         for field_value in doc.field_values() {
             if field_value.field() == self.field {
-                if let Value::Bytes(ref bytes) = *field_value.value() {
+                if let &Value::Bytes(ref bytes) = field_value.value() {
                     self.vals.extend_from_slice(bytes);
                 } else {
                     panic!(

src/fastfield/delete.rs

@@ -2,7 +2,6 @@ use bit_set::BitSet;
 use common::HasLen;
 use directory::ReadOnlySource;
 use directory::WritePtr;
-use space_usage::ByteCount;
 use std::io;
 use std::io::Write;
 use DocId;
@@ -42,8 +41,7 @@ pub struct DeleteBitSet {
 impl DeleteBitSet {
     /// Opens a delete bitset given its data source.
     pub fn open(data: ReadOnlySource) -> DeleteBitSet {
-        let num_deleted: usize = data
-            .as_slice()
+        let num_deleted: usize = data.as_slice()
             .iter()
             .map(|b| b.count_ones() as usize)
             .sum();
@@ -64,11 +62,6 @@ impl DeleteBitSet {
             b & (1u8 << shift) != 0
         }
     }
-
-    /// Summarize total space usage of this bitset.
-    pub fn space_usage(&self) -> ByteCount {
-        self.data.len()
-    }
 }
 
 impl HasLen for DeleteBitSet {

src/fastfield/facet_reader.rs

@@ -1,6 +1,5 @@
 use super::MultiValueIntFastFieldReader;
 use schema::Facet;
-use std::str;
 use termdict::TermDictionary;
 use termdict::TermOrdinal;
 use DocId;
@@ -21,7 +20,6 @@ use DocId;
 pub struct FacetReader {
     term_ords: MultiValueIntFastFieldReader<u64>,
     term_dict: TermDictionary,
-    buffer: Vec<u8>,
 }
 
 impl FacetReader {
@@ -39,7 +37,6 @@ impl FacetReader {
         FacetReader {
             term_ords,
             term_dict,
-            buffer: vec![],
         }
     }
 
@@ -58,18 +55,10 @@ impl FacetReader {
     }
 
     /// Given a term ordinal returns the term associated to it.
-    pub fn facet_from_ord(
-        &mut self,
-        facet_ord: TermOrdinal,
-        output: &mut Facet,
-    ) -> Result<(), str::Utf8Error> {
-        let found_term = self
-            .term_dict
-            .ord_to_term(facet_ord as u64, &mut self.buffer);
+    pub fn facet_from_ord(&self, facet_ord: TermOrdinal, output: &mut Facet) {
+        let found_term = self.term_dict
+            .ord_to_term(facet_ord as u64, output.inner_buffer_mut());
         assert!(found_term, "Term ordinal {} no found.", facet_ord);
-        let facet_str = str::from_utf8(&self.buffer[..])?;
-        output.set_facet_str(facet_str);
-        Ok(())
     }
 
     /// Return the list of facet ordinals associated to a document.

src/fastfield/mod.rs

@@ -127,19 +127,19 @@ mod tests {
     use common::CompositeFile;
     use directory::{Directory, RAMDirectory, WritePtr};
     use fastfield::FastFieldReader;
-    use rand::prelude::SliceRandom;
-    use rand::rngs::StdRng;
+    use rand::Rng;
     use rand::SeedableRng;
+    use rand::XorShiftRng;
     use schema::Document;
     use schema::Field;
-    use schema::Schema;
     use schema::FAST;
+    use schema::{Schema, SchemaBuilder};
     use std::collections::HashMap;
     use std::path::Path;
 
     lazy_static! {
         pub static ref SCHEMA: Schema = {
-            let mut schema_builder = Schema::builder();
+            let mut schema_builder = SchemaBuilder::default();
             schema_builder.add_u64_field("field", FAST);
             schema_builder.build()
         };
@@ -298,7 +298,7 @@ mod tests {
     fn test_signed_intfastfield() {
         let path = Path::new("test");
         let mut directory: RAMDirectory = RAMDirectory::create();
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
 
         let i64_field = schema_builder.add_i64_field("field", FAST);
         let schema = schema_builder.build();
@@ -342,7 +342,7 @@ mod tests {
     fn test_signed_intfastfield_default_val() {
         let path = Path::new("test");
         let mut directory: RAMDirectory = RAMDirectory::create();
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let i64_field = schema_builder.add_i64_field("field", FAST);
         let schema = schema_builder.build();
 
@@ -367,10 +367,11 @@ mod tests {
         }
     }
 
-    // Warning: this generates the same permutation at each call
     pub fn generate_permutation() -> Vec<u64> {
-        let mut permutation: Vec<u64> = (0u64..100_000u64).collect();
-        permutation.shuffle(&mut StdRng::from_seed([1u8; 32]));
+        let seed: [u8; 16] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16];
+        let mut rng = XorShiftRng::from_seed(seed);
+        let mut permutation: Vec<u64> = (0u64..1_000_000u64).collect();
+        rng.shuffle(&mut permutation);
         permutation
     }
 

src/fastfield/multivalued/mod.rs

@@ -9,12 +9,12 @@ mod tests {
 
     use schema::Cardinality;
     use schema::IntOptions;
-    use schema::Schema;
+    use schema::SchemaBuilder;
     use Index;
 
     #[test]
     fn test_multivalued_u64() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let field = schema_builder.add_u64_field(
             "multifield",
             IntOptions::default().set_fast(Cardinality::MultiValues),
@@ -49,7 +49,7 @@ mod tests {
 
     #[test]
     fn test_multivalued_i64() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let field = schema_builder.add_i64_field(
             "multifield",
             IntOptions::default().set_fast(Cardinality::MultiValues),

src/fastfield/multivalued/reader.rs

@@ -47,11 +47,11 @@ impl<Item: FastValue> MultiValueIntFastFieldReader<Item> {
 mod tests {
 
     use core::Index;
-    use schema::{Document, Facet, Schema};
+    use schema::{Document, Facet, SchemaBuilder};
 
     #[test]
     fn test_multifastfield_reader() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let facet_field = schema_builder.add_facet_field("facets");
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -82,20 +82,20 @@ mod tests {
 
         let mut facet = Facet::root();
         {
-            facet_reader.facet_from_ord(1, &mut facet).unwrap();
+            facet_reader.facet_from_ord(1, &mut facet);
             assert_eq!(facet, Facet::from("/category"));
         }
         {
-            facet_reader.facet_from_ord(2, &mut facet).unwrap();
+            facet_reader.facet_from_ord(2, &mut facet);
             assert_eq!(facet, Facet::from("/category/cat1"));
         }
         {
-            facet_reader.facet_from_ord(3, &mut facet).unwrap();
+            facet_reader.facet_from_ord(3, &mut facet);
             assert_eq!(format!("{}", facet), "/category/cat2");
             assert_eq!(facet, Facet::from("/category/cat2"));
         }
         {
-            facet_reader.facet_from_ord(4, &mut facet).unwrap();
+            facet_reader.facet_from_ord(4, &mut facet);
             assert_eq!(facet, Facet::from("/category/cat3"));
         }
 

src/fastfield/multivalued/writer.rs

@@ -132,8 +132,7 @@ impl MultiValueIntFastFieldWriter {
                     );
 
                     let mut doc_vals: Vec<u64> = Vec::with_capacity(100);
-                    for (start, stop) in self
-                        .doc_index
+                    for (start, stop) in self.doc_index
                         .windows(2)
                         .map(|interval| (interval[0], interval[1]))
                         .chain(Some(last_interval).into_iter())
@@ -149,6 +148,7 @@ impl MultiValueIntFastFieldWriter {
                             value_serializer.add_val(val)?;
                         }
                     }
+
                 }
                 None => {
                     let val_min_max = self.vals.iter().cloned().minmax();

src/fastfield/reader.rs

@@ -7,10 +7,11 @@ use directory::ReadOnlySource;
 use directory::{Directory, RAMDirectory, WritePtr};
 use fastfield::{FastFieldSerializer, FastFieldsWriter};
 use owning_ref::OwningRef;
-use schema::Schema;
+use schema::SchemaBuilder;
 use schema::FAST;
 use std::collections::HashMap;
 use std::marker::PhantomData;
+use std::mem;
 use std::path::Path;
 use DocId;
 
@@ -79,8 +80,10 @@ impl<Item: FastValue> FastFieldReader<Item> {
     // TODO change start to `u64`.
     // For multifastfield, start is an index in a second fastfield, not a `DocId`
     pub fn get_range(&self, start: u32, output: &mut [Item]) {
-        for (i, out) in output.iter_mut().enumerate() {
-            *out = self.get(start + i as u32);
+        let output_u64: &mut [u64] = unsafe { mem::transmute(output) }; // ok: Item is either `u64` or `i64`
+        self.bit_unpacker.get_range(start, output_u64);
+        for out in output_u64.iter_mut() {
+            *out = Item::from_u64(*out + self.min_value_u64).as_u64();
         }
     }
 
@@ -105,7 +108,7 @@ impl<Item: FastValue> FastFieldReader<Item> {
 
 impl<Item: FastValue> From<Vec<Item>> for FastFieldReader<Item> {
     fn from(vals: Vec<Item>) -> FastFieldReader<Item> {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let field = schema_builder.add_u64_field("field", FAST);
         let schema = schema_builder.build();
         let path = Path::new("__dummy__");

src/fieldnorm/code.rs

@@ -10,28 +10,27 @@ pub fn fieldnorm_to_id(fieldnorm: u32) -> u8 {
         .unwrap_or_else(|idx| idx - 1) as u8
 }
 
-#[cfg_attr(feature = "cargo-clippy", allow(clippy::unreadable_literal))]
 pub const FIELD_NORMS_TABLE: [u32; 256] = [
     0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,
     26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 42, 44, 46, 48, 50, 52, 54, 56, 60,
     64, 68, 72, 76, 80, 84, 88, 96, 104, 112, 120, 128, 136, 144, 152, 168, 184, 200, 216, 232,
-    248, 264, 280, 312, 344, 376, 408, 440, 472, 504, 536, 600, 664, 728, 792, 856, 920, 984,
-    1_048, 1176, 1304, 1432, 1560, 1688, 1816, 1944, 2072, 2328, 2584, 2840, 3096, 3352, 3608,
-    3864, 4120, 4632, 5144, 5656, 6168, 6680, 7192, 7704, 8216, 9240, 10264, 11288, 12312, 13336,
-    14360, 15384, 16408, 18456, 20504, 22552, 24600, 26648, 28696, 30744, 32792, 36888, 40984,
-    45080, 49176, 53272, 57368, 61464, 65560, 73752, 81944, 90136, 98328, 106520, 114712, 122904,
-    131096, 147480, 163864, 180248, 196632, 213016, 229400, 245784, 262168, 294936, 327704, 360472,
-    393240, 426008, 458776, 491544, 524312, 589848, 655384, 720920, 786456, 851992, 917528, 983064,
-    1048600, 1179672, 1310744, 1441816, 1572888, 1703960, 1835032, 1966104, 2097176, 2359320,
-    2621464, 2883608, 3145752, 3407896, 3670040, 3932184, 4194328, 4718616, 5242904, 5767192,
-    6291480, 6815768, 7340056, 7864344, 8388632, 9437208, 10485784, 11534360, 12582936, 13631512,
-    14680088, 15728664, 16777240, 18874392, 20971544, 23068696, 25165848, 27263000, 29360152,
-    31457304, 33554456, 37748760, 41943064, 46137368, 50331672, 54525976, 58720280, 62914584,
-    67108888, 75497496, 83886104, 92274712, 100663320, 109051928, 117440536, 125829144, 134217752,
-    150994968, 167772184, 184549400, 201326616, 218103832, 234881048, 251658264, 268435480,
-    301989912, 335544344, 369098776, 402653208, 436207640, 469762072, 503316504, 536870936,
-    603979800, 671088664, 738197528, 805306392, 872415256, 939524120, 1006632984, 1073741848,
-    1207959576, 1342177304, 1476395032, 1610612760, 1744830488, 1879048216, 2013265944,
+    248, 264, 280, 312, 344, 376, 408, 440, 472, 504, 536, 600, 664, 728, 792, 856, 920, 984, 1048,
+    1176, 1304, 1432, 1560, 1688, 1816, 1944, 2072, 2328, 2584, 2840, 3096, 3352, 3608, 3864, 4120,
+    4632, 5144, 5656, 6168, 6680, 7192, 7704, 8216, 9240, 10264, 11288, 12312, 13336, 14360, 15384,
+    16408, 18456, 20504, 22552, 24600, 26648, 28696, 30744, 32792, 36888, 40984, 45080, 49176,
+    53272, 57368, 61464, 65560, 73752, 81944, 90136, 98328, 106520, 114712, 122904, 131096, 147480,
+    163864, 180248, 196632, 213016, 229400, 245784, 262168, 294936, 327704, 360472, 393240, 426008,
+    458776, 491544, 524312, 589848, 655384, 720920, 786456, 851992, 917528, 983064, 1048600,
+    1179672, 1310744, 1441816, 1572888, 1703960, 1835032, 1966104, 2097176, 2359320, 2621464,
+    2883608, 3145752, 3407896, 3670040, 3932184, 4194328, 4718616, 5242904, 5767192, 6291480,
+    6815768, 7340056, 7864344, 8388632, 9437208, 10485784, 11534360, 12582936, 13631512, 14680088,
+    15728664, 16777240, 18874392, 20971544, 23068696, 25165848, 27263000, 29360152, 31457304,
+    33554456, 37748760, 41943064, 46137368, 50331672, 54525976, 58720280, 62914584, 67108888,
+    75497496, 83886104, 92274712, 100663320, 109051928, 117440536, 125829144, 134217752, 150994968,
+    167772184, 184549400, 201326616, 218103832, 234881048, 251658264, 268435480, 301989912,
+    335544344, 369098776, 402653208, 436207640, 469762072, 503316504, 536870936, 603979800,
+    671088664, 738197528, 805306392, 872415256, 939524120, 1006632984, 1073741848, 1207959576,
+    1342177304, 1476395032, 1610612760, 1744830488, 1879048216, 2013265944,
 ];
 
 #[cfg(test)]

src/fieldnorm/mod.rs

@@ -15,7 +15,7 @@
 //! precompute computationally expensive functions of the fieldnorm
 //! in a very short array.
 //!
-//! This trick is used by the BM25 similarity.
+//! This trick is used by the [BM25 similarity]().
 mod code;
 mod reader;
 mod serializer;

src/functional_test.rs

@@ -2,6 +2,7 @@ use rand::thread_rng;
 use std::collections::HashSet;
 
 use rand::Rng;
+use rand::distributions::Range;
 use schema::*;
 use Index;
 use Searcher;
@@ -15,7 +16,7 @@ fn check_index_content(searcher: &Searcher, vals: &HashSet<u64>) {
 #[ignore]
 #[cfg(feature = "mmap")]
 fn test_indexing() {
-    let mut schema_builder = Schema::builder();
+    let mut schema_builder = SchemaBuilder::default();
 
     let id_field = schema_builder.add_u64_field("id", INT_INDEXED);
     let multiples_field = schema_builder.add_u64_field("multiples", INT_INDEXED);
@@ -23,6 +24,7 @@ fn test_indexing() {
 
     let index = Index::create_from_tempdir(schema).unwrap();
 
+    let universe = Range::new(0u64, 20u64);
     let mut rng = thread_rng();
 
     let mut index_writer = index.writer_with_num_threads(3, 120_000_000).unwrap();
@@ -31,7 +33,7 @@ fn test_indexing() {
     let mut uncommitted_docs: HashSet<u64> = HashSet::new();
 
     for _ in 0..200 {
-        let random_val = rng.gen_range(0, 20);
+        let random_val = rng.sample(&universe);
         if random_val == 0 {
             index_writer.commit().expect("Commit failed");
             committed_docs.extend(&uncommitted_docs);

src/indexer/delete_queue.rs

@@ -52,8 +52,7 @@ impl DeleteQueue {
     //
     // Past delete operations are not accessible.
     pub fn cursor(&self) -> DeleteCursor {
-        let last_block = self
-            .inner
+        let last_block = self.inner
             .read()
             .expect("Read lock poisoned when opening delete queue cursor")
             .last_block
@@ -93,8 +92,7 @@ impl DeleteQueue {
     // be some unflushed operations.
     //
     fn flush(&self) -> Option<Arc<Block>> {
-        let mut self_wlock = self
-            .inner
+        let mut self_wlock = self.inner
             .write()
             .expect("Failed to acquire write lock on delete queue writer");
 
@@ -134,8 +132,7 @@ impl From<DeleteQueue> for NextBlock {
 impl NextBlock {
     fn next_block(&self) -> Option<Arc<Block>> {
         {
-            let next_read_lock = self
-                .0
+            let next_read_lock = self.0
                 .read()
                 .expect("Failed to acquire write lock in delete queue");
             if let InnerNextBlock::Closed(ref block) = *next_read_lock {
@@ -144,8 +141,7 @@ impl NextBlock {
         }
         let next_block;
         {
-            let mut next_write_lock = self
-                .0
+            let mut next_write_lock = self.0
                 .write()
                 .expect("Failed to acquire write lock in delete queue");
             match *next_write_lock {
@@ -186,18 +182,19 @@ impl DeleteCursor {
     /// `opstamp >= target_opstamp`.
     pub fn skip_to(&mut self, target_opstamp: u64) {
         // TODO Can be optimize as we work with block.
-        while self.is_behind_opstamp(target_opstamp) {
+        #[cfg_attr(feature = "cargo-clippy", allow(while_let_loop))]
+        loop {
+            if let Some(operation) = self.get() {
+                if operation.opstamp >= target_opstamp {
+                    break;
+                }
+            } else {
+                break;
+            }
             self.advance();
         }
     }
 
-    #[cfg_attr(feature = "cargo-clippy", allow(clippy::wrong_self_convention))]
-    fn is_behind_opstamp(&mut self, target_opstamp: u64) -> bool {
-        self.get()
-            .map(|operation| operation.opstamp < target_opstamp)
-            .unwrap_or(false)
-    }
-
     /// If the current block has been entirely
     /// consumed, try to load the next one.
     ///

src/indexer/directory_lock.rs

@@ -0,0 +1,27 @@
+use core::LOCKFILE_FILEPATH;
+use directory::error::OpenWriteError;
+use Directory;
+
+/// The directory lock is a mechanism used to
+/// prevent the creation of two [`IndexWriter`](struct.IndexWriter.html)
+///
+/// Only one lock can exist at a time for a given directory.
+/// The lock is release automatically on `Drop`.
+pub struct DirectoryLock {
+    directory: Box<Directory>,
+}
+
+impl DirectoryLock {
+    pub fn lock(mut directory: Box<Directory>) -> Result<DirectoryLock, OpenWriteError> {
+        directory.open_write(&*LOCKFILE_FILEPATH)?;
+        Ok(DirectoryLock { directory })
+    }
+}
+
+impl Drop for DirectoryLock {
+    fn drop(&mut self) {
+        if let Err(e) = self.directory.delete(&*LOCKFILE_FILEPATH) {
+            error!("Failed to remove the lock file. {:?}", e);
+        }
+    }
+}

src/indexer/index_writer.rs

@@ -8,16 +8,16 @@ use core::SegmentComponent;
 use core::SegmentId;
 use core::SegmentMeta;
 use core::SegmentReader;
-use crossbeam::channel;
-use directory::DirectoryLock;
+use crossbeam_channel as channel;
 use docset::DocSet;
 use error::TantivyError;
 use fastfield::write_delete_bitset;
-use futures::{Canceled, Future};
+use futures::sync::oneshot::Receiver;
 use indexer::delete_queue::{DeleteCursor, DeleteQueue};
 use indexer::doc_opstamp_mapping::DocToOpstampMapping;
 use indexer::operation::DeleteOperation;
 use indexer::stamper::Stamper;
+use indexer::DirectoryLock;
 use indexer::MergePolicy;
 use indexer::SegmentEntry;
 use indexer::SegmentWriter;
@@ -26,7 +26,7 @@ use schema::Document;
 use schema::IndexRecordOption;
 use schema::Term;
 use std::mem;
-use std::sync::Arc;
+use std::mem::swap;
 use std::thread;
 use std::thread::JoinHandle;
 use Result;
@@ -52,19 +52,16 @@ type DocumentReceiver = channel::Receiver<AddOperation>;
 ///
 /// Returns (the heap size in bytes, the hash table size in number of bits)
 fn initial_table_size(per_thread_memory_budget: usize) -> usize {
-    assert!(per_thread_memory_budget > 1_000);
     let table_size_limit: usize = per_thread_memory_budget / 3;
-    if let Some(limit) = (1..)
+    (1..)
+        .into_iter()
         .take_while(|num_bits: &usize| compute_table_size(*num_bits) < table_size_limit)
         .last()
-    {
-        limit.min(19) // we cap it at 2^19 = 512K.
-    } else {
-        unreachable!(
+        .expect(&format!(
             "Per thread memory is too small: {}",
             per_thread_memory_budget
-        );
-    }
+        ))
+        .min(19) // we cap it at 512K
 }
 
 /// `IndexWriter` is the user entry-point to add document to an index.
@@ -142,7 +139,7 @@ pub fn open_index_writer(
     let stamper = Stamper::new(current_opstamp);
 
     let segment_updater =
-        SegmentUpdater::create(index.clone(), stamper.clone(), &delete_queue.cursor())?;
+        SegmentUpdater::new(index.clone(), stamper.clone(), &delete_queue.cursor())?;
 
     let mut index_writer = IndexWriter {
         _directory_lock: Some(directory_lock),
@@ -180,7 +177,7 @@ pub fn compute_deleted_bitset(
 ) -> Result<bool> {
     let mut might_have_changed = false;
 
-    #[cfg_attr(feature = "cargo-clippy", allow(clippy::while_let_loop))]
+    #[cfg_attr(feature = "cargo-clippy", allow(while_let_loop))]
     loop {
         if let Some(delete_op) = delete_cursor.get() {
             if delete_op.opstamp > target_opstamp {
@@ -304,28 +301,25 @@ fn index_documents(
 
     let last_docstamp: u64 = *(doc_opstamps.last().unwrap());
 
-    let delete_bitset_opt = if delete_cursor.get().is_some() {
-        let doc_to_opstamps = DocToOpstampMapping::from(doc_opstamps);
-        let segment_reader = SegmentReader::open(segment)?;
-        let mut deleted_bitset = BitSet::with_capacity(num_docs as usize);
-        let may_have_deletes = compute_deleted_bitset(
-            &mut deleted_bitset,
-            &segment_reader,
-            &mut delete_cursor,
-            &doc_to_opstamps,
-            last_docstamp,
-        )?;
+    let doc_to_opstamps = DocToOpstampMapping::from(doc_opstamps);
+    let segment_reader = SegmentReader::open(segment)?;
+    let mut deleted_bitset = BitSet::with_capacity(num_docs as usize);
+    let may_have_deletes = compute_deleted_bitset(
+        &mut deleted_bitset,
+        &segment_reader,
+        &mut delete_cursor,
+        &doc_to_opstamps,
+        last_docstamp,
+    )?;
+
+    let segment_entry = SegmentEntry::new(segment_meta, delete_cursor, {
         if may_have_deletes {
             Some(deleted_bitset)
         } else {
             None
         }
-    } else {
-        // if there are no delete operation in the queue, no need
-        // to even open the segment.
-        None
-    };
-    let segment_entry = SegmentEntry::new(segment_meta, delete_cursor, delete_bitset_opt);
+    });
+
     Ok(segment_updater.add_segment(generation, segment_entry))
 }
 
@@ -347,8 +341,7 @@ impl IndexWriter {
         }
         drop(self.workers_join_handle);
 
-        let result = self
-            .segment_updater
+        let result = self.segment_updater
             .wait_merging_thread()
             .map_err(|_| TantivyError::ErrorInThread("Failed to join merging thread.".into()));
 
@@ -367,16 +360,13 @@ impl IndexWriter {
             .add_segment(self.generation, segment_entry);
     }
 
-    /// Creates a new segment.
+    /// *Experimental & Advanced API* Creates a new segment.
+    /// and marks it as currently in write.
     ///
     /// This method is useful only for users trying to do complex
     /// operations, like converting an index format to another.
-    ///
-    /// It is safe to start writing file associated to the new `Segment`.
-    /// These will not be garbage collected as long as an instance object of
-    /// `SegmentMeta` object associated to the new `Segment` is "alive".
     pub fn new_segment(&self) -> Segment {
-        self.index.new_segment()
+        self.segment_updater.new_segment()
     }
 
     /// Spawns a new worker thread for indexing.
@@ -391,10 +381,9 @@ impl IndexWriter {
         let mut delete_cursor = self.delete_queue.cursor();
 
         let mem_budget = self.heap_size_in_bytes_per_thread;
-        let index = self.index.clone();
         let join_handle: JoinHandle<Result<()>> = thread::Builder::new()
             .name(format!(
-                "thrd-tantivy-index{}-gen{}",
+                "indexing thread {} for gen {}",
                 self.worker_id, generation
             ))
             .spawn(move || {
@@ -417,7 +406,7 @@ impl IndexWriter {
                         // was dropped.
                         return Ok(());
                     }
-                    let segment = index.new_segment();
+                    let segment = segment_updater.new_segment();
                     index_documents(
                         mem_budget,
                         &segment,
@@ -434,7 +423,7 @@ impl IndexWriter {
     }
 
     /// Accessor to the merge policy.
-    pub fn get_merge_policy(&self) -> Arc<Box<MergePolicy>> {
+    pub fn get_merge_policy(&self) -> Box<MergePolicy> {
         self.segment_updater.get_merge_policy()
     }
 
@@ -459,10 +448,7 @@ impl IndexWriter {
     /// Merges a given list of segments
     ///
     /// `segment_ids` is required to be non-empty.
-    pub fn merge(
-        &mut self,
-        segment_ids: &[SegmentId],
-    ) -> Result<impl Future<Item = SegmentMeta, Error = Canceled>> {
+    pub fn merge(&mut self, segment_ids: &[SegmentId]) -> Result<Receiver<SegmentMeta>> {
         self.segment_updater.start_merge(segment_ids)
     }
 
@@ -475,10 +461,13 @@ impl IndexWriter {
     ///
     /// Returns the former segment_ready channel.
     fn recreate_document_channel(&mut self) -> DocumentReceiver {
-        let (document_sender, document_receiver): (DocumentSender, DocumentReceiver) =
-            channel::bounded(PIPELINE_MAX_SIZE_IN_DOCS);
-        mem::replace(&mut self.document_sender, document_sender);
-        mem::replace(&mut self.document_receiver, document_receiver)
+        let (mut document_sender, mut document_receiver): (
+            DocumentSender,
+            DocumentReceiver,
+        ) = channel::bounded(PIPELINE_MAX_SIZE_IN_DOCS);
+        swap(&mut self.document_sender, &mut document_sender);
+        swap(&mut self.document_receiver, &mut document_receiver);
+        document_receiver
     }
 
     /// Rollback to the last commit
@@ -499,8 +488,7 @@ impl IndexWriter {
         let document_receiver = self.document_receiver.clone();
 
         // take the directory lock to create a new index_writer.
-        let directory_lock = self
-            ._directory_lock
+        let directory_lock = self._directory_lock
             .take()
             .expect("The IndexWriter does not have any lock. This is a bug, please report.");
 
@@ -565,12 +553,17 @@ impl IndexWriter {
         // and recreate a new one channels.
         self.recreate_document_channel();
 
-        let former_workers_join_handle = mem::replace(&mut self.workers_join_handle, Vec::new());
+        let mut former_workers_join_handle = Vec::new();
+        swap(
+            &mut former_workers_join_handle,
+            &mut self.workers_join_handle,
+        );
 
         for worker_handle in former_workers_join_handle {
             let indexing_worker_result = worker_handle
                 .join()
                 .map_err(|e| TantivyError::ErrorInThread(format!("{:?}", e)))?;
+
             indexing_worker_result?;
             // add a new worker for the next generation.
             self.add_indexing_worker()?;
@@ -643,10 +636,7 @@ impl IndexWriter {
     pub fn add_document(&mut self, document: Document) -> u64 {
         let opstamp = self.stamper.stamp();
         let add_operation = AddOperation { opstamp, document };
-        let send_result = self.document_sender.send(add_operation);
-        if let Err(e) = send_result {
-            panic!("Failed to index document. Sending to indexing channel failed. This probably means all of the indexing threads have panicked. {:?}", e);
-        }
+        self.document_sender.send(add_operation);
         opstamp
     }
 }
@@ -655,7 +645,6 @@ impl IndexWriter {
 mod tests {
 
     use super::initial_table_size;
-    use directory::error::LockError;
     use error::*;
     use indexer::NoMergePolicy;
     use schema::{self, Document};
@@ -664,34 +653,20 @@ mod tests {
 
     #[test]
     fn test_lockfile_stops_duplicates() {
-        let schema_builder = schema::Schema::builder();
+        let schema_builder = schema::SchemaBuilder::default();
         let index = Index::create_in_ram(schema_builder.build());
-        let _index_writer = index.writer(3_000_000).unwrap();
-        match index.writer(3_000_000) {
-            Err(TantivyError::LockFailure(LockError::LockBusy, _)) => {}
-            _ => panic!("Expected a `LockFailure` error"),
-        }
-    }
-
-    #[test]
-    fn test_lockfile_already_exists_error_msg() {
-        let schema_builder = schema::Schema::builder();
-        let index = Index::create_in_ram(schema_builder.build());
-        let _index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-        match index.writer_with_num_threads(1, 3_000_000) {
-            Err(err) => {
-                let err_msg = err.to_string();
-                assert!(err_msg.contains("already an `IndexWriter`"));
-            }
-            _ => panic!("Expected LockfileAlreadyExists error"),
+        let _index_writer = index.writer(40_000_000).unwrap();
+        match index.writer(40_000_000) {
+            Err(TantivyError::FileAlreadyExists(_)) => {}
+            _ => panic!("Expected FileAlreadyExists error"),
         }
     }
 
     #[test]
     fn test_set_merge_policy() {
-        let schema_builder = schema::Schema::builder();
+        let schema_builder = schema::SchemaBuilder::default();
         let index = Index::create_in_ram(schema_builder.build());
-        let index_writer = index.writer(3_000_000).unwrap();
+        let index_writer = index.writer(40_000_000).unwrap();
         assert_eq!(
             format!("{:?}", index_writer.get_merge_policy()),
             "LogMergePolicy { min_merge_size: 8, min_layer_size: 10000, \
@@ -707,19 +682,19 @@ mod tests {
 
     #[test]
     fn test_lockfile_released_on_drop() {
-        let schema_builder = schema::Schema::builder();
+        let schema_builder = schema::SchemaBuilder::default();
         let index = Index::create_in_ram(schema_builder.build());
         {
-            let _index_writer = index.writer(3_000_000).unwrap();
+            let _index_writer = index.writer(40_000_000).unwrap();
             // the lock should be released when the
             // index_writer leaves the scope.
         }
-        let _index_writer_two = index.writer(3_000_000).unwrap();
+        let _index_writer_two = index.writer(40_000_000).unwrap();
     }
 
     #[test]
     fn test_commit_and_rollback() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
 
@@ -741,7 +716,7 @@ mod tests {
                 index_writer.add_document(doc!(text_field=>"b"));
                 index_writer.add_document(doc!(text_field=>"c"));
             }
-            assert!(index_writer.commit().is_ok());
+            assert_eq!(index_writer.commit().unwrap(), 2u64);
             index.load_searchers().unwrap();
             assert_eq!(num_docs_containing("a"), 0);
             assert_eq!(num_docs_containing("b"), 1);
@@ -753,7 +728,7 @@ mod tests {
 
     #[test]
     fn test_with_merges() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
         let num_docs_containing = |s: &str| {
@@ -790,7 +765,7 @@ mod tests {
 
     #[test]
     fn test_prepare_with_commit_message() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
 
@@ -804,6 +779,7 @@ mod tests {
             {
                 let mut prepared_commit = index_writer.prepare_commit().expect("commit failed");
                 prepared_commit.set_payload("first commit");
+                assert_eq!(prepared_commit.opstamp(), 100);
                 prepared_commit.commit().expect("commit failed");
             }
             {
@@ -823,7 +799,7 @@ mod tests {
 
     #[test]
     fn test_prepare_but_rollback() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_in_ram(schema_builder.build());
 
@@ -837,6 +813,7 @@ mod tests {
             {
                 let mut prepared_commit = index_writer.prepare_commit().expect("commit failed");
                 prepared_commit.set_payload("first commit");
+                assert_eq!(prepared_commit.opstamp(), 100);
                 prepared_commit.abort().expect("commit failed");
             }
             {
@@ -866,32 +843,4 @@ mod tests {
         assert_eq!(initial_table_size(1_000_000_000), 19);
     }
 
-    #[cfg(not(feature = "no_fail"))]
-    #[test]
-    fn test_write_commit_fails() {
-        use fail;
-        let mut schema_builder = schema::Schema::builder();
-        let text_field = schema_builder.add_text_field("text", schema::TEXT);
-        let index = Index::create_in_ram(schema_builder.build());
-
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-        for _ in 0..100 {
-            index_writer.add_document(doc!(text_field => "a"));
-        }
-        index_writer.commit().unwrap();
-        fail::cfg("RAMDirectory::atomic_write", "return(error_write_failed)").unwrap();
-        for _ in 0..100 {
-            index_writer.add_document(doc!(text_field => "b"));
-        }
-        assert!(index_writer.commit().is_err());
-        index.load_searchers().unwrap();
-        let num_docs_containing = |s: &str| {
-            let searcher = index.searcher();
-            let term_a = Term::from_field_text(text_field, s);
-            searcher.doc_freq(&term_a)
-        };
-        assert_eq!(num_docs_containing("a"), 100);
-        assert_eq!(num_docs_containing("b"), 0);
-        fail::cfg("RAMDirectory::atomic_write", "off").unwrap();
-    }
 }

src/indexer/merge_operation.rs

@@ -1,64 +0,0 @@
-use census::{Inventory, TrackedObject};
-use std::collections::HashSet;
-use SegmentId;
-
-#[derive(Default)]
-pub struct MergeOperationInventory(Inventory<InnerMergeOperation>);
-
-impl MergeOperationInventory {
-    pub fn segment_in_merge(&self) -> HashSet<SegmentId> {
-        let mut segment_in_merge = HashSet::default();
-        for merge_op in self.0.list() {
-            for &segment_id in &merge_op.segment_ids {
-                segment_in_merge.insert(segment_id);
-            }
-        }
-        segment_in_merge
-    }
-}
-
-/// A `MergeOperation` has two role.
-/// It carries all of the information required to describe a merge :
-/// - `target_opstamp` is the opstamp up to which we want to consume the
-/// delete queue and reflect their deletes.
-/// - `segment_ids` is the list of segment to be merged.
-///
-/// The second role is to ensure keep track of the fact that these
-/// segments are in merge and avoid starting a merge operation that
-/// may conflict with this one.
-///
-/// This works by tracking merge operations. When considering computing
-/// merge candidates, we simply list tracked merge operations and remove
-/// their segments from possible merge candidates.
-pub struct MergeOperation {
-    inner: TrackedObject<InnerMergeOperation>,
-}
-
-struct InnerMergeOperation {
-    target_opstamp: u64,
-    segment_ids: Vec<SegmentId>,
-}
-
-impl MergeOperation {
-    pub fn new(
-        inventory: &MergeOperationInventory,
-        target_opstamp: u64,
-        segment_ids: Vec<SegmentId>,
-    ) -> MergeOperation {
-        let inner_merge_operation = InnerMergeOperation {
-            target_opstamp,
-            segment_ids,
-        };
-        MergeOperation {
-            inner: inventory.0.track(inner_merge_operation),
-        }
-    }
-
-    pub fn target_opstamp(&self) -> u64 {
-        self.inner.target_opstamp
-    }
-
-    pub fn segment_ids(&self) -> &[SegmentId] {
-        &self.inner.segment_ids[..]
-    }
-}

src/indexer/merge_policy.rs

@@ -11,7 +11,7 @@ pub struct MergeCandidate(pub Vec<SegmentId>);
 ///
 /// Every time a the list of segments changes, the segment updater
 /// asks the merge policy if some segments should be merged.
-pub trait MergePolicy: marker::Send + marker::Sync + Debug {
+pub trait MergePolicy: MergePolicyClone + marker::Send + marker::Sync + Debug {
     /// Given the list of segment metas, returns the list of merge candidates.
     ///
     /// This call happens on the segment updater thread, and will block
@@ -19,6 +19,21 @@ pub trait MergePolicy: marker::Send + marker::Sync + Debug {
     fn compute_merge_candidates(&self, segments: &[SegmentMeta]) -> Vec<MergeCandidate>;
 }
 
+/// MergePolicyClone
+pub trait MergePolicyClone {
+  /// Returns a boxed clone of the MergePolicy.
+  fn box_clone(&self) -> Box<MergePolicy>;
+}
+
+impl<T> MergePolicyClone for T
+where
+  T: 'static + MergePolicy + Clone,
+{
+  fn box_clone(&self) -> Box<MergePolicy> {
+    Box::new(self.clone())
+  }
+}
+
 /// Never merge segments.
 #[derive(Debug, Clone)]
 pub struct NoMergePolicy;

src/indexer/merger.rs

@@ -46,7 +46,7 @@ fn compute_total_num_tokens(readers: &[SegmentReader], field: Field) -> u64 {
             .cloned()
             .enumerate()
             .map(|(fieldnorm_ord, count)| {
-                count as u64 * u64::from(FieldNormReader::id_to_fieldnorm(fieldnorm_ord as u8))
+                count as u64 * FieldNormReader::id_to_fieldnorm(fieldnorm_ord as u8) as u64
             })
             .sum::<u64>()
 }
@@ -111,7 +111,7 @@ impl TermOrdinalMapping {
             .iter()
             .flat_map(|term_ordinals| term_ordinals.iter().cloned().max())
             .max()
-            .unwrap_or_else(TermOrdinal::default)
+            .unwrap_or(TermOrdinal::default())
     }
 }
 
@@ -190,7 +190,7 @@ impl IndexMerger {
                         `term_ordinal_mapping`.");
                     self.write_hierarchical_facet_field(
                         field,
-                        &term_ordinal_mapping,
+                        term_ordinal_mapping,
                         fast_field_serializer,
                     )?;
                 }
@@ -314,7 +314,7 @@ impl IndexMerger {
     fn write_hierarchical_facet_field(
         &self,
         field: Field,
-        term_ordinal_mappings: &TermOrdinalMapping,
+        term_ordinal_mappings: TermOrdinalMapping,
         fast_field_serializer: &mut FastFieldSerializer,
     ) -> Result<()> {
         // Multifastfield consists in 2 fastfields.
@@ -393,8 +393,8 @@ impl IndexMerger {
 
         // We can now initialize our serializer, and push it the different values
         {
-            let mut serialize_vals = fast_field_serializer
-                .new_u64_fast_field_with_idx(field, min_value, max_value, 1)?;
+            let mut serialize_vals =
+                fast_field_serializer.new_u64_fast_field_with_idx(field, min_value, max_value, 1)?;
             for reader in &self.readers {
                 let ff_reader: MultiValueIntFastFieldReader<u64> =
                     reader.multi_fast_field_reader(field)?;
@@ -440,8 +440,7 @@ impl IndexMerger {
     ) -> Result<Option<TermOrdinalMapping>> {
         let mut positions_buffer: Vec<u32> = Vec::with_capacity(1_000);
         let mut delta_computer = DeltaComputer::new();
-        let field_readers = self
-            .readers
+        let field_readers = self.readers
             .iter()
             .map(|reader| reader.inverted_index(indexed_field))
             .collect::<Vec<_>>();
@@ -617,7 +616,7 @@ impl IndexMerger {
                     store_writer.store(&doc)?;
                 }
             } else {
-                store_writer.stack(&store_reader)?;
+                store_writer.stack(store_reader)?;
             }
         }
         Ok(())
@@ -638,9 +637,10 @@ impl SerializableSegment for IndexMerger {
 #[cfg(test)]
 mod tests {
     use byteorder::{BigEndian, ReadBytesExt, WriteBytesExt};
+    use collector::chain;
     use collector::tests::TestCollector;
     use collector::tests::{BytesFastFieldTestCollector, FastFieldTestCollector};
-    use collector::{Count, FacetCollector};
+    use collector::FacetCollector;
     use core::Index;
     use futures::Future;
     use query::AllQuery;
@@ -649,12 +649,10 @@ mod tests {
     use schema;
     use schema::Cardinality;
     use schema::Document;
-    use schema::Facet;
     use schema::IndexRecordOption;
     use schema::IntOptions;
     use schema::Term;
     use schema::TextFieldIndexing;
-    use schema::INT_INDEXED;
     use std::io::Cursor;
     use DocAddress;
     use IndexWriter;
@@ -662,7 +660,7 @@ mod tests {
 
     #[test]
     fn test_index_merger_no_deletes() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let text_fieldtype = schema::TextOptions::default()
             .set_indexing_options(
                 TextFieldIndexing::default()
@@ -747,69 +745,63 @@ mod tests {
             index.load_searchers().unwrap();
             let searcher = index.searcher();
             let get_doc_ids = |terms: Vec<Term>| {
+                let mut collector = TestCollector::default();
                 let query = BooleanQuery::new_multiterms_query(terms);
-                let top_docs = searcher.search(&query, &TestCollector).unwrap();
-                top_docs.docs().to_vec()
+                assert!(searcher.search(&query, &mut collector).is_ok());
+                collector.docs()
             };
             {
                 assert_eq!(
                     get_doc_ids(vec![Term::from_field_text(text_field, "a")]),
-                    vec![DocAddress(0, 1), DocAddress(0, 2), DocAddress(0, 4)]
+                    vec![1, 2, 4]
                 );
                 assert_eq!(
                     get_doc_ids(vec![Term::from_field_text(text_field, "af")]),
-                    vec![DocAddress(0, 0), DocAddress(0, 3)]
+                    vec![0, 3]
                 );
                 assert_eq!(
                     get_doc_ids(vec![Term::from_field_text(text_field, "g")]),
-                    vec![DocAddress(0, 4)]
+                    vec![4]
                 );
                 assert_eq!(
                     get_doc_ids(vec![Term::from_field_text(text_field, "b")]),
-                    vec![
-                        DocAddress(0, 0),
-                        DocAddress(0, 1),
-                        DocAddress(0, 2),
-                        DocAddress(0, 3),
-                        DocAddress(0, 4)
-                    ]
+                    vec![0, 1, 2, 3, 4]
                 );
             }
             {
-                let doc = searcher.doc(DocAddress(0, 0)).unwrap();
-                assert_eq!(doc.get_first(text_field).unwrap().text(), Some("af b"));
+                let doc = searcher.doc(&DocAddress(0, 0)).unwrap();
+                assert_eq!(doc.get_first(text_field).unwrap().text(), "af b");
             }
             {
-                let doc = searcher.doc(DocAddress(0, 1)).unwrap();
-                assert_eq!(doc.get_first(text_field).unwrap().text(), Some("a b c"));
+                let doc = searcher.doc(&DocAddress(0, 1)).unwrap();
+                assert_eq!(doc.get_first(text_field).unwrap().text(), "a b c");
             }
             {
-                let doc = searcher.doc(DocAddress(0, 2)).unwrap();
-                assert_eq!(doc.get_first(text_field).unwrap().text(), Some("a b c d"));
+                let doc = searcher.doc(&DocAddress(0, 2)).unwrap();
+                assert_eq!(doc.get_first(text_field).unwrap().text(), "a b c d");
             }
             {
-                let doc = searcher.doc(DocAddress(0, 3)).unwrap();
-                assert_eq!(doc.get_first(text_field).unwrap().text(), Some("af b"));
+                let doc = searcher.doc(&DocAddress(0, 3)).unwrap();
+                assert_eq!(doc.get_first(text_field).unwrap().text(), "af b");
             }
             {
-                let doc = searcher.doc(DocAddress(0, 4)).unwrap();
-                assert_eq!(doc.get_first(text_field).unwrap().text(), Some("a b c g"));
+                let doc = searcher.doc(&DocAddress(0, 4)).unwrap();
+                assert_eq!(doc.get_first(text_field).unwrap().text(), "a b c g");
             }
             {
                 let get_fast_vals = |terms: Vec<Term>| {
                     let query = BooleanQuery::new_multiterms_query(terms);
-                    searcher
-                        .search(&query, &FastFieldTestCollector::for_field(score_field))
-                        .unwrap()
+                    let mut collector = FastFieldTestCollector::for_field(score_field);
+                    assert!(searcher.search(&query, &mut collector).is_ok());
+                    collector.vals()
                 };
                 let get_fast_vals_bytes = |terms: Vec<Term>| {
                     let query = BooleanQuery::new_multiterms_query(terms);
+                    let mut collector = BytesFastFieldTestCollector::for_field(bytes_score_field);
                     searcher
-                        .search(
-                            &query,
-                            &BytesFastFieldTestCollector::for_field(bytes_score_field),
-                        )
-                        .expect("failed to search")
+                        .search(&query, &mut collector)
+                        .expect("failed to search");
+                    collector.vals()
                 };
                 assert_eq!(
                     get_fast_vals(vec![Term::from_field_text(text_field, "a")]),
@@ -825,7 +817,7 @@ mod tests {
 
     #[test]
     fn test_index_merger_with_deletes() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let text_fieldtype = schema::TextOptions::default()
             .set_indexing_options(
                 TextFieldIndexing::default().set_index_option(IndexRecordOption::WithFreqs),
@@ -836,16 +828,24 @@ mod tests {
         let score_field = schema_builder.add_u64_field("score", score_fieldtype);
         let bytes_score_field = schema_builder.add_bytes_field("score_bytes");
         let index = Index::create_in_ram(schema_builder.build());
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
 
         let search_term = |searcher: &Searcher, term: Term| {
-            let collector = FastFieldTestCollector::for_field(score_field);
-            let bytes_collector = BytesFastFieldTestCollector::for_field(bytes_score_field);
+            let mut collector = FastFieldTestCollector::for_field(score_field);
+            let mut bytes_collector = BytesFastFieldTestCollector::for_field(bytes_score_field);
             let term_query = TermQuery::new(term, IndexRecordOption::Basic);
-            let (scores, bytes) = searcher
-                .search(&term_query, &(collector, bytes_collector))
-                .unwrap();
-            let mut score_bytes = Cursor::new(bytes);
+
+            {
+                let mut combined_collector =
+                    chain().push(&mut collector).push(&mut bytes_collector);
+                searcher
+                    .search(&term_query, &mut combined_collector)
+                    .unwrap();
+            }
+
+            let scores = collector.vals();
+
+            let mut score_bytes = Cursor::new(bytes_collector.vals());
             for &score in &scores {
                 assert_eq!(score as u32, score_bytes.read_u32::<BigEndian>().unwrap());
             }
@@ -858,21 +858,21 @@ mod tests {
         {
             // a first commit
             index_writer.add_document(doc!(
-                text_field => "a b d",
-                score_field => 1u64,
-                bytes_score_field => vec![0u8, 0, 0, 1],
-            ));
+                    text_field => "a b d",
+                    score_field => 1u64,
+                    bytes_score_field => vec![0u8, 0, 0, 1],
+                ));
             index_writer.add_document(doc!(
-                text_field => "b c",
-                score_field => 2u64,
-                bytes_score_field => vec![0u8, 0, 0, 2],
-            ));
+                    text_field => "b c",
+                    score_field => 2u64,
+                    bytes_score_field => vec![0u8, 0, 0, 2],
+                ));
             index_writer.delete_term(Term::from_field_text(text_field, "c"));
             index_writer.add_document(doc!(
-                text_field => "c d",
-                score_field => 3u64,
-                bytes_score_field => vec![0u8, 0, 0, 3],
-            ));
+                    text_field => "c d",
+                    score_field => 3u64,
+                    bytes_score_field => vec![0u8, 0, 0, 3],
+                ));
             index_writer.commit().expect("committed");
             index.load_searchers().unwrap();
             let ref searcher = *index.searcher();
@@ -899,37 +899,37 @@ mod tests {
         {
             // a second commit
             index_writer.add_document(doc!(
-                text_field => "a d e",
-                score_field => 4_000u64,
-                bytes_score_field => vec![0u8, 0, 0, 4],
-            ));
+                    text_field => "a d e",
+                    score_field => 4_000u64,
+                    bytes_score_field => vec![0u8, 0, 0, 4],
+                ));
             index_writer.add_document(doc!(
-                text_field => "e f",
-                score_field => 5_000u64,
-                bytes_score_field => vec![0u8, 0, 0, 5],
-            ));
+                    text_field => "e f",
+                    score_field => 5_000u64,
+                    bytes_score_field => vec![0u8, 0, 0, 5],
+                ));
             index_writer.delete_term(Term::from_field_text(text_field, "a"));
             index_writer.delete_term(Term::from_field_text(text_field, "f"));
             index_writer.add_document(doc!(
-                text_field => "f g",
-                score_field => 6_000u64,
-                bytes_score_field => vec![0u8, 0, 23, 112],
-            ));
+                    text_field => "f g",
+                    score_field => 6_000u64,
+                    bytes_score_field => vec![0u8, 0, 23, 112],
+                ));
             index_writer.add_document(doc!(
-                text_field => "g h",
-                score_field => 7_000u64,
-                bytes_score_field => vec![0u8, 0, 27, 88],
-            ));
+                    text_field => "g h",
+                    score_field => 7_000u64,
+                    bytes_score_field => vec![0u8, 0, 27, 88],
+                ));
             index_writer.commit().expect("committed");
             index.load_searchers().unwrap();
             let searcher = index.searcher();
 
             assert_eq!(searcher.segment_readers().len(), 2);
             assert_eq!(searcher.num_docs(), 3);
-            assert_eq!(searcher.segment_readers()[0].num_docs(), 2);
-            assert_eq!(searcher.segment_readers()[0].max_doc(), 4);
-            assert_eq!(searcher.segment_readers()[1].num_docs(), 1);
-            assert_eq!(searcher.segment_readers()[1].max_doc(), 3);
+            assert_eq!(searcher.segment_readers()[0].num_docs(), 1);
+            assert_eq!(searcher.segment_readers()[0].max_doc(), 3);
+            assert_eq!(searcher.segment_readers()[1].num_docs(), 2);
+            assert_eq!(searcher.segment_readers()[1].max_doc(), 4);
             assert_eq!(
                 search_term(&searcher, Term::from_field_text(text_field, "a")),
                 empty_vec
@@ -963,15 +963,15 @@ mod tests {
                 .segment_reader(0)
                 .fast_field_reader::<u64>(score_field)
                 .unwrap();
-            assert_eq!(score_field_reader.min_value(), 4000);
-            assert_eq!(score_field_reader.max_value(), 7000);
+            assert_eq!(score_field_reader.min_value(), 1);
+            assert_eq!(score_field_reader.max_value(), 3);
 
             let score_field_reader = searcher
                 .segment_reader(1)
                 .fast_field_reader::<u64>(score_field)
                 .unwrap();
-            assert_eq!(score_field_reader.min_value(), 1);
-            assert_eq!(score_field_reader.max_value(), 3);
+            assert_eq!(score_field_reader.min_value(), 4000);
+            assert_eq!(score_field_reader.max_value(), 7000);
         }
         {
             // merging the segments
@@ -984,7 +984,7 @@ mod tests {
                 .wait()
                 .expect("Merging failed");
             index.load_searchers().unwrap();
-            let searcher = index.searcher();
+            let ref searcher = *index.searcher();
             assert_eq!(searcher.segment_readers().len(), 1);
             assert_eq!(searcher.num_docs(), 3);
             assert_eq!(searcher.segment_readers()[0].num_docs(), 3);
@@ -1030,7 +1030,7 @@ mod tests {
             index_writer.commit().unwrap();
 
             index.load_searchers().unwrap();
-            let searcher = index.searcher();
+            let ref searcher = *index.searcher();
             assert_eq!(searcher.segment_readers().len(), 1);
             assert_eq!(searcher.num_docs(), 2);
             assert_eq!(searcher.segment_readers()[0].num_docs(), 2);
@@ -1126,26 +1126,30 @@ mod tests {
         {
             // Test removing all docs
             index_writer.delete_term(Term::from_field_text(text_field, "g"));
-            index_writer.commit().unwrap();
             let segment_ids = index
                 .searchable_segment_ids()
                 .expect("Searchable segments failed.");
+            index_writer
+                .merge(&segment_ids)
+                .expect("Failed to initiate merge")
+                .wait()
+                .expect("Merging failed");
             index.load_searchers().unwrap();
 
             let ref searcher = *index.searcher();
-            assert!(segment_ids.is_empty());
-            assert!(searcher.segment_readers().is_empty());
+            assert_eq!(searcher.segment_readers().len(), 1);
             assert_eq!(searcher.num_docs(), 0);
         }
     }
 
     #[test]
     fn test_merge_facets() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let facet_field = schema_builder.add_facet_field("facet");
         let index = Index::create_in_ram(schema_builder.build());
+        use schema::Facet;
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             let index_doc = |index_writer: &mut IndexWriter, doc_facets: &[&str]| {
                 let mut doc = Document::default();
                 for facet in doc_facets {
@@ -1172,16 +1176,20 @@ mod tests {
             index_doc(&mut index_writer, &["/top/e", "/top/f"]);
             index_writer.commit().expect("committed");
         }
-
         index.load_searchers().unwrap();
         let test_searcher = |expected_num_docs: usize, expected: &[(&str, u64)]| {
             let searcher = index.searcher();
             let mut facet_collector = FacetCollector::for_field(facet_field);
             facet_collector.add_facet(Facet::from("/top"));
-            let (count, facet_counts) = searcher
-                .search(&AllQuery, &(Count, facet_collector))
-                .unwrap();
-            assert_eq!(count, expected_num_docs);
+            use collector::{CountCollector, MultiCollector};
+            let mut count_collector = CountCollector::default();
+            {
+                let mut multi_collectors =
+                    MultiCollector::from(vec![&mut count_collector, &mut facet_collector]);
+                searcher.search(&AllQuery, &mut multi_collectors).unwrap();
+            }
+            assert_eq!(count_collector.count(), expected_num_docs);
+            let facet_counts = facet_collector.harvest();
             let facets: Vec<(String, u64)> = facet_counts
                 .get("/top")
                 .map(|(facet, count)| (facet.to_string(), count))
@@ -1205,18 +1213,20 @@ mod tests {
                 ("/top/f", 1),
             ],
         );
+
         // Merging the segments
         {
             let segment_ids = index
                 .searchable_segment_ids()
                 .expect("Searchable segments failed.");
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             index_writer
                 .merge(&segment_ids)
                 .expect("Failed to initiate merge")
                 .wait()
                 .expect("Merging failed");
             index_writer.wait_merging_threads().unwrap();
+
             index.load_searchers().unwrap();
             test_searcher(
                 11,
@@ -1233,7 +1243,7 @@ mod tests {
 
         // Deleting one term
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             let facet = Facet::from_path(vec!["top", "a", "firstdoc"]);
             let facet_term = Term::from_facet(facet_field, &facet);
             index_writer.delete_term(facet_term);
@@ -1253,37 +1263,9 @@ mod tests {
         }
     }
 
-    #[test]
-    fn test_bug_merge() {
-        let mut schema_builder = schema::Schema::builder();
-        let int_field = schema_builder.add_u64_field("intvals", INT_INDEXED);
-        let index = Index::create_in_ram(schema_builder.build());
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-        index_writer.add_document(doc!(int_field => 1u64));
-        index_writer.commit().expect("commit failed");
-        index_writer.add_document(doc!(int_field => 1u64));
-        index_writer.commit().expect("commit failed");
-        index.load_searchers().unwrap();
-        let searcher = index.searcher();
-        assert_eq!(searcher.num_docs(), 2);
-        index_writer.delete_term(Term::from_field_u64(int_field, 1));
-        let segment_ids = index
-            .searchable_segment_ids()
-            .expect("Searchable segments failed.");
-        index_writer
-            .merge(&segment_ids)
-            .expect("Failed to initiate merge")
-            .wait()
-            .expect("Merging failed");
-        index.load_searchers().unwrap();
-        // commit has not been called yet. The document should still be
-        // there.
-        assert_eq!(index.searcher().num_docs(), 2);
-    }
-
     #[test]
     fn test_merge_multivalued_int_fields_all_deleted() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let int_options = IntOptions::default()
             .set_fast(Cardinality::MultiValues)
             .set_indexed();
@@ -1291,7 +1273,7 @@ mod tests {
         let index = Index::create_in_ram(schema_builder.build());
 
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             let mut doc = Document::default();
             doc.add_u64(int_field, 1);
             index_writer.add_document(doc.clone());
@@ -1299,26 +1281,24 @@ mod tests {
             index_writer.add_document(doc);
             index_writer.commit().expect("commit failed");
             index_writer.delete_term(Term::from_field_u64(int_field, 1));
-
+            index_writer.commit().expect("commit failed");
+        }
+        index.load_searchers().unwrap();
+        let searcher = index.searcher();
+        assert_eq!(searcher.num_docs(), 0);
+        // Merging the segments
+        {
             let segment_ids = index
                 .searchable_segment_ids()
                 .expect("Searchable segments failed.");
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             index_writer
                 .merge(&segment_ids)
                 .expect("Failed to initiate merge")
                 .wait()
                 .expect("Merging failed");
-
-            // assert delete has not been committed
-            index.load_searchers().unwrap();
-            let searcher = index.searcher();
-            assert_eq!(searcher.num_docs(), 2);
-
-            index_writer.commit().unwrap();
-
             index_writer.wait_merging_threads().unwrap();
         }
-
         index.load_searchers().unwrap();
         let searcher = index.searcher();
         assert_eq!(searcher.num_docs(), 0);
@@ -1326,7 +1306,7 @@ mod tests {
 
     #[test]
     fn test_merge_multivalued_int_fields() {
-        let mut schema_builder = schema::Schema::builder();
+        let mut schema_builder = schema::SchemaBuilder::default();
         let int_options = IntOptions::default()
             .set_fast(Cardinality::MultiValues)
             .set_indexed();
@@ -1334,7 +1314,7 @@ mod tests {
         let index = Index::create_in_ram(schema_builder.build());
 
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             let index_doc = |index_writer: &mut IndexWriter, int_vals: &[u64]| {
                 let mut doc = Document::default();
                 for &val in int_vals {
@@ -1392,38 +1372,29 @@ mod tests {
             assert_eq!(&vals, &[17]);
         }
 
-        println!(
-            "{:?}",
-            searcher
-                .segment_readers()
-                .iter()
-                .map(|reader| reader.max_doc())
-                .collect::<Vec<_>>()
-        );
-
         {
             let segment = searcher.segment_reader(1u32);
             let ff_reader = segment.multi_fast_field_reader(int_field).unwrap();
             ff_reader.get_vals(0, &mut vals);
+            assert_eq!(&vals, &[20]);
+        }
+
+        {
+            let segment = searcher.segment_reader(2u32);
+            let ff_reader = segment.multi_fast_field_reader(int_field).unwrap();
+            ff_reader.get_vals(0, &mut vals);
             assert_eq!(&vals, &[28, 27]);
 
             ff_reader.get_vals(1, &mut vals);
             assert_eq!(&vals, &[1_000]);
         }
 
-        {
-            let segment = searcher.segment_reader(2u32);
-            let ff_reader = segment.multi_fast_field_reader(int_field).unwrap();
-            ff_reader.get_vals(0, &mut vals);
-            assert_eq!(&vals, &[20]);
-        }
-
         // Merging the segments
         {
             let segment_ids = index
                 .searchable_segment_ids()
                 .expect("Searchable segments failed.");
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             index_writer
                 .merge(&segment_ids)
                 .expect("Failed to initiate merge")
@@ -1436,14 +1407,6 @@ mod tests {
 
         {
             let searcher = index.searcher();
-            println!(
-                "{:?}",
-                searcher
-                    .segment_readers()
-                    .iter()
-                    .map(|reader| reader.max_doc())
-                    .collect::<Vec<_>>()
-            );
             let segment = searcher.segment_reader(0u32);
             let ff_reader = segment.multi_fast_field_reader(int_field).unwrap();
 
@@ -1469,13 +1432,13 @@ mod tests {
             assert_eq!(&vals, &[17]);
 
             ff_reader.get_vals(7, &mut vals);
-            assert_eq!(&vals, &[28, 27]);
+            assert_eq!(&vals, &[20]);
 
             ff_reader.get_vals(8, &mut vals);
-            assert_eq!(&vals, &[1_000]);
+            assert_eq!(&vals, &[28, 27]);
 
             ff_reader.get_vals(9, &mut vals);
-            assert_eq!(&vals, &[20]);
+            assert_eq!(&vals, &[1_000]);
         }
     }
 }

src/indexer/mod.rs

@@ -1,9 +1,8 @@
 pub mod delete_queue;
-
+mod directory_lock;
 mod doc_opstamp_mapping;
 pub mod index_writer;
 mod log_merge_policy;
-mod merge_operation;
 pub mod merge_policy;
 pub mod merger;
 pub mod operation;
@@ -16,12 +15,12 @@ pub mod segment_updater;
 mod segment_writer;
 mod stamper;
 
+pub(crate) use self::directory_lock::DirectoryLock;
 pub use self::index_writer::IndexWriter;
 pub use self::log_merge_policy::LogMergePolicy;
-pub use self::merge_operation::{MergeOperation, MergeOperationInventory};
 pub use self::merge_policy::{MergeCandidate, MergePolicy, NoMergePolicy};
 pub use self::prepared_commit::PreparedCommit;
-pub use self::segment_entry::SegmentEntry;
+pub use self::segment_entry::{SegmentEntry, SegmentState};
 pub use self::segment_manager::SegmentManager;
 pub use self::segment_serializer::SegmentSerializer;
 pub use self::segment_writer::SegmentWriter;

src/indexer/segment_entry.rs

@@ -4,6 +4,21 @@ use core::SegmentMeta;
 use indexer::delete_queue::DeleteCursor;
 use std::fmt;
 
+#[derive(Clone, Copy, PartialEq, Eq, Debug)]
+pub enum SegmentState {
+    Ready,
+    InMerge,
+}
+
+impl SegmentState {
+    pub fn letter_code(&self) -> char {
+        match *self {
+            SegmentState::InMerge => 'M',
+            SegmentState::Ready => 'R',
+        }
+    }
+}
+
 /// A segment entry describes the state of
 /// a given segment, at a given instant.
 ///
@@ -20,6 +35,7 @@ use std::fmt;
 #[derive(Clone)]
 pub struct SegmentEntry {
     meta: SegmentMeta,
+    state: SegmentState,
     delete_bitset: Option<BitSet>,
     delete_cursor: DeleteCursor,
 }
@@ -33,6 +49,7 @@ impl SegmentEntry {
     ) -> SegmentEntry {
         SegmentEntry {
             meta: segment_meta,
+            state: SegmentState::Ready,
             delete_bitset,
             delete_cursor,
         }
@@ -55,6 +72,14 @@ impl SegmentEntry {
         &mut self.delete_cursor
     }
 
+    /// Return the `SegmentEntry`.
+    ///
+    /// The state describes whether the segment is available for
+    /// a merge or not.
+    pub fn state(&self) -> SegmentState {
+        self.state
+    }
+
     /// Returns the segment id.
     pub fn segment_id(&self) -> SegmentId {
         self.meta.id()
@@ -64,10 +89,33 @@ impl SegmentEntry {
     pub fn meta(&self) -> &SegmentMeta {
         &self.meta
     }
+
+    /// Mark the `SegmentEntry` as in merge.
+    ///
+    /// Only segments that are not already
+    /// in a merge are elligible for future merge.
+    pub fn start_merge(&mut self) {
+        self.state = SegmentState::InMerge;
+    }
+
+    /// Cancel a merge
+    ///
+    /// If a merge fails, it is important to switch
+    /// the segment back to a idle state, so that it
+    /// may be elligible for future merges.
+    pub fn cancel_merge(&mut self) {
+        self.state = SegmentState::Ready;
+    }
+
+    /// Returns true iff a segment should
+    /// be considered for a merge.
+    pub fn is_ready(&self) -> bool {
+        self.state == SegmentState::Ready
+    }
 }
 
 impl fmt::Debug for SegmentEntry {
     fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
-        write!(formatter, "SegmentEntry({:?})", self.meta)
+        write!(formatter, "SegmentEntry({:?}, {:?})", self.meta, self.state)
     }
 }

src/indexer/segment_manager.rs

@@ -1,7 +1,7 @@
 use super::segment_register::SegmentRegister;
 use core::SegmentId;
 use core::SegmentMeta;
-use core::META_FILEPATH;
+use core::{LOCKFILE_FILEPATH, META_FILEPATH};
 use error::TantivyError;
 use indexer::delete_queue::DeleteCursor;
 use indexer::SegmentEntry;
@@ -16,6 +16,7 @@ use Result as TantivyResult;
 struct SegmentRegisters {
     uncommitted: SegmentRegister,
     committed: SegmentRegister,
+    writing: HashSet<SegmentId>,
 }
 
 /// The segment manager stores the list of segments
@@ -40,17 +41,12 @@ impl Debug for SegmentManager {
 }
 
 pub fn get_mergeable_segments(
-    in_merge_segment_ids: &HashSet<SegmentId>,
     segment_manager: &SegmentManager,
 ) -> (Vec<SegmentMeta>, Vec<SegmentMeta>) {
     let registers_lock = segment_manager.read();
     (
-        registers_lock
-            .committed
-            .get_mergeable_segments(in_merge_segment_ids),
-        registers_lock
-            .uncommitted
-            .get_mergeable_segments(in_merge_segment_ids),
+        registers_lock.committed.get_mergeable_segments(),
+        registers_lock.uncommitted.get_mergeable_segments(),
     )
 }
 
@@ -63,6 +59,7 @@ impl SegmentManager {
             registers: RwLock::new(SegmentRegisters {
                 uncommitted: SegmentRegister::default(),
                 committed: SegmentRegister::new(segment_metas, delete_cursor),
+                writing: HashSet::new(),
             }),
         }
     }
@@ -75,13 +72,16 @@ impl SegmentManager {
         segment_entries
     }
 
-    /// List the files that are useful to the index.
-    ///
-    /// This does not include lock files, or files that are obsolete
-    /// but have not yet been deleted by the garbage collector.
+    /// Returns the overall number of segments in the `SegmentManager`
+    pub fn num_segments(&self) -> usize {
+        let registers_lock = self.read();
+        registers_lock.committed.len() + registers_lock.uncommitted.len()
+    }
+
     pub fn list_files(&self) -> HashSet<PathBuf> {
         let mut files = HashSet::new();
         files.insert(META_FILEPATH.clone());
+        files.insert(LOCKFILE_FILEPATH.clone());
         for segment_meta in SegmentMeta::all() {
             files.extend(segment_meta.list_files());
         }
@@ -103,21 +103,6 @@ impl SegmentManager {
             .expect("Failed to acquire write lock on SegmentManager.")
     }
 
-    /// Deletes all empty segments
-    fn remove_empty_segments(&self) {
-        let mut registers_lock = self.write();
-        registers_lock
-            .committed
-            .segment_entries()
-            .iter()
-            .filter(|segment| segment.meta().num_docs() == 0)
-            .for_each(|segment| {
-                registers_lock
-                    .committed
-                    .remove_segment(&segment.segment_id())
-            });
-    }
-
     pub fn commit(&self, segment_entries: Vec<SegmentEntry>) {
         let mut registers_lock = self.write();
         registers_lock.committed.clear();
@@ -133,22 +118,25 @@ impl SegmentManager {
     /// the `segment_ids` are not either all committed or all
     /// uncommitted.
     pub fn start_merge(&self, segment_ids: &[SegmentId]) -> TantivyResult<Vec<SegmentEntry>> {
-        let registers_lock = self.read();
+        let mut registers_lock = self.write();
         let mut segment_entries = vec![];
         if registers_lock.uncommitted.contains_all(segment_ids) {
             for segment_id in segment_ids {
                 let segment_entry = registers_lock.uncommitted
-                    .get(segment_id)
+                    .start_merge(segment_id)
                     .expect("Segment id not found {}. Should never happen because of the contains all if-block.");
                 segment_entries.push(segment_entry);
             }
         } else if registers_lock.committed.contains_all(segment_ids) {
             for segment_id in segment_ids {
                 let segment_entry = registers_lock.committed
-                    .get(segment_id)
+                    .start_merge(segment_id)
                     .expect("Segment id not found {}. Should never happen because of the contains all if-block.");
                 segment_entries.push(segment_entry);
             }
+            for segment_id in segment_ids {
+                registers_lock.committed.start_merge(segment_id);
+            }
         } else {
             let error_msg = "Merge operation sent for segments that are not \
                              all uncommited or commited."
@@ -158,8 +146,50 @@ impl SegmentManager {
         Ok(segment_entries)
     }
 
+    pub fn cancel_merge(
+        &self,
+        before_merge_segment_ids: &[SegmentId],
+        after_merge_segment_id: SegmentId,
+    ) {
+        let mut registers_lock = self.write();
+
+        // we mark all segments are ready for merge.
+        {
+            let target_segment_register: &mut SegmentRegister;
+            target_segment_register = {
+                if registers_lock
+                    .uncommitted
+                    .contains_all(before_merge_segment_ids)
+                {
+                    &mut registers_lock.uncommitted
+                } else if registers_lock
+                    .committed
+                    .contains_all(before_merge_segment_ids)
+                {
+                    &mut registers_lock.committed
+                } else {
+                    warn!("couldn't find segment in SegmentManager");
+                    return;
+                }
+            };
+            for segment_id in before_merge_segment_ids {
+                target_segment_register.cancel_merge(segment_id);
+            }
+        }
+
+        // ... and we make sure the target segment entry
+        // can be garbage collected.
+        registers_lock.writing.remove(&after_merge_segment_id);
+    }
+
+    pub fn write_segment(&self, segment_id: SegmentId) {
+        let mut registers_lock = self.write();
+        registers_lock.writing.insert(segment_id);
+    }
+
     pub fn add_segment(&self, segment_entry: SegmentEntry) {
         let mut registers_lock = self.write();
+        registers_lock.writing.remove(&segment_entry.segment_id());
         registers_lock.uncommitted.add_segment_entry(segment_entry);
     }
 
@@ -169,6 +199,10 @@ impl SegmentManager {
         after_merge_segment_entry: SegmentEntry,
     ) {
         let mut registers_lock = self.write();
+        registers_lock
+            .writing
+            .remove(&after_merge_segment_entry.segment_id());
+
         let target_register: &mut SegmentRegister = {
             if registers_lock
                 .uncommitted
@@ -192,7 +226,6 @@ impl SegmentManager {
     }
 
     pub fn committed_segment_metas(&self) -> Vec<SegmentMeta> {
-        self.remove_empty_segments();
         let registers_lock = self.read();
         registers_lock.committed.segment_metas()
     }

src/indexer/segment_register.rs

@@ -3,7 +3,6 @@ use core::SegmentMeta;
 use indexer::delete_queue::DeleteCursor;
 use indexer::segment_entry::SegmentEntry;
 use std::collections::HashMap;
-use std::collections::HashSet;
 use std::fmt::{self, Debug, Formatter};
 
 /// The segment register keeps track
@@ -22,8 +21,8 @@ pub struct SegmentRegister {
 impl Debug for SegmentRegister {
     fn fmt(&self, f: &mut Formatter) -> Result<(), fmt::Error> {
         write!(f, "SegmentRegister(")?;
-        for k in self.segment_states.keys() {
-            write!(f, "{}, ", k.short_uuid_string())?;
+        for (k, v) in &self.segment_states {
+            write!(f, "{}:{}, ", k.short_uuid_string(), v.state().letter_code())?;
         }
         write!(f, ")")?;
         Ok(())
@@ -35,13 +34,14 @@ impl SegmentRegister {
         self.segment_states.clear();
     }
 
-    pub fn get_mergeable_segments(
-        &self,
-        in_merge_segment_ids: &HashSet<SegmentId>,
-    ) -> Vec<SegmentMeta> {
+    pub fn len(&self) -> usize {
+        self.segment_states.len()
+    }
+
+    pub fn get_mergeable_segments(&self) -> Vec<SegmentMeta> {
         self.segment_states
             .values()
-            .filter(|segment_entry| !in_merge_segment_ids.contains(&segment_entry.segment_id()))
+            .filter(|segment_entry| segment_entry.is_ready())
             .map(|segment_entry| segment_entry.meta().clone())
             .collect()
     }
@@ -51,8 +51,7 @@ impl SegmentRegister {
     }
 
     pub fn segment_metas(&self) -> Vec<SegmentMeta> {
-        let mut segment_ids: Vec<SegmentMeta> = self
-            .segment_states
+        let mut segment_ids: Vec<SegmentMeta> = self.segment_states
             .values()
             .map(|segment_entry| segment_entry.meta().clone())
             .collect();
@@ -60,7 +59,7 @@ impl SegmentRegister {
         segment_ids
     }
 
-    pub fn contains_all(&self, segment_ids: &[SegmentId]) -> bool {
+    pub fn contains_all(&mut self, segment_ids: &[SegmentId]) -> bool {
         segment_ids
             .iter()
             .all(|segment_id| self.segment_states.contains_key(segment_id))
@@ -75,8 +74,20 @@ impl SegmentRegister {
         self.segment_states.remove(segment_id);
     }
 
-    pub fn get(&self, segment_id: &SegmentId) -> Option<SegmentEntry> {
-        self.segment_states.get(segment_id).cloned()
+    pub fn cancel_merge(&mut self, segment_id: &SegmentId) {
+        self.segment_states
+            .get_mut(segment_id)
+            .expect("Received a merge notification for a segment that is not registered")
+            .cancel_merge();
+    }
+
+    pub fn start_merge(&mut self, segment_id: &SegmentId) -> Option<SegmentEntry> {
+        if let Some(segment_entry) = self.segment_states.get_mut(segment_id) {
+            segment_entry.start_merge();
+            Some(segment_entry.clone())
+        } else {
+            None
+        }
     }
 
     pub fn new(segment_metas: Vec<SegmentMeta>, delete_cursor: &DeleteCursor) -> SegmentRegister {
@@ -88,6 +99,11 @@ impl SegmentRegister {
         }
         SegmentRegister { segment_states }
     }
+
+    #[cfg(test)]
+    pub fn segment_entry(&self, segment_id: &SegmentId) -> Option<SegmentEntry> {
+        self.segment_states.get(segment_id).cloned()
+    }
 }
 
 #[cfg(test)]
@@ -96,6 +112,7 @@ mod tests {
     use core::SegmentId;
     use core::SegmentMeta;
     use indexer::delete_queue::*;
+    use indexer::SegmentState;
 
     fn segment_ids(segment_register: &SegmentRegister) -> Vec<SegmentId> {
         segment_register
@@ -119,12 +136,42 @@ mod tests {
             let segment_entry = SegmentEntry::new(segment_meta, delete_queue.cursor(), None);
             segment_register.add_segment_entry(segment_entry);
         }
+        assert_eq!(
+            segment_register
+                .segment_entry(&segment_id_a)
+                .unwrap()
+                .state(),
+            SegmentState::Ready
+        );
         assert_eq!(segment_ids(&segment_register), vec![segment_id_a]);
         {
             let segment_meta = SegmentMeta::new(segment_id_b, 0u32);
             let segment_entry = SegmentEntry::new(segment_meta, delete_queue.cursor(), None);
             segment_register.add_segment_entry(segment_entry);
         }
+        assert_eq!(
+            segment_register
+                .segment_entry(&segment_id_b)
+                .unwrap()
+                .state(),
+            SegmentState::Ready
+        );
+        segment_register.start_merge(&segment_id_a);
+        segment_register.start_merge(&segment_id_b);
+        assert_eq!(
+            segment_register
+                .segment_entry(&segment_id_a)
+                .unwrap()
+                .state(),
+            SegmentState::InMerge
+        );
+        assert_eq!(
+            segment_register
+                .segment_entry(&segment_id_b)
+                .unwrap()
+                .state(),
+            SegmentState::InMerge
+        );
         segment_register.remove_segment(&segment_id_a);
         segment_register.remove_segment(&segment_id_b);
         {

src/indexer/segment_updater.rs

@@ -16,10 +16,9 @@ use futures_cpupool::CpuFuture;
 use futures_cpupool::CpuPool;
 use indexer::delete_queue::DeleteCursor;
 use indexer::index_writer::advance_deletes;
-use indexer::merge_operation::MergeOperationInventory;
 use indexer::merger::IndexMerger;
 use indexer::stamper::Stamper;
-use indexer::MergeOperation;
+use indexer::MergeCandidate;
 use indexer::SegmentEntry;
 use indexer::SegmentSerializer;
 use indexer::{DefaultMergePolicy, MergePolicy};
@@ -27,7 +26,6 @@ use schema::Schema;
 use serde_json;
 use std::borrow::BorrowMut;
 use std::collections::HashMap;
-use std::collections::HashSet;
 use std::io::Write;
 use std::mem;
 use std::ops::DerefMut;
@@ -47,16 +45,8 @@ use Result;
 /// and flushed.
 ///
 /// This method is not part of tantivy's public API
-pub fn save_new_metas(schema: Schema, directory: &mut Directory) -> Result<()> {
-    save_metas(
-        &IndexMeta {
-            segments: Vec::new(),
-            schema,
-            opstamp: 0u64,
-            payload: None,
-        },
-        directory,
-    )
+pub fn save_new_metas(schema: Schema, opstamp: u64, directory: &mut Directory) -> Result<()> {
+    save_metas(vec![], schema, opstamp, None, directory)
 }
 
 /// Save the index meta file.
@@ -68,10 +58,21 @@ pub fn save_new_metas(schema: Schema, directory: &mut Directory) -> Result<()> {
 /// and flushed.
 ///
 /// This method is not part of tantivy's public API
-fn save_metas(metas: &IndexMeta, directory: &mut Directory) -> Result<()> {
-    let mut buffer = serde_json::to_vec_pretty(metas)?;
-    // Just adding a new line at the end of the buffer.
-    writeln!(&mut buffer)?;
+pub fn save_metas(
+    segment_metas: Vec<SegmentMeta>,
+    schema: Schema,
+    opstamp: u64,
+    payload: Option<String>,
+    directory: &mut Directory,
+) -> Result<()> {
+    let metas = IndexMeta {
+        segments: segment_metas,
+        schema,
+        opstamp,
+        payload,
+    };
+    let mut buffer = serde_json::to_vec_pretty(&metas)?;
+    write!(&mut buffer, "\n")?;
     directory.atomic_write(&META_FILEPATH, &buffer[..])?;
     debug!("Saved metas {:?}", serde_json::to_string_pretty(&metas));
     Ok(())
@@ -82,21 +83,16 @@ fn save_metas(metas: &IndexMeta, directory: &mut Directory) -> Result<()> {
 //
 // All this processing happens on a single thread
 // consuming a common queue.
-//
-// We voluntarily pass a merge_operation ref to guarantee that
-// the merge_operation is alive during the process
 #[derive(Clone)]
 pub struct SegmentUpdater(Arc<InnerSegmentUpdater>);
 
 fn perform_merge(
-    merge_operation: &MergeOperation,
     index: &Index,
     mut segment_entries: Vec<SegmentEntry>,
+    mut merged_segment: Segment,
+    target_opstamp: u64,
 ) -> Result<SegmentEntry> {
-    let target_opstamp = merge_operation.target_opstamp();
-
     // first we need to apply deletes to our segment.
-    let mut merged_segment = index.new_segment();
 
     // TODO add logging
     let schema = index.schema();
@@ -130,27 +126,19 @@ fn perform_merge(
 }
 
 struct InnerSegmentUpdater {
-    // we keep a copy of the current active IndexMeta to
-    // avoid loading the file everytime we need it in the
-    // `SegmentUpdater`.
-    //
-    // This should be up to date as all update happen through
-    // the unique active `SegmentUpdater`.
-    active_metas: RwLock<Arc<IndexMeta>>,
     pool: CpuPool,
     index: Index,
     segment_manager: SegmentManager,
-    merge_policy: RwLock<Arc<Box<MergePolicy>>>,
+    merge_policy: RwLock<Box<MergePolicy>>,
     merging_thread_id: AtomicUsize,
     merging_threads: RwLock<HashMap<usize, JoinHandle<Result<()>>>>,
     generation: AtomicUsize,
     killed: AtomicBool,
     stamper: Stamper,
-    merge_operations: MergeOperationInventory,
 }
 
 impl SegmentUpdater {
-    pub fn create(
+    pub fn new(
         index: Index,
         stamper: Stamper,
         delete_cursor: &DeleteCursor,
@@ -161,29 +149,32 @@ impl SegmentUpdater {
             .name_prefix("segment_updater")
             .pool_size(1)
             .create();
-        let index_meta = index.load_metas()?;
         Ok(SegmentUpdater(Arc::new(InnerSegmentUpdater {
-            active_metas: RwLock::new(Arc::new(index_meta)),
             pool,
             index,
             segment_manager,
-            merge_policy: RwLock::new(Arc::new(Box::new(DefaultMergePolicy::default()))),
+            merge_policy: RwLock::new(Box::new(DefaultMergePolicy::default())),
             merging_thread_id: AtomicUsize::default(),
             merging_threads: RwLock::new(HashMap::new()),
             generation: AtomicUsize::default(),
             killed: AtomicBool::new(false),
             stamper,
-            merge_operations: Default::default(),
         })))
     }
 
-    pub fn get_merge_policy(&self) -> Arc<Box<MergePolicy>> {
-        self.0.merge_policy.read().unwrap().clone()
+    pub fn new_segment(&self) -> Segment {
+        let new_segment = self.0.index.new_segment();
+        let segment_id = new_segment.id();
+        self.0.segment_manager.write_segment(segment_id);
+        new_segment
+    }
+
+    pub fn get_merge_policy(&self) -> Box<MergePolicy> {
+        self.0.merge_policy.read().unwrap().box_clone()
     }
 
     pub fn set_merge_policy(&self, merge_policy: Box<MergePolicy>) {
-        let arc_merge_policy = Arc::new(merge_policy);
-        *self.0.merge_policy.write().unwrap() = arc_merge_policy;
+        *self.0.merge_policy.write().unwrap() = merge_policy;
     }
 
     fn get_merging_thread_id(&self) -> usize {
@@ -204,8 +195,7 @@ impl SegmentUpdater {
                 segment_updater.0.segment_manager.add_segment(segment_entry);
                 segment_updater.consider_merge_options();
                 true
-            })
-            .forget();
+            }).forget();
             true
         } else {
             false
@@ -237,39 +227,20 @@ impl SegmentUpdater {
         if self.is_alive() {
             let index = &self.0.index;
             let directory = index.directory();
-            let mut commited_segment_metas = self.0.segment_manager.committed_segment_metas();
-
-            // We sort segment_readers by number of documents.
-            // This is an heuristic to make multithreading more efficient.
-            //
-            // This is not done at the searcher level because I had a strange
-            // use case in which I was dealing with a large static index,
-            // dispatched over 5 SSD drives.
-            //
-            // A `UnionDirectory` makes it possible to read from these
-            // 5 different drives and creates a meta.json on the fly.
-            // In order to optimize the throughput, it creates a lasagna of segments
-            // from the different drives.
-            //
-            // Segment 1 from disk 1, Segment 1 from disk 2, etc.
-            commited_segment_metas.sort_by_key(|segment_meta| -(segment_meta.max_doc() as i32));
-            let index_meta = IndexMeta {
-                segments: commited_segment_metas,
-                schema: index.schema(),
+            save_metas(
+                self.0.segment_manager.committed_segment_metas(),
+                index.schema(),
                 opstamp,
-                payload: commit_message,
-            };
-            save_metas(&index_meta, directory.box_clone().borrow_mut())
-                .expect("Could not save metas.");
-            self.store_meta(&index_meta);
+                commit_message,
+                directory.box_clone().borrow_mut(),
+            ).expect("Could not save metas.");
         }
     }
 
     pub fn garbage_collect_files(&self) -> Result<()> {
         self.run_async(move |segment_updater| {
             segment_updater.garbage_collect_files_exec();
-        })
-        .wait()
+        }).wait()
     }
 
     fn garbage_collect_files_exec(&self) {
@@ -291,67 +262,54 @@ impl SegmentUpdater {
                 segment_updater.garbage_collect_files_exec();
                 segment_updater.consider_merge_options();
             }
-        })
-        .wait()
+        }).wait()
     }
 
     pub fn start_merge(&self, segment_ids: &[SegmentId]) -> Result<Receiver<SegmentMeta>> {
-        let commit_opstamp = self.load_metas().opstamp;
-        let merge_operation = MergeOperation::new(
-            &self.0.merge_operations,
-            commit_opstamp,
-            segment_ids.to_vec(),
-        );
-        self.run_async(move |segment_updater| segment_updater.start_merge_impl(merge_operation))
-            .wait()?
-    }
-
-    fn store_meta(&self, index_meta: &IndexMeta) {
-        *self.0.active_metas.write().unwrap() = Arc::new(index_meta.clone());
-    }
-    fn load_metas(&self) -> Arc<IndexMeta> {
-        self.0.active_metas.read().unwrap().clone()
+        //let future_merged_segment = */
+        let segment_ids_vec = segment_ids.to_vec();
+        self.run_async(move |segment_updater| {
+            segment_updater.start_merge_impl(&segment_ids_vec[..])
+        }).wait()?
     }
 
     // `segment_ids` is required to be non-empty.
-    fn start_merge_impl(&self, merge_operation: MergeOperation) -> Result<Receiver<SegmentMeta>> {
-        assert!(
-            !merge_operation.segment_ids().is_empty(),
-            "Segment_ids cannot be empty."
-        );
+    fn start_merge_impl(&self, segment_ids: &[SegmentId]) -> Result<Receiver<SegmentMeta>> {
+        assert!(!segment_ids.is_empty(), "Segment_ids cannot be empty.");
 
         let segment_updater_clone = self.clone();
-        let segment_entries: Vec<SegmentEntry> = self
-            .0
-            .segment_manager
-            .start_merge(merge_operation.segment_ids())?;
+        let segment_entries: Vec<SegmentEntry> = self.0.segment_manager.start_merge(segment_ids)?;
 
-        //        let segment_ids_vec = merge_operation.segment_ids.to_vec();
+        let segment_ids_vec = segment_ids.to_vec();
 
         let merging_thread_id = self.get_merging_thread_id();
         info!(
             "Starting merge thread #{} - {:?}",
-            merging_thread_id,
-            merge_operation.segment_ids()
+            merging_thread_id, segment_ids
         );
         let (merging_future_send, merging_future_recv) = oneshot();
 
+        let target_opstamp = self.0.stamper.stamp();
+
         // first we need to apply deletes to our segment.
         let merging_join_handle = thread::Builder::new()
             .name(format!("mergingthread-{}", merging_thread_id))
             .spawn(move || {
                 // first we need to apply deletes to our segment.
+                let merged_segment = segment_updater_clone.new_segment();
+                let merged_segment_id = merged_segment.id();
                 let merge_result = perform_merge(
-                    &merge_operation,
                     &segment_updater_clone.0.index,
                     segment_entries,
+                    merged_segment,
+                    target_opstamp,
                 );
 
                 match merge_result {
                     Ok(after_merge_segment_entry) => {
                         let merged_segment_meta = after_merge_segment_entry.meta().clone();
                         segment_updater_clone
-                            .end_merge(merge_operation, after_merge_segment_entry)
+                            .end_merge(segment_ids_vec, after_merge_segment_entry)
                             .expect("Segment updater thread is corrupted.");
 
                         // the future may fail if the listener of the oneshot future
@@ -362,18 +320,13 @@ impl SegmentUpdater {
                         let _merging_future_res = merging_future_send.send(merged_segment_meta);
                     }
                     Err(e) => {
-                        warn!(
-                            "Merge of {:?} was cancelled: {:?}",
-                            merge_operation.segment_ids(),
-                            e
-                        );
+                        warn!("Merge of {:?} was cancelled: {:?}", segment_ids_vec, e);
                         // ... cancel merge
                         if cfg!(test) {
                             panic!("Merge failed.");
                         }
-                        // As `merge_operation` will be dropped, the segment in merge state will
-                        // be available for merge again.
-                        // `merging_future_send` will be dropped, sending an error to the future.
+                        segment_updater_clone.cancel_merge(&segment_ids_vec, merged_segment_id);
+                        // merging_future_send will be dropped, sending an error to the future.
                     }
                 }
                 segment_updater_clone
@@ -394,34 +347,16 @@ impl SegmentUpdater {
     }
 
     fn consider_merge_options(&self) {
-        let merge_segment_ids: HashSet<SegmentId> = self.0.merge_operations.segment_in_merge();
         let (committed_segments, uncommitted_segments) =
-            get_mergeable_segments(&merge_segment_ids, &self.0.segment_manager);
-
+            get_mergeable_segments(&self.0.segment_manager);
         // Committed segments cannot be merged with uncommitted_segments.
         // We therefore consider merges using these two sets of segments independently.
         let merge_policy = self.get_merge_policy();
-
-        let current_opstamp = self.0.stamper.stamp();
-        let mut merge_candidates: Vec<MergeOperation> = merge_policy
-            .compute_merge_candidates(&uncommitted_segments)
-            .into_iter()
-            .map(|merge_candidate| {
-                MergeOperation::new(&self.0.merge_operations, current_opstamp, merge_candidate.0)
-            })
-            .collect();
-
-        let commit_opstamp = self.load_metas().opstamp;
-        let committed_merge_candidates = merge_policy
-            .compute_merge_candidates(&committed_segments)
-            .into_iter()
-            .map(|merge_candidate| {
-                MergeOperation::new(&self.0.merge_operations, commit_opstamp, merge_candidate.0)
-            })
-            .collect::<Vec<_>>();
-        merge_candidates.extend(committed_merge_candidates.into_iter());
-        for merge_operation in merge_candidates {
-            match self.start_merge_impl(merge_operation) {
+        let mut merge_candidates = merge_policy.compute_merge_candidates(&uncommitted_segments);
+        let committed_merge_candidates = merge_policy.compute_merge_candidates(&committed_segments);
+        merge_candidates.extend_from_slice(&committed_merge_candidates[..]);
+        for MergeCandidate(segment_metas) in merge_candidates {
+            match self.start_merge_impl(&segment_metas) {
                 Ok(merge_future) => {
                     if let Err(e) = merge_future.fuse().poll() {
                         error!("The merge task failed quickly after starting: {:?}", e);
@@ -437,16 +372,31 @@ impl SegmentUpdater {
         }
     }
 
+    fn cancel_merge(
+        &self,
+        before_merge_segment_ids: &[SegmentId],
+        after_merge_segment_entry: SegmentId,
+    ) {
+        self.0
+            .segment_manager
+            .cancel_merge(before_merge_segment_ids, after_merge_segment_entry);
+    }
+
     fn end_merge(
         &self,
-        merge_operation: MergeOperation,
+        before_merge_segment_ids: Vec<SegmentId>,
         mut after_merge_segment_entry: SegmentEntry,
     ) -> Result<()> {
         self.run_async(move |segment_updater| {
             info!("End merge {:?}", after_merge_segment_entry.meta());
             let mut delete_cursor = after_merge_segment_entry.delete_cursor().clone();
             if let Some(delete_operation) = delete_cursor.get() {
-                let committed_opstamp = segment_updater.load_metas().opstamp;
+                let committed_opstamp = segment_updater
+                    .0
+                    .index
+                    .load_metas()
+                    .expect("Failed to read opstamp")
+                    .opstamp;
                 if delete_operation.opstamp < committed_opstamp {
                     let index = &segment_updater.0.index;
                     let segment = index.segment(after_merge_segment_entry.meta().clone());
@@ -455,15 +405,16 @@ impl SegmentUpdater {
                     {
                         error!(
                             "Merge of {:?} was cancelled (advancing deletes failed): {:?}",
-                            merge_operation.segment_ids(),
-                            e
+                            before_merge_segment_ids, e
                         );
+                        // ... cancel merge
                         if cfg!(test) {
                             panic!("Merge failed.");
                         }
-                        // ... cancel merge
-                        // `merge_operations` are tracked. As it is dropped, the
-                        // the segment_ids will be available again for merge.
+                        segment_updater.cancel_merge(
+                            &before_merge_segment_ids,
+                            after_merge_segment_entry.segment_id(),
+                        );
                         return;
                     }
                 }
@@ -471,14 +422,13 @@ impl SegmentUpdater {
             segment_updater
                 .0
                 .segment_manager
-                .end_merge(merge_operation.segment_ids(), after_merge_segment_entry);
+                .end_merge(&before_merge_segment_ids, after_merge_segment_entry);
             segment_updater.consider_merge_options();
             info!("save metas");
-            let previous_metas = segment_updater.load_metas();
-            segment_updater.save_metas(previous_metas.opstamp, previous_metas.payload.clone());
+            let previous_metas = segment_updater.0.index.load_metas().unwrap();
+            segment_updater.save_metas(previous_metas.opstamp, previous_metas.payload);
             segment_updater.garbage_collect_files_exec();
-        })
-        .wait()
+        }).wait()
     }
 
     /// Wait for current merging threads.
@@ -497,25 +447,32 @@ impl SegmentUpdater {
     /// Obsolete files will eventually be cleaned up
     /// by the directory garbage collector.
     pub fn wait_merging_thread(&self) -> Result<()> {
+        let mut num_segments: usize;
         loop {
-            let merging_threads: HashMap<usize, JoinHandle<Result<()>>> = {
+            num_segments = self.0.segment_manager.num_segments();
+
+            let mut new_merging_threads = HashMap::new();
+            {
                 let mut merging_threads = self.0.merging_threads.write().unwrap();
-                mem::replace(merging_threads.deref_mut(), HashMap::new())
-            };
-            if merging_threads.is_empty() {
-                return Ok(());
+                mem::swap(&mut new_merging_threads, merging_threads.deref_mut());
             }
-            debug!("wait merging thread {}", merging_threads.len());
-            for (_, merging_thread_handle) in merging_threads {
+            debug!("wait merging thread {}", new_merging_threads.len());
+            for (_, merging_thread_handle) in new_merging_threads {
                 merging_thread_handle
                     .join()
                     .map(|_| ())
                     .map_err(|_| TantivyError::ErrorInThread("Merging thread failed.".into()))?;
             }
-            // Our merging thread may have queued their completed merged segment.
-            // Let's wait for that too.
+            // Our merging thread may have queued their completed
             self.run_async(move |_| {}).wait()?;
+
+            let new_num_segments = self.0.segment_manager.num_segments();
+
+            if new_num_segments >= num_segments {
+                break;
+            }
         }
+        Ok(())
     }
 }
 
@@ -528,14 +485,14 @@ mod tests {
 
     #[test]
     fn test_delete_during_merge() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
 
         let index = Index::create_in_ram(schema);
 
         // writing the segment
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
         index_writer.set_merge_policy(Box::new(MergeWheneverPossible));
 
         {
@@ -579,75 +536,4 @@ mod tests {
         assert_eq!(index.searcher().segment_readers().len(), 1);
         assert_eq!(index.searcher().num_docs(), 302);
     }
-
-    #[test]
-    fn delete_all_docs() {
-        let mut schema_builder = Schema::builder();
-        let text_field = schema_builder.add_text_field("text", TEXT);
-        let schema = schema_builder.build();
-
-        let index = Index::create_in_ram(schema);
-
-        // writing the segment
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-
-        {
-            for _ in 0..100 {
-                index_writer.add_document(doc!(text_field=>"a"));
-                index_writer.add_document(doc!(text_field=>"b"));
-            }
-            assert!(index_writer.commit().is_ok());
-        }
-
-        {
-            for _ in 0..100 {
-                index_writer.add_document(doc!(text_field=>"c"));
-                index_writer.add_document(doc!(text_field=>"d"));
-            }
-            assert!(index_writer.commit().is_ok());
-        }
-
-        {
-            index_writer.add_document(doc!(text_field=>"e"));
-            index_writer.add_document(doc!(text_field=>"f"));
-            assert!(index_writer.commit().is_ok());
-        }
-
-        {
-            let seg_ids = index
-                .searchable_segment_ids()
-                .expect("Searchable segments failed.");
-            // docs exist, should have at least 1 segment
-            assert!(seg_ids.len() > 0);
-        }
-
-        {
-            let term_vals = vec!["a", "b", "c", "d", "e", "f"];
-            for term_val in term_vals {
-                let term = Term::from_field_text(text_field, term_val);
-                index_writer.delete_term(term);
-                assert!(index_writer.commit().is_ok());
-            }
-        }
-
-        {
-            index_writer
-                .wait_merging_threads()
-                .expect("waiting for merging threads");
-        }
-
-        index.load_searchers().unwrap();
-        assert_eq!(index.searcher().num_docs(), 0);
-
-        let seg_ids = index
-            .searchable_segment_ids()
-            .expect("Searchable segments failed.");
-        assert!(seg_ids.is_empty());
-
-        index.load_searchers().unwrap();
-        assert_eq!(index.searcher().num_docs(), 0);
-        // empty segments should be erased
-        assert!(index.searchable_segment_metas().unwrap().is_empty());
-        assert!(index.searcher().segment_readers().is_empty());
-    }
 }

src/indexer/segment_writer.rs

@@ -49,21 +49,20 @@ impl SegmentWriter {
     ) -> Result<SegmentWriter> {
         let segment_serializer = SegmentSerializer::for_segment(&mut segment)?;
         let multifield_postings = MultiFieldPostingsWriter::new(schema, table_bits);
-        let tokenizers =
-            schema
-                .fields()
-                .iter()
-                .map(|field_entry| field_entry.field_type())
-                .map(|field_type| match *field_type {
-                    FieldType::Str(ref text_options) => text_options
-                        .get_indexing_options()
-                        .and_then(|text_index_option| {
-                            let tokenizer_name = &text_index_option.tokenizer();
-                            segment.index().tokenizers().get(tokenizer_name)
-                        }),
-                    _ => None,
-                })
-                .collect();
+        let tokenizers = schema
+            .fields()
+            .iter()
+            .map(|field_entry| field_entry.field_type())
+            .map(|field_type| match *field_type {
+                FieldType::Str(ref text_options) => text_options.get_indexing_options().and_then(
+                    |text_index_option| {
+                        let tokenizer_name = &text_index_option.tokenizer();
+                        segment.index().tokenizers().get(tokenizer_name)
+                    },
+                ),
+                _ => None,
+            })
+            .collect();
         Ok(SegmentWriter {
             max_doc: 0,
             multifield_postings,
@@ -111,18 +110,19 @@ impl SegmentWriter {
             }
             match *field_options.field_type() {
                 FieldType::HierarchicalFacet => {
-                    let facets: Vec<&str> = field_values
+                    let facets: Vec<&[u8]> = field_values
                         .iter()
                         .flat_map(|field_value| match *field_value.value() {
-                            Value::Facet(ref facet) => Some(facet.encoded_str()),
+                            Value::Facet(ref facet) => Some(facet.encoded_bytes()),
                             _ => {
                                 panic!("Expected hierarchical facet");
                             }
                         })
                         .collect();
                     let mut term = Term::for_field(field); // we set the Term
-                    for fake_str in facets {
+                    for facet_bytes in facets {
                         let mut unordered_term_id_opt = None;
+                        let fake_str = unsafe { str::from_utf8_unchecked(facet_bytes) };
                         FacetTokenizer.token_stream(fake_str).process(&mut |token| {
                             term.set_text(&token.text);
                             let unordered_term_id =

src/indexer/stamper.rs

@@ -1,66 +1,50 @@
-use std::sync::atomic::Ordering;
-use std::sync::Arc;
-
 // AtomicU64 have not landed in stable.
 // For the moment let's just use AtomicUsize on
 // x86/64 bit platform, and a mutex on other platform.
-#[cfg(target_arch = "x86_64")]
+
+#[cfg(target = "x86_64")]
 mod archicture_impl {
 
     use std::sync::atomic::{AtomicUsize, Ordering};
+    use std::sync::Arc;
 
-    #[derive(Default)]
-    pub struct AtomicU64Ersatz(AtomicUsize);
+    #[derive(Clone, Default)]
+    pub struct Stamper(Arc<AtomicU64>);
 
-    impl AtomicU64Ersatz {
-        pub fn new(first_opstamp: u64) -> AtomicU64Ersatz {
-            AtomicU64Ersatz(AtomicUsize::new(first_opstamp as usize))
+    impl Stamper {
+        pub fn new(first_opstamp: u64) -> Stamper {
+            Stamper(Arc::new(AtomicU64::new(first_opstamp)))
         }
 
-        pub fn fetch_add(&self, val: u64, order: Ordering) -> u64 {
-            self.0.fetch_add(val as usize, order) as u64
+        pub fn stamp(&self) -> u64 {
+            self.0.fetch_add(1u64, Ordering::SeqCst) as u64
         }
     }
 }
 
-#[cfg(not(target_arch = "x86_64"))]
+#[cfg(not(target = "x86_64"))]
 mod archicture_impl {
 
-    use std::sync::atomic::Ordering;
-    /// Under other architecture, we rely on a mutex.
-    use std::sync::RwLock;
+    use std::sync::{Arc, Mutex};
 
-    #[derive(Default)]
-    pub struct AtomicU64Ersatz(RwLock<u64>);
+    #[derive(Clone, Default)]
+    pub struct Stamper(Arc<Mutex<u64>>);
 
-    impl AtomicU64Ersatz {
-        pub fn new(first_opstamp: u64) -> AtomicU64Ersatz {
-            AtomicU64Ersatz(RwLock::new(first_opstamp))
+    impl Stamper {
+        pub fn new(first_opstamp: u64) -> Stamper {
+            Stamper(Arc::new(Mutex::new(first_opstamp)))
         }
 
-        pub fn fetch_add(&self, incr: u64, _order: Ordering) -> u64 {
-            let mut lock = self.0.write().unwrap();
-            let previous_val = *lock;
-            *lock = previous_val + incr;
+        pub fn stamp(&self) -> u64 {
+            let mut guard = self.0.lock().expect("Failed to lock the stamper");
+            let previous_val = *guard;
+            *guard = previous_val + 1;
             previous_val
         }
     }
 }
 
-use self::archicture_impl::AtomicU64Ersatz;
-
-#[derive(Clone, Default)]
-pub struct Stamper(Arc<AtomicU64Ersatz>);
-
-impl Stamper {
-    pub fn new(first_opstamp: u64) -> Stamper {
-        Stamper(Arc::new(AtomicU64Ersatz::new(first_opstamp)))
-    }
-
-    pub fn stamp(&self) -> u64 {
-        self.0.fetch_add(1u64, Ordering::SeqCst) as u64
-    }
-}
+pub use self::archicture_impl::Stamper;
 
 #[cfg(test)]
 mod test {

src/lib.rs

@@ -1,7 +1,11 @@
 #![doc(html_logo_url = "http://fulmicoton.com/tantivy-logo/tantivy-logo.png")]
+#![cfg_attr(feature = "cargo-clippy", allow(module_inception))]
+#![cfg_attr(feature = "cargo-clippy", allow(inline_always))]
 #![cfg_attr(all(feature = "unstable", test), feature(test))]
-#![cfg_attr(feature = "cargo-clippy", allow(clippy::module_inception))]
 #![doc(test(attr(allow(unused_variables), deny(warnings))))]
+#![allow(unknown_lints)]
+#![allow(new_without_default)]
+#![allow(decimal_literal_representation)]
 #![warn(missing_docs)]
 #![recursion_limit = "80"]
 
@@ -23,8 +27,7 @@
 //! # use tempdir::TempDir;
 //! # use tantivy::Index;
 //! # use tantivy::schema::*;
-//! # use tantivy::{Score, DocAddress};
-//! # use tantivy::collector::TopDocs;
+//! # use tantivy::collector::TopCollector;
 //! # use tantivy::query::QueryParser;
 //! #
 //! # fn main() {
@@ -46,7 +49,7 @@
 //! // in a compressed, row-oriented key-value store.
 //! // This store is useful to reconstruct the
 //! // documents that were selected during the search phase.
-//! let mut schema_builder = Schema::builder();
+//! let mut schema_builder = SchemaBuilder::default();
 //! let title = schema_builder.add_text_field("title", TEXT | STORED);
 //! let body = schema_builder.add_text_field("body", TEXT);
 //! let schema = schema_builder.build();
@@ -86,14 +89,14 @@
 //! // A ticket has been opened regarding this problem.
 //! let query = query_parser.parse_query("sea whale")?;
 //!
-//! // Perform search.
-//! // `topdocs` contains the 10 most relevant doc ids, sorted by decreasing scores...
-//! let top_docs: Vec<(Score, DocAddress)> =
-//!     searcher.search(&query, &TopDocs::with_limit(10))?;
+//! let mut top_collector = TopCollector::with_limit(10);
+//! searcher.search(&*query, &mut top_collector)?;
 //!
-//! for (_score, doc_address) in top_docs {
-//!     // Retrieve the actual content of documents given its `doc_address`.
-//!     let retrieved_doc = searcher.doc(doc_address)?;
+//! // Our top collector now contains the 10
+//! // most relevant doc ids...
+//! let doc_addresses = top_collector.docs();
+//! for doc_address in doc_addresses {
+//!     let retrieved_doc = searcher.doc(&doc_address)?;
 //!     println!("{}", schema.to_json(&retrieved_doc));
 //! }
 //!
@@ -129,21 +132,23 @@ extern crate base64;
 extern crate bit_set;
 extern crate bitpacking;
 extern crate byteorder;
+
+#[macro_use]
 extern crate combine;
+
 extern crate crossbeam;
+extern crate crossbeam_channel;
 extern crate fnv;
 extern crate fst;
 extern crate fst_regex;
 extern crate futures;
 extern crate futures_cpupool;
-extern crate htmlescape;
 extern crate itertools;
 extern crate levenshtein_automata;
 extern crate num_cpus;
 extern crate owning_ref;
 extern crate regex;
 extern crate rust_stemmers;
-extern crate scoped_pool;
 extern crate serde;
 extern crate stable_deref_trait;
 extern crate tempdir;
@@ -160,18 +165,13 @@ extern crate winapi;
 #[cfg(test)]
 extern crate rand;
 
-#[cfg(test)]
-#[macro_use]
-extern crate maplit;
-
 #[cfg(all(test, feature = "unstable"))]
 extern crate test;
 
-#[macro_use]
-extern crate downcast_rs;
+extern crate tinysegmenter;
 
 #[macro_use]
-extern crate fail;
+extern crate downcast;
 
 #[cfg(test)]
 mod functional_test;
@@ -181,7 +181,7 @@ mod macros;
 
 pub use error::TantivyError;
 
-#[deprecated(since = "0.7.0", note = "please use `tantivy::TantivyError` instead")]
+#[deprecated(since="0.7.0", note="please use `tantivy::TantivyError` instead")]
 pub use error::TantivyError as Error;
 
 extern crate census;
@@ -206,13 +206,9 @@ pub(crate) mod positions;
 pub mod postings;
 pub mod query;
 pub mod schema;
-pub mod space_usage;
 pub mod store;
 pub mod termdict;
 
-mod snippet;
-pub use self::snippet::{Snippet, SnippetGenerator};
-
 mod docset;
 pub use self::docset::{DocSet, SkipResult};
 
@@ -229,7 +225,11 @@ pub use common::{i64_to_u64, u64_to_i64};
 /// Expose the current version of tantivy, as well
 /// whether it was compiled with the simd compression.
 pub fn version() -> &'static str {
-    env!("CARGO_PKG_VERSION")
+    if cfg!(feature = "simdcompression") {
+        concat!(env!("CARGO_PKG_VERSION"), "-simd")
+    } else {
+        concat!(env!("CARGO_PKG_VERSION"), "-nosimd")
+    }
 }
 
 /// Defines tantivy's merging strategy
@@ -261,12 +261,12 @@ impl DocAddress {
     /// The segment ordinal is an id identifying the segment
     /// hosting the document. It is only meaningful, in the context
     /// of a searcher.
-    pub fn segment_ord(self) -> SegmentLocalId {
+    pub fn segment_ord(&self) -> SegmentLocalId {
         self.0
     }
 
     /// Return the segment local `DocId`
-    pub fn doc(self) -> DocId {
+    pub fn doc(&self) -> DocId {
         self.1
     }
 }
@@ -290,11 +290,9 @@ mod tests {
     use docset::DocSet;
     use query::BooleanQuery;
     use rand::distributions::Bernoulli;
-    use rand::distributions::Uniform;
-    use rand::rngs::StdRng;
-    use rand::{Rng, SeedableRng};
+    use rand::distributions::Range;
+    use rand::{Rng, SeedableRng, XorShiftRng};
     use schema::*;
-    use DocAddress;
     use Index;
     use IndexWriter;
     use Postings;
@@ -313,15 +311,16 @@ mod tests {
     }
 
     pub fn generate_nonunique_unsorted(max_value: u32, n_elems: usize) -> Vec<u32> {
-        let seed: [u8; 32] = [1; 32];
-        StdRng::from_seed(seed)
-            .sample_iter(&Uniform::new(0u32, max_value))
+        let seed: [u8; 16] = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15];
+        XorShiftRng::from_seed(seed)
+            .sample_iter(&Range::new(0u32, max_value))
             .take(n_elems)
             .collect::<Vec<u32>>()
     }
 
     pub fn sample_with_seed(n: u32, ratio: f64, seed_val: u8) -> Vec<u32> {
-        StdRng::from_seed([seed_val; 32])
+        let seed: [u8; 16] = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, seed_val];
+        XorShiftRng::from_seed(seed)
             .sample_iter(&Bernoulli::new(ratio))
             .take(n as usize)
             .enumerate()
@@ -336,13 +335,13 @@ mod tests {
     #[test]
     #[cfg(feature = "mmap")]
     fn test_indexing() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_from_tempdir(schema).unwrap();
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             {
                 let doc = doc!(text_field=>"af b");
                 index_writer.add_document(doc);
@@ -361,10 +360,10 @@ mod tests {
 
     #[test]
     fn test_docfreq1() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let index = Index::create_in_ram(schema_builder.build());
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
         {
             index_writer.add_document(doc!(text_field=>"a b c"));
             index_writer.commit().unwrap();
@@ -401,12 +400,12 @@ mod tests {
 
     #[test]
     fn test_fieldnorm_no_docs_with_field() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let title_field = schema_builder.add_text_field("title", TEXT);
         let text_field = schema_builder.add_text_field("text", TEXT);
         let index = Index::create_in_ram(schema_builder.build());
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             {
                 let doc = doc!(text_field=>"a b c");
                 index_writer.add_document(doc);
@@ -430,11 +429,11 @@ mod tests {
 
     #[test]
     fn test_fieldnorm() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let index = Index::create_in_ram(schema_builder.build());
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             {
                 let doc = doc!(text_field=>"a b c");
                 index_writer.add_document(doc);
@@ -471,7 +470,7 @@ mod tests {
 
     #[test]
     fn test_delete_postings1() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let term_abcd = Term::from_field_text(text_field, "abcd");
         let term_a = Term::from_field_text(text_field, "a");
@@ -481,22 +480,43 @@ mod tests {
         let index = Index::create_in_ram(schema);
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-            // 0
-            index_writer.add_document(doc!(text_field=>"a b"));
-            // 1
-            index_writer.add_document(doc!(text_field=>" a c"));
-            // 2
-            index_writer.add_document(doc!(text_field=>" b c"));
-            // 3
-            index_writer.add_document(doc!(text_field=>" b d"));
-
-            index_writer.delete_term(Term::from_field_text(text_field, "c"));
-            index_writer.delete_term(Term::from_field_text(text_field, "a"));
-            // 4
-            index_writer.add_document(doc!(text_field=>" b c"));
-            // 5
-            index_writer.add_document(doc!(text_field=>" a"));
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
+            {
+                // 0
+                let doc = doc!(text_field=>"a b");
+                index_writer.add_document(doc);
+            }
+            {
+                // 1
+                let doc = doc!(text_field=>" a c");
+                index_writer.add_document(doc);
+            }
+            {
+                // 2
+                let doc = doc!(text_field=>" b c");
+                index_writer.add_document(doc);
+            }
+            {
+                // 3
+                let doc = doc!(text_field=>" b d");
+                index_writer.add_document(doc);
+            }
+            {
+                index_writer.delete_term(Term::from_field_text(text_field, "c"));
+            }
+            {
+                index_writer.delete_term(Term::from_field_text(text_field, "a"));
+            }
+            {
+                // 4
+                let doc = doc!(text_field=>" b c");
+                index_writer.add_document(doc);
+            }
+            {
+                // 5
+                let doc = doc!(text_field=>" a");
+                index_writer.add_document(doc);
+            }
             index_writer.commit().unwrap();
         }
         {
@@ -504,9 +524,11 @@ mod tests {
             let searcher = index.searcher();
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(text_field);
-            assert!(inverted_index
-                .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
-                .is_none());
+            assert!(
+                inverted_index
+                    .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
+                    .is_none()
+            );
             {
                 let mut postings = inverted_index
                     .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
@@ -528,11 +550,16 @@ mod tests {
         }
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-            // 0
-            index_writer.add_document(doc!(text_field=>"a b"));
-            // 1
-            index_writer.delete_term(Term::from_field_text(text_field, "c"));
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
+            {
+                // 0
+                let doc = doc!(text_field=>"a b");
+                index_writer.add_document(doc);
+            }
+            {
+                // 1
+                index_writer.delete_term(Term::from_field_text(text_field, "c"));
+            }
             index_writer.rollback().unwrap();
         }
         {
@@ -541,9 +568,11 @@ mod tests {
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(term_abcd.field());
 
-            assert!(inverted_index
-                .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
-                .is_none());
+            assert!(
+                inverted_index
+                    .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
+                    .is_none()
+            );
             {
                 let mut postings = inverted_index
                     .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
@@ -565,9 +594,14 @@ mod tests {
         }
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-            index_writer.add_document(doc!(text_field=>"a b"));
-            index_writer.delete_term(Term::from_field_text(text_field, "c"));
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
+            {
+                let doc = doc!(text_field=>"a b");
+                index_writer.add_document(doc);
+            }
+            {
+                index_writer.delete_term(Term::from_field_text(text_field, "c"));
+            }
             index_writer.rollback().unwrap();
             index_writer.delete_term(Term::from_field_text(text_field, "a"));
             index_writer.commit().unwrap();
@@ -577,9 +611,11 @@ mod tests {
             let searcher = index.searcher();
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(term_abcd.field());
-            assert!(inverted_index
-                .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
-                .is_none());
+            assert!(
+                inverted_index
+                    .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
+                    .is_none()
+            );
             {
                 let mut postings = inverted_index
                     .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
@@ -609,12 +645,12 @@ mod tests {
 
     #[test]
     fn test_indexed_u64() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let field = schema_builder.add_u64_field("value", INT_INDEXED);
         let schema = schema_builder.build();
 
         let index = Index::create_in_ram(schema);
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
         index_writer.add_document(doc!(field=>1u64));
         index_writer.commit().unwrap();
         index.load_searchers().unwrap();
@@ -632,12 +668,12 @@ mod tests {
 
     #[test]
     fn test_indexed_i64() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let value_field = schema_builder.add_i64_field("value", INT_INDEXED);
         let schema = schema_builder.build();
 
         let index = Index::create_in_ram(schema);
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
         let negative_val = -1i64;
         index_writer.add_document(doc!(value_field => negative_val));
         index_writer.commit().unwrap();
@@ -656,12 +692,12 @@ mod tests {
 
     #[test]
     fn test_indexedfield_not_in_documents() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let absent_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
-        let mut index_writer = index.writer_with_num_threads(2, 6_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(2, 40_000_000).unwrap();
         index_writer.add_document(doc!(text_field=>"a"));
         assert!(index_writer.commit().is_ok());
         assert!(index.load_searchers().is_ok());
@@ -672,13 +708,13 @@ mod tests {
 
     #[test]
     fn test_delete_postings2() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
 
         // writing the segment
-        let mut index_writer = index.writer_with_num_threads(2, 6_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(2, 40_000_000).unwrap();
 
         let add_document = |index_writer: &mut IndexWriter, val: &'static str| {
             let doc = doc!(text_field=>val);
@@ -708,13 +744,13 @@ mod tests {
 
     #[test]
     fn test_termfreq() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             {
                 let doc = doc!(text_field=>"af af af bc bc");
                 index_writer.add_document(doc);
@@ -727,9 +763,11 @@ mod tests {
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(text_field);
             let term_abcd = Term::from_field_text(text_field, "abcd");
-            assert!(inverted_index
-                .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
-                .is_none());
+            assert!(
+                inverted_index
+                    .read_postings(&term_abcd, IndexRecordOption::WithFreqsAndPositions)
+                    .is_none()
+            );
             let term_af = Term::from_field_text(text_field, "af");
             let mut postings = inverted_index
                 .read_postings(&term_af, IndexRecordOption::WithFreqsAndPositions)
@@ -743,17 +781,26 @@ mod tests {
 
     #[test]
     fn test_searcher_1() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
 
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
-            index_writer.add_document(doc!(text_field=>"af af af b"));
-            index_writer.add_document(doc!(text_field=>"a b c"));
-            index_writer.add_document(doc!(text_field=>"a b c d"));
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
+            {
+                let doc = doc!(text_field=>"af af af b");
+                index_writer.add_document(doc);
+            }
+            {
+                let doc = doc!(text_field=>"a b c");
+                index_writer.add_document(doc);
+            }
+            {
+                let doc = doc!(text_field=>"a b c d");
+                index_writer.add_document(doc);
+            }
             index_writer.commit().unwrap();
         }
         {
@@ -761,49 +808,62 @@ mod tests {
             let searcher = index.searcher();
             let get_doc_ids = |terms: Vec<Term>| {
                 let query = BooleanQuery::new_multiterms_query(terms);
-                let topdocs = searcher.search(&query, &TestCollector).unwrap();
-                topdocs.docs().to_vec()
+                let mut collector = TestCollector::default();
+                assert!(searcher.search(&query, &mut collector).is_ok());
+                collector.docs()
             };
-            assert_eq!(
-                get_doc_ids(vec![Term::from_field_text(text_field, "a")]),
-                vec![DocAddress(0, 1), DocAddress(0, 2)]
-            );
-            assert_eq!(
-                get_doc_ids(vec![Term::from_field_text(text_field, "af")]),
-                vec![DocAddress(0, 0)]
-            );
-            assert_eq!(
-                get_doc_ids(vec![Term::from_field_text(text_field, "b")]),
-                vec![DocAddress(0, 0), DocAddress(0, 1), DocAddress(0, 2)]
-            );
-            assert_eq!(
-                get_doc_ids(vec![Term::from_field_text(text_field, "c")]),
-                vec![DocAddress(0, 1), DocAddress(0, 2)]
-            );
-            assert_eq!(
-                get_doc_ids(vec![Term::from_field_text(text_field, "d")]),
-                vec![DocAddress(0, 2)]
-            );
-            assert_eq!(
-                get_doc_ids(vec![
-                    Term::from_field_text(text_field, "b"),
-                    Term::from_field_text(text_field, "a"),
-                ]),
-                vec![DocAddress(0, 0), DocAddress(0, 1), DocAddress(0, 2)]
-            );
+            {
+                assert_eq!(
+                    get_doc_ids(vec![Term::from_field_text(text_field, "a")]),
+                    vec![1, 2]
+                );
+            }
+            {
+                assert_eq!(
+                    get_doc_ids(vec![Term::from_field_text(text_field, "af")]),
+                    vec![0]
+                );
+            }
+            {
+                assert_eq!(
+                    get_doc_ids(vec![Term::from_field_text(text_field, "b")]),
+                    vec![0, 1, 2]
+                );
+            }
+            {
+                assert_eq!(
+                    get_doc_ids(vec![Term::from_field_text(text_field, "c")]),
+                    vec![1, 2]
+                );
+            }
+            {
+                assert_eq!(
+                    get_doc_ids(vec![Term::from_field_text(text_field, "d")]),
+                    vec![2]
+                );
+            }
+            {
+                assert_eq!(
+                    get_doc_ids(vec![
+                        Term::from_field_text(text_field, "b"),
+                        Term::from_field_text(text_field, "a"),
+                    ]),
+                    vec![0, 1, 2]
+                );
+            }
         }
     }
 
     #[test]
     fn test_searcher_2() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
 
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             {
                 let doc = doc!(text_field=>"af b");
                 index_writer.add_document(doc);
@@ -823,7 +883,7 @@ mod tests {
 
     #[test]
     fn test_doc_macro() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let other_text_field = schema_builder.add_text_field("text2", TEXT);
         let document = doc!(text_field => "tantivy",
@@ -832,16 +892,16 @@ mod tests {
         assert_eq!(document.len(), 3);
         let values = document.get_all(text_field);
         assert_eq!(values.len(), 2);
-        assert_eq!(values[0].text(), Some("tantivy"));
-        assert_eq!(values[1].text(), Some("some other value"));
+        assert_eq!(values[0].text(), "tantivy");
+        assert_eq!(values[1].text(), "some other value");
         let values = document.get_all(other_text_field);
         assert_eq!(values.len(), 1);
-        assert_eq!(values[0].text(), Some("short"));
+        assert_eq!(values[0].text(), "short");
     }
 
     #[test]
     fn test_wrong_fast_field_type() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let fast_field_unsigned = schema_builder.add_u64_field("unsigned", FAST);
         let fast_field_signed = schema_builder.add_i64_field("signed", FAST);
         let text_field = schema_builder.add_text_field("text", TEXT);

src/macros.rs

@@ -26,12 +26,12 @@
 /// #[macro_use]
 /// extern crate tantivy;
 ///
-/// use tantivy::schema::{Schema, TEXT, FAST};
+/// use tantivy::schema::{SchemaBuilder, TEXT, FAST};
 ///
 /// //...
 ///
 /// # fn main() {
-/// let mut schema_builder = Schema::builder();
+/// let mut schema_builder = SchemaBuilder::new();
 /// let title = schema_builder.add_text_field("title", TEXT);
 /// let author = schema_builder.add_text_field("text", TEXT);
 /// let likes = schema_builder.add_u64_field("num_u64", FAST);
@@ -67,33 +67,33 @@ macro_rules! doc(
 
 #[cfg(test)]
 mod test {
-    use schema::{Schema, FAST, TEXT};
+    use schema::{SchemaBuilder, FAST, TEXT};
 
     #[test]
     fn test_doc_basic() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let title = schema_builder.add_text_field("title", TEXT);
         let author = schema_builder.add_text_field("text", TEXT);
         let likes = schema_builder.add_u64_field("num_u64", FAST);
         let _schema = schema_builder.build();
         let _doc = doc!(
-        title => "Life Aquatic",
-        author => "Wes Anderson",
-        likes => 4u64
-        );
+            title => "Life Aquatic",
+            author => "Wes Anderson",
+            likes => 4u64
+            );
     }
 
     #[test]
     fn test_doc_trailing_comma() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let title = schema_builder.add_text_field("title", TEXT);
         let author = schema_builder.add_text_field("text", TEXT);
         let likes = schema_builder.add_u64_field("num_u64", FAST);
         let _schema = schema_builder.build();
         let _doc = doc!(
-        title => "Life Aquatic",
-        author => "Wes Anderson",
-        likes => 4u64,
-        );
+            title => "Life Aquatic",
+            author => "Wes Anderson",
+            likes => 4u64,
+            );
     }
 }

src/positions/mod.rs

@@ -1,3 +1,4 @@
+
 /// Positions are stored in three parts and over two files.
 //
 /// The `SegmentComponent::POSITIONS` file contains all of the bitpacked positions delta,
@@ -23,12 +24,13 @@
 /// The long skip structure makes it possible to skip rapidly to the a checkpoint close to this
 /// value, and then skip normally.
 ///
+
 mod reader;
 mod serializer;
 
 pub use self::reader::PositionReader;
 pub use self::serializer::PositionSerializer;
-use bitpacking::{BitPacker, BitPacker4x};
+use bitpacking::{BitPacker4x, BitPacker};
 
 const COMPRESSION_BLOCK_SIZE: usize = BitPacker4x::BLOCK_LEN;
 const LONG_SKIP_IN_BLOCKS: usize = 1_024;
@@ -41,10 +43,10 @@ lazy_static! {
 #[cfg(test)]
 pub mod tests {
 
-    use super::{PositionReader, PositionSerializer};
+    use std::iter;
+    use super::{PositionSerializer, PositionReader};
     use directory::ReadOnlySource;
     use positions::COMPRESSION_BLOCK_SIZE;
-    use std::iter;
 
     fn create_stream_buffer(vals: &[u32]) -> (ReadOnlySource, ReadOnlySource) {
         let mut skip_buffer = vec![];
@@ -57,10 +59,7 @@ pub mod tests {
             }
             serializer.close().unwrap();
         }
-        (
-            ReadOnlySource::from(stream_buffer),
-            ReadOnlySource::from(skip_buffer),
-        )
+        (ReadOnlySource::from(stream_buffer), ReadOnlySource::from(skip_buffer))
     }
 
     #[test]
@@ -104,7 +103,7 @@ pub mod tests {
         assert_eq!(skip.len(), 12);
         assert_eq!(stream.len(), 1168);
 
-        let mut position_reader = PositionReader::new(stream, skip, 0u64);
+        let mut position_reader = PositionReader::new(stream,skip, 0u64);
         let mut buf = [0u32; 7];
         let mut c = 0;
         for _ in 0..100 {
@@ -126,7 +125,7 @@ pub mod tests {
         let (stream, skip) = create_stream_buffer(&v[..]);
         assert_eq!(skip.len(), 15_749);
         assert_eq!(stream.len(), 1_000_000);
-        let mut position_reader = PositionReader::new(stream, skip, 128 * 1024);
+        let mut position_reader = PositionReader::new(stream,skip, 128 * 1024);
         let mut buf = [0u32; 1];
         position_reader.read(&mut buf);
         assert_eq!(buf[0], CONST_VAL);
@@ -138,17 +137,12 @@ pub mod tests {
         let (stream, skip) = create_stream_buffer(&v[..]);
         assert_eq!(skip.len(), 15_749);
         assert_eq!(stream.len(), 4_987_872);
-        for &offset in &[
-            10,
-            128 * 1024,
-            128 * 1024 - 1,
-            128 * 1024 + 7,
-            128 * 10 * 1024 + 10,
-        ] {
-            let mut position_reader = PositionReader::new(stream.clone(), skip.clone(), offset);
+        for &offset in &[10, 128 * 1024, 128 * 1024 - 1, 128 * 1024 + 7, 128 * 10 * 1024 + 10] {
+            let mut position_reader = PositionReader::new(stream.clone(),skip.clone(), offset);
             let mut buf = [0u32; 1];
             position_reader.read(&mut buf);
             assert_eq!(buf[0], offset as u32);
         }
     }
 }
+

src/positions/reader.rs

@@ -1,12 +1,12 @@
-use super::BIT_PACKER;
-use bitpacking::{BitPacker, BitPacker4x};
-use common::{BinarySerializable, FixedSize};
-use directory::ReadOnlySource;
+use bitpacking::{BitPacker4x, BitPacker};
 use owned_read::OwnedRead;
-use positions::COMPRESSION_BLOCK_SIZE;
-use positions::LONG_SKIP_INTERVAL;
-use positions::LONG_SKIP_IN_BLOCKS;
+use common::{BinarySerializable, FixedSize};
 use postings::compression::compressed_block_size;
+use directory::ReadOnlySource;
+use positions::COMPRESSION_BLOCK_SIZE;
+use positions::LONG_SKIP_IN_BLOCKS;
+use positions::LONG_SKIP_INTERVAL;
+use super::BIT_PACKER;
 
 pub struct PositionReader {
     skip_read: OwnedRead,
@@ -18,6 +18,7 @@ pub struct PositionReader {
                           // of the block of the next int to read.
 }
 
+
 // `ahead` represents the offset of the block currently loaded
 // compared to the cursor of the actual stream.
 //
@@ -31,8 +32,7 @@ fn read_impl(
     buffer: &mut [u32; 128],
     mut inner_offset: usize,
     num_bits: &[u8],
-    output: &mut [u32],
-) -> usize {
+    output: &mut [u32]) -> usize {
     let mut output_start = 0;
     let mut output_len = output.len();
     let mut ahead = 0;
@@ -47,7 +47,8 @@ fn read_impl(
             output_start += available_len;
             inner_offset = 0;
             let num_bits = num_bits[ahead];
-            BitPacker4x::new().decompress(position, &mut buffer[..], num_bits);
+            BitPacker4x::new()
+                .decompress(position, &mut buffer[..], num_bits);
             let block_len = compressed_block_size(num_bits);
             position = &position[block_len..];
             ahead += 1;
@@ -55,12 +56,11 @@ fn read_impl(
     }
 }
 
+
 impl PositionReader {
-    pub fn new(
-        position_source: ReadOnlySource,
-        skip_source: ReadOnlySource,
-        offset: u64,
-    ) -> PositionReader {
+    pub fn new(position_source: ReadOnlySource,
+               skip_source: ReadOnlySource,
+               offset: u64) -> PositionReader {
         let skip_len = skip_source.len();
         let (body, footer) = skip_source.split(skip_len - u32::SIZE_IN_BYTES);
         let num_long_skips = u32::deserialize(&mut footer.as_slice()).expect("Index corrupted");
@@ -70,8 +70,7 @@ impl PositionReader {
         let small_skip = (offset - (long_skip_id as u64) * (LONG_SKIP_INTERVAL as u64)) as usize;
         let offset_num_bytes: u64 = {
             if long_skip_id > 0 {
-                let mut long_skip_blocks: &[u8] =
-                    &long_skips.as_slice()[(long_skip_id - 1) * 8..][..8];
+                let mut long_skip_blocks: &[u8] = &long_skips.as_slice()[(long_skip_id - 1) * 8..][..8];
                 u64::deserialize(&mut long_skip_blocks).expect("Index corrupted") * 16
             } else {
                 0
@@ -80,13 +79,13 @@ impl PositionReader {
         let mut position_read = OwnedRead::new(position_source);
         position_read.advance(offset_num_bytes as usize);
         let mut skip_read = OwnedRead::new(skip_body);
-        skip_read.advance(long_skip_id * LONG_SKIP_IN_BLOCKS);
+        skip_read.advance(long_skip_id  * LONG_SKIP_IN_BLOCKS);
         let mut position_reader = PositionReader {
             skip_read,
             position_read,
             inner_offset: 0,
             buffer: Box::new([0u32; 128]),
-            ahead: None,
+            ahead: None
         };
         position_reader.skip(small_skip);
         position_reader
@@ -109,8 +108,7 @@ impl PositionReader {
             self.buffer.as_mut(),
             self.inner_offset,
             &skip_data[1..],
-            output,
-        ));
+            output));
     }
 
     /// Skip the next `skip_len` integer.
@@ -120,25 +118,27 @@ impl PositionReader {
     ///
     /// May panic if the end of the stream is reached.
     pub fn skip(&mut self, skip_len: usize) {
+
         let skip_len_plus_inner_offset = skip_len + self.inner_offset;
 
         let num_blocks_to_advance = skip_len_plus_inner_offset / COMPRESSION_BLOCK_SIZE;
         self.inner_offset = skip_len_plus_inner_offset % COMPRESSION_BLOCK_SIZE;
 
-        self.ahead = self.ahead.and_then(|num_blocks| {
-            if num_blocks >= num_blocks_to_advance {
-                Some(num_blocks - num_blocks_to_advance)
-            } else {
-                None
-            }
-        });
+        self.ahead = self.ahead
+            .and_then(|num_blocks| {
+                if num_blocks >= num_blocks_to_advance {
+                    Some(num_blocks_to_advance - num_blocks_to_advance)
+                } else {
+                    None
+                }
+            });
 
-        let skip_len = self.skip_read.as_ref()[..num_blocks_to_advance]
+        let skip_len = self.skip_read
+            .as_ref()[..num_blocks_to_advance]
             .iter()
             .cloned()
             .map(|num_bit| num_bit as usize)
-            .sum::<usize>()
-            * (COMPRESSION_BLOCK_SIZE / 8);
+            .sum::<usize>() * (COMPRESSION_BLOCK_SIZE / 8);
 
         self.skip_read.advance(num_blocks_to_advance);
         self.position_read.advance(skip_len);

src/positions/serializer.rs

@@ -1,8 +1,8 @@
-use super::BIT_PACKER;
-use bitpacking::BitPacker;
-use common::BinarySerializable;
-use positions::{COMPRESSION_BLOCK_SIZE, LONG_SKIP_INTERVAL};
 use std::io;
+use bitpacking::BitPacker;
+use positions::{COMPRESSION_BLOCK_SIZE, LONG_SKIP_INTERVAL};
+use common::BinarySerializable;
+use super::BIT_PACKER;
 
 pub struct PositionSerializer<W: io::Write> {
     write_stream: W,
@@ -23,7 +23,7 @@ impl<W: io::Write> PositionSerializer<W> {
             buffer: vec![0u8; 128 * 4],
             num_ints: 0u64,
             long_skips: Vec::new(),
-            cumulated_num_bits: 0u64,
+            cumulated_num_bits: 0u64
         }
     }
 
@@ -31,6 +31,7 @@ impl<W: io::Write> PositionSerializer<W> {
         self.num_ints
     }
 
+
     fn remaining_block_len(&self) -> usize {
         COMPRESSION_BLOCK_SIZE - self.block.len()
     }
@@ -51,8 +52,8 @@ impl<W: io::Write> PositionSerializer<W> {
 
     fn flush_block(&mut self) -> io::Result<()> {
         let num_bits = BIT_PACKER.num_bits(&self.block[..]);
-        self.cumulated_num_bits += u64::from(num_bits);
-        self.write_skiplist.write_all(&[num_bits])?;
+        self.cumulated_num_bits += num_bits as u64;
+        self.write_skiplist.write(&[num_bits])?;
         let written_len = BIT_PACKER.compress(&self.block[..], &mut self.buffer, num_bits);
         self.write_stream.write_all(&self.buffer[..written_len])?;
         self.block.clear();

src/postings/compression/mod.rs

@@ -28,16 +28,14 @@ impl BlockEncoder {
 
     pub fn compress_block_sorted(&mut self, block: &[u32], offset: u32) -> (u8, &[u8]) {
         let num_bits = self.bitpacker.num_bits_sorted(offset, block);
-        let written_size =
-            self.bitpacker
-                .compress_sorted(offset, block, &mut self.output[..], num_bits);
+        let written_size = self.bitpacker
+            .compress_sorted(offset, block, &mut self.output[..], num_bits);
         (num_bits, &self.output[..written_size])
     }
 
     pub fn compress_block_unsorted(&mut self, block: &[u32]) -> (u8, &[u8]) {
         let num_bits = self.bitpacker.num_bits(block);
-        let written_size = self
-            .bitpacker
+        let written_size = self.bitpacker
             .compress(block, &mut self.output[..], num_bits);
         (num_bits, &self.output[..written_size])
     }
@@ -64,21 +62,19 @@ impl BlockDecoder {
         }
     }
 
-    pub fn uncompress_block_sorted(
-        &mut self,
-        compressed_data: &[u8],
-        offset: u32,
-        num_bits: u8,
-    ) -> usize {
+    pub fn uncompress_block_sorted(&mut self, compressed_data: &[u8], offset: u32, num_bits: u8) -> usize {
         self.output_len = COMPRESSION_BLOCK_SIZE;
-        self.bitpacker
-            .decompress_sorted(offset, &compressed_data, &mut self.output, num_bits)
+        self.bitpacker.decompress_sorted(
+            offset,
+            &compressed_data,
+            &mut self.output,
+            num_bits,
+        )
     }
 
     pub fn uncompress_block_unsorted(&mut self, compressed_data: &[u8], num_bits: u8) -> usize {
         self.output_len = COMPRESSION_BLOCK_SIZE;
-        self.bitpacker
-            .decompress(&compressed_data, &mut self.output, num_bits)
+        self.bitpacker.decompress(&compressed_data, &mut self.output, num_bits)
     }
 
     #[inline]
@@ -92,6 +88,7 @@ impl BlockDecoder {
     }
 }
 
+
 pub trait VIntEncoder {
     /// Compresses an array of `u32` integers,
     /// using [delta-encoding](https://en.wikipedia.org/wiki/Delta_ encoding)
@@ -266,17 +263,21 @@ pub mod tests {
 mod bench {
 
     use super::*;
+    use rand::Rng;
     use rand::SeedableRng;
-    use rand::{Rng, XorShiftRng};
+    use rand::XorShiftRng;
     use test::Bencher;
 
-    fn generate_array_with_seed(n: usize, ratio: f64, seed_val: u8) -> Vec<u32> {
-        let seed: &[u8; 16] = &[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, seed_val];
+    fn generate_array_with_seed(n: usize, ratio: f32, seed_val: u32) -> Vec<u32> {
+        let seed: &[u32; 4] = &[1, 2, 3, seed_val];
         let mut rng: XorShiftRng = XorShiftRng::from_seed(*seed);
-        (0u32..).filter(|_| rng.gen_bool(ratio)).take(n).collect()
+        (0..u32::max_value())
+            .filter(|_| rng.next_f32() < ratio)
+            .take(n)
+            .collect()
     }
 
-    pub fn generate_array(n: usize, ratio: f64) -> Vec<u32> {
+    pub fn generate_array(n: usize, ratio: f32) -> Vec<u32> {
         generate_array_with_seed(n, ratio, 4)
     }
 
@@ -293,23 +294,24 @@ mod bench {
     fn bench_uncompress(b: &mut Bencher) {
         let mut encoder = BlockEncoder::new();
         let data = generate_array(COMPRESSION_BLOCK_SIZE, 0.1);
-        let (num_bits, compressed) = encoder.compress_block_sorted(&data, 0u32);
+        let (_, compressed) = encoder.compress_block_sorted(&data, 0u32);
         let mut decoder = BlockDecoder::new();
         b.iter(|| {
-            decoder.uncompress_block_sorted(compressed, 0u32, num_bits);
+            decoder.uncompress_block_sorted(compressed, 0u32);
         });
     }
 
     #[test]
     fn test_all_docs_compression_numbits() {
-        for expected_num_bits in 0u8.. {
+        for num_bits in 0..33 {
             let mut data = [0u32; 128];
-            if expected_num_bits > 0 {
-                data[0] = (1u64 << (expected_num_bits as usize) - 1) as u32;
+            if num_bits > 0 {
+                data[0] = 1 << (num_bits - 1);
             }
             let mut encoder = BlockEncoder::new();
-            let (num_bits, compressed) = encoder.compress_block_unsorted(&data);
-            assert_eq!(compressed.len(), compressed_block_size(num_bits));
+            let compressed = encoder.compress_block_unsorted(&data);
+            assert_eq!(compressed[0] as usize, num_bits);
+            assert_eq!(compressed.len(), compressed_block_size(compressed[0]));
         }
     }
 

src/postings/compression/vint.rs

@@ -1,5 +1,9 @@
 #[inline(always)]
-pub fn compress_sorted<'a>(input: &[u32], output: &'a mut [u8], mut offset: u32) -> &'a [u8] {
+pub fn compress_sorted<'a>(
+    input: &[u32],
+    output: &'a mut [u8],
+    mut offset: u32,
+) -> &'a [u8] {
     let mut byte_written = 0;
     for &v in input {
         let mut to_encode: u32 = v - offset;
@@ -42,41 +46,47 @@ pub(crate) fn compress_unsorted<'a>(input: &[u32], output: &'a mut [u8]) -> &'a
 }
 
 #[inline(always)]
-pub fn uncompress_sorted<'a>(compressed_data: &'a [u8], output: &mut [u32], offset: u32) -> usize {
+pub fn uncompress_sorted<'a>(
+    compressed_data: &'a [u8],
+    output: &mut [u32],
+    offset: u32,
+) -> usize {
     let mut read_byte = 0;
     let mut result = offset;
-    for output_mut in output.iter_mut() {
+    let num_els = output.len();
+    for i in 0..num_els {
         let mut shift = 0u32;
         loop {
             let cur_byte = compressed_data[read_byte];
             read_byte += 1;
-            result += u32::from(cur_byte % 128u8) << shift;
+            result += ((cur_byte % 128u8) as u32) << shift;
             if cur_byte & 128u8 != 0u8 {
                 break;
             }
             shift += 7;
         }
-        *output_mut = result;
+        output[i] = result;
     }
     read_byte
 }
 
 #[inline(always)]
-pub(crate) fn uncompress_unsorted(compressed_data: &[u8], output_arr: &mut [u32]) -> usize {
+pub(crate) fn uncompress_unsorted<'a>(compressed_data: &'a [u8], output: &mut [u32]) -> usize {
     let mut read_byte = 0;
-    for output_mut in output_arr.iter_mut() {
+    let num_els = output.len();
+    for i in 0..num_els {
         let mut result = 0u32;
         let mut shift = 0u32;
         loop {
             let cur_byte = compressed_data[read_byte];
             read_byte += 1;
-            result += u32::from(cur_byte % 128u8) << shift;
+            result += ((cur_byte % 128u8) as u32) << shift;
             if cur_byte & 128u8 != 0u8 {
                 break;
             }
             shift += 7;
         }
-        *output_mut = result;
+        output[i] = result;
     }
     read_byte
 }

src/postings/mod.rs

@@ -2,7 +2,6 @@
 Postings module (also called inverted index)
 */
 
-pub(crate) mod compression;
 /// Postings module
 ///
 /// Postings, also called inverted lists, is the key datastructure
@@ -12,17 +11,18 @@ mod postings_writer;
 mod recorder;
 mod segment_postings;
 mod serializer;
-mod skip;
+pub(crate) mod compression;
 mod stacker;
 mod term_info;
+mod skip;
 
 pub(crate) use self::postings_writer::MultiFieldPostingsWriter;
 pub use self::serializer::{FieldSerializer, InvertedIndexSerializer};
 
-use self::compression::COMPRESSION_BLOCK_SIZE;
 pub use self::postings::Postings;
-pub(crate) use self::skip::SkipReader;
 pub use self::term_info::TermInfo;
+pub(crate) use self::skip::SkipReader;
+use self::compression::{COMPRESSION_BLOCK_SIZE};
 
 pub use self::segment_postings::{BlockSegmentPostings, SegmentPostings};
 
@@ -34,7 +34,7 @@ pub(crate) const USE_SKIP_INFO_LIMIT: u32 = COMPRESSION_BLOCK_SIZE as u32;
 
 pub(crate) type UnorderedTermId = u64;
 
-#[cfg_attr(feature = "cargo-clippy", allow(clippy::enum_variant_names))]
+#[allow(enum_variant_names)]
 #[derive(Debug, PartialEq, Clone, Copy, Eq)]
 pub(crate) enum FreqReadingOption {
     NoFreq,
@@ -54,25 +54,25 @@ pub mod tests {
     use indexer::operation::AddOperation;
     use indexer::SegmentWriter;
     use query::Scorer;
-    use rand::rngs::StdRng;
-    use rand::{Rng, SeedableRng};
+    use rand::{Rng, SeedableRng, XorShiftRng};
     use schema::Field;
     use schema::IndexRecordOption;
-    use schema::{Document, Schema, Term, INT_INDEXED, STRING, TEXT};
+    use schema::{Document, SchemaBuilder, Term, INT_INDEXED, STRING, TEXT};
     use std::iter;
     use DocId;
     use Score;
 
     #[test]
     pub fn test_position_write() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         let mut segment = index.new_segment();
         let mut posting_serializer = InvertedIndexSerializer::open(&mut segment).unwrap();
         {
-            let mut field_serializer = posting_serializer.new_field(text_field, 120 * 4).unwrap();
+            let mut field_serializer = posting_serializer
+                .new_field(text_field, 120 * 4).unwrap();
             field_serializer.new_term("abc".as_bytes()).unwrap();
             for doc_id in 0u32..120u32 {
                 let delta_positions = vec![1, 2, 3, 2];
@@ -89,7 +89,7 @@ pub mod tests {
 
     #[test]
     pub fn test_skip_positions() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::new();
         let title = schema_builder.add_text_field("title", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
@@ -164,7 +164,7 @@ pub mod tests {
     #[test]
     pub fn test_position_and_fieldnorm1() {
         let mut positions = Vec::new();
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema.clone());
@@ -221,10 +221,12 @@ pub mod tests {
             }
             {
                 let term_a = Term::from_field_text(text_field, "abcdef");
-                assert!(segment_reader
-                    .inverted_index(term_a.field())
-                    .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
-                    .is_none());
+                assert!(
+                    segment_reader
+                        .inverted_index(term_a.field())
+                        .read_postings(&term_a, IndexRecordOption::WithFreqsAndPositions)
+                        .is_none()
+                );
             }
             {
                 let term_a = Term::from_field_text(text_field, "a");
@@ -275,12 +277,12 @@ pub mod tests {
     #[test]
     pub fn test_position_and_fieldnorm2() {
         let mut positions: Vec<u32> = Vec::new();
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             {
                 let mut doc = Document::default();
                 doc.add_text(text_field, "g b b d c g c");
@@ -316,13 +318,13 @@ pub mod tests {
         let num_docs = 300u32;
 
         let index = {
-            let mut schema_builder = Schema::builder();
+            let mut schema_builder = SchemaBuilder::default();
             let value_field = schema_builder.add_u64_field("value", INT_INDEXED);
             let schema = schema_builder.build();
 
             let index = Index::create_in_ram(schema);
             {
-                let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+                let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
                 for i in 0..num_docs {
                     let mut doc = Document::default();
                     doc.add_u64(value_field, 2);
@@ -399,7 +401,7 @@ pub mod tests {
 
         // delete some of the documents
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             index_writer.delete_term(term_0);
             assert!(index_writer.commit().is_ok());
         }
@@ -449,7 +451,7 @@ pub mod tests {
 
         // delete everything else
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
             index_writer.delete_term(term_1);
 
             assert!(index_writer.commit().is_ok());
@@ -457,14 +459,25 @@ pub mod tests {
         index.load_searchers().unwrap();
 
         let searcher = index.searcher();
+        let segment_reader = searcher.segment_reader(0);
 
         // finally, check that it's empty
         {
-            let searchable_segment_ids = index
-                .searchable_segment_ids()
-                .expect("could not get index segment ids");
-            assert!(searchable_segment_ids.is_empty());
-            assert_eq!(searcher.num_docs(), 0);
+            let mut segment_postings = segment_reader
+                .inverted_index(term_2.field())
+                .read_postings(&term_2, IndexRecordOption::Basic)
+                .unwrap();
+
+            assert_eq!(segment_postings.skip_next(0), SkipResult::Reached);
+            assert_eq!(segment_postings.doc(), 0);
+            assert!(segment_reader.is_deleted(0));
+
+            let mut segment_postings = segment_reader
+                .inverted_index(term_2.field())
+                .read_postings(&term_2, IndexRecordOption::Basic)
+                .unwrap();
+
+            assert_eq!(segment_postings.skip_next(num_docs), SkipResult::End);
         }
     }
 
@@ -486,25 +499,26 @@ pub mod tests {
             Term::from_field_text(field, "d")
         };
         pub static ref INDEX: Index = {
-            let mut schema_builder = Schema::builder();
+            let mut schema_builder = SchemaBuilder::default();
             let text_field = schema_builder.add_text_field("text", STRING);
             let schema = schema_builder.build();
 
-            let mut rng: StdRng = StdRng::from_seed([1u8; 32]);
+            let seed: [u8; 16] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16];
+            let mut rng: XorShiftRng = XorShiftRng::from_seed(seed);
 
             let index = Index::create_in_ram(schema);
             let posting_list_size = 1_000_000;
             {
-                let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+                let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
                 for _ in 0..posting_list_size {
                     let mut doc = Document::default();
-                    if rng.gen_bool(1f64 / 15f64) {
+                    if rng.gen_bool(1f64/ 15f64) {
                         doc.add_text(text_field, "a");
                     }
-                    if rng.gen_bool(1f64 / 10f64) {
+                    if rng.gen_bool(1f64/ 10f64) {
                         doc.add_text(text_field, "b");
                     }
-                    if rng.gen_bool(1f64 / 5f64) {
+                    if rng.gen_bool(1f64/ 5f64) {
                         doc.add_text(text_field, "c");
                     }
                     doc.add_text(text_field, "d");
@@ -641,7 +655,7 @@ mod bench {
         });
     }
 
-    fn bench_skip_next(p: f64, b: &mut Bencher) {
+    fn bench_skip_next(p: f32, b: &mut Bencher) {
         let searcher = INDEX.searcher();
         let segment_reader = searcher.segment_reader(0);
         let docs = tests::sample(segment_reader.num_docs(), p);

src/postings/postings_writer.rs

@@ -1,8 +1,6 @@
 use super::stacker::{Addr, MemoryArena, TermHashMap};
 
-use postings::recorder::{
-    BufferLender, NothingRecorder, Recorder, TFAndPositionRecorder, TermFrequencyRecorder,
-};
+use postings::recorder::{NothingRecorder, Recorder, TFAndPositionRecorder, TermFrequencyRecorder};
 use postings::UnorderedTermId;
 use postings::{FieldSerializer, InvertedIndexSerializer};
 use schema::IndexRecordOption;
@@ -17,7 +15,7 @@ use tokenizer::TokenStream;
 use DocId;
 use Result;
 
-fn posting_from_field_entry(field_entry: &FieldEntry) -> Box<PostingsWriter> {
+fn posting_from_field_entry<'a>(field_entry: &FieldEntry) -> Box<PostingsWriter> {
     match *field_entry.field_type() {
         FieldType::Str(ref text_options) => text_options
             .get_indexing_options()
@@ -96,12 +94,11 @@ impl MultiFieldPostingsWriter {
         &self,
         serializer: &mut InvertedIndexSerializer,
     ) -> Result<HashMap<Field, HashMap<UnorderedTermId, TermOrdinal>>> {
-        let mut term_offsets: Vec<(&[u8], Addr, UnorderedTermId)> = self
-            .term_index
+        let mut term_offsets: Vec<(&[u8], Addr, UnorderedTermId)> = self.term_index
             .iter()
             .map(|(term_bytes, addr, bucket_id)| (term_bytes, addr, bucket_id as UnorderedTermId))
             .collect();
-        term_offsets.sort_unstable_by_key(|&(k, _, _)| k);
+        term_offsets.sort_by_key(|&(k, _, _)| k);
 
         let mut offsets: Vec<(Field, usize)> = vec![];
         let term_offsets_it = term_offsets
@@ -110,8 +107,10 @@ impl MultiFieldPostingsWriter {
             .map(|(key, _, _)| Term::wrap(key).field())
             .enumerate();
 
-        let mut unordered_term_mappings: HashMap<Field, HashMap<UnorderedTermId, TermOrdinal>> =
-            HashMap::new();
+        let mut unordered_term_mappings: HashMap<
+            Field,
+            HashMap<UnorderedTermId, TermOrdinal>,
+        > = HashMap::new();
 
         let mut prev_field = Field(u32::max_value());
         for (offset, field) in term_offsets_it {
@@ -128,8 +127,8 @@ impl MultiFieldPostingsWriter {
 
             let field_entry = self.schema.get_field_entry(field);
 
-            match *field_entry.field_type() {
-                FieldType::Str(_) | FieldType::HierarchicalFacet => {
+            match field_entry.field_type() {
+                &FieldType::Str(_) | &FieldType::HierarchicalFacet => {
                     // populating the (unordered term ord) -> (ordered term ord) mapping
                     // for the field.
                     let mut unordered_term_ids = term_offsets[start..stop]
@@ -143,8 +142,8 @@ impl MultiFieldPostingsWriter {
                         .collect();
                     unordered_term_mappings.insert(field, mapping);
                 }
-                FieldType::U64(_) | FieldType::I64(_) => {}
-                FieldType::Bytes => {}
+                &FieldType::U64(_) | &FieldType::I64(_) => {}
+                &FieldType::Bytes => {}
             }
 
             let postings_writer = &self.per_field_postings_writers[field.0 as usize];
@@ -203,11 +202,14 @@ pub trait PostingsWriter {
         heap: &mut MemoryArena,
     ) -> u32 {
         let mut term = Term::for_field(field);
-        let mut sink = |token: &Token| {
-            term.set_text(token.text.as_str());
-            self.subscribe(term_index, doc_id, token.position as u32, &term, heap);
+        let num_tokens = {
+            let mut sink = |token: &Token| {
+                term.set_text(token.text.as_str());
+                self.subscribe(term_index, doc_id, token.position as u32, &term, heap);
+            };
+            token_stream.process(&mut sink)
         };
-        token_stream.process(&mut sink)
+        num_tokens
     }
 
     fn total_num_tokens(&self) -> u64;
@@ -215,7 +217,7 @@ pub trait PostingsWriter {
 
 /// The `SpecializedPostingsWriter` is just here to remove dynamic
 /// dispatch to the recorder information.
-pub(crate) struct SpecializedPostingsWriter<Rec: Recorder + 'static> {
+pub struct SpecializedPostingsWriter<Rec: Recorder + 'static> {
     total_num_tokens: u64,
     _recorder_type: PhantomData<Rec>,
 }
@@ -247,7 +249,8 @@ impl<Rec: Recorder + 'static> PostingsWriter for SpecializedPostingsWriter<Rec>
         debug_assert!(term.as_slice().len() >= 4);
         self.total_num_tokens += 1;
         term_index.mutate_or_create(term, |opt_recorder: Option<Rec>| {
-            if let Some(mut recorder) = opt_recorder {
+            if opt_recorder.is_some() {
+                let mut recorder = opt_recorder.unwrap();
                 let current_doc = recorder.current_doc();
                 if current_doc != doc {
                     recorder.close_doc(heap);
@@ -256,7 +259,7 @@ impl<Rec: Recorder + 'static> PostingsWriter for SpecializedPostingsWriter<Rec>
                 recorder.record_position(position, heap);
                 recorder
             } else {
-                let mut recorder = Rec::new();
+                let mut recorder = Rec::new(heap);
                 recorder.new_doc(doc, heap);
                 recorder.record_position(position, heap);
                 recorder
@@ -271,11 +274,10 @@ impl<Rec: Recorder + 'static> PostingsWriter for SpecializedPostingsWriter<Rec>
         termdict_heap: &MemoryArena,
         heap: &MemoryArena,
     ) -> io::Result<()> {
-        let mut buffer_lender = BufferLender::default();
         for &(term_bytes, addr, _) in term_addrs {
-            let recorder: Rec = termdict_heap.read(addr);
+            let recorder: Rec = unsafe { termdict_heap.read(addr) };
             serializer.new_term(&term_bytes[4..])?;
-            recorder.serialize(&mut buffer_lender, serializer, heap)?;
+            recorder.serialize(serializer, heap)?;
             serializer.close_term()?;
         }
         Ok(())

src/postings/recorder.rs

@@ -1,51 +1,10 @@
 use super::stacker::{ExpUnrolledLinkedList, MemoryArena};
-use common::{read_u32_vint, write_u32_vint};
 use postings::FieldSerializer;
-use std::io;
+use std::{self, io};
 use DocId;
 
 const EMPTY_ARRAY: [u32; 0] = [0u32; 0];
-const POSITION_END: u32 = 0;
-
-#[derive(Default)]
-pub(crate) struct BufferLender {
-    buffer_u8: Vec<u8>,
-    buffer_u32: Vec<u32>,
-}
-
-impl BufferLender {
-    pub fn lend_u8(&mut self) -> &mut Vec<u8> {
-        self.buffer_u8.clear();
-        &mut self.buffer_u8
-    }
-    pub fn lend_all(&mut self) -> (&mut Vec<u8>, &mut Vec<u32>) {
-        self.buffer_u8.clear();
-        self.buffer_u32.clear();
-        (&mut self.buffer_u8, &mut self.buffer_u32)
-    }
-}
-
-pub struct VInt32Reader<'a> {
-    data: &'a [u8],
-}
-
-impl<'a> VInt32Reader<'a> {
-    fn new(data: &'a [u8]) -> VInt32Reader<'a> {
-        VInt32Reader { data }
-    }
-}
-
-impl<'a> Iterator for VInt32Reader<'a> {
-    type Item = u32;
-
-    fn next(&mut self) -> Option<u32> {
-        if self.data.is_empty() {
-            None
-        } else {
-            Some(read_u32_vint(&mut self.data))
-        }
-    }
-}
+const POSITION_END: u32 = std::u32::MAX;
 
 /// Recorder is in charge of recording relevant information about
 /// the presence of a term in a document.
@@ -56,9 +15,9 @@ impl<'a> Iterator for VInt32Reader<'a> {
 ///   * the document id
 ///   * the term frequency
 ///   * the term positions
-pub(crate) trait Recorder: Copy + 'static {
+pub trait Recorder: Copy {
     ///
-    fn new() -> Self;
+    fn new(heap: &mut MemoryArena) -> Self;
     /// Returns the current document
     fn current_doc(&self) -> u32;
     /// Starts recording information about a new document
@@ -70,12 +29,7 @@ pub(crate) trait Recorder: Copy + 'static {
     /// Close the document. It will help record the term frequency.
     fn close_doc(&mut self, heap: &mut MemoryArena);
     /// Pushes the postings information to the serializer.
-    fn serialize(
-        &self,
-        buffer_lender: &mut BufferLender,
-        serializer: &mut FieldSerializer,
-        heap: &MemoryArena,
-    ) -> io::Result<()>;
+    fn serialize(&self, serializer: &mut FieldSerializer, heap: &MemoryArena) -> io::Result<()>;
 }
 
 /// Only records the doc ids
@@ -86,9 +40,9 @@ pub struct NothingRecorder {
 }
 
 impl Recorder for NothingRecorder {
-    fn new() -> Self {
+    fn new(heap: &mut MemoryArena) -> Self {
         NothingRecorder {
-            stack: ExpUnrolledLinkedList::new(),
+            stack: ExpUnrolledLinkedList::new(heap),
             current_doc: u32::max_value(),
         }
     }
@@ -99,23 +53,16 @@ impl Recorder for NothingRecorder {
 
     fn new_doc(&mut self, doc: DocId, heap: &mut MemoryArena) {
         self.current_doc = doc;
-        let _ = write_u32_vint(doc, &mut self.stack.writer(heap));
+        self.stack.push(doc, heap);
     }
 
     fn record_position(&mut self, _position: u32, _heap: &mut MemoryArena) {}
 
     fn close_doc(&mut self, _heap: &mut MemoryArena) {}
 
-    fn serialize(
-        &self,
-        buffer_lender: &mut BufferLender,
-        serializer: &mut FieldSerializer,
-        heap: &MemoryArena,
-    ) -> io::Result<()> {
-        let buffer = buffer_lender.lend_u8();
-        self.stack.read_to_end(heap, buffer);
-        for doc in VInt32Reader::new(&buffer[..]) {
-            serializer.write_doc(doc as u32, 0u32, &EMPTY_ARRAY)?;
+    fn serialize(&self, serializer: &mut FieldSerializer, heap: &MemoryArena) -> io::Result<()> {
+        for doc in self.stack.iter(heap) {
+            serializer.write_doc(doc, 0u32, &EMPTY_ARRAY)?;
         }
         Ok(())
     }
@@ -130,9 +77,9 @@ pub struct TermFrequencyRecorder {
 }
 
 impl Recorder for TermFrequencyRecorder {
-    fn new() -> Self {
+    fn new(heap: &mut MemoryArena) -> Self {
         TermFrequencyRecorder {
-            stack: ExpUnrolledLinkedList::new(),
+            stack: ExpUnrolledLinkedList::new(heap),
             current_doc: u32::max_value(),
             current_tf: 0u32,
         }
@@ -144,7 +91,7 @@ impl Recorder for TermFrequencyRecorder {
 
     fn new_doc(&mut self, doc: DocId, heap: &mut MemoryArena) {
         self.current_doc = doc;
-        let _ = write_u32_vint(doc, &mut self.stack.writer(heap));
+        self.stack.push(doc, heap);
     }
 
     fn record_position(&mut self, _position: u32, _heap: &mut MemoryArena) {
@@ -153,24 +100,23 @@ impl Recorder for TermFrequencyRecorder {
 
     fn close_doc(&mut self, heap: &mut MemoryArena) {
         debug_assert!(self.current_tf > 0);
-        let _ = write_u32_vint(self.current_tf, &mut self.stack.writer(heap));
+        self.stack.push(self.current_tf, heap);
         self.current_tf = 0;
     }
 
-    fn serialize(
-        &self,
-        buffer_lender: &mut BufferLender,
-        serializer: &mut FieldSerializer,
-        heap: &MemoryArena,
-    ) -> io::Result<()> {
-        let buffer = buffer_lender.lend_u8();
-        self.stack.read_to_end(heap, buffer);
-        let mut u32_it = VInt32Reader::new(&buffer[..]);
-        while let Some(doc) = u32_it.next() {
-            let term_freq = u32_it.next().unwrap_or(self.current_tf);
-            serializer.write_doc(doc as u32, term_freq, &EMPTY_ARRAY)?;
-        }
+    fn serialize(&self, serializer: &mut FieldSerializer, heap: &MemoryArena) -> io::Result<()> {
+        // the last document has not been closed...
+        // its term freq is self.current_tf.
+        let mut doc_iter = self.stack
+            .iter(heap)
+            .chain(Some(self.current_tf).into_iter());
 
+        while let Some(doc) = doc_iter.next() {
+            let term_freq = doc_iter
+                .next()
+                .expect("The IndexWriter recorded a doc without a term freq.");
+            serializer.write_doc(doc, term_freq, &EMPTY_ARRAY)?;
+        }
         Ok(())
     }
 }
@@ -181,10 +127,11 @@ pub struct TFAndPositionRecorder {
     stack: ExpUnrolledLinkedList,
     current_doc: DocId,
 }
+
 impl Recorder for TFAndPositionRecorder {
-    fn new() -> Self {
+    fn new(heap: &mut MemoryArena) -> Self {
         TFAndPositionRecorder {
-            stack: ExpUnrolledLinkedList::new(),
+            stack: ExpUnrolledLinkedList::new(heap),
             current_doc: u32::max_value(),
         }
     }
@@ -195,88 +142,33 @@ impl Recorder for TFAndPositionRecorder {
 
     fn new_doc(&mut self, doc: DocId, heap: &mut MemoryArena) {
         self.current_doc = doc;
-        let _ = write_u32_vint(doc, &mut self.stack.writer(heap));
+        self.stack.push(doc, heap);
     }
 
     fn record_position(&mut self, position: u32, heap: &mut MemoryArena) {
-        let _ = write_u32_vint(position + 1u32, &mut self.stack.writer(heap));
+        self.stack.push(position, heap);
     }
 
     fn close_doc(&mut self, heap: &mut MemoryArena) {
-        let _ = write_u32_vint(POSITION_END, &mut self.stack.writer(heap));
+        self.stack.push(POSITION_END, heap);
     }
 
-    fn serialize(
-        &self,
-        buffer_lender: &mut BufferLender,
-        serializer: &mut FieldSerializer,
-        heap: &MemoryArena,
-    ) -> io::Result<()> {
-        let (buffer_u8, buffer_positions) = buffer_lender.lend_all();
-        self.stack.read_to_end(heap, buffer_u8);
-        let mut u32_it = VInt32Reader::new(&buffer_u8[..]);
-        while let Some(doc) = u32_it.next() {
-            let mut prev_position_plus_one = 1u32;
-            buffer_positions.clear();
-            loop {
-                match u32_it.next() {
-                    Some(POSITION_END) | None => {
-                        break;
-                    }
-                    Some(position_plus_one) => {
-                        let delta_position = position_plus_one - prev_position_plus_one;
-                        buffer_positions.push(delta_position);
-                        prev_position_plus_one = position_plus_one;
-                    }
+    fn serialize(&self, serializer: &mut FieldSerializer, heap: &MemoryArena) -> io::Result<()> {
+        let mut doc_positions = Vec::with_capacity(100);
+        let mut positions_iter = self.stack.iter(heap);
+        while let Some(doc) = positions_iter.next() {
+            let mut prev_position = 0;
+            doc_positions.clear();
+            for position in &mut positions_iter {
+                if position == POSITION_END {
+                    break;
+                } else {
+                    doc_positions.push(position - prev_position);
+                    prev_position = position;
                 }
             }
-            serializer.write_doc(doc, buffer_positions.len() as u32, &buffer_positions)?;
+            serializer.write_doc(doc, doc_positions.len() as u32, &doc_positions)?;
         }
         Ok(())
     }
 }
-
-#[cfg(test)]
-mod tests {
-
-    use super::write_u32_vint;
-    use super::BufferLender;
-    use super::VInt32Reader;
-
-    #[test]
-    fn test_buffer_lender() {
-        let mut buffer_lender = BufferLender::default();
-        {
-            let buf = buffer_lender.lend_u8();
-            assert!(buf.is_empty());
-            buf.push(1u8);
-        }
-        {
-            let buf = buffer_lender.lend_u8();
-            assert!(buf.is_empty());
-            buf.push(1u8);
-        }
-        {
-            let (_, buf) = buffer_lender.lend_all();
-            assert!(buf.is_empty());
-            buf.push(1u32);
-        }
-        {
-            let (_, buf) = buffer_lender.lend_all();
-            assert!(buf.is_empty());
-            buf.push(1u32);
-        }
-    }
-
-    #[test]
-    fn test_vint_u32() {
-        let mut buffer = vec![];
-        let vals = [0, 1, 324_234_234, u32::max_value()];
-        for &i in &vals {
-            assert!(write_u32_vint(i, &mut buffer).is_ok());
-        }
-        assert_eq!(buffer.len(), 1 + 1 + 5 + 5);
-        let res: Vec<u32> = VInt32Reader::new(&buffer[..]).collect();
-        assert_eq!(&res[..], &vals[..]);
-    }
-}

src/postings/segment_postings.rs

@@ -1,20 +1,20 @@
+use postings::compression::{BlockDecoder, VIntDecoder, COMPRESSION_BLOCK_SIZE};
+use DocId;
 use common::BitSet;
 use common::HasLen;
-use common::{BinarySerializable, VInt};
+use postings::compression::compressed_block_size;
 use docset::{DocSet, SkipResult};
 use fst::Streamer;
-use owned_read::OwnedRead;
-use positions::PositionReader;
-use postings::compression::compressed_block_size;
-use postings::compression::{BlockDecoder, VIntDecoder, COMPRESSION_BLOCK_SIZE};
 use postings::serializer::PostingsSerializer;
 use postings::FreqReadingOption;
 use postings::Postings;
-use postings::SkipReader;
+use owned_read::OwnedRead;
+use common::{VInt, BinarySerializable};
 use postings::USE_SKIP_INFO_LIMIT;
+use postings::SkipReader;
 use schema::IndexRecordOption;
+use positions::PositionReader;
 use std::cmp::Ordering;
-use DocId;
 
 const EMPTY_ARR: [u8; 0] = [];
 
@@ -98,7 +98,7 @@ impl SegmentPostings {
             docs.len() as u32,
             OwnedRead::new(buffer),
             IndexRecordOption::Basic,
-            IndexRecordOption::Basic,
+            IndexRecordOption::Basic
         );
         SegmentPostings::from_block_postings(block_segment_postings, None)
     }
@@ -124,17 +124,22 @@ impl SegmentPostings {
 }
 
 fn exponential_search(target: u32, arr: &[u32]) -> (usize, usize) {
+    let mut start = 0;
     let end = arr.len();
-    debug_assert!(arr.len() <= 128);
+    debug_assert!(target >= arr[start]);
     debug_assert!(target <= arr[end - 1]);
-    let mut begin = 0;
-    for &pivot in [1,3,7,15,31,63].iter().take_while(|&&el| el < end) {
-        if arr[pivot] > target {
-            return (begin, pivot);
+    let mut jump = 1;
+    loop {
+        let new = start + jump;
+        if new >= end {
+            return (start, end);
         }
-        begin = pivot;
+        if arr[new] > target {
+            return (start, new);
+        }
+        start = new;
+        jump *= 2;
     }
-    (begin, end)
 }
 
 /// Search the first index containing an element greater or equal to the target.
@@ -146,11 +151,7 @@ fn exponential_search(target: u32, arr: &[u32]) -> (usize, usize) {
 /// The target is assumed smaller or equal to the last element.
 fn search_within_block(block_docs: &[u32], target: u32) -> usize {
     let (start, end) = exponential_search(target, block_docs);
-    start.wrapping_add(
-        block_docs[start..end]
-            .binary_search(&target)
-            .unwrap_or_else(|e| e),
-    )
+    start.wrapping_add(block_docs[start..end].binary_search(&target).unwrap_or_else(|e| e))
 }
 
 impl DocSet for SegmentPostings {
@@ -178,20 +179,21 @@ impl DocSet for SegmentPostings {
         // check if we need to go to the next block
         let need_positions = self.position_computer.is_some();
         let mut sum_freqs_skipped: u32 = 0;
-        if !self
-            .block_cursor
-            .docs()
-            .last()
-            .map(|doc| *doc >= target)
-            .unwrap_or(false)
-        // there should always be at least a document in the block
-        // since advance returned.
+        if !self.block_cursor
+               .docs()
+               .last()
+               .map(|doc| *doc >= target)
+               .unwrap_or(false)  // there should always be at least a document in the block
+                                       // since advance returned.
         {
             // we are not in the right block.
             //
             // First compute all of the freqs skipped from the current block.
             if need_positions {
-                sum_freqs_skipped = self.block_cursor.freqs()[self.cur..].iter().sum();
+                sum_freqs_skipped = self.block_cursor
+                    .freqs()[self.cur..]
+                    .iter()
+                    .sum();
                 match self.block_cursor.skip_to(target) {
                     BlockSegmentPostingsSkipResult::Success(block_skip_freqs) => {
                         sum_freqs_skipped += block_skip_freqs;
@@ -200,25 +202,22 @@ impl DocSet for SegmentPostings {
                         return SkipResult::End;
                     }
                 }
-            } else if self.block_cursor.skip_to(target)
-                == BlockSegmentPostingsSkipResult::Terminated
-            {
+            } else {
                 // no positions needed. no need to sum freqs.
-                return SkipResult::End;
+                if self.block_cursor.skip_to(target) == BlockSegmentPostingsSkipResult::Terminated {
+                    return SkipResult::End;
+                }
             }
             self.cur = 0;
         }
 
         // we're in the right block now, start with an exponential search
         let block_docs = self.block_cursor.docs();
-        let new_cur = self
-            .cur
-            .wrapping_add(search_within_block(&block_docs[self.cur..], target));
 
+        debug_assert!(target >= self.doc());
+        let new_cur = self.cur.wrapping_add(search_within_block(&block_docs[self.cur..], target));
         if need_positions {
-            sum_freqs_skipped += self.block_cursor.freqs()[self.cur..new_cur]
-                .iter()
-                .sum::<u32>();
+            sum_freqs_skipped += self.block_cursor.freqs()[self.cur..new_cur].iter().sum::<u32>();
             self.position_computer
                 .as_mut()
                 .unwrap()
@@ -230,9 +229,9 @@ impl DocSet for SegmentPostings {
         let doc = block_docs[new_cur];
         debug_assert!(doc >= target);
         if doc == target {
-            SkipResult::Reached
+            return SkipResult::Reached;
         } else {
-            SkipResult::OverStep
+            return SkipResult::OverStep;
         }
     }
 
@@ -331,10 +330,7 @@ pub struct BlockSegmentPostings {
     skip_reader: SkipReader,
 }
 
-fn split_into_skips_and_postings(
-    doc_freq: u32,
-    mut data: OwnedRead,
-) -> (Option<OwnedRead>, OwnedRead) {
+fn split_into_skips_and_postings(doc_freq: u32, mut data: OwnedRead) -> (Option<OwnedRead>, OwnedRead) {
     if doc_freq >= USE_SKIP_INFO_LIMIT {
         let skip_len = VInt::deserialize(&mut data).expect("Data corrupted").0 as usize;
         let mut postings_data = data.clone();
@@ -349,7 +345,7 @@ fn split_into_skips_and_postings(
 #[derive(Debug, Eq, PartialEq)]
 pub enum BlockSegmentPostingsSkipResult {
     Terminated,
-    Success(u32), //< number of term freqs to skip
+    Success(u32) //< number of term freqs to skip
 }
 
 impl BlockSegmentPostings {
@@ -357,7 +353,7 @@ impl BlockSegmentPostings {
         doc_freq: u32,
         data: OwnedRead,
         record_option: IndexRecordOption,
-        requested_option: IndexRecordOption,
+        requested_option: IndexRecordOption
     ) -> BlockSegmentPostings {
         let freq_reading_option = match (record_option, requested_option) {
             (IndexRecordOption::Basic, _) => FreqReadingOption::NoFreq,
@@ -366,10 +362,11 @@ impl BlockSegmentPostings {
         };
 
         let (skip_data_opt, postings_data) = split_into_skips_and_postings(doc_freq, data);
-        let skip_reader = match skip_data_opt {
-            Some(skip_data) => SkipReader::new(skip_data, record_option),
-            None => SkipReader::new(OwnedRead::new(&EMPTY_ARR[..]), record_option),
-        };
+        let skip_reader =
+            match skip_data_opt {
+                Some(skip_data) => SkipReader::new(skip_data, record_option),
+                None => SkipReader::new(OwnedRead::new(&EMPTY_ARR[..]), record_option)
+            };
         let doc_freq = doc_freq as usize;
         let num_vint_docs = doc_freq % COMPRESSION_BLOCK_SIZE;
         BlockSegmentPostings {
@@ -453,6 +450,7 @@ impl BlockSegmentPostings {
         self.doc_decoder.output_len
     }
 
+
     /// position on a block that may contains `doc_id`.
     /// Always advance the current block.
     ///
@@ -463,7 +461,9 @@ impl BlockSegmentPostings {
     /// Returns false iff all of the document remaining are smaller than
     /// `doc_id`. In that case, all of these document are consumed.
     ///
-    pub fn skip_to(&mut self, target_doc: DocId) -> BlockSegmentPostingsSkipResult {
+    pub fn skip_to(&mut self,
+                   target_doc: DocId) -> BlockSegmentPostingsSkipResult {
+
         let mut skip_freqs = 0u32;
         while self.skip_reader.advance() {
             if self.skip_reader.doc() >= target_doc {
@@ -472,11 +472,11 @@ impl BlockSegmentPostings {
                 //
                 // We found our block!
                 let num_bits = self.skip_reader.doc_num_bits();
-                let num_consumed_bytes = self.doc_decoder.uncompress_block_sorted(
-                    self.remaining_data.as_ref(),
-                    self.doc_offset,
-                    num_bits,
-                );
+                let num_consumed_bytes = self.doc_decoder
+                    .uncompress_block_sorted(
+                        self.remaining_data.as_ref(),
+                        self.doc_offset,
+                        num_bits);
                 self.remaining_data.advance(num_consumed_bytes);
                 let tf_num_bits = self.skip_reader.tf_num_bits();
                 match self.freq_reading_option {
@@ -486,9 +486,9 @@ impl BlockSegmentPostings {
                         self.remaining_data.advance(num_bytes_to_skip);
                     }
                     FreqReadingOption::ReadFreq => {
-                        let num_consumed_bytes = self
-                            .freq_decoder
-                            .uncompress_block_unsorted(self.remaining_data.as_ref(), tf_num_bits);
+                        let num_consumed_bytes = self.freq_decoder
+                            .uncompress_block_unsorted(self.remaining_data.as_ref(),
+                                                       tf_num_bits);
                         self.remaining_data.advance(num_consumed_bytes);
                     }
                 }
@@ -518,8 +518,7 @@ impl BlockSegmentPostings {
                 }
             }
             self.num_vint_docs = 0;
-            return self
-                .docs()
+            return self.docs()
                 .last()
                 .map(|last_doc| {
                     if *last_doc >= target_doc {
@@ -539,11 +538,11 @@ impl BlockSegmentPostings {
     pub fn advance(&mut self) -> bool {
         if self.skip_reader.advance() {
             let num_bits = self.skip_reader.doc_num_bits();
-            let num_consumed_bytes = self.doc_decoder.uncompress_block_sorted(
-                self.remaining_data.as_ref(),
-                self.doc_offset,
-                num_bits,
-            );
+            let num_consumed_bytes = self.doc_decoder
+                .uncompress_block_sorted(
+                    self.remaining_data.as_ref(),
+                        self.doc_offset,
+                        num_bits);
             self.remaining_data.advance(num_consumed_bytes);
             let tf_num_bits = self.skip_reader.tf_num_bits();
             match self.freq_reading_option {
@@ -553,9 +552,9 @@ impl BlockSegmentPostings {
                     self.remaining_data.advance(num_bytes_to_skip);
                 }
                 FreqReadingOption::ReadFreq => {
-                    let num_consumed_bytes = self
-                        .freq_decoder
-                        .uncompress_block_unsorted(self.remaining_data.as_ref(), tf_num_bits);
+                    let num_consumed_bytes = self.freq_decoder
+                        .uncompress_block_unsorted(self.remaining_data.as_ref(),
+                                                   tf_num_bits);
                     self.remaining_data.advance(num_consumed_bytes);
                 }
             }
@@ -595,6 +594,7 @@ impl BlockSegmentPostings {
             doc_offset: 0,
             doc_freq: 0,
 
+
             remaining_data: OwnedRead::new(vec![]),
             skip_reader: SkipReader::new(OwnedRead::new(vec![]), IndexRecordOption::Basic),
         }
@@ -616,21 +616,19 @@ impl<'b> Streamer<'b> for BlockSegmentPostings {
 #[cfg(test)]
 mod tests {
 
-    use super::exponential_search;
-    use super::search_within_block;
     use super::BlockSegmentPostings;
-    use super::BlockSegmentPostingsSkipResult;
     use super::SegmentPostings;
     use common::HasLen;
     use core::Index;
     use docset::DocSet;
     use fst::Streamer;
     use schema::IndexRecordOption;
-    use schema::Schema;
+    use schema::SchemaBuilder;
     use schema::Term;
     use schema::INT_INDEXED;
+    use super::BlockSegmentPostingsSkipResult;
     use DocId;
-    use SkipResult;
+    use super::search_within_block;
 
     #[test]
     fn test_empty_segment_postings() {
@@ -647,6 +645,7 @@ mod tests {
         assert_eq!(postings.doc_freq(), 0);
     }
 
+
     fn search_within_block_trivial_but_slow(block: &[u32], target: u32) -> usize {
         block
             .iter()
@@ -654,25 +653,11 @@ mod tests {
             .enumerate()
             .filter(|&(_, ref val)| *val >= target)
             .next()
-            .unwrap()
-            .0
-    }
-
-    #[test]
-    fn test_exponentiel_search() {
-        assert_eq!(exponential_search(0, &[1, 2]), (0, 1));
-        assert_eq!(exponential_search(1, &[1, 2]), (0, 1));
-        assert_eq!(
-            exponential_search(7, &[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]),
-            (3, 7)
-        );
+            .unwrap().0
     }
 
     fn util_test_search_within_block(block: &[u32], target: u32) {
-        assert_eq!(
-            search_within_block(block, target),
-            search_within_block_trivial_but_slow(block, target)
-        );
+        assert_eq!(search_within_block(block, target), search_within_block_trivial_but_slow(block, target));
     }
 
     fn util_test_search_within_block_all(block: &[u32]) {
@@ -692,14 +677,14 @@ mod tests {
     #[test]
     fn test_search_within_block() {
         for len in 1u32..128u32 {
-            let v: Vec<u32> = (0..len).map(|i| i * 2).collect();
+            let v: Vec<u32> = (0..len).map(|i| i*2).collect();
             util_test_search_within_block_all(&v[..]);
         }
     }
 
     #[test]
     fn test_block_segment_postings() {
-        let mut block_segments = build_block_postings(&(0..100_000).collect::<Vec<u32>>());
+        let mut block_segments = build_block_postings((0..100_000).collect::<Vec<u32>>());
         let mut offset: u32 = 0u32;
         // checking that the block before calling advance is empty
         assert!(block_segments.docs().is_empty());
@@ -713,44 +698,14 @@ mod tests {
         }
     }
 
-    #[test]
-    fn test_skip_right_at_new_block() {
-        let mut doc_ids = (0..128).collect::<Vec<u32>>();
-        doc_ids.push(129);
-        doc_ids.push(130);
-        {
-            let block_segments = build_block_postings(&doc_ids);
-            let mut docset = SegmentPostings::from_block_postings(block_segments, None);
-            assert_eq!(docset.skip_next(128), SkipResult::OverStep);
-            assert_eq!(docset.doc(), 129);
-            assert!(docset.advance());
-            assert_eq!(docset.doc(), 130);
-            assert!(!docset.advance());
-        }
-        {
-            let block_segments = build_block_postings(&doc_ids);
-            let mut docset = SegmentPostings::from_block_postings(block_segments, None);
-            assert_eq!(docset.skip_next(129), SkipResult::Reached);
-            assert_eq!(docset.doc(), 129);
-            assert!(docset.advance());
-            assert_eq!(docset.doc(), 130);
-            assert!(!docset.advance());
-        }
-        {
-            let block_segments = build_block_postings(&doc_ids);
-            let mut docset = SegmentPostings::from_block_postings(block_segments, None);
-            assert_eq!(docset.skip_next(131), SkipResult::End);
-        }
-    }
-
-    fn build_block_postings(docs: &[DocId]) -> BlockSegmentPostings {
-        let mut schema_builder = Schema::builder();
+    fn build_block_postings(docs: Vec<DocId>) -> BlockSegmentPostings {
+        let mut schema_builder = SchemaBuilder::default();
         let int_field = schema_builder.add_u64_field("id", INT_INDEXED);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
         let mut last_doc = 0u32;
-        for &doc in docs {
+        for doc in docs {
             for _ in last_doc..doc {
                 index_writer.add_document(doc!(int_field=>1u64));
             }
@@ -770,56 +725,39 @@ mod tests {
     #[test]
     fn test_block_segment_postings_skip() {
         for i in 0..4 {
-            let mut block_postings = build_block_postings(&[3]);
-            assert_eq!(
-                block_postings.skip_to(i),
-                BlockSegmentPostingsSkipResult::Success(0u32)
-            );
-            assert_eq!(
-                block_postings.skip_to(i),
-                BlockSegmentPostingsSkipResult::Terminated
-            );
+            let mut block_postings = build_block_postings(vec![3]);
+            assert_eq!(block_postings.skip_to(i), BlockSegmentPostingsSkipResult::Success(0u32));
+            assert_eq!(block_postings.skip_to(i), BlockSegmentPostingsSkipResult::Terminated);
         }
-        let mut block_postings = build_block_postings(&[3]);
-        assert_eq!(
-            block_postings.skip_to(4u32),
-            BlockSegmentPostingsSkipResult::Terminated
-        );
+        let mut block_postings = build_block_postings(vec![3]);
+        assert_eq!(block_postings.skip_to(4u32), BlockSegmentPostingsSkipResult::Terminated);
     }
 
+
     #[test]
     fn test_block_segment_postings_skip2() {
         let mut docs = vec![0];
         for i in 0..1300 {
             docs.push((i * i / 100) + i);
         }
-        let mut block_postings = build_block_postings(&docs[..]);
-        for i in vec![0, 424, 10000] {
-            assert_eq!(
-                block_postings.skip_to(i),
-                BlockSegmentPostingsSkipResult::Success(0u32)
-            );
+        let mut block_postings = build_block_postings(docs.clone());
+        for i in vec![0,  424, 10000] {
+            assert_eq!(block_postings.skip_to(i), BlockSegmentPostingsSkipResult::Success(0u32));
             let docs = block_postings.docs();
             assert!(docs[0] <= i);
             assert!(docs.last().cloned().unwrap_or(0u32) >= i);
         }
-        assert_eq!(
-            block_postings.skip_to(100_000),
-            BlockSegmentPostingsSkipResult::Terminated
-        );
-        assert_eq!(
-            block_postings.skip_to(101_000),
-            BlockSegmentPostingsSkipResult::Terminated
-        );
+        assert_eq!(block_postings.skip_to(100_000), BlockSegmentPostingsSkipResult::Terminated);
+        assert_eq!(block_postings.skip_to(101_000), BlockSegmentPostingsSkipResult::Terminated);
     }
 
     #[test]
     fn test_reset_block_segment_postings() {
-        let mut schema_builder = Schema::builder();
+        let mut schema_builder = SchemaBuilder::default();
         let int_field = schema_builder.add_u64_field("id", INT_INDEXED);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_with_num_threads(1, 40_000_000).unwrap();
         // create two postings list, one containg even number,
         // the other containing odd numbers.
         for i in 0..6 {

src/postings/serializer.rs

@@ -1,18 +1,18 @@
 use super::TermInfo;
-use common::{BinarySerializable, VInt};
+use common::{VInt, BinarySerializable};
 use common::{CompositeWrite, CountingWriter};
+use postings::compression::{VIntEncoder, BlockEncoder, COMPRESSION_BLOCK_SIZE};
 use core::Segment;
 use directory::WritePtr;
-use positions::PositionSerializer;
-use postings::compression::{BlockEncoder, VIntEncoder, COMPRESSION_BLOCK_SIZE};
-use postings::skip::SkipSerializer;
-use postings::USE_SKIP_INFO_LIMIT;
 use schema::Schema;
 use schema::{Field, FieldEntry, FieldType};
 use std::io::{self, Write};
 use termdict::{TermDictionaryBuilder, TermOrdinal};
 use DocId;
 use Result;
+use postings::USE_SKIP_INFO_LIMIT;
+use postings::skip::SkipSerializer;
+use positions::PositionSerializer;
 
 /// `PostingsSerializer` is in charge of serializing
 /// postings on disk, in the
@@ -55,7 +55,7 @@ pub struct InvertedIndexSerializer {
 
 impl InvertedIndexSerializer {
     /// Open a new `PostingsSerializer` for the given segment
-    fn create(
+    fn new(
         terms_write: CompositeWrite<WritePtr>,
         postings_write: CompositeWrite<WritePtr>,
         positions_write: CompositeWrite<WritePtr>,
@@ -74,7 +74,7 @@ impl InvertedIndexSerializer {
     /// Open a new `PostingsSerializer` for the given segment
     pub fn open(segment: &mut Segment) -> Result<InvertedIndexSerializer> {
         use SegmentComponent::{POSITIONS, POSITIONSSKIP, POSTINGS, TERMS};
-        InvertedIndexSerializer::create(
+        InvertedIndexSerializer::new(
             CompositeWrite::wrap(segment.open_write(TERMS)?),
             CompositeWrite::wrap(segment.open_write(POSTINGS)?),
             CompositeWrite::wrap(segment.open_write(POSITIONS)?),
@@ -99,12 +99,12 @@ impl InvertedIndexSerializer {
         let positions_write = self.positions_write.for_field(field);
         let positionsidx_write = self.positionsidx_write.for_field(field);
         let field_type: FieldType = (*field_entry.field_type()).clone();
-        FieldSerializer::create(
-            &field_type,
+        FieldSerializer::new(
+            field_type,
             term_dictionary_write,
             postings_write,
             positions_write,
-            positionsidx_write,
+            positionsidx_write
         )
     }
 
@@ -130,12 +130,12 @@ pub struct FieldSerializer<'a> {
 }
 
 impl<'a> FieldSerializer<'a> {
-    fn create(
-        field_type: &FieldType,
+    fn new(
+        field_type: FieldType,
         term_dictionary_write: &'a mut CountingWriter<WritePtr>,
         postings_write: &'a mut CountingWriter<WritePtr>,
         positions_write: &'a mut CountingWriter<WritePtr>,
-        positionsidx_write: &'a mut CountingWriter<WritePtr>,
+        positionsidx_write: &'a mut CountingWriter<WritePtr>
     ) -> io::Result<FieldSerializer<'a>> {
         let (term_freq_enabled, position_enabled): (bool, bool) = match field_type {
             FieldType::Str(ref text_options) => {
@@ -152,9 +152,8 @@ impl<'a> FieldSerializer<'a> {
             _ => (false, false),
         };
         let term_dictionary_builder =
-            TermDictionaryBuilder::create(term_dictionary_write, &field_type)?;
-        let postings_serializer =
-            PostingsSerializer::new(postings_write, term_freq_enabled, position_enabled);
+            TermDictionaryBuilder::new(term_dictionary_write, field_type)?;
+        let postings_serializer = PostingsSerializer::new(postings_write, term_freq_enabled, position_enabled);
         let positions_serializer_opt = if position_enabled {
             Some(PositionSerializer::new(positions_write, positionsidx_write))
         } else {
@@ -172,15 +171,14 @@ impl<'a> FieldSerializer<'a> {
     }
 
     fn current_term_info(&self) -> TermInfo {
-        let positions_idx = self
-            .positions_serializer_opt
+        let positions_idx = self.positions_serializer_opt
             .as_ref()
             .map(|positions_serializer| positions_serializer.positions_idx())
             .unwrap_or(0u64);
         TermInfo {
             doc_freq: 0,
             postings_offset: self.postings_serializer.addr(),
-            positions_idx,
+            positions_idx
         }
     }
 
@@ -255,7 +253,7 @@ impl<'a> FieldSerializer<'a> {
 struct Block {
     doc_ids: [DocId; COMPRESSION_BLOCK_SIZE],
     term_freqs: [u32; COMPRESSION_BLOCK_SIZE],
-    len: usize,
+    len: usize
 }
 
 impl Block {
@@ -263,7 +261,7 @@ impl Block {
         Block {
             doc_ids: [0u32; COMPRESSION_BLOCK_SIZE],
             term_freqs: [0u32; COMPRESSION_BLOCK_SIZE],
-            len: 0,
+            len: 0
         }
     }
 
@@ -314,12 +312,9 @@ pub struct PostingsSerializer<W: Write> {
     termfreq_sum_enabled: bool,
 }
 
+
 impl<W: Write> PostingsSerializer<W> {
-    pub fn new(
-        write: W,
-        termfreq_enabled: bool,
-        termfreq_sum_enabled: bool,
-    ) -> PostingsSerializer<W> {
+    pub fn new(write: W, termfreq_enabled: bool, termfreq_sum_enabled: bool) -> PostingsSerializer<W> {
         PostingsSerializer {
             output_write: CountingWriter::wrap(write),
 
@@ -342,16 +337,14 @@ impl<W: Write> PostingsSerializer<W> {
                 .block_encoder
                 .compress_block_sorted(&self.block.doc_ids(), self.last_doc_id_encoded);
             self.last_doc_id_encoded = self.block.last_doc();
-            self.skip_write
-                .write_doc(self.last_doc_id_encoded, num_bits);
+            self.skip_write.write_doc(self.last_doc_id_encoded, num_bits);
             // last el block 0, offset block 1,
             self.postings_write.extend(block_encoded);
         }
         if self.termfreq_enabled {
             // encode the term_freqs
-            let (num_bits, block_encoded): (u8, &[u8]) = self
-                .block_encoder
-                .compress_block_unsorted(&self.block.term_freqs());
+            let (num_bits, block_encoded): (u8, &[u8]) =
+                self.block_encoder.compress_block_unsorted(&self.block.term_freqs());
             self.postings_write.extend(block_encoded);
             self.skip_write.write_term_freq(num_bits);
             if self.termfreq_sum_enabled {
@@ -382,15 +375,13 @@ impl<W: Write> PostingsSerializer<W> {
             // In that case, the remaining part is encoded
             // using variable int encoding.
             {
-                let block_encoded = self
-                    .block_encoder
+                let block_encoded = self.block_encoder
                     .compress_vint_sorted(&self.block.doc_ids(), self.last_doc_id_encoded);
                 self.postings_write.write_all(block_encoded)?;
             }
             // ... Idem for term frequencies
             if self.termfreq_enabled {
-                let block_encoded = self
-                    .block_encoder
+                let block_encoded = self.block_encoder
                     .compress_vint_unsorted(self.block.term_freqs());
                 self.postings_write.write_all(block_encoded)?;
             }
@@ -401,6 +392,7 @@ impl<W: Write> PostingsSerializer<W> {
             VInt(skip_data.len() as u64).serialize(&mut self.output_write)?;
             self.output_write.write_all(skip_data)?;
             self.output_write.write_all(&self.postings_write[..])?;
+
         } else {
             self.output_write.write_all(&self.postings_write[..])?;
         }

src/postings/skip.rs

@@ -1,8 +1,8 @@
+use DocId;
 use common::BinarySerializable;
 use owned_read::OwnedRead;
 use postings::compression::COMPRESSION_BLOCK_SIZE;
 use schema::IndexRecordOption;
-use DocId;
 
 pub struct SkipSerializer {
     buffer: Vec<u8>,
@@ -18,11 +18,8 @@ impl SkipSerializer {
     }
 
     pub fn write_doc(&mut self, last_doc: DocId, doc_num_bits: u8) {
-        assert!(
-            last_doc > self.prev_doc,
-            "write_doc(...) called with non-increasing doc ids. \
-             Did you forget to call clear maybe?"
-        );
+        assert!(last_doc > self.prev_doc, "write_doc(...) called with non-increasing doc ids. \
+                Did you forget to call clear maybe?");
         let delta_doc = last_doc - self.prev_doc;
         self.prev_doc = last_doc;
         delta_doc.serialize(&mut self.buffer).unwrap();
@@ -33,10 +30,9 @@ impl SkipSerializer {
         self.buffer.push(tf_num_bits);
     }
 
+
     pub fn write_total_term_freq(&mut self, tf_sum: u32) {
-        tf_sum
-            .serialize(&mut self.buffer)
-            .expect("Should never fail");
+        tf_sum.serialize(&mut self.buffer).expect("Should never fail");
     }
 
     pub fn data(&self) -> &[u8] {
@@ -107,32 +103,33 @@ impl SkipReader {
         } else {
             let doc_delta = u32::deserialize(&mut self.owned_read).expect("Skip data corrupted");
             self.doc += doc_delta as DocId;
-            self.doc_num_bits = self.owned_read.get(0);
+            self.doc_num_bits =  self.owned_read.get(0);
             match self.skip_info {
                 IndexRecordOption::Basic => {
                     self.owned_read.advance(1);
                 }
-                IndexRecordOption::WithFreqs => {
+                IndexRecordOption::WithFreqs=> {
                     self.tf_num_bits = self.owned_read.get(1);
                     self.owned_read.advance(2);
                 }
                 IndexRecordOption::WithFreqsAndPositions => {
                     self.tf_num_bits = self.owned_read.get(1);
                     self.owned_read.advance(2);
-                    self.tf_sum =
-                        u32::deserialize(&mut self.owned_read).expect("Failed reading tf_sum");
+                    self.tf_sum = u32::deserialize(&mut self.owned_read)
+                        .expect("Failed reading tf_sum");
                 }
             }
             true
         }
+
     }
 }
 
 #[cfg(test)]
 mod tests {
 
-    use super::IndexRecordOption;
     use super::{SkipReader, SkipSerializer};
+    use super::IndexRecordOption;
     use owned_read::OwnedRead;
 
     #[test]
@@ -174,4 +171,4 @@ mod tests {
         assert_eq!(skip_reader.doc_num_bits(), 5u8);
         assert!(!skip_reader.advance());
     }
-}
+}

src/postings/stacker/expull.rs

@@ -1,37 +1,28 @@
 use super::{Addr, MemoryArena};
 
-use postings::stacker::memory_arena::load;
-use postings::stacker::memory_arena::store;
-use std::io;
+use common::is_power_of_2;
 use std::mem;
 
 const MAX_BLOCK_LEN: u32 = 1u32 << 15;
-const FIRST_BLOCK: usize = 16;
-const INLINED_BLOCK_LEN: usize = FIRST_BLOCK + mem::size_of::<Addr>();
 
-enum CapacityResult {
-    Available(u32),
-    NeedAlloc(u32),
-}
+const FIRST_BLOCK: u32 = 4u32;
 
-fn len_to_capacity(len: u32) -> CapacityResult {
+#[inline]
+pub fn jump_needed(len: u32) -> Option<usize> {
     match len {
-        0...15 => CapacityResult::Available(FIRST_BLOCK as u32 - len),
-        16...MAX_BLOCK_LEN => {
-            let cap = 1 << (32u32 - (len - 1u32).leading_zeros());
-            let available = cap - len;
-            if available == 0 {
-                CapacityResult::NeedAlloc(len)
+        0...3 => None,
+        4...MAX_BLOCK_LEN => {
+            if is_power_of_2(len as usize) {
+                Some(len as usize)
             } else {
-                CapacityResult::Available(available)
+                None
             }
         }
         n => {
-            let available = n % MAX_BLOCK_LEN;
-            if available == 0 {
-                CapacityResult::NeedAlloc(MAX_BLOCK_LEN)
+            if n % MAX_BLOCK_LEN == 0 {
+                Some(MAX_BLOCK_LEN as usize)
             } else {
-                CapacityResult::Available(MAX_BLOCK_LEN - available)
+                None
             }
         }
     }
@@ -61,119 +52,82 @@ fn len_to_capacity(len: u32) -> CapacityResult {
 #[derive(Debug, Clone, Copy)]
 pub struct ExpUnrolledLinkedList {
     len: u32,
+    head: Addr,
     tail: Addr,
-    inlined_data: [u8; INLINED_BLOCK_LEN as usize],
-}
-
-pub struct ExpUnrolledLinkedListWriter<'a> {
-    eull: &'a mut ExpUnrolledLinkedList,
-    heap: &'a mut MemoryArena,
-}
-
-fn ensure_capacity<'a>(
-    eull: &'a mut ExpUnrolledLinkedList,
-    heap: &'a mut MemoryArena,
-) -> &'a mut [u8] {
-    if eull.len <= FIRST_BLOCK as u32 {
-        // We are still hitting the inline block.
-        if eull.len < FIRST_BLOCK as u32 {
-            return &mut eull.inlined_data[eull.len as usize..FIRST_BLOCK];
-        }
-        // We need to allocate a new block!
-        let new_block_addr: Addr = heap.allocate_space(FIRST_BLOCK + mem::size_of::<Addr>());
-        store(&mut eull.inlined_data[FIRST_BLOCK..], new_block_addr);
-        eull.tail = new_block_addr;
-        return heap.slice_mut(eull.tail, FIRST_BLOCK);
-    }
-    let len = match len_to_capacity(eull.len) {
-        CapacityResult::NeedAlloc(new_block_len) => {
-            let new_block_addr: Addr =
-                heap.allocate_space(new_block_len as usize + mem::size_of::<Addr>());
-            heap.write_at(eull.tail, new_block_addr);
-            eull.tail = new_block_addr;
-            new_block_len
-        }
-        CapacityResult::Available(available) => available,
-    };
-    heap.slice_mut(eull.tail, len as usize)
-}
-
-impl<'a> ExpUnrolledLinkedListWriter<'a> {
-    pub fn extend_from_slice(&mut self, mut buf: &[u8]) {
-        if buf.is_empty() {
-            // we need to cut early, because `ensure_capacity`
-            // allocates if there is no capacity at all right now.
-            return;
-        }
-        while !buf.is_empty() {
-            let add_len: usize;
-            {
-                let output_buf = ensure_capacity(self.eull, self.heap);
-                add_len = buf.len().min(output_buf.len());
-                output_buf[..add_len].copy_from_slice(&buf[..add_len]);
-            }
-            self.eull.len += add_len as u32;
-            self.eull.tail = self.eull.tail.offset(add_len as u32);
-            buf = &buf[add_len..];
-        }
-    }
-}
-
-impl<'a> io::Write for ExpUnrolledLinkedListWriter<'a> {
-    fn write(&mut self, buf: &[u8]) -> io::Result<usize> {
-        // There is no use case to only write the capacity.
-        // This is not IO after all, so we write the whole
-        // buffer even if the contract of `.write` is looser.
-        self.extend_from_slice(buf);
-        Ok(buf.len())
-    }
-
-    fn write_all(&mut self, buf: &[u8]) -> io::Result<()> {
-        self.extend_from_slice(buf);
-        Ok(())
-    }
-
-    fn flush(&mut self) -> io::Result<()> {
-        Ok(())
-    }
 }
 
 impl ExpUnrolledLinkedList {
-    pub fn new() -> ExpUnrolledLinkedList {
+    pub fn new(heap: &mut MemoryArena) -> ExpUnrolledLinkedList {
+        let addr = heap.allocate_space((FIRST_BLOCK as usize) * mem::size_of::<u32>());
         ExpUnrolledLinkedList {
             len: 0u32,
-            tail: Addr::null_pointer(),
-            inlined_data: [0u8; INLINED_BLOCK_LEN as usize],
+            head: addr,
+            tail: addr,
         }
     }
 
-    #[inline(always)]
-    pub fn writer<'a>(&'a mut self, heap: &'a mut MemoryArena) -> ExpUnrolledLinkedListWriter<'a> {
-        ExpUnrolledLinkedListWriter { eull: self, heap }
+    pub fn iter<'a>(&self, heap: &'a MemoryArena) -> ExpUnrolledLinkedListIterator<'a> {
+        ExpUnrolledLinkedListIterator {
+            heap,
+            addr: self.head,
+            len: self.len,
+            consumed: 0,
+        }
     }
 
-    pub fn read_to_end(&self, heap: &MemoryArena, output: &mut Vec<u8>) {
-        let len = self.len as usize;
-        if len <= FIRST_BLOCK {
-            output.extend_from_slice(&self.inlined_data[..len]);
-            return;
+    /// Appends a new element to the current stack.
+    ///
+    /// If the current block end is reached, a new block is allocated.
+    pub fn push(&mut self, val: u32, heap: &mut MemoryArena) {
+        self.len += 1;
+        if let Some(new_block_len) = jump_needed(self.len) {
+            // We need to allocate another block.
+            // We also allocate an extra `u32` to store the pointer
+            // to the future next block.
+            let new_block_size: usize = (new_block_len + 1) * mem::size_of::<u32>();
+            let new_block_addr: Addr = heap.allocate_space(new_block_size);
+            unsafe {
+                // logic
+                heap.write(self.tail, new_block_addr)
+            };
+            self.tail = new_block_addr;
         }
-        output.extend_from_slice(&self.inlined_data[..FIRST_BLOCK]);
-        let mut cur = FIRST_BLOCK;
-        let mut addr = load(&self.inlined_data[FIRST_BLOCK..]);
-        loop {
-            let cap = match len_to_capacity(cur as u32) {
-                CapacityResult::Available(capacity) => capacity,
-                CapacityResult::NeedAlloc(capacity) => capacity,
-            } as usize;
-            let data = heap.slice(addr, cap);
-            if cur + cap >= len {
-                output.extend_from_slice(&data[..(len - cur)]);
-                return;
-            }
-            output.extend_from_slice(data);
-            cur += cap;
-            addr = heap.read(addr.offset(cap as u32));
+        unsafe {
+            // logic
+            heap.write(self.tail, val);
+            self.tail = self.tail.offset(mem::size_of::<u32>() as u32);
+        }
+    }
+}
+
+pub struct ExpUnrolledLinkedListIterator<'a> {
+    heap: &'a MemoryArena,
+    addr: Addr,
+    len: u32,
+    consumed: u32,
+}
+
+impl<'a> Iterator for ExpUnrolledLinkedListIterator<'a> {
+    type Item = u32;
+
+    fn next(&mut self) -> Option<u32> {
+        if self.consumed == self.len {
+            None
+        } else {
+            self.consumed += 1;
+            let addr: Addr = if jump_needed(self.consumed).is_some() {
+                unsafe {
+                    // logic
+                    self.heap.read(self.addr)
+                }
+            } else {
+                self.addr
+            };
+            self.addr = addr.offset(mem::size_of::<u32>() as u32);
+            Some(unsafe {
+                // logic
+                self.heap.read(addr)
+            })
         }
     }
 }
@@ -182,134 +136,46 @@ impl ExpUnrolledLinkedList {
 mod tests {
 
     use super::super::MemoryArena;
-    use super::len_to_capacity;
+    use super::jump_needed;
     use super::*;
-    use byteorder::{ByteOrder, LittleEndian, WriteBytesExt};
 
-    #[test]
     #[test]
     fn test_stack() {
         let mut heap = MemoryArena::new();
-        let mut stack = ExpUnrolledLinkedList::new();
-        stack.writer(&mut heap).extend_from_slice(&[1u8]);
-        stack.writer(&mut heap).extend_from_slice(&[2u8]);
-        stack.writer(&mut heap).extend_from_slice(&[3u8, 4u8]);
-        stack.writer(&mut heap).extend_from_slice(&[5u8]);
+        let mut stack = ExpUnrolledLinkedList::new(&mut heap);
+        stack.push(1u32, &mut heap);
+        stack.push(2u32, &mut heap);
+        stack.push(4u32, &mut heap);
+        stack.push(8u32, &mut heap);
         {
-            let mut buffer = Vec::new();
-            stack.read_to_end(&heap, &mut buffer);
-            assert_eq!(&buffer[..], &[1u8, 2u8, 3u8, 4u8, 5u8]);
+            let mut it = stack.iter(&heap);
+            assert_eq!(it.next().unwrap(), 1u32);
+            assert_eq!(it.next().unwrap(), 2u32);
+            assert_eq!(it.next().unwrap(), 4u32);
+            assert_eq!(it.next().unwrap(), 8u32);
+            assert!(it.next().is_none());
         }
     }
 
-    #[test]
-    fn test_stack_long() {
-        let mut heap = MemoryArena::new();
-        let mut stack = ExpUnrolledLinkedList::new();
-        let source: Vec<u32> = (0..100).collect();
-        for &el in &source {
-            assert!(stack
-                .writer(&mut heap)
-                .write_u32::<LittleEndian>(el)
-                .is_ok());
-        }
-        let mut buffer = Vec::new();
-        stack.read_to_end(&heap, &mut buffer);
-        let mut result = vec![];
-        let mut remaining = &buffer[..];
-        while !remaining.is_empty() {
-            result.push(LittleEndian::read_u32(&remaining[..4]));
-            remaining = &remaining[4..];
-        }
-        assert_eq!(&result[..], &source[..]);
-    }
-
-    #[test]
-    fn test_stack_interlaced() {
-        let mut heap = MemoryArena::new();
-        let mut stack = ExpUnrolledLinkedList::new();
-        let mut stack2 = ExpUnrolledLinkedList::new();
-
-        let mut vec1: Vec<u8> = vec![];
-        let mut vec2: Vec<u8> = vec![];
-
-        for i in 0..9 {
-            assert!(stack.writer(&mut heap).write_u32::<LittleEndian>(i).is_ok());
-            assert!(vec1.write_u32::<LittleEndian>(i).is_ok());
-            if i % 2 == 0 {
-                assert!(stack2
-                    .writer(&mut heap)
-                    .write_u32::<LittleEndian>(i)
-                    .is_ok());
-                assert!(vec2.write_u32::<LittleEndian>(i).is_ok());
-            }
-        }
-        let mut res1 = vec![];
-        let mut res2 = vec![];
-        stack.read_to_end(&heap, &mut res1);
-        stack2.read_to_end(&heap, &mut res2);
-        assert_eq!(&vec1[..], &res1[..]);
-        assert_eq!(&vec2[..], &res2[..]);
-    }
-
     #[test]
     fn test_jump_if_needed() {
-        let mut available = 16u32;
-        for i in 0..10_000_000 {
-            match len_to_capacity(i) {
-                CapacityResult::NeedAlloc(cap) => {
-                    assert_eq!(available, 0, "Failed len={}: Expected 0 got {}", i, cap);
-                    available = cap;
-                }
-                CapacityResult::Available(cap) => {
-                    assert_eq!(
-                        available, cap,
-                        "Failed len={}: Expected {} Got {}",
-                        i, available, cap
-                    );
-                }
-            }
-            available -= 1;
+        let mut block_len = 4u32;
+        let mut i = 0;
+        while i < 10_000_000 {
+            assert!(jump_needed(i + block_len - 1).is_none());
+            assert!(jump_needed(i + block_len + 1).is_none());
+            assert!(jump_needed(i + block_len).is_some());
+            let new_block_len = jump_needed(i + block_len).unwrap();
+            i += block_len;
+            block_len = new_block_len as u32;
         }
     }
-
-    #[test]
-    fn test_jump_if_needed_progression() {
-        let mut v = vec![];
-        for i in 0.. {
-            if v.len() >= 10 {
-                break;
-            }
-            match len_to_capacity(i) {
-                CapacityResult::NeedAlloc(cap) => {
-                    v.push((i, cap));
-                }
-                _ => {}
-            }
-        }
-        assert_eq!(
-            &v[..],
-            &[
-                (16, 16),
-                (32, 32),
-                (64, 64),
-                (128, 128),
-                (256, 256),
-                (512, 512),
-                (1024, 1024),
-                (2048, 2048),
-                (4096, 4096),
-                (8192, 8192)
-            ]
-        );
-    }
 }
 
 #[cfg(all(test, feature = "unstable"))]
 mod bench {
-    use super::super::MemoryArena;
     use super::ExpUnrolledLinkedList;
-    use byteorder::{NativeEndian, WriteBytesExt};
+    use tantivy_memory_arena::MemoryArena;
     use test::Bencher;
 
     const NUM_STACK: usize = 10_000;
@@ -333,19 +199,20 @@ mod bench {
 
     #[bench]
     fn bench_push_stack(bench: &mut Bencher) {
+        let heap = MemoryArena::new();
         bench.iter(|| {
-            let mut heap = MemoryArena::new();
             let mut stacks = Vec::with_capacity(100);
             for _ in 0..NUM_STACK {
-                let mut stack = ExpUnrolledLinkedList::new();
+                let (_, stack) = heap.allocate_object::<ExpUnrolledLinkedList>();
                 stacks.push(stack);
             }
             for s in 0..NUM_STACK {
                 for i in 0u32..STACK_SIZE {
                     let t = s * 392017 % NUM_STACK;
-                    let _ = stacks[t].writer(&mut heap).write_u32::<NativeEndian>(i);
+                    stacks[t].push(i, &heap);
                 }
             }
+            heap.clear();
         });
     }
 }

src/postings/stacker/memory_arena.rs

@@ -37,7 +37,7 @@ const PAGE_SIZE: usize = 1 << NUM_BITS_PAGE_ADDR; // pages are 1 MB large
 /// page of memory.
 ///
 /// The last 20 bits are an address within this page of memory.
-#[derive(Copy, Clone, Debug)]
+#[derive(Clone, Copy, Debug)]
 pub struct Addr(u32);
 
 impl Addr {
@@ -47,7 +47,7 @@ impl Addr {
     }
 
     /// Returns the `Addr` object for `addr + offset`
-    pub fn offset(self, offset: u32) -> Addr {
+    pub fn offset(&self, offset: u32) -> Addr {
         Addr(self.0.wrapping_add(offset))
     }
 
@@ -55,30 +55,46 @@ impl Addr {
         Addr((page_id << NUM_BITS_PAGE_ADDR | local_addr) as u32)
     }
 
-    fn page_id(self) -> usize {
+    fn page_id(&self) -> usize {
         (self.0 as usize) >> NUM_BITS_PAGE_ADDR
     }
 
-    fn page_local_addr(self) -> usize {
+    fn page_local_addr(&self) -> usize {
         (self.0 as usize) & (PAGE_SIZE - 1)
     }
 
     /// Returns true if and only if the `Addr` is null.
-    pub fn is_null(self) -> bool {
+    pub fn is_null(&self) -> bool {
         self.0 == u32::max_value()
     }
 }
 
-pub fn store<Item: Copy + 'static>(dest: &mut [u8], val: Item) {
-    assert_eq!(dest.len(), std::mem::size_of::<Item>());
-    unsafe {
-        ptr::write_unaligned(dest.as_mut_ptr() as *mut Item, val);
-    }
+/// Trait required for an object to be `storable`.
+///
+/// # Warning
+///
+/// Most of the time you should not implement this trait,
+/// and only use the `MemoryArena` with object implementing `Copy`.
+///
+/// `ArenaStorable` is used in `tantivy` to force
+/// a `Copy` object and a `slice` of data to be stored contiguously.
+pub trait ArenaStorable {
+    fn num_bytes(&self) -> usize;
+    unsafe fn write_into(self, arena: &mut MemoryArena, addr: Addr);
 }
 
-pub fn load<Item: Copy + 'static>(data: &[u8]) -> Item {
-    assert_eq!(data.len(), std::mem::size_of::<Item>());
-    unsafe { ptr::read_unaligned(data.as_ptr() as *const Item) }
+impl<V> ArenaStorable for V
+where
+    V: Copy,
+{
+    fn num_bytes(&self) -> usize {
+        mem::size_of::<V>()
+    }
+
+    unsafe fn write_into(self, arena: &mut MemoryArena, addr: Addr) {
+        let dst_ptr = arena.get_mut_ptr(addr) as *mut V;
+        ptr::write_unaligned(dst_ptr, self);
+    }
 }
 
 /// The `MemoryArena`
@@ -110,9 +126,47 @@ impl MemoryArena {
         self.pages.len() * PAGE_SIZE
     }
 
-    pub fn write_at<Item: Copy + 'static>(&mut self, addr: Addr, val: Item) {
-        let dest = self.slice_mut(addr, std::mem::size_of::<Item>());
-        store(dest, val);
+    /// Writes a slice at the given address, assuming the
+    /// memory was allocated beforehands.
+    ///
+    ///  # Panics
+    ///
+    /// May panic or corrupt the heap if he space was not
+    /// properly allocated beforehands.
+    pub fn write_bytes<B: AsRef<[u8]>>(&mut self, addr: Addr, data: B) {
+        let bytes = data.as_ref();
+        self.pages[addr.page_id()]
+            .get_mut_slice(addr.page_local_addr(), bytes.len())
+            .copy_from_slice(bytes);
+    }
+
+    /// Returns the `len` bytes starting at `addr`
+    ///
+    /// # Panics
+    ///
+    /// Panics if the memory has not been allocated beforehands.
+    pub fn read_slice(&self, addr: Addr, len: usize) -> &[u8] {
+        self.pages[addr.page_id()].get_slice(addr.page_local_addr(), len)
+    }
+
+    unsafe fn get_mut_ptr(&mut self, addr: Addr) -> *mut u8 {
+        self.pages[addr.page_id()].get_mut_ptr(addr.page_local_addr())
+    }
+
+    /// Stores an item's data in the heap
+    ///
+    /// It allocates the `Item` beforehands.
+    pub fn store<Item: ArenaStorable>(&mut self, val: Item) -> Addr {
+        let num_bytes = val.num_bytes();
+        let addr = self.allocate_space(num_bytes);
+        unsafe {
+            self.write(addr, val);
+        };
+        addr
+    }
+
+    pub unsafe fn write<Item: ArenaStorable>(&mut self, addr: Addr, val: Item) {
+        val.write_into(self, addr)
     }
 
     /// Read an item in the heap at the given `address`.
@@ -120,21 +174,9 @@ impl MemoryArena {
     /// # Panics
     ///
     /// If the address is erroneous
-    pub fn read<Item: Copy + 'static>(&self, addr: Addr) -> Item {
-        load(self.slice(addr, mem::size_of::<Item>()))
-    }
-
-    pub fn slice(&self, addr: Addr, len: usize) -> &[u8] {
-        self.pages[addr.page_id()].slice(addr.page_local_addr(), len)
-    }
-
-    pub fn slice_from(&self, addr: Addr) -> &[u8] {
-        self.pages[addr.page_id()].slice_from(addr.page_local_addr())
-    }
-
-    #[inline(always)]
-    pub fn slice_mut(&mut self, addr: Addr, len: usize) -> &mut [u8] {
-        self.pages[addr.page_id()].slice_mut(addr.page_local_addr(), len)
+    pub unsafe fn read<Item: Copy>(&self, addr: Addr) -> Item {
+        let ptr = self.pages[addr.page_id()].get_ptr(addr.page_local_addr());
+        ptr::read_unaligned(ptr as *const Item)
     }
 
     /// Allocates `len` bytes and returns the allocated address.
@@ -155,10 +197,14 @@ struct Page {
 
 impl Page {
     fn new(page_id: usize) -> Page {
+        let mut data: Vec<u8> = Vec::with_capacity(PAGE_SIZE);
+        unsafe {
+            data.set_len(PAGE_SIZE);
+        } // avoid initializing page
         Page {
             page_id,
             len: 0,
-            data: vec![0u8; PAGE_SIZE].into_boxed_slice(),
+            data: data.into_boxed_slice(),
         }
     }
 
@@ -167,18 +213,14 @@ impl Page {
         len + self.len <= PAGE_SIZE
     }
 
-    fn slice(&self, local_addr: usize, len: usize) -> &[u8] {
-        &self.slice_from(local_addr)[..len]
-    }
-
-    fn slice_from(&self, local_addr: usize) -> &[u8] {
-        &self.data[local_addr..]
-    }
-
-    fn slice_mut(&mut self, local_addr: usize, len: usize) -> &mut [u8] {
+    fn get_mut_slice(&mut self, local_addr: usize, len: usize) -> &mut [u8] {
         &mut self.data[local_addr..][..len]
     }
 
+    fn get_slice(&self, local_addr: usize, len: usize) -> &[u8] {
+        &self.data[local_addr..][..len]
+    }
+
     fn allocate_space(&mut self, len: usize) -> Option<Addr> {
         if self.is_available(len) {
             let addr = Addr::new(self.page_id, self.len);
@@ -188,6 +230,16 @@ impl Page {
             None
         }
     }
+
+    #[inline(always)]
+    pub(crate) unsafe fn get_ptr(&self, addr: usize) -> *const u8 {
+        self.data.as_ptr().offset(addr as isize)
+    }
+
+    #[inline(always)]
+    pub(crate) unsafe fn get_mut_ptr(&mut self, addr: usize) -> *mut u8 {
+        self.data.as_mut_ptr().offset(addr as isize)
+    }
 }
 
 #[cfg(test)]
@@ -202,13 +254,13 @@ mod tests {
         let b = b"happy tax payer";
 
         let addr_a = arena.allocate_space(a.len());
-        arena.slice_mut(addr_a, a.len()).copy_from_slice(a);
+        arena.write_bytes(addr_a, a);
 
         let addr_b = arena.allocate_space(b.len());
-        arena.slice_mut(addr_b, b.len()).copy_from_slice(b);
+        arena.write_bytes(addr_b, b);
 
-        assert_eq!(arena.slice(addr_a, a.len()), a);
-        assert_eq!(arena.slice(addr_b, b.len()), b);
+        assert_eq!(arena.read_slice(addr_a, a.len()), a);
+        assert_eq!(arena.read_slice(addr_b, b.len()), b);
     }
 
     #[derive(Clone, Copy, Debug, Eq, PartialEq)]
@@ -231,15 +283,9 @@ mod tests {
             b: 221,
             c: 12,
         };
-
-        let num_bytes = std::mem::size_of::<MyTest>();
-        let addr_a = arena.allocate_space(num_bytes);
-        arena.write_at(addr_a, a);
-
-        let addr_b = arena.allocate_space(num_bytes);
-        arena.write_at(addr_b, b);
-
-        assert_eq!(arena.read::<MyTest>(addr_a), a);
-        assert_eq!(arena.read::<MyTest>(addr_b), b);
+        let addr_a = arena.store(a);
+        let addr_b = arena.store(b);
+        assert_eq!(unsafe { arena.read::<MyTest>(addr_a) }, a);
+        assert_eq!(unsafe { arena.read::<MyTest>(addr_b) }, b);
     }
 }

src/postings/stacker/mod.rs

@@ -1,7 +1,9 @@
 mod expull;
 mod memory_arena;
+mod murmurhash2;
 mod term_hashmap;
 
 pub use self::expull::ExpUnrolledLinkedList;
-pub use self::memory_arena::{Addr, MemoryArena};
+pub use self::memory_arena::{Addr, ArenaStorable, MemoryArena};
+use self::murmurhash2::murmurhash2;
 pub use self::term_hashmap::{compute_table_size, TermHashMap};

src/postings/stacker/murmurhash2.rs

@@ -0,0 +1,86 @@
+use std::ptr;
+const SEED: u32 = 3_242_157_231u32;
+const M: u32 = 0x5bd1_e995;
+
+#[inline(always)]
+pub fn murmurhash2(key: &[u8]) -> u32 {
+    let mut key_ptr: *const u32 = key.as_ptr() as *const u32;
+    let len = key.len() as u32;
+    let mut h: u32 = SEED ^ len;
+
+    let num_blocks = len >> 2;
+    for _ in 0..num_blocks {
+        let mut k: u32 = unsafe { ptr::read_unaligned(key_ptr) }; // ok because of num_blocks definition
+        k = k.wrapping_mul(M);
+        k ^= k >> 24;
+        k = k.wrapping_mul(M);
+        h = h.wrapping_mul(M);
+        h ^= k;
+        key_ptr = key_ptr.wrapping_offset(1);
+    }
+
+    // Handle the last few bytes of the input array
+    let remaining: &[u8] = &key[key.len() & !3..];
+    match remaining.len() {
+        3 => {
+            h ^= u32::from(remaining[2]) << 16;
+            h ^= u32::from(remaining[1]) << 8;
+            h ^= u32::from(remaining[0]);
+            h = h.wrapping_mul(M);
+        }
+        2 => {
+            h ^= u32::from(remaining[1]) << 8;
+            h ^= u32::from(remaining[0]);
+            h = h.wrapping_mul(M);
+        }
+        1 => {
+            h ^= u32::from(remaining[0]);
+            h = h.wrapping_mul(M);
+        }
+        _ => {}
+    }
+    h ^= h >> 13;
+    h = h.wrapping_mul(M);
+    h ^ (h >> 15)
+}
+
+#[cfg(test)]
+mod test {
+
+    use super::murmurhash2;
+    use std::collections::HashSet;
+
+    #[test]
+    fn test_murmur() {
+        let s1 = "abcdef";
+        let s2 = "abcdeg";
+        for i in 0..5 {
+            assert_eq!(
+                murmurhash2(&s1[i..5].as_bytes()),
+                murmurhash2(&s2[i..5].as_bytes())
+            );
+        }
+    }
+
+    #[test]
+    fn test_murmur_against_reference_impl() {
+        assert_eq!(murmurhash2("".as_bytes()), 3632506080);
+        assert_eq!(murmurhash2("a".as_bytes()), 455683869);
+        assert_eq!(murmurhash2("ab".as_bytes()), 2448092234);
+        assert_eq!(murmurhash2("abc".as_bytes()), 2066295634);
+        assert_eq!(murmurhash2("abcd".as_bytes()), 2588571162);
+        assert_eq!(murmurhash2("abcde".as_bytes()), 2988696942);
+        assert_eq!(murmurhash2("abcdefghijklmnop".as_bytes()), 2350868870);
+    }
+
+    #[test]
+    fn test_murmur_collisions() {
+        let mut set: HashSet<u32> = HashSet::default();
+        for i in 0..10_000 {
+            let s = format!("hash{}", i);
+            let hash = murmurhash2(s.as_bytes());
+            set.insert(hash);
+        }
+        assert_eq!(set.len(), 10_000);
+    }
+}

src/postings/stacker/term_hashmap.rs

@@ -1,16 +1,37 @@
-extern crate murmurhash32;
-
-use self::murmurhash32::murmurhash2;
-
-use super::{Addr, MemoryArena};
-use byteorder::{ByteOrder, NativeEndian};
-use postings::stacker::memory_arena::store;
+use super::murmurhash2;
+use super::{Addr, ArenaStorable, MemoryArena};
 use std::iter;
 use std::mem;
 use std::slice;
 
 pub type BucketId = usize;
 
+struct KeyBytesValue<'a, V> {
+    key: &'a [u8],
+    value: V,
+}
+
+impl<'a, V> KeyBytesValue<'a, V> {
+    fn new(key: &'a [u8], value: V) -> KeyBytesValue<'a, V> {
+        KeyBytesValue { key, value }
+    }
+}
+
+impl<'a, V> ArenaStorable for KeyBytesValue<'a, V>
+where
+    V: ArenaStorable,
+{
+    fn num_bytes(&self) -> usize {
+        0u16.num_bytes() + self.key.len() + self.value.num_bytes()
+    }
+
+    unsafe fn write_into(self, arena: &mut MemoryArena, addr: Addr) {
+        arena.write(addr, self.key.len() as u16);
+        arena.write_bytes(addr.offset(2), self.key);
+        arena.write(addr.offset(2 + self.key.len() as u32), self.value);
+    }
+}
+
 /// Returns the actual memory size in bytes
 /// required to create a table of size $2^num_bits$.
 pub fn compute_table_size(num_bits: usize) -> usize {
@@ -40,7 +61,7 @@ impl Default for KeyValue {
 }
 
 impl KeyValue {
-    fn is_empty(self) -> bool {
+    fn is_empty(&self) -> bool {
         self.key_value_addr.is_null()
     }
 }
@@ -90,7 +111,8 @@ impl<'a> Iterator for Iter<'a> {
     fn next(&mut self) -> Option<Self::Item> {
         self.inner.next().cloned().map(move |bucket: usize| {
             let kv = self.hashmap.table[bucket];
-            let (key, offset): (&'a [u8], Addr) = self.hashmap.get_key_value(kv.key_value_addr);
+            let (key, offset): (&'a [u8], Addr) =
+                unsafe { self.hashmap.get_key_value(kv.key_value_addr) };
             (key, offset, bucket as BucketId)
         })
     }
@@ -121,22 +143,12 @@ impl TermHashMap {
         self.table.len() < self.occupied.len() * 3
     }
 
-    #[inline(always)]
-    fn get_key_value(&self, addr: Addr) -> (&[u8], Addr) {
-        let data = self.heap.slice_from(addr);
-        let key_bytes_len = NativeEndian::read_u16(data) as usize;
-        let key_bytes: &[u8] = &data[2..][..key_bytes_len];
-        (key_bytes, addr.offset(2u32 + key_bytes_len as u32))
-    }
-
-    #[inline(always)]
-    fn get_value_addr_if_key_match(&self, target_key: &[u8], addr: Addr) -> Option<Addr> {
-        let (stored_key, value_addr) = self.get_key_value(addr);
-        if stored_key == target_key {
-            Some(value_addr)
-        } else {
-            None
-        }
+    unsafe fn get_key_value(&self, addr: Addr) -> (&[u8], Addr) {
+        let key_bytes_len = self.heap.read::<u16>(addr) as usize;
+        let key_addr = addr.offset(2u32);
+        let key_bytes: &[u8] = self.heap.read_slice(key_addr, key_bytes_len);
+        let val_addr: Addr = key_addr.offset(key_bytes.len() as u32);
+        (key_bytes, val_addr)
     }
 
     pub fn set_bucket(&mut self, hash: u32, key_value_addr: Addr, bucket: usize) {
@@ -187,39 +199,36 @@ impl TermHashMap {
     pub fn mutate_or_create<S, V, TMutator>(&mut self, key: S, mut updater: TMutator) -> BucketId
     where
         S: AsRef<[u8]>,
-        V: Copy + 'static,
+        V: Copy,
         TMutator: FnMut(Option<V>) -> V,
     {
         if self.is_saturated() {
             self.resize();
         }
         let key_bytes: &[u8] = key.as_ref();
-        let hash = murmurhash2(key.as_ref());
+        let hash = murmurhash2::murmurhash2(key.as_ref());
         let mut probe = self.probe(hash);
         loop {
             let bucket = probe.next_probe();
             let kv: KeyValue = self.table[bucket];
             if kv.is_empty() {
                 let val = updater(None);
-                let num_bytes =
-                    std::mem::size_of::<u16>() + key_bytes.len() + std::mem::size_of::<V>();
-                let key_addr = self.heap.allocate_space(num_bytes);
-                {
-                    let data = self.heap.slice_mut(key_addr, num_bytes);
-                    NativeEndian::write_u16(data, key_bytes.len() as u16);
-                    let stop = 2 + key_bytes.len();
-                    data[2..stop].copy_from_slice(key_bytes);
-                    store(&mut data[stop..], val);
-                }
+                let key_addr = self.heap.store(KeyBytesValue::new(key_bytes, val));
                 self.set_bucket(hash, key_addr, bucket);
                 return bucket as BucketId;
             } else if kv.hash == hash {
-                if let Some(val_addr) =
-                    self.get_value_addr_if_key_match(key_bytes, kv.key_value_addr)
-                {
-                    let v = self.heap.read(val_addr);
-                    let new_v = updater(Some(v));
-                    self.heap.write_at(val_addr, new_v);
+                let (key_matches, val_addr) = {
+                    let (stored_key, val_addr): (&[u8], Addr) =
+                        unsafe { self.get_key_value(kv.key_value_addr) };
+                    (stored_key == key_bytes, val_addr)
+                };
+                if key_matches {
+                    unsafe {
+                        // logic
+                        let v = self.heap.read(val_addr);
+                        let new_v = updater(Some(v));
+                        self.heap.write(val_addr, new_v);
+                    };
                     return bucket as BucketId;
                 }
             }
@@ -227,6 +236,24 @@ impl TermHashMap {
     }
 }
 
+#[cfg(all(test, feature = "unstable"))]
+mod bench {
+    use super::murmurhash2::murmurhash2;
+    use test::Bencher;
+
+    #[bench]
+    fn bench_murmurhash2(b: &mut Bencher) {
+        let keys: [&'static str; 3] = ["wer qwe qwe qwe ", "werbq weqweqwe2 ", "weraq weqweqwe3 "];
+        b.iter(|| {
+            let mut s = 0;
+            for &key in &keys {
+                s ^= murmurhash2(key.as_bytes());
+            }
+            s
+        });
+    }
+}
+
 #[cfg(test)]
 mod tests {
 
@@ -258,7 +285,10 @@ mod tests {
         let mut vanilla_hash_map = HashMap::new();
         let mut iter_values = hash_map.iter();
         while let Some((key, addr, _)) = iter_values.next() {
-            let val: u32 = hash_map.heap.read(addr);
+            let val: u32 = unsafe {
+                // test
+                hash_map.heap.read(addr)
+            };
             vanilla_hash_map.insert(key.to_owned(), val);
         }
         assert_eq!(vanilla_hash_map.len(), 2);