TermInfo contain the end_offset of the postings.

We slice the ReadOnlySource tightly.

.gitignore

@@ -12,3 +12,4 @@ cpp/simdcomp/bitpackingbenchmark
 *.bk
 .idea
 trace.dat
+cargo-timing*

CHANGELOG.md

@@ -1,14 +1,12 @@
-Tantivy 0.13.2
+Tantivy 0.14.0
-===================
+=========================
-Bugfix. Acquiring a facet reader on a segment that does not contain any 
+- Remove dependency to atomicwrites #833 .Implemented by @pmasurel upon suggestion and research from @asafigan). 
-doc with this facet returns `None`. (#896)
 
 Tantivy 0.13.1
-======================
+===================
 Made `Query` and `Collector` `Send + Sync`.
 Updated misc dependency versions.
 
-
 Tantivy 0.13.0
 ======================
 Tantivy 0.13 introduce a change in the index format that will require

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tantivy"
-version = "0.13.2"
+version = "0.14.0-dev"
 authors = ["Paul Masurel <paul.masurel@gmail.com>"]
 license = "MIT"
 categories = ["database-implementations", "data-structures"]
@@ -22,8 +22,7 @@ tantivy-fst = "0.3"
 memmap = {version = "0.7", optional=true}
 lz4 = {version="1", optional=true}
 snap = "1"
-atomicwrites = {version="0.2", optional=true}
+tempfile = {version="3", optional=true}
-tempfile = "3"
 log = "0.4"
 serde = {version="1", features=["derive"]}
 serde_json = "1"
@@ -35,10 +34,10 @@ uuid = { version = "0.8", features = ["v4", "serde"] }
 crossbeam = "0.7"
 futures = {version = "0.3",  features=["thread-pool"] }
 owning_ref = "0.4"
+tantivy-query-grammar = { version="0.14.0-dev", path="./query-grammar" }
 stable_deref_trait = "1"
 rust-stemmers = "1"
 downcast-rs = "1"
-tantivy-query-grammar = { version="0.13", path="./query-grammar" }
 bitpacking = {version="0.8", default-features = false, features=["bitpacker4x"]}
 census = "0.4"
 fnv = "1"
@@ -75,12 +74,11 @@ overflow-checks = true
 
 [features]
 default = ["mmap"]
-mmap = ["atomicwrites", "fs2", "memmap", "notify"]
+mmap = ["fs2", "tempfile", "memmap", "notify"]
 lz4-compression = ["lz4"]
 failpoints = ["fail/failpoints"]
 unstable = [] # useful for benches.
 wasm-bindgen = ["uuid/wasm-bindgen"]
-scoref64 = [] # scores are f64 instead of f32. was introduced to debug blockwand.
 
 [workspace]
 members = ["query-grammar"]

README.md

@@ -34,11 +34,6 @@ Tantivy is, in fact, strongly inspired by Lucene's design.
 The following [benchmark](https://tantivy-search.github.io/bench/) break downs 
 performance for different type of queries / collection.
 
-
-In general, Tantivy tends to be 
-- slower than Lucene on union with a Top-K due to Block-WAND optimization.
-- faster than Lucene on intersection and phrase queries. 
-
 Your mileage WILL vary depending on the nature of queries and their load.
 
 # Features

doc/src/index-format.md

@@ -0,0 +1,50 @@
+
+# Managed files 
++----------+-----------+-------------------+
+| content  |  footer   |  footer_len: u32  |
++----------+-----------+-------------------+
+
+# Term Dictionary (Composite File)
+
++---------+---------------------------+------------------------+
+|   fst   | term_info_store           |   footer_len:  u64     |
++---------+---------------------------+------------------------+
+
+During a merge the term info store need to fit in memory.
+It has a cost of n bytes per term.
+
+# term_info_store
++-------------------+---------------------------+------------------------+
+|  len_block_meta   |  block_meta               |  term_infos            |
++-------------------+---------------------------+------------------------+
+
+# inverted_index
++------------------------+---------------------------+------------------------+
+|  total_num_tokens: u64 |  posting_lists.. |  term_infos            |
++------------------------+---------------------------+------------------------+
+
+# postings lists
++------------------------+---------------------------+------------------------+
+|    
++
+
+# composite file
++----------------+-----+----------------+----------------------+----------------+
+| field file 1   | ... | field field n  |composite file footer | footer len: u32| 
++----------------+-----+----------------+----------------------+----------------+
+
+# composite file footer
+
++-----------------+---------------------------------------+
+|num fields: vint | (file_addr, offset_delta: vint) []... |
++-----------------+---------------------------------------+
+
+# FileAddr 
++--------------+--------------+
+|  field: u32  |  idx: VInt   |
++--------------+--------------+
+
+# Posting lists 
++-----------------------------------------+
+| skip_reader 
++-----------------------------------------+

examples/basic_search.rs

@@ -112,18 +112,6 @@ fn main() -> tantivy::Result<()> {
             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"
-    ));
-
     // Multivalued field just need to be repeated.
     index_writer.add_document(doc!(
     title => "Frankenstein",

examples/faceted_search_with_tweaked_score.rs

@@ -56,7 +56,7 @@ fn main() -> tantivy::Result<()> {
         );
         let top_docs_by_custom_score =
             TopDocs::with_limit(2).tweak_score(move |segment_reader: &SegmentReader| {
-                let ingredient_reader = segment_reader.facet_reader(ingredient).unwrap();
+                let mut ingredient_reader = segment_reader.facet_reader(ingredient).unwrap();
                 let facet_dict = ingredient_reader.facet_dict();
 
                 let query_ords: HashSet<u64> = facets

query-grammar/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "tantivy-query-grammar"
-version = "0.13.0"
+version = "0.14.0-dev"
 authors = ["Paul Masurel <paul.masurel@gmail.com>"]
 license = "MIT"
 categories = ["database-implementations", "data-structures"]

query-grammar/src/occur.rs

@@ -52,7 +52,7 @@ mod test {
     use crate::Occur;
 
     #[test]
-    fn test_Occur_compose() {
+    fn test_occur_compose() {
         assert_eq!(Occur::compose(Occur::Should, Occur::Should), Occur::Should);
         assert_eq!(Occur::compose(Occur::Should, Occur::Must), Occur::Must);
         assert_eq!(

query-grammar/src/query_grammar.rs

@@ -9,8 +9,10 @@ use combine::{
 
 fn field<'a>() -> impl Parser<&'a str, Output = String> {
     (
-        letter(),
+        (letter().or(char('_'))),
-        many(satisfy(|c: char| c.is_alphanumeric() || c == '_')),
+        many(satisfy(|c: char| {
+            c.is_alphanumeric() || c == '_' || c == '-'
+        })),
     )
         .skip(char(':'))
         .map(|(s1, s2): (char, String)| format!("{}{}", s1, s2))
@@ -279,6 +281,8 @@ pub fn parse_to_ast<'a>() -> impl Parser<&'a str, Output = UserInputAST> {
 #[cfg(test)]
 mod test {
 
+    type TestParseResult = Result<(), StringStreamError>;
+
     use super::*;
     use combine::parser::Parser;
 
@@ -296,9 +300,10 @@ mod test {
     }
 
     #[test]
-    fn test_occur_symbol() {
+    fn test_occur_symbol() -> TestParseResult {
-        assert_eq!(super::occur_symbol().parse("-"), Ok((Occur::MustNot, "")));
+        assert_eq!(super::occur_symbol().parse("-")?, (Occur::MustNot, ""));
-        assert_eq!(super::occur_symbol().parse("+"), Ok((Occur::Must, "")));
+        assert_eq!(super::occur_symbol().parse("+")?, (Occur::Must, ""));
+        Ok(())
     }
 
     #[test]
@@ -410,6 +415,25 @@ mod test {
         assert_eq!(format!("{:?}", ast), "\"abc\"");
     }
 
+    #[test]
+    fn test_field_name() -> TestParseResult {
+        assert_eq!(
+            super::field().parse("my-field-name:a")?,
+            ("my-field-name".to_string(), "a")
+        );
+        assert_eq!(
+            super::field().parse("my_field_name:a")?,
+            ("my_field_name".to_string(), "a")
+        );
+        assert!(super::field().parse(":a").is_err());
+        assert!(super::field().parse("-my_field:a").is_err());
+        assert_eq!(
+            super::field().parse("_my_field:a")?,
+            ("_my_field".to_string(), "a")
+        );
+        Ok(())
+    }
+
     #[test]
     fn test_range_parser() {
         // testing the range() parser separately

src/collector/facet_collector.rs

@@ -472,7 +472,7 @@ mod tests {
         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_for_tests().unwrap();
         let num_facets: usize = 3 * 4 * 5;
         let facets: Vec<Facet> = (0..num_facets)
             .map(|mut n| {
@@ -531,7 +531,7 @@ mod tests {
         let facet_field = schema_builder.add_facet_field("facets");
         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_for_tests().unwrap();
         index_writer.add_document(doc!(
             facet_field => Facet::from_text(&"/subjects/A/a"),
             facet_field => Facet::from_text(&"/subjects/B/a"),
@@ -550,12 +550,12 @@ mod tests {
     }
 
     #[test]
-    fn test_doc_search_by_facet() {
+    fn test_doc_search_by_facet() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         let facet_field = schema_builder.add_facet_field("facet");
         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_for_tests()?;
         index_writer.add_document(doc!(
             facet_field => Facet::from_text(&"/A/A"),
         ));
@@ -568,8 +568,8 @@ mod tests {
         index_writer.add_document(doc!(
             facet_field => Facet::from_text(&"/D/C/A"),
         ));
-        index_writer.commit().unwrap();
+        index_writer.commit()?;
-        let reader = index.reader().unwrap();
+        let reader = index.reader()?;
         let searcher = reader.searcher();
         assert_eq!(searcher.num_docs(), 4);
 
@@ -586,17 +586,17 @@ mod tests {
         assert_eq!(count_facet("/A/C"), 1);
         assert_eq!(count_facet("/A/C/A"), 1);
         assert_eq!(count_facet("/C/A"), 0);
+
+        let query_parser = QueryParser::for_index(&index, vec![]);
         {
-            let query_parser = QueryParser::for_index(&index, vec![]);
+            let query = query_parser.parse_query("facet:/A/B")?;
-            {
+            assert_eq!(1, searcher.search(&query, &Count).unwrap());
-                let query = query_parser.parse_query("facet:/A/B").unwrap();
-                assert_eq!(1, searcher.search(&query, &Count).unwrap());
-            }
-            {
-                let query = query_parser.parse_query("facet:/A").unwrap();
-                assert_eq!(3, searcher.search(&query, &Count).unwrap());
-            }
         }
+        {
+            let query = query_parser.parse_query("facet:/A")?;
+            assert_eq!(3, searcher.search(&query, &Count)?);
+        }
+        Ok(())
     }
 
     #[test]
@@ -631,7 +631,7 @@ mod tests {
             .collect();
         docs[..].shuffle(&mut thread_rng());
 
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests().unwrap();
         for doc in docs {
             index_writer.add_document(doc);
         }
@@ -684,7 +684,7 @@ mod bench {
         // 40425 docs
         docs[..].shuffle(&mut thread_rng());
 
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests().unwrap();
         for doc in docs {
             index_writer.add_document(doc);
         }

src/collector/int_facet_collector.rs

@@ -89,7 +89,7 @@ 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_for_tests().unwrap();
             {
                 for i in 0u64..10u64 {
                     index_writer.add_document(doc!(

src/collector/multi_collector.rs

@@ -259,7 +259,7 @@ mod tests {
 
         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_for_tests().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"));

src/collector/top_score_collector.rs

@@ -38,7 +38,7 @@ use std::fmt;
 /// 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, 10_000_000).unwrap();
 /// 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"));
@@ -123,7 +123,7 @@ impl TopDocs {
     /// 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, 10_000_000).unwrap();
     /// 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"));
@@ -163,7 +163,7 @@ impl TopDocs {
     /// #   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)?;
+    /// #   let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
     /// #   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));
@@ -264,7 +264,7 @@ impl TopDocs {
     /// fn create_index() -> tantivy::Result<Index> {
     ///   let schema = create_schema();
     ///   let index = Index::create_in_ram(schema);
-    ///   let mut index_writer = index.writer_with_num_threads(1, 3_000_000)?;
+    ///   let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
     ///   let product_name = index.schema().get_field("product_name").unwrap();
     ///   let popularity: Field = index.schema().get_field("popularity").unwrap();
     ///   index_writer.add_document(doc!(product_name => "The Diary of Muadib", popularity => 1u64));
@@ -371,7 +371,7 @@ impl TopDocs {
     /// # fn main() -> tantivy::Result<()> {
     /// #   let schema = create_schema();
     /// #   let index = Index::create_in_ram(schema);
-    /// #   let mut index_writer = index.writer_with_num_threads(1, 3_000_000)?;
+    /// #   let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
     /// #   let product_name = index.schema().get_field("product_name").unwrap();
     /// #   
     /// let popularity: Field = index.schema().get_field("popularity").unwrap();
@@ -561,7 +561,7 @@ 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();
+            let mut index_writer = index.writer_with_num_threads(1, 10_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"));
@@ -821,7 +821,7 @@ mod tests {
     ) -> (Index, Box<dyn Query>) {
         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, 10_000_000).unwrap();
         doc_adder(&mut index_writer);
         index_writer.commit().unwrap();
         let query_parser = QueryParser::for_index(&index, vec![query_field]);

src/core/index.rs

@@ -300,6 +300,15 @@ impl Index {
         )
     }
 
+    /// Helper to create an index writer for tests.
+    ///
+    /// That index writer only simply has a single thread and a heap of 5 MB.
+    /// Using a single thread gives us a deterministic allocation of DocId.
+    #[cfg(test)]
+    pub fn writer_for_tests(&self) -> crate::Result<IndexWriter> {
+        self.writer_with_num_threads(1, 10_000_000)
+    }
+
     /// Creates a multithreaded writer
     ///
     /// Tantivy will automatically define the number of threads to use.
@@ -502,7 +511,7 @@ mod tests {
             let schema = throw_away_schema();
             let field = schema.get_field("num_likes").unwrap();
             let mut index = Index::create_from_tempdir(schema).unwrap();
-            let mut writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut writer = index.writer_for_tests().unwrap();
             writer.commit().unwrap();
             let reader = index
                 .reader_builder()
@@ -539,23 +548,33 @@ mod tests {
             test_index_on_commit_reload_policy_aux(field, &write_index, &reader);
         }
     }
-
     fn test_index_on_commit_reload_policy_aux(field: Field, index: &Index, reader: &IndexReader) {
         let mut reader_index = reader.index();
         let (sender, receiver) = crossbeam::channel::unbounded();
         let _watch_handle = reader_index.directory_mut().watch(Box::new(move || {
             let _ = sender.send(());
         }));
-        let mut writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut writer = index.writer_for_tests().unwrap();
         assert_eq!(reader.searcher().num_docs(), 0);
         writer.add_document(doc!(field=>1u64));
         writer.commit().unwrap();
-        assert!(receiver.recv().is_ok());
+        // We need a loop here because it is possible for notify to send more than
-        assert_eq!(reader.searcher().num_docs(), 1);
+        // one modify event. It was observed on CI on MacOS.
+        loop {
+            assert!(receiver.recv().is_ok());
+            if reader.searcher().num_docs() == 1 {
+                break;
+            }
+        }
         writer.add_document(doc!(field=>2u64));
         writer.commit().unwrap();
-        assert!(receiver.recv().is_ok());
+        // ... Same as above
-        assert_eq!(reader.searcher().num_docs(), 2);
+        loop {
+            assert!(receiver.recv().is_ok());
+            if reader.searcher().num_docs() == 2 {
+                break;
+            }
+        }
     }
 
     // This test will not pass on windows, because windows

src/core/index_meta.rs

@@ -116,6 +116,7 @@ impl SegmentMeta {
             SegmentComponent::FASTFIELDS => ".fast".to_string(),
             SegmentComponent::FIELDNORMS => ".fieldnorm".to_string(),
             SegmentComponent::DELETE => format!(".{}.del", self.delete_opstamp().unwrap_or(0)),
+            SegmentComponent::FIELDSTATS => ".fieldstats".to_string(),
         });
         PathBuf::from(path)
     }

src/core/inverted_index_reader.rs

@@ -1,9 +1,7 @@
-use crate::common::BinarySerializable;
 use crate::directory::ReadOnlySource;
 use crate::positions::PositionReader;
 use crate::postings::TermInfo;
 use crate::postings::{BlockSegmentPostings, SegmentPostings};
-use crate::schema::FieldType;
 use crate::schema::IndexRecordOption;
 use crate::schema::Term;
 use crate::termdict::TermDictionary;
@@ -37,14 +35,12 @@ impl InvertedIndexReader {
         postings_source: ReadOnlySource,
         positions_source: ReadOnlySource,
         positions_idx_source: ReadOnlySource,
+        total_num_tokens: u64,
         record_option: IndexRecordOption,
     ) -> InvertedIndexReader {
-        let total_num_tokens_data = postings_source.slice(0, 8);
-        let mut total_num_tokens_cursor = total_num_tokens_data.as_slice();
-        let total_num_tokens = u64::deserialize(&mut total_num_tokens_cursor).unwrap_or(0u64);
         InvertedIndexReader {
             termdict,
-            postings_source: postings_source.slice_from(8),
+            postings_source,
             positions_source,
             positions_idx_source,
             record_option,
@@ -54,10 +50,7 @@ impl InvertedIndexReader {
 
     /// Creates an empty `InvertedIndexReader` object, which
     /// contains no terms at all.
-    pub fn empty(field_type: &FieldType) -> InvertedIndexReader {
+    pub fn empty(record_option: IndexRecordOption) -> InvertedIndexReader {
-        let record_option = field_type
-            .get_index_record_option()
-            .unwrap_or(IndexRecordOption::Basic);
         InvertedIndexReader {
             termdict: TermDictionary::empty(),
             postings_source: ReadOnlySource::empty(),
@@ -93,7 +86,7 @@ impl InvertedIndexReader {
         term_info: &TermInfo,
         block_postings: &mut BlockSegmentPostings,
     ) {
-        let offset = term_info.postings_offset as usize;
+        let offset = term_info.postings_start_offset as usize;
         let end_source = self.postings_source.len();
         let postings_slice = self.postings_source.slice(offset, end_source);
         block_postings.reset(term_info.doc_freq, postings_slice);
@@ -121,8 +114,10 @@ impl InvertedIndexReader {
         term_info: &TermInfo,
         requested_option: IndexRecordOption,
     ) -> BlockSegmentPostings {
-        let offset = term_info.postings_offset as usize;
+        let postings_data = self.postings_source.slice(
-        let postings_data = self.postings_source.slice_from(offset);
+            term_info.postings_start_offset as usize,
+            term_info.postings_end_offset as usize,
+        );
         BlockSegmentPostings::from_data(
             term_info.doc_freq,
             postings_data,

src/core/segment_component.rs

@@ -24,14 +24,17 @@ pub enum SegmentComponent {
     /// Accessing a document from the store is relatively slow, as it
     /// requires to decompress the entire block it belongs to.
     STORE,
+
     /// Bitset describing which document of the segment is deleted.
     DELETE,
+
+    FIELDSTATS,
 }
 
 impl SegmentComponent {
     /// Iterates through the components.
     pub fn iterator() -> slice::Iter<'static, SegmentComponent> {
-        static SEGMENT_COMPONENTS: [SegmentComponent; 8] = [
+        static SEGMENT_COMPONENTS: [SegmentComponent; 9] = [
             SegmentComponent::POSTINGS,
             SegmentComponent::POSITIONS,
             SegmentComponent::POSITIONSSKIP,
@@ -40,6 +43,7 @@ impl SegmentComponent {
             SegmentComponent::TERMS,
             SegmentComponent::STORE,
             SegmentComponent::DELETE,
+            SegmentComponent::FIELDSTATS,
         ];
         SEGMENT_COMPONENTS.iter()
     }

src/core/segment_reader.rs

@@ -1,4 +1,3 @@
-use crate::common::CompositeFile;
 use crate::common::HasLen;
 use crate::core::InvertedIndexReader;
 use crate::core::Segment;
@@ -9,13 +8,14 @@ use crate::fastfield::DeleteBitSet;
 use crate::fastfield::FacetReader;
 use crate::fastfield::FastFieldReaders;
 use crate::fieldnorm::{FieldNormReader, FieldNormReaders};
-use crate::schema::Field;
 use crate::schema::FieldType;
 use crate::schema::Schema;
+use crate::schema::{Field, IndexRecordOption};
 use crate::space_usage::SegmentSpaceUsage;
 use crate::store::StoreReader;
 use crate::termdict::TermDictionary;
 use crate::DocId;
+use crate::{common::CompositeFile, postings::FieldStats};
 use fail::fail_point;
 use std::collections::HashMap;
 use std::fmt;
@@ -49,6 +49,7 @@ pub struct SegmentReader {
     positions_idx_composite: CompositeFile,
     fast_fields_readers: Arc<FastFieldReaders>,
     fieldnorm_readers: FieldNormReaders,
+    field_stats: FieldStats,
 
     store_source: ReadOnlySource,
     delete_bitset_opt: Option<DeleteBitSet>,
@@ -112,10 +113,8 @@ impl SegmentReader {
             return None;
         }
         let term_ords_reader = self.fast_fields().u64s(field)?;
-        let termdict = self.termdict_composite
+        let termdict_source = self.termdict_composite.open_read(field)?;
-            .open_read(field)
+        let termdict = TermDictionary::from_source(&termdict_source);
-            .map(|source| TermDictionary::from_source(&source))
-            .unwrap_or_else(TermDictionary::empty);
         let facet_reader = FacetReader::new(term_ords_reader, termdict);
         Some(facet_reader)
     }
@@ -127,17 +126,15 @@ impl SegmentReader {
     ///
     /// They are simply stored as a fast field, serialized in
     /// the `.fieldnorm` file of the segment.
-    pub fn get_fieldnorms_reader(&self, field: Field) -> FieldNormReader {
+    pub fn get_fieldnorms_reader(&self, field: Field) -> crate::Result<FieldNormReader> {
-        if let Some(fieldnorm_reader) = self.fieldnorm_readers.get_field(field) {
+        self.fieldnorm_readers.get_field(field).ok_or_else(|| {
-            fieldnorm_reader
-        } else {
             let field_name = self.schema.get_field_name(field);
             let err_msg = format!(
                 "Field norm not found for field {:?}. Was it market as indexed during indexing.",
                 field_name
             );
-            panic!(err_msg);
+            crate::TantivyError::SchemaError(err_msg)
-        }
+        })
     }
 
     /// Accessor to the segment's `StoreReader`.
@@ -183,6 +180,9 @@ impl SegmentReader {
         let fieldnorm_data = segment.open_read(SegmentComponent::FIELDNORMS)?;
         let fieldnorm_readers = FieldNormReaders::open(fieldnorm_data)?;
 
+        let field_stats_data = segment.open_read(SegmentComponent::FIELDSTATS)?;
+        let field_stats = FieldStats::from_source(field_stats_data.as_slice())?;
+
         let delete_bitset_opt = if segment.meta().has_deletes() {
             let delete_data = segment.open_read(SegmentComponent::DELETE)?;
             Some(DeleteBitSet::open(delete_data))
@@ -198,6 +198,7 @@ impl SegmentReader {
             postings_composite,
             fast_fields_readers: fast_field_readers,
             fieldnorm_readers,
+            field_stats,
             segment_id: segment.id(),
             store_source,
             delete_bitset_opt,
@@ -214,6 +215,11 @@ impl SegmentReader {
     /// The field reader is in charge of iterating through the
     /// term dictionary associated to a specific field,
     /// and opening the posting list associated to any term.
+    ///
+    /// If the field is marked as index, a warn is logged and an empty `InvertedIndexReader`
+    /// is returned.
+    /// Similarly if the field is marked as indexed but no term has been indexed for the given
+    /// index. an empty `InvertedIndexReader` is returned (but no warning is logged).
     pub fn inverted_index(&self, field: Field) -> Arc<InvertedIndexReader> {
         if let Some(inv_idx_reader) = self
             .inv_idx_reader_cache
@@ -228,21 +234,21 @@ impl SegmentReader {
         let record_option_opt = field_type.get_index_record_option();
 
         if record_option_opt.is_none() {
-            panic!("Field {:?} does not seem indexed.", field_entry.name());
+            warn!("Field {:?} does not seem indexed.", field_entry.name());
         }
 
-        let record_option = record_option_opt.unwrap();
-
         let postings_source_opt = self.postings_composite.open_read(field);
 
-        if postings_source_opt.is_none() {
+        if postings_source_opt.is_none() || record_option_opt.is_none() {
             // no documents in the segment contained this field.
             // As a result, no data is associated to the inverted index.
             //
             // Returns an empty inverted index.
-            return Arc::new(InvertedIndexReader::empty(field_type));
+            let record_option = record_option_opt.unwrap_or(IndexRecordOption::Basic);
+            return Arc::new(InvertedIndexReader::empty(record_option));
         }
 
+        let record_option = record_option_opt.unwrap();
         let postings_source = postings_source_opt.unwrap();
 
         let termdict_source = self.termdict_composite.open_read(field).expect(
@@ -259,11 +265,17 @@ impl SegmentReader {
             .open_read(field)
             .expect("Index corrupted. Failed to open field positions in composite file.");
 
+        let total_num_tokens = self
+            .field_stats
+            .get(field)
+            .map(|field_stat| field_stat.num_tokens())
+            .unwrap_or(0u64);
         let inv_idx_reader = Arc::new(InvertedIndexReader::new(
             TermDictionary::from_source(&termdict_source),
             postings_source,
             positions_source,
             positions_idx_source,
+            total_num_tokens,
             record_option,
         ));
 
@@ -341,7 +353,7 @@ mod test {
         let name = schema.get_field("name").unwrap();
 
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             index_writer.add_document(doc!(name => "tantivy"));
             index_writer.add_document(doc!(name => "horse"));
             index_writer.add_document(doc!(name => "jockey"));

src/directory/footer.rs

@@ -271,7 +271,11 @@ mod tests {
         let mut vec = Vec::new();
         let footer_proxy = FooterProxy::new(&mut vec);
         assert!(footer_proxy.terminate().is_ok());
-        assert_eq!(vec.len(), 167);
+        if crate::store::COMPRESSION == "lz4" {
+            assert_eq!(vec.len(), 158);
+        } else {
+            assert_eq!(vec.len(), 167);
+        }
         let footer = Footer::deserialize(&mut &vec[..]).unwrap();
         assert!(matches!(
            footer.versioned_footer,

src/directory/managed_directory.rs

@@ -1,4 +1,4 @@
-use crate::core::MANAGED_FILEPATH;
+use crate::core::{MANAGED_FILEPATH, META_FILEPATH};
 use crate::directory::error::{DeleteError, IOError, LockError, OpenReadError, OpenWriteError};
 use crate::directory::footer::{Footer, FooterProxy};
 use crate::directory::DirectoryLock;
@@ -246,13 +246,15 @@ impl ManagedDirectory {
     /// List files for which checksum does not match content
     pub fn list_damaged(&self) -> result::Result<HashSet<PathBuf>, OpenReadError> {
         let mut hashset = HashSet::new();
-        let managed_paths = self
+        let mut managed_paths = self
             .meta_informations
             .read()
             .expect("Managed directory rlock poisoned in list damaged.")
             .managed_paths
             .clone();
 
+        managed_paths.remove(*META_FILEPATH);
+
         for path in managed_paths.into_iter() {
             if !self.validate_checksum(&path)? {
                 hashset.insert(path);

src/directory/mmap_directory.rs

@@ -487,11 +487,13 @@ impl Directory for MmapDirectory {
         }
     }
 
-    fn atomic_write(&mut self, path: &Path, data: &[u8]) -> io::Result<()> {
+    fn atomic_write(&mut self, path: &Path, content: &[u8]) -> io::Result<()> {
         debug!("Atomic Write {:?}", path);
+        let mut tempfile = tempfile::Builder::new().tempfile_in(&self.inner.root_path)?;
+        tempfile.write_all(content)?;
+        tempfile.flush()?;
         let full_path = self.resolve_path(path);
-        let meta_file = atomicwrites::AtomicFile::new(full_path, atomicwrites::AllowOverwrite);
+        tempfile.into_temp_path().persist(full_path)?;
-        meta_file.write(|f| f.write_all(data))?;
         Ok(())
     }
 
@@ -652,7 +654,7 @@ mod tests {
         {
             let index = Index::create(mmap_directory.clone(), schema).unwrap();
 
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             let mut log_merge_policy = LogMergePolicy::default();
             log_merge_policy.set_min_merge_size(3);
             index_writer.set_merge_policy(Box::new(log_merge_policy));

src/directory/tests.rs

@@ -211,19 +211,19 @@ fn test_watch(directory: &mut dyn Directory) {
         .unwrap();
 
     for i in 0..10 {
-        assert_eq!(i, counter.load(SeqCst));
+        assert!(i <= counter.load(SeqCst));
         assert!(directory
             .atomic_write(Path::new("meta.json"), b"random_test_data_2")
             .is_ok());
         assert_eq!(receiver.recv_timeout(Duration::from_millis(500)), Ok(i));
-        assert_eq!(i + 1, counter.load(SeqCst));
+        assert!(i + 1 <= counter.load(SeqCst)); // notify can trigger more than once.
     }
     mem::drop(watch_handle);
     assert!(directory
         .atomic_write(Path::new("meta.json"), b"random_test_data")
         .is_ok());
     assert!(receiver.recv_timeout(Duration::from_millis(500)).is_ok());
-    assert_eq!(10, counter.load(SeqCst));
+    assert!(10 <= counter.load(SeqCst));
 }
 
 fn test_lock_non_blocking(directory: &mut dyn Directory) {

src/fastfield/bytes/mod.rs

@@ -15,7 +15,7 @@ mod tests {
         let field = schema_builder.add_bytes_field("bytesfield");
         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_for_tests().unwrap();
         index_writer.add_document(doc!(field=>vec![0u8, 1, 2, 3]));
         index_writer.add_document(doc!(field=>vec![]));
         index_writer.add_document(doc!(field=>vec![255u8]));

src/fastfield/facet_reader.rs

@@ -73,52 +73,7 @@ impl FacetReader {
     }
 
     /// Return the list of facet ordinals associated to a document.
-    pub fn facet_ords(&self, doc: DocId, output: &mut Vec<u64>) {
+    pub fn facet_ords(&mut self, doc: DocId, output: &mut Vec<u64>) {
         self.term_ords.get_vals(doc, output);
     }
 }
-
-#[cfg(test)]
-mod tests {
-    use crate::{Document, schema::{Facet, SchemaBuilder}};
-    use crate::Index;
-
-    #[test]
-    fn test_facet_not_populated_for_all_docs() -> crate::Result<()> {
-        let mut schema_builder = SchemaBuilder::default();
-        let facet_field = schema_builder.add_facet_field("facet");
-        let schema = schema_builder.build();
-        let index = Index::create_in_ram(schema);
-        let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
-        index_writer.add_document(doc!(facet_field=>Facet::from_text("/a/b")));
-        index_writer.add_document(Document::default());
-        index_writer.commit()?;
-        let searcher = index.reader()?.searcher();
-        let facet_reader = searcher.segment_reader(0u32).facet_reader(facet_field).unwrap();
-        let mut facet_ords = Vec::new();
-        facet_reader.facet_ords(0u32, &mut facet_ords);
-        assert_eq!(&facet_ords, &[2u64]);
-        facet_reader.facet_ords(1u32, &mut facet_ords);
-        assert!(facet_ords.is_empty());
-        Ok(())
-    }
-    #[test]
-    fn test_facet_not_populated_for_any_docs() -> crate::Result<()> {
-        let mut schema_builder = SchemaBuilder::default();
-        let facet_field = schema_builder.add_facet_field("facet");
-        let schema = schema_builder.build();
-        let index = Index::create_in_ram(schema);
-        let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
-        index_writer.add_document(Document::default());
-        index_writer.add_document(Document::default());
-        index_writer.commit()?;
-        let searcher = index.reader()?.searcher();
-        let facet_reader = searcher.segment_reader(0u32).facet_reader(facet_field).unwrap();
-        let mut facet_ords = Vec::new();
-        facet_reader.facet_ords(0u32, &mut facet_ords);
-        assert!(facet_ords.is_empty());
-        facet_reader.facet_ords(1u32, &mut facet_ords);
-        assert!(facet_ords.is_empty());
-        Ok(())
-    }
-}

src/fastfield/mod.rs

@@ -474,7 +474,7 @@ mod tests {
         let date_field = schema_builder.add_date_field("date", 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();
+        let mut index_writer = index.writer_for_tests().unwrap();
         index_writer.set_merge_policy(Box::new(NoMergePolicy));
         index_writer.add_document(doc!(date_field =>crate::chrono::prelude::Utc::now()));
         index_writer.commit().unwrap();
@@ -511,7 +511,7 @@ mod tests {
         );
         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_for_tests().unwrap();
         index_writer.set_merge_policy(Box::new(NoMergePolicy));
         index_writer.add_document(doc!(
             date_field => crate::DateTime::from_u64(1i64.to_u64()),

src/fastfield/multivalued/mod.rs

@@ -25,7 +25,7 @@ mod tests {
         );
         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_for_tests().unwrap();
         index_writer.add_document(doc!(field=>1u64, field=>3u64));
         index_writer.add_document(doc!());
         index_writer.add_document(doc!(field=>4u64));
@@ -64,7 +64,7 @@ mod tests {
             schema_builder.add_i64_field("time_stamp_i", IntOptions::default().set_stored());
         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_for_tests().unwrap();
         let first_time_stamp = chrono::Utc::now();
         index_writer.add_document(
             doc!(date_field=>first_time_stamp, date_field=>first_time_stamp, time_i=>1i64),
@@ -186,7 +186,7 @@ mod tests {
         );
         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_for_tests().unwrap();
         index_writer.add_document(doc!(field=> 1i64, field => 3i64));
         index_writer.add_document(doc!());
         index_writer.add_document(doc!(field=> -4i64));
@@ -221,7 +221,7 @@ mod tests {
         let field = schema_builder.add_facet_field("facetfield");
         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_for_tests().unwrap();
         for i in 0..100_000 {
             index_writer.add_document(doc!(field=> Facet::from(format!("/lang/{}", i).as_str())));
         }

src/fastfield/multivalued/reader.rs

@@ -74,7 +74,7 @@ mod tests {
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         let mut index_writer = index
-            .writer_with_num_threads(1, 30_000_000)
+            .writer_for_tests()
             .expect("Failed to create index writer.");
         index_writer.add_document(doc!(
             facet_field => Facet::from("/category/cat2"),

src/fastfield/writer.rs

@@ -126,7 +126,6 @@ impl FastFieldsWriter {
         for field_writer in &self.single_value_writers {
             field_writer.serialize(serializer)?;
         }
-        
         for field_writer in &self.multi_values_writers {
             let field = field_writer.field();
             field_writer.serialize(serializer, mapping.get(&field))?;

src/indexer/index_writer.rs

@@ -800,7 +800,7 @@ mod tests {
         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 index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let index_writer = index.writer_for_tests().unwrap();
         let operations = vec![
             UserOperation::Add(doc!(text_field=>"a")),
             UserOperation::Add(doc!(text_field=>"b")),
@@ -815,7 +815,7 @@ mod tests {
         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();
+        let mut index_writer = index.writer_for_tests().unwrap();
         index_writer.add_document(doc!(text_field => "hello1"));
         index_writer.add_document(doc!(text_field => "hello2"));
         assert!(index_writer.commit().is_ok());
@@ -864,7 +864,7 @@ mod tests {
             .reload_policy(ReloadPolicy::Manual)
             .try_into()
             .unwrap();
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests().unwrap();
         let a_term = Term::from_field_text(text_field, "a");
         let b_term = Term::from_field_text(text_field, "b");
         let operations = vec![
@@ -926,8 +926,8 @@ mod tests {
     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();
+        let _index_writer = index.writer_for_tests().unwrap();
-        match index.writer_with_num_threads(1, 3_000_000) {
+        match index.writer_for_tests() {
             Err(err) => {
                 let err_msg = err.to_string();
                 assert!(err_msg.contains("already an `IndexWriter`"));
@@ -1261,7 +1261,7 @@ mod tests {
         let idfield = schema_builder.add_text_field("id", STRING);
         schema_builder.add_text_field("optfield", STRING);
         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_for_tests().unwrap();
         index_writer.add_document(doc!(idfield=>"myid"));
         let commit = index_writer.commit();
         assert!(commit.is_ok());

src/indexer/merger.rs

@@ -25,14 +25,14 @@ use std::cmp;
 use std::collections::HashMap;
 use std::sync::Arc;
 
-fn compute_total_num_tokens(readers: &[SegmentReader], field: Field) -> u64 {
+fn compute_total_num_tokens(readers: &[SegmentReader], field: Field) -> crate::Result<u64> {
     let mut total_tokens = 0u64;
     let mut count: [usize; 256] = [0; 256];
     for reader in readers {
         if reader.has_deletes() {
             // if there are deletes, then we use an approximation
             // using the fieldnorm
-            let fieldnorms_reader = reader.get_fieldnorms_reader(field);
+            let fieldnorms_reader = reader.get_fieldnorms_reader(field)?;
             for doc in reader.doc_ids_alive() {
                 let fieldnorm_id = fieldnorms_reader.fieldnorm_id(doc);
                 count[fieldnorm_id as usize] += 1;
@@ -41,7 +41,7 @@ fn compute_total_num_tokens(readers: &[SegmentReader], field: Field) -> u64 {
             total_tokens += reader.inverted_index(field).total_num_tokens();
         }
     }
-    total_tokens
+    Ok(total_tokens
         + count
             .iter()
             .cloned()
@@ -49,7 +49,7 @@ fn compute_total_num_tokens(readers: &[SegmentReader], field: Field) -> u64 {
             .map(|(fieldnorm_ord, count)| {
                 count as u64 * u64::from(FieldNormReader::id_to_fieldnorm(fieldnorm_ord as u8))
             })
-            .sum::<u64>()
+            .sum::<u64>())
 }
 
 pub struct IndexMerger {
@@ -175,7 +175,7 @@ impl IndexMerger {
         for field in fields {
             fieldnorms_data.clear();
             for reader in &self.readers {
-                let fieldnorms_reader = reader.get_fieldnorms_reader(field);
+                let fieldnorms_reader = reader.get_fieldnorms_reader(field)?;
                 for doc_id in reader.doc_ids_alive() {
                     let fieldnorm_id = fieldnorms_reader.fieldnorm_id(doc_id);
                     fieldnorms_data.push(fieldnorm_id);
@@ -541,7 +541,7 @@ impl IndexMerger {
         // The total number of tokens will only be exact when there has been no deletes.
         //
         // Otherwise, we approximate by removing deleted documents proportionally.
-        let total_num_tokens: u64 = compute_total_num_tokens(&self.readers, indexed_field);
+        let total_num_tokens: u64 = compute_total_num_tokens(&self.readers, indexed_field)?;
 
         // Create the total list of doc ids
         // by stacking the doc ids from the different segment.
@@ -751,7 +751,7 @@ mod tests {
         };
 
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             {
                 // writing the segment
                 {
@@ -803,7 +803,7 @@ mod tests {
             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_for_tests().unwrap();
             block_on(index_writer.merge(&segment_ids)).expect("Merging failed");
             index_writer.wait_merging_threads().unwrap();
         }
@@ -904,7 +904,7 @@ 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_for_tests().unwrap();
         let reader = index.reader().unwrap();
         let search_term = |searcher: &Searcher, term: Term| {
             let collector = FastFieldTestCollector::for_field(score_field);
@@ -1211,7 +1211,7 @@ mod tests {
         let index = Index::create_in_ram(schema_builder.build());
         let reader = index.reader().unwrap();
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             let index_doc = |index_writer: &mut IndexWriter, doc_facets: &[&str]| {
                 let mut doc = Document::default();
                 for facet in doc_facets {
@@ -1276,7 +1276,7 @@ mod tests {
             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_for_tests().unwrap();
             block_on(index_writer.merge(&segment_ids)).expect("Merging failed");
             index_writer.wait_merging_threads().unwrap();
             reader.reload().unwrap();
@@ -1295,7 +1295,7 @@ 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_for_tests().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);
@@ -1320,7 +1320,7 @@ mod tests {
         let mut schema_builder = schema::Schema::builder();
         let int_field = schema_builder.add_u64_field("intvals", INDEXED);
         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_for_tests().unwrap();
         index_writer.add_document(doc!(int_field => 1u64));
         index_writer.commit().expect("commit failed");
         index_writer.add_document(doc!(int_field => 1u64));
@@ -1349,7 +1349,7 @@ mod tests {
         let index = Index::create_in_ram(schema_builder.build());
         let reader = index.reader().unwrap();
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             let mut doc = Document::default();
             doc.add_u64(int_field, 1);
             index_writer.add_document(doc.clone());
@@ -1388,7 +1388,7 @@ 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_for_tests().unwrap();
             let index_doc = |index_writer: &mut IndexWriter, int_vals: &[u64]| {
                 let mut doc = Document::default();
                 for &val in int_vals {
@@ -1462,7 +1462,7 @@ mod tests {
             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_for_tests().unwrap();
             assert!(block_on(index_writer.merge(&segment_ids)).is_ok());
             assert!(index_writer.wait_merging_threads().is_ok());
         }
@@ -1516,7 +1516,7 @@ mod tests {
 
         let index = Index::create_in_ram(builder.build());
 
-        let mut writer = index.writer_with_num_threads(1, 3_000_000)?;
+        let mut writer = index.writer_for_tests()?;
 
         // Make sure we'll attempt to merge every created segment
         let mut policy = crate::indexer::LogMergePolicy::default();
@@ -1548,7 +1548,7 @@ mod tests {
         let mut builder = schema::SchemaBuilder::new();
         let text = builder.add_text_field("text", TEXT);
         let index = Index::create_in_ram(builder.build());
-        let mut writer = index.writer_with_num_threads(1, 3_000_000)?;
+        let mut writer = index.writer_for_tests()?;
         let happy_term = Term::from_field_text(text, "happy");
         let term_query = TermQuery::new(happy_term, IndexRecordOption::WithFreqs);
         for _ in 0..62 {

src/indexer/mod.rs

@@ -29,8 +29,9 @@ pub use self::segment_writer::SegmentWriter;
 /// Alias for the default merge policy, which is the `LogMergePolicy`.
 pub type DefaultMergePolicy = LogMergePolicy;
 
+#[cfg(feature = "mmap")]
 #[cfg(test)]
-mod tests {
+mod tests_mmap {
     use crate::schema::{self, Schema};
     use crate::{Index, Term};
 
@@ -39,7 +40,7 @@ mod tests {
         let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", schema::TEXT);
         let index = Index::create_from_tempdir(schema_builder.build()).unwrap();
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests().unwrap();
         // there must be one deleted document in the segment
         index_writer.add_document(doc!(text_field=>"b"));
         index_writer.delete_term(Term::from_field_text(text_field, "b"));

src/indexer/segment_updater.rs

@@ -555,7 +555,7 @@ 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();
+        let mut index_writer = index.writer_for_tests().unwrap();
         index_writer.set_merge_policy(Box::new(MergeWheneverPossible));
 
         {
@@ -608,7 +608,7 @@ 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();
+        let mut index_writer = index.writer_for_tests().unwrap();
 
         {
             for _ in 0..100 {
@@ -679,7 +679,7 @@ 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();
+        let mut index_writer = index.writer_for_tests().unwrap();
 
         {
             for _ in 0..100 {

src/indexer/segment_writer.rs

@@ -17,7 +17,6 @@ use crate::tokenizer::{TokenStreamChain, Tokenizer};
 use crate::Opstamp;
 use crate::{DocId, SegmentComponent};
 use std::io;
-use std::str;
 
 /// Computes the initial size of the hash table.
 ///
@@ -48,6 +47,7 @@ pub struct SegmentWriter {
     fieldnorms_writer: FieldNormsWriter,
     doc_opstamps: Vec<Opstamp>,
     tokenizers: Vec<Option<TextAnalyzer>>,
+    term_buffer: Term,
 }
 
 impl SegmentWriter {
@@ -91,6 +91,7 @@ impl SegmentWriter {
             fast_field_writers: FastFieldsWriter::from_schema(schema),
             doc_opstamps: Vec::with_capacity(1_000),
             tokenizers,
+            term_buffer: Term::new(),
         })
     }
 
@@ -128,30 +129,32 @@ impl SegmentWriter {
             if !field_options.is_indexed() {
                 continue;
             }
+            let (term_buffer, multifield_postings) =
+                (&mut self.term_buffer, &mut self.multifield_postings);
             match *field_options.field_type() {
                 FieldType::HierarchicalFacet => {
-                    let facets: Vec<&str> = field_values
+                    term_buffer.set_field(field);
-                        .iter()
+                    let facets =
-                        .flat_map(|field_value| match *field_value.value() {
+                        field_values
-                            Value::Facet(ref facet) => Some(facet.encoded_str()),
+                            .iter()
-                            _ => {
+                            .flat_map(|field_value| match *field_value.value() {
-                                panic!("Expected hierarchical facet");
+                                Value::Facet(ref facet) => Some(facet.encoded_str()),
-                            }
+                                _ => {
-                        })
+                                    panic!("Expected hierarchical facet");
-                        .collect();
+                                }
-                    let mut term = Term::for_field(field); // we set the Term
+                            });
                     for fake_str in facets {
                         let mut unordered_term_id_opt = None;
                         FacetTokenizer.token_stream(fake_str).process(&mut |token| {
-                            term.set_text(&token.text);
+                            term_buffer.set_text(&token.text);
                             let unordered_term_id =
-                                self.multifield_postings.subscribe(doc_id, &term);
+                                multifield_postings.subscribe(doc_id, &term_buffer);
                             unordered_term_id_opt = Some(unordered_term_id);
                         });
                         if let Some(unordered_term_id) = unordered_term_id_opt {
                             self.fast_field_writers
                                 .get_multivalue_writer(field)
-                                .expect("writer for facet missing")
+                                .expect("multified writer for facet missing")
                                 .add_val(unordered_term_id);
                         }
                     }
@@ -168,7 +171,6 @@ impl SegmentWriter {
                                 if let Some(last_token) = tok_str.tokens.last() {
                                     total_offset += last_token.offset_to;
                                 }
-
                                 token_streams
                                     .push(PreTokenizedStream::from(tok_str.clone()).into());
                             }
@@ -178,7 +180,6 @@ impl SegmentWriter {
                                 {
                                     offsets.push(total_offset);
                                     total_offset += text.len();
-
                                     token_streams.push(tokenizer.token_stream(text));
                                 }
                             }
@@ -190,8 +191,12 @@ impl SegmentWriter {
                         0
                     } else {
                         let mut token_stream = TokenStreamChain::new(offsets, token_streams);
-                        self.multifield_postings
+                        multifield_postings.index_text(
-                            .index_text(doc_id, field, &mut token_stream)
+                            doc_id,
+                            field,
+                            &mut token_stream,
+                            term_buffer,
+                        )
                     };
 
                     self.fieldnorms_writer.record(doc_id, field, num_tokens);
@@ -199,44 +204,36 @@ impl SegmentWriter {
                 FieldType::U64(ref int_option) => {
                     if int_option.is_indexed() {
                         for field_value in field_values {
-                            let term = Term::from_field_u64(
+                            term_buffer.set_field(field_value.field());
-                                field_value.field(),
+                            term_buffer.set_u64(field_value.value().u64_value());
-                                field_value.value().u64_value(),
+                            multifield_postings.subscribe(doc_id, &term_buffer);
-                            );
-                            self.multifield_postings.subscribe(doc_id, &term);
                         }
                     }
                 }
                 FieldType::Date(ref int_option) => {
                     if int_option.is_indexed() {
                         for field_value in field_values {
-                            let term = Term::from_field_i64(
+                            term_buffer.set_field(field_value.field());
-                                field_value.field(),
+                            term_buffer.set_i64(field_value.value().date_value().timestamp());
-                                field_value.value().date_value().timestamp(),
+                            multifield_postings.subscribe(doc_id, &term_buffer);
-                            );
-                            self.multifield_postings.subscribe(doc_id, &term);
                         }
                     }
                 }
                 FieldType::I64(ref int_option) => {
                     if int_option.is_indexed() {
                         for field_value in field_values {
-                            let term = Term::from_field_i64(
+                            term_buffer.set_field(field_value.field());
-                                field_value.field(),
+                            term_buffer.set_i64(field_value.value().i64_value());
-                                field_value.value().i64_value(),
+                            multifield_postings.subscribe(doc_id, &term_buffer);
-                            );
-                            self.multifield_postings.subscribe(doc_id, &term);
                         }
                     }
                 }
                 FieldType::F64(ref int_option) => {
                     if int_option.is_indexed() {
                         for field_value in field_values {
-                            let term = Term::from_field_f64(
+                            term_buffer.set_field(field_value.field());
-                                field_value.field(),
+                            term_buffer.set_f64(field_value.value().f64_value());
-                                field_value.value().f64_value(),
+                            multifield_postings.subscribe(doc_id, &term_buffer);
-                            );
-                            self.multifield_postings.subscribe(doc_id, &term);
                         }
                     }
                 }

src/lib.rs

@@ -245,18 +245,10 @@ pub type DocId = u32;
 /// with opstamp `n+1`.
 pub type Opstamp = u64;
 
-/// A Score that represents the relevance of the document to the query
-///
-/// This is modelled internally as a `f64`, because tantivy was compiled with the `scoref64`
-/// feature. The larger the number, the more relevant the document is to the search query.
-#[cfg(feature = "scoref64")]
-pub type Score = f64;
-
 /// A Score that represents the relevance of the document to the query
 ///
 /// This is modelled internally as a `f32`. The larger the number, the more relevant
 /// the document to the search query.
-#[cfg(not(feature = "scoref64"))]
 pub type Score = f32;
 
 /// A `SegmentLocalId` identifies a segment.
@@ -296,7 +288,6 @@ mod tests {
     use crate::schema::*;
     use crate::DocAddress;
     use crate::Index;
-    use crate::IndexWriter;
     use crate::Postings;
     use crate::ReloadPolicy;
     use rand::distributions::Bernoulli;
@@ -361,14 +352,14 @@ mod tests {
 
     #[test]
     #[cfg(feature = "mmap")]
-    fn test_indexing() {
+    fn test_indexing() -> crate::Result<()> {
         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_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_for_tests()?;
             {
                 let doc = doc!(text_field=>"af b");
                 index_writer.add_document(doc);
@@ -383,29 +374,30 @@ mod tests {
             }
             assert!(index_writer.commit().is_ok());
         }
+        Ok(())
     }
 
     #[test]
-    fn test_docfreq1() {
+    fn test_docfreq1() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         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_for_tests()?;
         {
             index_writer.add_document(doc!(text_field=>"a b c"));
-            index_writer.commit().unwrap();
+            index_writer.commit()?;
         }
         {
             index_writer.add_document(doc!(text_field=>"a"));
             index_writer.add_document(doc!(text_field=>"a a"));
-            index_writer.commit().unwrap();
+            index_writer.commit()?;
         }
         {
             index_writer.add_document(doc!(text_field=>"c"));
-            index_writer.commit().unwrap();
+            index_writer.commit()?;
         }
         {
-            let reader = index.reader().unwrap();
+            let reader = index.reader()?;
             let searcher = reader.searcher();
             let term_a = Term::from_field_text(text_field, "a");
             assert_eq!(searcher.doc_freq(&term_a), 3);
@@ -416,67 +408,50 @@ mod tests {
             let term_d = Term::from_field_text(text_field, "d");
             assert_eq!(searcher.doc_freq(&term_d), 0);
         }
+        Ok(())
     }
 
     #[test]
-    fn test_fieldnorm_no_docs_with_field() {
+    fn test_fieldnorm_no_docs_with_field() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         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_for_tests()?;
+        index_writer.add_document(doc!(text_field=>"a b c"));
+        index_writer.commit()?;
+        let index_reader = index.reader()?;
+        let searcher = index_reader.searcher();
+        let reader = searcher.segment_reader(0);
         {
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let fieldnorm_reader = reader.get_fieldnorms_reader(text_field)?;
-            {
+            assert_eq!(fieldnorm_reader.fieldnorm(0), 3);
-                let doc = doc!(text_field=>"a b c");
-                index_writer.add_document(doc);
-            }
-            index_writer.commit().unwrap();
         }
         {
-            let index_reader = index.reader().unwrap();
+            let fieldnorm_reader = reader.get_fieldnorms_reader(title_field)?;
-            let searcher = index_reader.searcher();
+            assert_eq!(fieldnorm_reader.fieldnorm_id(0), 0);
-            let reader = searcher.segment_reader(0);
-            {
-                let fieldnorm_reader = reader.get_fieldnorms_reader(text_field);
-                assert_eq!(fieldnorm_reader.fieldnorm(0), 3);
-            }
-            {
-                let fieldnorm_reader = reader.get_fieldnorms_reader(title_field);
-                assert_eq!(fieldnorm_reader.fieldnorm_id(0), 0);
-            }
         }
+        Ok(())
     }
 
     #[test]
-    fn test_fieldnorm() {
+    fn test_fieldnorm() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         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_for_tests()?;
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        index_writer.add_document(doc!(text_field=>"a b c"));
-            {
+        index_writer.add_document(doc!());
-                let doc = doc!(text_field=>"a b c");
+        index_writer.add_document(doc!(text_field=>"a b"));
-                index_writer.add_document(doc);
+        index_writer.commit()?;
-            }
+        let reader = index.reader()?;
-            {
+        let searcher = reader.searcher();
-                let doc = doc!();
+        let segment_reader: &SegmentReader = searcher.segment_reader(0);
-                index_writer.add_document(doc);
+        let fieldnorms_reader = segment_reader.get_fieldnorms_reader(text_field)?;
-            }
+        assert_eq!(fieldnorms_reader.fieldnorm(0), 3);
-            {
+        assert_eq!(fieldnorms_reader.fieldnorm(1), 0);
-                let doc = doc!(text_field=>"a b");
+        assert_eq!(fieldnorms_reader.fieldnorm(2), 2);
-                index_writer.add_document(doc);
+        Ok(())
-            }
-            index_writer.commit().unwrap();
-        }
-        {
-            let reader = index.reader().unwrap();
-            let searcher = reader.searcher();
-            let segment_reader: &SegmentReader = searcher.segment_reader(0);
-            let fieldnorms_reader = segment_reader.get_fieldnorms_reader(text_field);
-            assert_eq!(fieldnorms_reader.fieldnorm(0), 3);
-            assert_eq!(fieldnorms_reader.fieldnorm(1), 0);
-            assert_eq!(fieldnorms_reader.fieldnorm(2), 2);
-        }
     }
 
     fn advance_undeleted(docset: &mut dyn DocSet, reader: &SegmentReader) -> bool {
@@ -491,7 +466,7 @@ mod tests {
     }
 
     #[test]
-    fn test_delete_postings1() {
+    fn test_delete_postings1() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let term_abcd = Term::from_field_text(text_field, "abcd");
@@ -507,7 +482,7 @@ mod tests {
             .unwrap();
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests()?;
             // 0
             index_writer.add_document(doc!(text_field=>"a b"));
             // 1
@@ -523,10 +498,10 @@ mod tests {
             index_writer.add_document(doc!(text_field=>" b c"));
             // 5
             index_writer.add_document(doc!(text_field=>" a"));
-            index_writer.commit().unwrap();
+            index_writer.commit()?;
         }
         {
-            reader.reload().unwrap();
+            reader.reload()?;
             let searcher = reader.searcher();
             let segment_reader = searcher.segment_reader(0);
             let inverted_index = segment_reader.inverted_index(text_field);
@@ -554,15 +529,15 @@ mod tests {
         }
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests()?;
             // 0
             index_writer.add_document(doc!(text_field=>"a b"));
             // 1
             index_writer.delete_term(Term::from_field_text(text_field, "c"));
-            index_writer.rollback().unwrap();
+            index_writer.rollback()?;
         }
         {
-            reader.reload().unwrap();
+            reader.reload()?;
             let searcher = reader.searcher();
             let seg_reader = searcher.segment_reader(0);
             let inverted_index = seg_reader.inverted_index(term_abcd.field());
@@ -591,15 +566,15 @@ mod tests {
         }
         {
             // writing the segment
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests()?;
             index_writer.add_document(doc!(text_field=>"a b"));
             index_writer.delete_term(Term::from_field_text(text_field, "c"));
-            index_writer.rollback().unwrap();
+            index_writer.rollback()?;
             index_writer.delete_term(Term::from_field_text(text_field, "a"));
-            index_writer.commit().unwrap();
+            index_writer.commit()?;
         }
         {
-            reader.reload().unwrap();
+            reader.reload()?;
             let searcher = reader.searcher();
             let segment_reader = searcher.segment_reader(0);
             let inverted_index = segment_reader.inverted_index(term_abcd.field());
@@ -631,19 +606,20 @@ mod tests {
                 assert!(!advance_undeleted(&mut postings, segment_reader));
             }
         }
+        Ok(())
     }
 
     #[test]
-    fn test_indexed_u64() {
+    fn test_indexed_u64() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         let field = schema_builder.add_u64_field("value", 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_for_tests()?;
         index_writer.add_document(doc!(field=>1u64));
-        index_writer.commit().unwrap();
+        index_writer.commit()?;
-        let reader = index.reader().unwrap();
+        let reader = index.reader()?;
         let searcher = reader.searcher();
         let term = Term::from_field_u64(field, 1u64);
         let mut postings = searcher
@@ -653,20 +629,21 @@ mod tests {
             .unwrap();
         assert_eq!(postings.doc(), 0);
         assert_eq!(postings.advance(), TERMINATED);
+        Ok(())
     }
 
     #[test]
-    fn test_indexed_i64() {
+    fn test_indexed_i64() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         let value_field = schema_builder.add_i64_field("value", 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_for_tests()?;
         let negative_val = -1i64;
         index_writer.add_document(doc!(value_field => negative_val));
-        index_writer.commit().unwrap();
+        index_writer.commit()?;
-        let reader = index.reader().unwrap();
+        let reader = index.reader()?;
         let searcher = reader.searcher();
         let term = Term::from_field_i64(value_field, negative_val);
         let mut postings = searcher
@@ -676,20 +653,21 @@ mod tests {
             .unwrap();
         assert_eq!(postings.doc(), 0);
         assert_eq!(postings.advance(), TERMINATED);
+        Ok(())
     }
 
     #[test]
-    fn test_indexed_f64() {
+    fn test_indexed_f64() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         let value_field = schema_builder.add_f64_field("value", 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_for_tests()?;
         let val = std::f64::consts::PI;
         index_writer.add_document(doc!(value_field => val));
-        index_writer.commit().unwrap();
+        index_writer.commit()?;
-        let reader = index.reader().unwrap();
+        let reader = index.reader()?;
         let searcher = reader.searcher();
         let term = Term::from_field_f64(value_field, val);
         let mut postings = searcher
@@ -699,26 +677,29 @@ mod tests {
             .unwrap();
         assert_eq!(postings.doc(), 0);
         assert_eq!(postings.advance(), TERMINATED);
+        Ok(())
     }
 
     #[test]
-    fn test_indexedfield_not_in_documents() {
+    fn test_indexedfield_not_in_documents() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         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_for_tests()?;
         index_writer.add_document(doc!(text_field=>"a"));
         assert!(index_writer.commit().is_ok());
-        let reader = index.reader().unwrap();
+        let reader = index.reader()?;
         let searcher = reader.searcher();
         let segment_reader = searcher.segment_reader(0);
-        segment_reader.inverted_index(absent_field); //< should not panic
+        let inverted_index = segment_reader.inverted_index(absent_field); //< should not panic
+        assert_eq!(inverted_index.terms().num_terms(), 0);
+        Ok(())
     }
 
     #[test]
-    fn test_delete_postings2() {
+    fn test_delete_postings2() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
@@ -726,53 +707,40 @@ mod tests {
         let reader = index
             .reader_builder()
             .reload_policy(ReloadPolicy::Manual)
-            .try_into()
+            .try_into()?;
-            .unwrap();
 
         // writing the segment
-        let mut index_writer = index.writer_with_num_threads(2, 6_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests()?;
-
+        index_writer.add_document(doc!(text_field=>"63"));
-        let add_document = |index_writer: &mut IndexWriter, val: &'static str| {
+        index_writer.add_document(doc!(text_field=>"70"));
-            let doc = doc!(text_field=>val);
+        index_writer.add_document(doc!(text_field=>"34"));
-            index_writer.add_document(doc);
+        index_writer.add_document(doc!(text_field=>"1"));
-        };
+        index_writer.add_document(doc!(text_field=>"38"));
-
+        index_writer.add_document(doc!(text_field=>"33"));
-        let remove_document = |index_writer: &mut IndexWriter, val: &'static str| {
+        index_writer.add_document(doc!(text_field=>"40"));
-            let delterm = Term::from_field_text(text_field, val);
+        index_writer.add_document(doc!(text_field=>"17"));
-            index_writer.delete_term(delterm);
+        index_writer.delete_term(Term::from_field_text(text_field, "38"));
-        };
+        index_writer.delete_term(Term::from_field_text(text_field, "34"));
-
+        index_writer.commit()?;
-        add_document(&mut index_writer, "63");
+        reader.reload()?;
-        add_document(&mut index_writer, "70");
+        assert_eq!(reader.searcher().num_docs(), 6);
-        add_document(&mut index_writer, "34");
+        Ok(())
-        add_document(&mut index_writer, "1");
-        add_document(&mut index_writer, "38");
-        add_document(&mut index_writer, "33");
-        add_document(&mut index_writer, "40");
-        add_document(&mut index_writer, "17");
-        remove_document(&mut index_writer, "38");
-        remove_document(&mut index_writer, "34");
-        index_writer.commit().unwrap();
-        reader.reload().unwrap();
-        let searcher = reader.searcher();
-        assert_eq!(searcher.num_docs(), 6);
     }
 
     #[test]
-    fn test_termfreq() {
+    fn test_termfreq() -> crate::Result<()> {
         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();
+            let mut index_writer = index.writer_for_tests()?;
-            let doc = doc!(text_field=>"af af af bc bc");
+            index_writer.add_document(doc!(text_field=>"af af af bc bc"));
-            index_writer.add_document(doc);
+            index_writer.commit()?;
-            index_writer.commit().unwrap();
         }
         {
-            let index_reader = index.reader().unwrap();
+            let index_reader = index.reader()?;
             let searcher = index_reader.searcher();
             let reader = searcher.segment_reader(0);
             let inverted_index = reader.inverted_index(text_field);
@@ -788,63 +756,63 @@ mod tests {
             assert_eq!(postings.term_freq(), 3);
             assert_eq!(postings.advance(), TERMINATED);
         }
+        Ok(())
     }
 
     #[test]
-    fn test_searcher_1() {
+    fn test_searcher_1() -> crate::Result<()> {
         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);
-        let reader = index.reader().unwrap();
+        let reader = index.reader()?;
-        {
+        // writing the segment
-            // writing the segment
+        let mut index_writer = index.writer_for_tests()?;
-            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=>"af af af b"));
+        index_writer.add_document(doc!(text_field=>"a b c"));
-            index_writer.add_document(doc!(text_field=>"a b c"));
+        index_writer.add_document(doc!(text_field=>"a b c d"));
-            index_writer.add_document(doc!(text_field=>"a b c d"));
+        index_writer.commit()?;
-            index_writer.commit().unwrap();
+
-        }
+        reader.reload()?;
-        {
+        let searcher = reader.searcher();
-            reader.reload().unwrap();
+        let get_doc_ids = |terms: Vec<Term>| {
-            let searcher = reader.searcher();
+            let query = BooleanQuery::new_multiterms_query(terms);
-            let get_doc_ids = |terms: Vec<Term>| {
+            searcher
-                let query = BooleanQuery::new_multiterms_query(terms);
+                .search(&query, &TEST_COLLECTOR_WITH_SCORE)
-                let topdocs = searcher.search(&query, &TEST_COLLECTOR_WITH_SCORE).unwrap();
+                .map(|topdocs| topdocs.docs().to_vec())
-                topdocs.docs().to_vec()
+        };
-            };
+        assert_eq!(
-            assert_eq!(
+            get_doc_ids(vec![Term::from_field_text(text_field, "a")])?,
-                get_doc_ids(vec![Term::from_field_text(text_field, "a")]),
+            vec![DocAddress(0, 1), DocAddress(0, 2)]
-                vec![DocAddress(0, 1), DocAddress(0, 2)]
+        );
-            );
+        assert_eq!(
-            assert_eq!(
+            get_doc_ids(vec![Term::from_field_text(text_field, "af")])?,
-                get_doc_ids(vec![Term::from_field_text(text_field, "af")]),
+            vec![DocAddress(0, 0)]
-                vec![DocAddress(0, 0)]
+        );
-            );
+        assert_eq!(
-            assert_eq!(
+            get_doc_ids(vec![Term::from_field_text(text_field, "b")])?,
-                get_doc_ids(vec![Term::from_field_text(text_field, "b")]),
+            vec![DocAddress(0, 0), DocAddress(0, 1), DocAddress(0, 2)]
-                vec![DocAddress(0, 0), DocAddress(0, 1), DocAddress(0, 2)]
+        );
-            );
+        assert_eq!(
-            assert_eq!(
+            get_doc_ids(vec![Term::from_field_text(text_field, "c")])?,
-                get_doc_ids(vec![Term::from_field_text(text_field, "c")]),
+            vec![DocAddress(0, 1), DocAddress(0, 2)]
-                vec![DocAddress(0, 1), DocAddress(0, 2)]
+        );
-            );
+        assert_eq!(
-            assert_eq!(
+            get_doc_ids(vec![Term::from_field_text(text_field, "d")])?,
-                get_doc_ids(vec![Term::from_field_text(text_field, "d")]),
+            vec![DocAddress(0, 2)]
-                vec![DocAddress(0, 2)]
+        );
-            );
+        assert_eq!(
-            assert_eq!(
+            get_doc_ids(vec![
-                get_doc_ids(vec![
+                Term::from_field_text(text_field, "b"),
-                    Term::from_field_text(text_field, "b"),
+                Term::from_field_text(text_field, "a"),
-                    Term::from_field_text(text_field, "a"),
+            ])?,
-                ]),
+            vec![DocAddress(0, 0), DocAddress(0, 1), DocAddress(0, 2)]
-                vec![DocAddress(0, 0), DocAddress(0, 1), DocAddress(0, 2)]
+        );
-            );
+        Ok(())
-        }
     }
 
     #[test]
-    fn test_searcher_2() {
+    fn test_searcher_2() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
         let schema = schema_builder.build();
@@ -852,19 +820,17 @@ mod tests {
         let reader = index
             .reader_builder()
             .reload_policy(ReloadPolicy::Manual)
-            .try_into()
+            .try_into()?;
-            .unwrap();
         assert_eq!(reader.searcher().num_docs(), 0u64);
-        {
+        // writing the segment
-            // writing the segment
+        let mut index_writer = index.writer_for_tests()?;
-            let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        index_writer.add_document(doc!(text_field=>"af b"));
-            index_writer.add_document(doc!(text_field=>"af b"));
+        index_writer.add_document(doc!(text_field=>"a b c"));
-            index_writer.add_document(doc!(text_field=>"a b c"));
+        index_writer.add_document(doc!(text_field=>"a b c d"));
-            index_writer.add_document(doc!(text_field=>"a b c d"));
+        index_writer.commit()?;
-            index_writer.commit().unwrap();
+        reader.reload()?;
-        }
-        reader.reload().unwrap();
         assert_eq!(reader.searcher().num_docs(), 3u64);
+        Ok(())
     }
 
     #[test]
@@ -886,7 +852,7 @@ mod tests {
     }
 
     #[test]
-    fn test_wrong_fast_field_type() {
+    fn test_wrong_fast_field_type() -> crate::Result<()> {
         let mut schema_builder = Schema::builder();
         let fast_field_unsigned = schema_builder.add_u64_field("unsigned", FAST);
         let fast_field_signed = schema_builder.add_i64_field("signed", FAST);
@@ -896,14 +862,14 @@ mod tests {
         let schema = schema_builder.build();
 
         let index = Index::create_in_ram(schema);
-        let mut index_writer = index.writer_with_num_threads(1, 50_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests()?;
         {
             let document =
                 doc!(fast_field_unsigned => 4u64, fast_field_signed=>4i64, fast_field_float=>4f64);
             index_writer.add_document(document);
-            index_writer.commit().unwrap();
+            index_writer.commit()?;
         }
-        let reader = index.reader().unwrap();
+        let reader = index.reader()?;
         let searcher = reader.searcher();
         let segment_reader: &SegmentReader = searcher.segment_reader(0);
         {
@@ -942,11 +908,12 @@ mod tests {
             let fast_field_reader = fast_field_reader_opt.unwrap();
             assert_eq!(fast_field_reader.get(0), 4f64)
         }
+        Ok(())
     }
 
     // motivated by #729
     #[test]
-    fn test_update_via_delete_insert() {
+    fn test_update_via_delete_insert() -> crate::Result<()> {
         use crate::collector::Count;
         use crate::indexer::NoMergePolicy;
         use crate::query::AllQuery;
@@ -960,17 +927,17 @@ mod tests {
         let schema = schema_builder.build();
 
         let index = Index::create_in_ram(schema.clone());
-        let index_reader = index.reader().unwrap();
+        let index_reader = index.reader()?;
 
-        let mut index_writer = index.writer(3_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests()?;
         index_writer.set_merge_policy(Box::new(NoMergePolicy));
 
         for doc_id in 0u64..DOC_COUNT {
             index_writer.add_document(doc!(id => doc_id));
         }
-        index_writer.commit().unwrap();
+        index_writer.commit()?;
 
-        index_reader.reload().unwrap();
+        index_reader.reload()?;
         let searcher = index_reader.searcher();
 
         assert_eq!(
@@ -981,12 +948,11 @@ mod tests {
         // update the 10 elements by deleting and re-adding
         for doc_id in 0u64..DOC_COUNT {
             index_writer.delete_term(Term::from_field_u64(id, doc_id));
-            index_writer.commit().unwrap();
+            index_writer.commit()?;
-            index_reader.reload().unwrap();
+            index_reader.reload()?;
-            let doc = doc!(id =>  doc_id);
+            index_writer.add_document(doc!(id =>  doc_id));
-            index_writer.add_document(doc);
+            index_writer.commit()?;
-            index_writer.commit().unwrap();
+            index_reader.reload()?;
-            index_reader.reload().unwrap();
             let searcher = index_reader.searcher();
             // The number of document should be stable.
             assert_eq!(
@@ -995,7 +961,7 @@ mod tests {
             );
         }
 
-        index_reader.reload().unwrap();
+        index_reader.reload()?;
         let searcher = index_reader.searcher();
         let segment_ids: Vec<SegmentId> = searcher
             .segment_readers()
@@ -1004,12 +970,18 @@ mod tests {
             .collect();
         block_on(index_writer.merge(&segment_ids)).unwrap();
 
-        index_reader.reload().unwrap();
+        index_reader.reload()?;
         let searcher = index_reader.searcher();
+        assert_eq!(searcher.search(&AllQuery, &Count)?, DOC_COUNT as usize);
+        Ok(())
+    }
 
-        assert_eq!(
+    #[test]
-            searcher.search(&AllQuery, &Count).unwrap(),
+    fn test_validate_checksum() -> crate::Result<()> {
-            DOC_COUNT as usize
+        let index_path = tempfile::tempdir().expect("dir");
-        );
+        let schema = Schema::builder().build();
+        let index = Index::create_in_dir(&index_path, schema)?;
+        assert!(index.validate_checksum()?.is_empty());
+        Ok(())
     }
 }

src/postings/block_segment_postings.rs

@@ -455,7 +455,7 @@ mod tests {
         let int_field = schema_builder.add_u64_field("id", 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_for_tests().unwrap();
         let mut last_doc = 0u32;
         for &doc in docs {
             for _ in last_doc..doc {
@@ -496,7 +496,7 @@ mod tests {
         let int_field = schema_builder.add_u64_field("id", 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_for_tests().unwrap();
         // create two postings list, one containg even number,
         // the other containing odd numbers.
         for i in 0..6 {

src/postings/compression/mod.rs

@@ -310,6 +310,7 @@ pub mod tests {
 mod bench {
 
     use super::*;
+    use crate::TERMINATED;
     use rand::rngs::StdRng;
     use rand::Rng;
     use rand::SeedableRng;
@@ -340,7 +341,7 @@ mod bench {
         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 mut decoder = BlockDecoder::new();
+        let mut decoder = BlockDecoder::default();
         b.iter(|| {
             decoder.uncompress_block_sorted(compressed, 0u32, num_bits);
         });
@@ -375,9 +376,9 @@ mod bench {
         let mut encoder = BlockEncoder::new();
         let data = generate_array(NUM_INTS_BENCH_VINT, 0.001);
         let compressed = encoder.compress_vint_sorted(&data, 0u32);
-        let mut decoder = BlockDecoder::new();
+        let mut decoder = BlockDecoder::default();
         b.iter(|| {
-            decoder.uncompress_vint_sorted(compressed, 0u32, NUM_INTS_BENCH_VINT);
+            decoder.uncompress_vint_sorted(compressed, 0u32, NUM_INTS_BENCH_VINT, TERMINATED);
         });
     }
 }

src/postings/mod.rs

@@ -5,6 +5,7 @@ Postings module (also called inverted index)
 mod block_search;
 mod block_segment_postings;
 pub(crate) mod compression;
+mod field_stats;
 mod postings;
 mod postings_writer;
 mod recorder;
@@ -15,6 +16,7 @@ mod stacker;
 mod term_info;
 
 pub(crate) use self::block_search::BlockSearcher;
+pub(crate) use self::field_stats::{FieldStat, FieldStats};
 
 pub(crate) use self::postings_writer::MultiFieldPostingsWriter;
 pub use self::serializer::{FieldSerializer, InvertedIndexSerializer};
@@ -91,7 +93,7 @@ pub mod tests {
         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, 30_000_000)?;
+        let mut index_writer = index.writer_for_tests()?;
         index_writer.add_document(doc!(title => r#"abc abc abc"#));
         index_writer.add_document(doc!(title => r#"abc be be be be abc"#));
         for _ in 0..1_000 {
@@ -176,7 +178,7 @@ pub mod tests {
             .tokenizers()
             .register("simple_no_truncation", SimpleTokenizer);
         let reader = index.reader().unwrap();
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests().unwrap();
         index_writer.set_merge_policy(Box::new(NoMergePolicy));
         {
             index_writer.add_document(doc!(text_field=>exceeding_token_text));
@@ -205,7 +207,7 @@ pub mod tests {
     }
 
     #[test]
-    pub fn test_position_and_fieldnorm1() {
+    pub fn test_position_and_fieldnorm1() -> crate::Result<()> {
         let mut positions = Vec::new();
         let mut schema_builder = Schema::builder();
         let text_field = schema_builder.add_text_field("text", TEXT);
@@ -217,42 +219,38 @@ pub mod tests {
             let mut segment_writer =
                 SegmentWriter::for_segment(3_000_000, segment.clone(), &schema).unwrap();
             {
-                let mut doc = Document::default();
                 // checking that position works if the field has two values
-                doc.add_text(text_field, "a b a c a d a a.");
-                doc.add_text(text_field, "d d d d a");
                 let op = AddOperation {
                     opstamp: 0u64,
-                    document: doc,
+                    document: doc!(
+                       text_field => "a b a c a d a a.",
+                       text_field => "d d d d a"
+                    ),
                 };
-                segment_writer.add_document(op, &schema).unwrap();
+                segment_writer.add_document(op, &schema)?;
             }
             {
-                let mut doc = Document::default();
-                doc.add_text(text_field, "b a");
                 let op = AddOperation {
                     opstamp: 1u64,
-                    document: doc,
+                    document: doc!(text_field => "b a"),
                 };
                 segment_writer.add_document(op, &schema).unwrap();
             }
             for i in 2..1000 {
-                let mut doc = Document::default();
+                let mut text: String = iter::repeat("e ").take(i).collect();
-                let mut text = iter::repeat("e ").take(i).collect::<String>();
                 text.push_str(" a");
-                doc.add_text(text_field, &text);
                 let op = AddOperation {
                     opstamp: 2u64,
-                    document: doc,
+                    document: doc!(text_field => text),
                 };
                 segment_writer.add_document(op, &schema).unwrap();
             }
-            segment_writer.finalize().unwrap();
+            segment_writer.finalize()?;
         }
         {
-            let segment_reader = SegmentReader::open(&segment).unwrap();
+            let segment_reader = SegmentReader::open(&segment)?;
             {
-                let fieldnorm_reader = segment_reader.get_fieldnorms_reader(text_field);
+                let fieldnorm_reader = segment_reader.get_fieldnorms_reader(text_field)?;
                 assert_eq!(fieldnorm_reader.fieldnorm(0), 8 + 5);
                 assert_eq!(fieldnorm_reader.fieldnorm(1), 2);
                 for i in 2..1000 {
@@ -312,6 +310,7 @@ pub mod tests {
                 assert_eq!(postings_e.doc(), TERMINATED);
             }
         }
+        Ok(())
     }
 
     #[test]
@@ -322,7 +321,7 @@ pub mod tests {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(text_field => "g b b d c g c"));
             index_writer.add_document(doc!(text_field => "g a b b a d c g c"));
             assert!(index_writer.commit().is_ok());
@@ -354,7 +353,7 @@ pub mod tests {
 
             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_for_tests().unwrap();
                 for i in 0u64..num_docs as u64 {
                     let doc = doc!(value_field => 2u64, value_field => i % 2u64);
                     index_writer.add_document(doc);
@@ -425,7 +424,7 @@ 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_for_tests().unwrap();
             index_writer.delete_term(term_0);
             assert!(index_writer.commit().is_ok());
         }
@@ -479,7 +478,7 @@ 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_for_tests().unwrap();
             index_writer.delete_term(term_1);
             assert!(index_writer.commit().is_ok());
         }
@@ -522,7 +521,7 @@ pub mod tests {
         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_for_tests().unwrap();
             for _ in 0..posting_list_size {
                 let mut doc = Document::default();
                 if rng.gen_bool(1f64 / 15f64) {
@@ -730,7 +729,7 @@ mod bench {
             let mut s = 0u32;
             while segment_postings.doc() != TERMINATED {
                 s += (segment_postings.doc() & n) % 1024;
-                segment_postings.advance()
+                segment_postings.advance();
             }
             s
         });

src/postings/postings_writer.rs

@@ -105,6 +105,7 @@ impl MultiFieldPostingsWriter {
         doc: DocId,
         field: Field,
         token_stream: &mut dyn TokenStream,
+        term_buffer: &mut Term,
     ) -> u32 {
         let postings_writer =
             self.per_field_postings_writers[field.field_id() as usize].deref_mut();
@@ -114,6 +115,7 @@ impl MultiFieldPostingsWriter {
             field,
             token_stream,
             &mut self.heap,
+            term_buffer,
         )
     }
 
@@ -220,13 +222,20 @@ pub trait PostingsWriter {
         field: Field,
         token_stream: &mut dyn TokenStream,
         heap: &mut MemoryArena,
+        term_buffer: &mut Term,
     ) -> u32 {
-        let mut term = Term::for_field(field);
+        term_buffer.set_field(field);
         let mut sink = |token: &Token| {
             // We skip all tokens with a len greater than u16.
             if token.text.len() <= MAX_TOKEN_LEN {
-                term.set_text(token.text.as_str());
+                term_buffer.set_text(token.text.as_str());
-                self.subscribe(term_index, doc_id, token.position as u32, &term, heap);
+                self.subscribe(
+                    term_index,
+                    doc_id,
+                    token.position as u32,
+                    &term_buffer,
+                    heap,
+                );
             } else {
                 info!(
                     "A token exceeding MAX_TOKEN_LEN ({}>{}) was dropped. Search for \

src/postings/segment_postings.rs

@@ -114,7 +114,7 @@ impl SegmentPostings {
                     .iter()
                     .map(|&fieldnorm| fieldnorm as u64)
                     .sum::<u64>();
-                total_num_tokens as Score / fieldnorms.len() as f32
+                total_num_tokens as Score / fieldnorms.len() as Score
             })
             .unwrap_or(0.0);
         let mut postings_serializer = PostingsSerializer::new(

src/postings/serializer.rs

@@ -1,5 +1,4 @@
-use super::TermInfo;
+use super::{FieldStat, FieldStats, TermInfo};
-use crate::common::{BinarySerializable, VInt};
 use crate::common::{CompositeWrite, CountingWriter};
 use crate::core::Segment;
 use crate::directory::WritePtr;
@@ -11,6 +10,10 @@ use crate::query::BM25Weight;
 use crate::schema::Schema;
 use crate::schema::{Field, FieldEntry, FieldType};
 use crate::termdict::{TermDictionaryBuilder, TermOrdinal};
+use crate::{
+    common::{BinarySerializable, VInt},
+    directory::TerminatingWrite,
+};
 use crate::{DocId, Score};
 use std::cmp::Ordering;
 use std::io::{self, Write};
@@ -51,6 +54,8 @@ pub struct InvertedIndexSerializer {
     postings_write: CompositeWrite<WritePtr>,
     positions_write: CompositeWrite<WritePtr>,
     positionsidx_write: CompositeWrite<WritePtr>,
+    field_stats: FieldStats,
+    field_stats_write: WritePtr,
     schema: Schema,
 }
 
@@ -61,6 +66,7 @@ impl InvertedIndexSerializer {
         postings_write: CompositeWrite<WritePtr>,
         positions_write: CompositeWrite<WritePtr>,
         positionsidx_write: CompositeWrite<WritePtr>,
+        field_stats_write: WritePtr,
         schema: Schema,
     ) -> crate::Result<InvertedIndexSerializer> {
         Ok(InvertedIndexSerializer {
@@ -68,18 +74,21 @@ impl InvertedIndexSerializer {
             postings_write,
             positions_write,
             positionsidx_write,
+            field_stats: FieldStats::default(),
+            field_stats_write,
             schema,
         })
     }
 
     /// Open a new `PostingsSerializer` for the given segment
     pub fn open(segment: &mut Segment) -> crate::Result<InvertedIndexSerializer> {
-        use crate::SegmentComponent::{POSITIONS, POSITIONSSKIP, POSTINGS, TERMS};
+        use crate::SegmentComponent::{FIELDSTATS, POSITIONS, POSITIONSSKIP, POSTINGS, TERMS};
         InvertedIndexSerializer::create(
             CompositeWrite::wrap(segment.open_write(TERMS)?),
             CompositeWrite::wrap(segment.open_write(POSTINGS)?),
             CompositeWrite::wrap(segment.open_write(POSITIONS)?),
             CompositeWrite::wrap(segment.open_write(POSITIONSSKIP)?),
+            segment.open_write(FIELDSTATS)?,
             segment.schema(),
         )
     }
@@ -94,6 +103,8 @@ impl InvertedIndexSerializer {
         total_num_tokens: u64,
         fieldnorm_reader: Option<FieldNormReader>,
     ) -> io::Result<FieldSerializer<'_>> {
+        self.field_stats
+            .insert(field, FieldStat::new(total_num_tokens));
         let field_entry: &FieldEntry = self.schema.get_field_entry(field);
         let term_dictionary_write = self.terms_write.for_field(field);
         let postings_write = self.postings_write.for_field(field);
@@ -112,7 +123,10 @@ impl InvertedIndexSerializer {
     }
 
     /// Closes the serializer.
-    pub fn close(self) -> io::Result<()> {
+    pub fn close(mut self) -> io::Result<()> {
+        self.field_stats
+            .serialize(self.field_stats_write.get_mut())?;
+        self.field_stats_write.terminate()?;
         self.terms_write.close()?;
         self.postings_write.close()?;
         self.positions_write.close()?;
@@ -142,7 +156,6 @@ impl<'a> FieldSerializer<'a> {
         positionsidx_write: &'a mut CountingWriter<WritePtr>,
         fieldnorm_reader: Option<FieldNormReader>,
     ) -> io::Result<FieldSerializer<'a>> {
-        total_num_tokens.serialize(postings_write)?;
         let (term_freq_enabled, position_enabled): (bool, bool) = match field_type {
             FieldType::Str(ref text_options) => {
                 if let Some(text_indexing_options) = text_options.get_indexing_options() {
@@ -190,7 +203,8 @@ impl<'a> FieldSerializer<'a> {
             .unwrap_or(0u64);
         TermInfo {
             doc_freq: 0,
-            postings_offset: self.postings_serializer.addr(),
+            postings_start_offset: self.postings_serializer.addr(),
+            postings_end_offset: 0u64,
             positions_idx,
         }
     }
@@ -244,10 +258,12 @@ impl<'a> FieldSerializer<'a> {
     /// using `VInt` encoding.
     pub fn close_term(&mut self) -> io::Result<()> {
         if self.term_open {
-            self.term_dictionary_builder
-                .insert_value(&self.current_term_info)?;
             self.postings_serializer
                 .close_term(self.current_term_info.doc_freq)?;
+            let end_offset = self.postings_serializer.addr();
+            self.current_term_info.postings_end_offset = end_offset;
+            self.term_dictionary_builder
+                .insert_value(&self.current_term_info)?;
             self.term_open = false;
         }
         Ok(())

src/postings/term_info.rs

@@ -7,35 +7,49 @@ use std::io;
 pub struct TermInfo {
     /// Number of documents in the segment containing the term
     pub doc_freq: u32,
-    /// Start offset within the postings (`.idx`) file.
+    /// Start offset of the posting list within the postings (`.idx`) file.
-    pub postings_offset: u64,
+    pub postings_start_offset: u64,
+    /// End offset of the posting list within the postings (`.idx`) file.
+    pub postings_end_offset: u64,
     /// Start offset of the first block within the position (`.pos`) file.
     pub positions_idx: u64,
 }
 
+impl TermInfo {
+    pub(crate) fn posting_num_bytes(&self) -> u32 {
+        let num_bytes = self.postings_end_offset - self.postings_start_offset;
+        assert!(num_bytes <= std::u32::MAX as u64);
+        num_bytes as u32
+    }
+}
+
 impl FixedSize for TermInfo {
     /// Size required for the binary serialization of a `TermInfo` object.
     /// This is large, but in practise, `TermInfo` are encoded in blocks and
     /// only the first `TermInfo` of a block is serialized uncompressed.
     /// The subsequent `TermInfo` are delta encoded and bitpacked.
-    const SIZE_IN_BYTES: usize = u32::SIZE_IN_BYTES + 2 * u64::SIZE_IN_BYTES;
+    const SIZE_IN_BYTES: usize = 2 * u32::SIZE_IN_BYTES + 2 * u64::SIZE_IN_BYTES;
 }
 
 impl BinarySerializable for TermInfo {
     fn serialize<W: io::Write>(&self, writer: &mut W) -> io::Result<()> {
         self.doc_freq.serialize(writer)?;
-        self.postings_offset.serialize(writer)?;
+        self.postings_start_offset.serialize(writer)?;
+        self.posting_num_bytes().serialize(writer)?;
         self.positions_idx.serialize(writer)?;
         Ok(())
     }
 
     fn deserialize<R: io::Read>(reader: &mut R) -> io::Result<Self> {
         let doc_freq = u32::deserialize(reader)?;
-        let postings_offset = u64::deserialize(reader)?;
+        let postings_start_offset = u64::deserialize(reader)?;
+        let postings_num_bytes = u32::deserialize(reader)?;
+        let postings_end_offset = postings_start_offset + u64::from(postings_num_bytes);
         let positions_idx = u64::deserialize(reader)?;
         Ok(TermInfo {
             doc_freq,
-            postings_offset,
+            postings_start_offset,
+            postings_end_offset,
             positions_idx,
         })
     }

src/query/all_query.rs

@@ -83,7 +83,7 @@ mod tests {
         let 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, 10_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests().unwrap();
         index_writer.add_document(doc!(field=>"aaa"));
         index_writer.add_document(doc!(field=>"bbb"));
         index_writer.commit().unwrap();

src/query/automaton_weight.rs

@@ -5,7 +5,6 @@ use crate::query::{BitSetDocSet, Explanation};
 use crate::query::{Scorer, Weight};
 use crate::schema::{Field, IndexRecordOption};
 use crate::termdict::{TermDictionary, TermStreamer};
-use crate::Result;
 use crate::TantivyError;
 use crate::{DocId, Score};
 use std::sync::Arc;
@@ -40,7 +39,7 @@ impl<A> Weight for AutomatonWeight<A>
 where
     A: Automaton + Send + Sync + 'static,
 {
-    fn scorer(&self, reader: &SegmentReader, boost: Score) -> Result<Box<dyn Scorer>> {
+    fn scorer(&self, reader: &SegmentReader, boost: Score) -> crate::Result<Box<dyn Scorer>> {
         let max_doc = reader.max_doc();
         let mut doc_bitset = BitSet::with_max_value(max_doc);
         let inverted_index = reader.inverted_index(self.field);
@@ -66,7 +65,7 @@ where
         Ok(Box::new(const_scorer))
     }
 
-    fn explain(&self, reader: &SegmentReader, doc: DocId) -> Result<Explanation> {
+    fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
         let mut scorer = self.scorer(reader, 1.0)?;
         if scorer.seek(doc) == doc {
             Ok(Explanation::new("AutomatonScorer", 1.0))
@@ -91,7 +90,7 @@ mod tests {
         let mut schema = Schema::builder();
         let title = schema.add_text_field("title", STRING);
         let index = Index::create_in_ram(schema.build());
-        let mut index_writer = index.writer_with_num_threads(1, 3_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests().unwrap();
         index_writer.add_document(doc!(title=>"abc"));
         index_writer.add_document(doc!(title=>"bcd"));
         index_writer.add_document(doc!(title=>"abcd"));

src/query/boolean_query/block_wand.rs

@@ -4,19 +4,6 @@ use crate::{DocId, DocSet, Score, TERMINATED};
 use std::ops::Deref;
 use std::ops::DerefMut;
 
-fn is_sorted<I: Iterator<Item = DocId>>(mut it: I) -> bool {
-    if let Some(first) = it.next() {
-        let mut prev = first;
-        for doc in it {
-            if doc < prev {
-                return false;
-            }
-            prev = doc;
-        }
-    }
-    true
-}
-
 /// Takes a term_scorers sorted by their current doc() and a threshold and returns
 /// Returns (pivot_len, pivot_ord) defined as follows:
 /// - `pivot_doc` lowest document that has a chance of exceeding (>) the threshold score.
@@ -55,37 +42,12 @@ fn find_pivot_doc(
     Some((before_pivot_len, pivot_len, pivot_doc))
 }
 
-struct TermScorerWithMaxScore<'a> {
-    scorer: &'a mut TermScorer,
-    max_score: Score,
-}
-
-impl<'a> From<&'a mut TermScorer> for TermScorerWithMaxScore<'a> {
-    fn from(scorer: &'a mut TermScorer) -> Self {
-        let max_score = scorer.max_score();
-        TermScorerWithMaxScore { scorer, max_score }
-    }
-}
-
-impl<'a> Deref for TermScorerWithMaxScore<'a> {
-    type Target = TermScorer;
-
-    fn deref(&self) -> &Self::Target {
-        self.scorer
-    }
-}
-
-impl<'a> DerefMut for TermScorerWithMaxScore<'a> {
-    fn deref_mut(&mut self) -> &mut Self::Target {
-        self.scorer
-    }
-}
-
 // Before and after calling this method, scorers need to be sorted by their `.doc()`.
 fn block_max_was_too_low_advance_one_scorer(
     scorers: &mut Vec<TermScorerWithMaxScore>,
     pivot_len: usize,
 ) {
+    debug_assert!(is_sorted(scorers.iter().map(|scorer| scorer.doc())));
     let mut scorer_to_seek = pivot_len - 1;
     let mut doc_to_seek_after = scorers[scorer_to_seek].doc();
     for scorer_ord in (0..pivot_len - 1).rev() {
@@ -102,6 +64,7 @@ fn block_max_was_too_low_advance_one_scorer(
     }
     scorers[scorer_to_seek].seek(doc_to_seek_after + 1);
     restore_ordering(scorers, scorer_to_seek);
+    debug_assert!(is_sorted(scorers.iter().map(|scorer| scorer.doc())));
 }
 
 // Given a list of term_scorers and a `ord` and assuming that `term_scorers[ord]` is sorted
@@ -177,64 +140,99 @@ pub fn block_wand(
         .map(TermScorerWithMaxScore::from)
         .collect();
     scorers.sort_by_key(|scorer| scorer.doc());
-    loop {
+    // At this point we need to ensure that the scorers are sorted!
-        // At this point we need to ensure that the scorers are sorted!
+    debug_assert!(is_sorted(scorers.iter().map(|scorer| scorer.doc())));
+    while let Some((before_pivot_len, pivot_len, pivot_doc)) =
+        find_pivot_doc(&scorers[..], threshold)
+    {
         debug_assert!(is_sorted(scorers.iter().map(|scorer| scorer.doc())));
-        if let Some((before_pivot_len, pivot_len, pivot_doc)) =
+        debug_assert_ne!(pivot_doc, TERMINATED);
-            find_pivot_doc(&scorers[..], threshold)
+        debug_assert!(before_pivot_len < pivot_len);
-        {
-            debug_assert_ne!(pivot_doc, TERMINATED);
-            debug_assert!(before_pivot_len < pivot_len);
 
-            let block_max_score_upperbound: Score = scorers[..pivot_len]
+        let block_max_score_upperbound: Score = scorers[..pivot_len]
-                .iter_mut()
+            .iter_mut()
-                .map(|scorer| {
+            .map(|scorer| {
-                    scorer.shallow_seek(pivot_doc);
+                scorer.shallow_seek(pivot_doc);
-                    scorer.block_max_score()
+                scorer.block_max_score()
-                })
+            })
-                .sum();
+            .sum();
 
-            // Beware after shallow advance, skip readers can be in advance compared to
+        // Beware after shallow advance, skip readers can be in advance compared to
-            // the segment posting lists.
+        // the segment posting lists.
-            //
+        //
-            // `block_segment_postings.load_block()` need to be called separately.
+        // `block_segment_postings.load_block()` need to be called separately.
-            if block_max_score_upperbound <= threshold {
+        if block_max_score_upperbound <= threshold {
-                // Block max condition was not reached
+            // Block max condition was not reached
-                // We could get away by simply advancing the scorers to DocId + 1 but it would
+            // We could get away by simply advancing the scorers to DocId + 1 but it would
-                // be inefficient. The optimization requires proper explanation and was
+            // be inefficient. The optimization requires proper explanation and was
-                // isolated in a different function.
+            // isolated in a different function.
-                block_max_was_too_low_advance_one_scorer(&mut scorers, pivot_len);
+            block_max_was_too_low_advance_one_scorer(&mut scorers, pivot_len);
-                continue;
+            continue;
-            }
-
-            // Block max condition is observed.
-            //
-            // Let's try and advance all scorers before the pivot to the pivot.
-            if !align_scorers(&mut scorers, pivot_doc, before_pivot_len) {
-                // At least of the scorer does not contain the pivot.
-                //
-                // Let's stop scoring this pivot and go through the pivot selection again.
-                // Note that the current pivot is not necessarily a bad candidate and it
-                // may be picked again.
-                continue;
-            }
-
-            // At this point, all scorers are positioned on the doc.
-            let score = scorers[..pivot_len]
-                .iter_mut()
-                .map(|scorer| scorer.score())
-                .sum();
-            if score > threshold {
-                threshold = callback(pivot_doc, score);
-            }
-            // let's advance all of the scorers that are currently positioned on the pivot.
-            advance_all_scorers_on_pivot(&mut scorers, pivot_len);
-        } else {
-            return;
         }
+
+        // Block max condition is observed.
+        //
+        // Let's try and advance all scorers before the pivot to the pivot.
+        if !align_scorers(&mut scorers, pivot_doc, before_pivot_len) {
+            // At least of the scorer does not contain the pivot.
+            //
+            // Let's stop scoring this pivot and go through the pivot selection again.
+            // Note that the current pivot is not necessarily a bad candidate and it
+            // may be picked again.
+            continue;
+        }
+
+        // At this point, all scorers are positioned on the doc.
+        let score = scorers[..pivot_len]
+            .iter_mut()
+            .map(|scorer| scorer.score())
+            .sum();
+        if score > threshold {
+            threshold = callback(pivot_doc, score);
+        }
+        // let's advance all of the scorers that are currently positioned on the pivot.
+        advance_all_scorers_on_pivot(&mut scorers, pivot_len);
     }
 }
 
+struct TermScorerWithMaxScore<'a> {
+    scorer: &'a mut TermScorer,
+    max_score: Score,
+}
+
+impl<'a> From<&'a mut TermScorer> for TermScorerWithMaxScore<'a> {
+    fn from(scorer: &'a mut TermScorer) -> Self {
+        let max_score = scorer.max_score();
+        TermScorerWithMaxScore { scorer, max_score }
+    }
+}
+
+impl<'a> Deref for TermScorerWithMaxScore<'a> {
+    type Target = TermScorer;
+
+    fn deref(&self) -> &Self::Target {
+        self.scorer
+    }
+}
+
+impl<'a> DerefMut for TermScorerWithMaxScore<'a> {
+    fn deref_mut(&mut self) -> &mut Self::Target {
+        self.scorer
+    }
+}
+
+fn is_sorted<I: Iterator<Item = DocId>>(mut it: I) -> bool {
+    if let Some(first) = it.next() {
+        let mut prev = first;
+        for doc in it {
+            if doc < prev {
+                return false;
+            }
+            prev = doc;
+        }
+    }
+    true
+}
 #[cfg(test)]
 mod tests {
     use crate::query::score_combiner::SumCombiner;
@@ -248,17 +246,21 @@ mod tests {
     use std::iter;
 
     struct Float(Score);
+
     impl Eq for Float {}
+
     impl PartialEq for Float {
         fn eq(&self, other: &Self) -> bool {
             self.cmp(&other) == Ordering::Equal
         }
     }
+
     impl PartialOrd for Float {
         fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
             Some(self.cmp(other))
         }
     }
+
     impl Ord for Float {
         fn cmp(&self, other: &Self) -> Ordering {
             other.0.partial_cmp(&self.0).unwrap_or(Ordering::Equal)

src/query/boolean_query/mod.rs

@@ -32,7 +32,7 @@ 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();
+            let mut index_writer = index.writer_for_tests().unwrap();
             {
                 index_writer.add_document(doc!(text_field => "a b c"));
                 index_writer.add_document(doc!(text_field => "a c"));
@@ -224,7 +224,7 @@ mod tests {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(text_field => "a b c"));
             index_writer.add_document(doc!(text_field => "a c"));
             index_writer.add_document(doc!(text_field => "b c"));

src/query/boost_query.rs

@@ -144,7 +144,7 @@ mod tests {
     fn test_boost_query_explain() {
         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_for_tests().unwrap();
         index_writer.add_document(Document::new());
         assert!(index_writer.commit().is_ok());
         let reader = index.reader().unwrap();

src/query/fuzzy_query.rs

@@ -177,7 +177,7 @@ mod test {
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         {
-            let mut index_writer = index.writer_with_num_threads(1, 10_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             index_writer.add_document(doc!(
                 country_field => "japan",
             ));

src/query/phrase_query/mod.rs

@@ -24,7 +24,7 @@ pub mod tests {
         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_for_tests().unwrap();
             for &text in texts {
                 let doc = doc!(text_field=>text);
                 index_writer.add_document(doc);
@@ -135,7 +135,7 @@ pub mod tests {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(text_field=>"a b c"));
             assert!(index_writer.commit().is_ok());
         }
@@ -186,7 +186,7 @@ pub mod tests {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(text_field=>"b"));
             index_writer.add_document(doc!(text_field=>"a b"));
             index_writer.add_document(doc!(text_field=>"b a"));
@@ -217,7 +217,7 @@ pub mod tests {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(text_field=>"a b c d e f g h"));
             assert!(index_writer.commit().is_ok());
         }

src/query/phrase_query/phrase_weight.rs

@@ -9,8 +9,8 @@ use crate::query::Weight;
 use crate::query::{EmptyScorer, Explanation};
 use crate::schema::IndexRecordOption;
 use crate::schema::Term;
+use crate::Score;
 use crate::{DocId, DocSet};
-use crate::{Result, Score};
 
 pub struct PhraseWeight {
     phrase_terms: Vec<(usize, Term)>,
@@ -32,7 +32,7 @@ impl PhraseWeight {
         }
     }
 
-    fn fieldnorm_reader(&self, reader: &SegmentReader) -> FieldNormReader {
+    fn fieldnorm_reader(&self, reader: &SegmentReader) -> crate::Result<FieldNormReader> {
         let field = self.phrase_terms[0].1.field();
         reader.get_fieldnorms_reader(field)
     }
@@ -41,9 +41,9 @@ impl PhraseWeight {
         &self,
         reader: &SegmentReader,
         boost: Score,
-    ) -> Result<Option<PhraseScorer<SegmentPostings>>> {
+    ) -> crate::Result<Option<PhraseScorer<SegmentPostings>>> {
         let similarity_weight = self.similarity_weight.boost_by(boost);
-        let fieldnorm_reader = self.fieldnorm_reader(reader);
+        let fieldnorm_reader = self.fieldnorm_reader(reader)?;
         if reader.has_deletes() {
             let mut term_postings_list = Vec::new();
             for &(offset, ref term) in &self.phrase_terms {
@@ -85,7 +85,7 @@ impl PhraseWeight {
 }
 
 impl Weight for PhraseWeight {
-    fn scorer(&self, reader: &SegmentReader, boost: Score) -> Result<Box<dyn Scorer>> {
+    fn scorer(&self, reader: &SegmentReader, boost: Score) -> crate::Result<Box<dyn Scorer>> {
         if let Some(scorer) = self.phrase_scorer(reader, boost)? {
             Ok(Box::new(scorer))
         } else {
@@ -93,7 +93,7 @@ impl Weight for PhraseWeight {
         }
     }
 
-    fn explain(&self, reader: &SegmentReader, doc: DocId) -> Result<Explanation> {
+    fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
         let scorer_opt = self.phrase_scorer(reader, 1.0)?;
         if scorer_opt.is_none() {
             return Err(does_not_match(doc));
@@ -102,7 +102,7 @@ impl Weight for PhraseWeight {
         if scorer.seek(doc) != doc {
             return Err(does_not_match(doc));
         }
-        let fieldnorm_reader = self.fieldnorm_reader(reader);
+        let fieldnorm_reader = self.fieldnorm_reader(reader)?;
         let fieldnorm_id = fieldnorm_reader.fieldnorm_id(doc);
         let phrase_count = scorer.phrase_count();
         let mut explanation = Explanation::new("Phrase Scorer", scorer.score());

src/query/range_query.rs

@@ -9,7 +9,6 @@ use crate::query::{Query, Scorer, Weight};
 use crate::schema::Type;
 use crate::schema::{Field, IndexRecordOption, Term};
 use crate::termdict::{TermDictionary, TermStreamer};
-use crate::Result;
 use crate::{DocId, Score};
 use std::collections::Bound;
 use std::ops::Range;
@@ -48,7 +47,7 @@ fn map_bound<TFrom, TTo, Transform: Fn(&TFrom) -> TTo>(
 /// let schema = schema_builder.build();
 ///
 /// let index = Index::create_in_ram(schema);
-/// let mut index_writer = index.writer_with_num_threads(1, 6_000_000)?;
+/// let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
 /// for year in 1950u64..2017u64 {
 ///     let num_docs_within_year = 10 + (year - 1950) * (year - 1950);
 ///     for _ in 0..num_docs_within_year {
@@ -246,7 +245,11 @@ impl RangeQuery {
 }
 
 impl Query for RangeQuery {
-    fn weight(&self, searcher: &Searcher, _scoring_enabled: bool) -> Result<Box<dyn Weight>> {
+    fn weight(
+        &self,
+        searcher: &Searcher,
+        _scoring_enabled: bool,
+    ) -> crate::Result<Box<dyn Weight>> {
         let schema = searcher.schema();
         let value_type = schema.get_field_entry(self.field).field_type().value_type();
         if value_type != self.value_type {
@@ -289,7 +292,7 @@ impl RangeWeight {
 }
 
 impl Weight for RangeWeight {
-    fn scorer(&self, reader: &SegmentReader, boost: Score) -> Result<Box<dyn Scorer>> {
+    fn scorer(&self, reader: &SegmentReader, boost: Score) -> crate::Result<Box<dyn Scorer>> {
         let max_doc = reader.max_doc();
         let mut doc_bitset = BitSet::with_max_value(max_doc);
 
@@ -315,7 +318,7 @@ impl Weight for RangeWeight {
         Ok(Box::new(ConstScorer::new(doc_bitset, boost)))
     }
 
-    fn explain(&self, reader: &SegmentReader, doc: DocId) -> Result<Explanation> {
+    fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
         let mut scorer = self.scorer(reader, 1.0)?;
         if scorer.seek(doc) != doc {
             return Err(does_not_match(doc));
@@ -342,7 +345,7 @@ mod tests {
 
         let index = Index::create_in_ram(schema);
         {
-            let mut index_writer = index.writer_with_num_threads(1, 6_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             for year in 1950u64..2017u64 {
                 let num_docs_within_year = 10 + (year - 1950) * (year - 1950);
                 for _ in 0..num_docs_within_year {
@@ -485,7 +488,7 @@ mod tests {
         schema_builder.add_i64_field("year", INDEXED);
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema.clone());
-        let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
+        let mut index_writer = index.writer_for_tests()?;
         let title = schema.get_field("title").unwrap();
         let year = schema.get_field("year").unwrap();
         index_writer.add_document(doc!(

src/query/regex_query.rs

@@ -103,7 +103,7 @@ mod test {
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         {
-            let mut index_writer = index.writer_with_num_threads(1, 10_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             index_writer.add_document(doc!(
                 country_field => "japan",
             ));

src/query/term_query/mod.rs

@@ -25,7 +25,7 @@ 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();
+            let mut index_writer = index.writer_for_tests().unwrap();
             let doc = doc!(text_field => "a");
             index_writer.add_document(doc);
             assert!(index_writer.commit().is_ok());
@@ -50,7 +50,7 @@ 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)?;
+            let mut index_writer = index.writer_for_tests()?;
             for _ in 0..COMPRESSION_BLOCK_SIZE {
                 let doc = doc!(text_field => "a");
                 index_writer.add_document(doc);
@@ -86,7 +86,7 @@ mod tests {
         let schema = schema_builder.build();
         let index = Index::create_in_ram(schema);
         {
-            let mut index_writer = index.writer_with_num_threads(1, 10_000_000).unwrap();
+            let mut index_writer = index.writer_for_tests().unwrap();
             index_writer.add_document(doc!(
                 left_field => "left1 left2 left2 left2f2 left2f2 left3 abcde abcde abcde abcde abcde abcde abcde abcde abcde abcewde abcde abcde",
                 right_field => "right1 right2",
@@ -136,7 +136,7 @@ mod tests {
         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, 5_000_000).unwrap();
+        let mut index_writer = index.writer_for_tests().unwrap();
         index_writer.add_document(doc!(text_field=>"a b"));
         index_writer.add_document(doc!(text_field=>"a c"));
         index_writer.delete_term(Term::from_field_text(text_field, "b"));
@@ -153,7 +153,7 @@ mod tests {
         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_for_tests().unwrap();
         index_writer.add_document(doc!(text_field=>"a"));
         index_writer.add_document(doc!(text_field=>"a"));
         index_writer.commit()?;

src/query/term_query/term_weight.rs

@@ -4,11 +4,10 @@ use crate::docset::DocSet;
 use crate::postings::SegmentPostings;
 use crate::query::bm25::BM25Weight;
 use crate::query::explanation::does_not_match;
-use crate::query::weight::{for_each_pruning_scorer, for_each_scorer};
+use crate::query::weight::for_each_scorer;
 use crate::query::Weight;
 use crate::query::{Explanation, Scorer};
 use crate::schema::IndexRecordOption;
-use crate::Result;
 use crate::Term;
 use crate::{DocId, Score};
 
@@ -19,12 +18,12 @@ pub struct TermWeight {
 }
 
 impl Weight for TermWeight {
-    fn scorer(&self, reader: &SegmentReader, boost: Score) -> Result<Box<dyn Scorer>> {
+    fn scorer(&self, reader: &SegmentReader, boost: Score) -> crate::Result<Box<dyn Scorer>> {
         let term_scorer = self.specialized_scorer(reader, boost)?;
         Ok(Box::new(term_scorer))
     }
 
-    fn explain(&self, reader: &SegmentReader, doc: DocId) -> Result<Explanation> {
+    fn explain(&self, reader: &SegmentReader, doc: DocId) -> crate::Result<Explanation> {
         let mut scorer = self.specialized_scorer(reader, 1.0)?;
         if scorer.seek(doc) != doc {
             return Err(does_not_match(doc));
@@ -32,7 +31,7 @@ impl Weight for TermWeight {
         Ok(scorer.explain())
     }
 
-    fn count(&self, reader: &SegmentReader) -> Result<u32> {
+    fn count(&self, reader: &SegmentReader) -> crate::Result<u32> {
         if let Some(delete_bitset) = reader.delete_bitset() {
             Ok(self.scorer(reader, 1.0)?.count(delete_bitset))
         } else {
@@ -73,8 +72,8 @@ impl Weight for TermWeight {
         reader: &SegmentReader,
         callback: &mut dyn FnMut(DocId, Score) -> Score,
     ) -> crate::Result<()> {
-        let mut scorer = self.scorer(reader, 1.0)?;
+        let scorer = self.specialized_scorer(reader, 1.0)?;
-        for_each_pruning_scorer(&mut scorer, threshold, callback);
+        crate::query::boolean_query::block_wand(vec![scorer], threshold, callback);
         Ok(())
     }
 }
@@ -96,10 +95,10 @@ impl TermWeight {
         &self,
         reader: &SegmentReader,
         boost: Score,
-    ) -> Result<TermScorer> {
+    ) -> crate::Result<TermScorer> {
         let field = self.term.field();
         let inverted_index = reader.inverted_index(field);
-        let fieldnorm_reader = reader.get_fieldnorms_reader(field);
+        let fieldnorm_reader = reader.get_fieldnorms_reader(field)?;
         let similarity_weight = self.similarity_weight.boost_by(boost);
         let postings_opt: Option<SegmentPostings> =
             inverted_index.read_postings(&self.term, self.index_record_option);

src/query/union.rs

@@ -398,9 +398,9 @@ mod bench {
 
     use crate::query::score_combiner::DoNothingCombiner;
     use crate::query::{ConstScorer, Union, VecDocSet};
-    use crate::tests;
     use crate::DocId;
     use crate::DocSet;
+    use crate::{tests, TERMINATED};
     use test::Bencher;
 
     #[bench]
@@ -414,10 +414,12 @@ mod bench {
                 union_docset
                     .iter()
                     .map(|doc_ids| VecDocSet::from(doc_ids.clone()))
-                    .map(ConstScorer::new)
+                    .map(|docset| ConstScorer::new(docset, 1.0))
                     .collect::<Vec<_>>(),
             );
-            while v.advance() {}
+            while v.doc() != TERMINATED {
+                v.advance();
+            }
         });
     }
     #[bench]
@@ -432,10 +434,12 @@ mod bench {
                 union_docset
                     .iter()
                     .map(|doc_ids| VecDocSet::from(doc_ids.clone()))
-                    .map(ConstScorer::new)
+                    .map(|docset| ConstScorer::new(docset, 1.0))
                     .collect::<Vec<_>>(),
             );
-            while v.advance() {}
+            while v.doc() != TERMINATED {
+                v.advance();
+            }
         });
     }
 }

src/reader/mod.rs

@@ -138,9 +138,11 @@ impl InnerIndexReader {
                 .collect::<crate::Result<_>>()?
         };
         let schema = self.index.schema();
-        let searchers = (0..self.num_searchers)
+        let searchers = std::iter::repeat_with(|| {
-            .map(|_| Searcher::new(schema.clone(), self.index.clone(), segment_readers.clone()))
+            Searcher::new(schema.clone(), self.index.clone(), segment_readers.clone())
-            .collect();
+        })
+        .take(self.num_searchers)
+        .collect();
         self.searcher_pool.publish_new_generation(searchers);
         Ok(())
     }

src/schema/document.rs

@@ -74,9 +74,8 @@ impl Document {
     }
 
     /// Add a text field.
-    pub fn add_text(&mut self, field: Field, text: &str) {
+    pub fn add_text<S: ToString>(&mut self, field: Field, text: S) {
-        let value = Value::Str(String::from(text));
+        self.add(FieldValue::new(field, Value::Str(text.to_string())));
-        self.add(FieldValue::new(field, value));
     }
 
     /// Add a pre-tokenized text field.
@@ -110,8 +109,8 @@ impl Document {
     }
 
     /// Add a bytes field
-    pub fn add_bytes(&mut self, field: Field, value: Vec<u8>) {
+    pub fn add_bytes<T: Into<Vec<u8>>>(&mut self, field: Field, value: T) {
-        self.add(FieldValue::new(field, Value::Bytes(value)))
+        self.add(FieldValue::new(field, Value::Bytes(value.into())))
     }
 
     /// Add a field value

src/schema/field_entry.rs

@@ -1,5 +1,5 @@
-use crate::schema::IntOptions;
 use crate::schema::TextOptions;
+use crate::schema::{is_valid_field_name, IntOptions};
 
 use crate::schema::FieldType;
 use serde::de::{self, MapAccess, Visitor};
@@ -24,6 +24,7 @@ impl FieldEntry {
     /// Creates a new u64 field entry in the schema, given
     /// a name, and some options.
     pub fn new_text(field_name: String, text_options: TextOptions) -> FieldEntry {
+        assert!(is_valid_field_name(&field_name));
         FieldEntry {
             name: field_name,
             field_type: FieldType::Str(text_options),
@@ -33,6 +34,7 @@ impl FieldEntry {
     /// Creates a new u64 field entry in the schema, given
     /// a name, and some options.
     pub fn new_u64(field_name: String, field_type: IntOptions) -> FieldEntry {
+        assert!(is_valid_field_name(&field_name));
         FieldEntry {
             name: field_name,
             field_type: FieldType::U64(field_type),
@@ -42,6 +44,7 @@ impl FieldEntry {
     /// Creates a new i64 field entry in the schema, given
     /// a name, and some options.
     pub fn new_i64(field_name: String, field_type: IntOptions) -> FieldEntry {
+        assert!(is_valid_field_name(&field_name));
         FieldEntry {
             name: field_name,
             field_type: FieldType::I64(field_type),
@@ -51,6 +54,7 @@ impl FieldEntry {
     /// Creates a new f64 field entry in the schema, given
     /// a name, and some options.
     pub fn new_f64(field_name: String, field_type: IntOptions) -> FieldEntry {
+        assert!(is_valid_field_name(&field_name));
         FieldEntry {
             name: field_name,
             field_type: FieldType::F64(field_type),
@@ -60,6 +64,7 @@ impl FieldEntry {
     /// Creates a new date field entry in the schema, given
     /// a name, and some options.
     pub fn new_date(field_name: String, field_type: IntOptions) -> FieldEntry {
+        assert!(is_valid_field_name(&field_name));
         FieldEntry {
             name: field_name,
             field_type: FieldType::Date(field_type),
@@ -68,6 +73,7 @@ impl FieldEntry {
 
     /// Creates a field entry for a facet.
     pub fn new_facet(field_name: String) -> FieldEntry {
+        assert!(is_valid_field_name(&field_name));
         FieldEntry {
             name: field_name,
             field_type: FieldType::HierarchicalFacet,
@@ -76,6 +82,7 @@ impl FieldEntry {
 
     /// Creates a field entry for a bytes field
     pub fn new_bytes(field_name: String) -> FieldEntry {
+        assert!(is_valid_field_name(&field_name));
         FieldEntry {
             name: field_name,
             field_type: FieldType::Bytes,
@@ -268,6 +275,12 @@ mod tests {
     use crate::schema::TEXT;
     use serde_json;
 
+    #[test]
+    #[should_panic]
+    fn test_invalid_field_name_should_panic() {
+        FieldEntry::new_text("-hello".to_string(), TEXT);
+    }
+
     #[test]
     fn test_json_serialization() {
         let field_value = FieldEntry::new_text(String::from("title"), TEXT);

src/schema/mod.rs

@@ -149,14 +149,16 @@ pub use self::int_options::IntOptions;
 use once_cell::sync::Lazy;
 use regex::Regex;
 
+/// Regular expression representing the restriction on a valid field names.
+pub const FIELD_NAME_PATTERN: &str = r#"^[_a-zA-Z][_\-a-zA-Z0-9]*$"#;
+
 /// Validator for a potential `field_name`.
 /// Returns true iff the name can be use for a field name.
 ///
 /// A field name must start by a letter `[a-zA-Z]`.
 /// The other characters can be any alphanumic character `[a-ZA-Z0-9]` or `_`.
 pub fn is_valid_field_name(field_name: &str) -> bool {
-    static FIELD_NAME_PTN: Lazy<Regex> =
+    static FIELD_NAME_PTN: Lazy<Regex> = Lazy::new(|| Regex::new(FIELD_NAME_PATTERN).unwrap());
-        Lazy::new(|| Regex::new("^[a-zA-Z][_a-zA-Z0-9]*$").unwrap());
     FIELD_NAME_PTN.is_match(field_name)
 }
 
@@ -170,6 +172,11 @@ mod tests {
         assert!(is_valid_field_name("text"));
         assert!(is_valid_field_name("text0"));
         assert!(!is_valid_field_name("0text"));
+        assert!(is_valid_field_name("field-name"));
+        assert!(is_valid_field_name("field_name"));
+        assert!(!is_valid_field_name("field!name"));
+        assert!(!is_valid_field_name("-fieldname"));
+        assert!(is_valid_field_name("_fieldname"));
         assert!(!is_valid_field_name(""));
         assert!(!is_valid_field_name("シャボン玉"));
         assert!(is_valid_field_name("my_text_field"));

src/schema/term.rs

@@ -4,7 +4,6 @@ use super::Field;
 use crate::common;
 use crate::schema::Facet;
 use crate::DateTime;
-use byteorder::{BigEndian, ByteOrder};
 use std::str;
 
 /// Size (in bytes) of the buffer of a int field.
@@ -19,6 +18,10 @@ where
     B: AsRef<[u8]>;
 
 impl Term {
+    pub(crate) fn new() -> Term {
+        Term(Vec::with_capacity(100))
+    }
+
     /// Builds a term given a field, and a i64-value
     ///
     /// Assuming the term has a field id of 1, and a i64 value of 3234,
@@ -93,6 +96,12 @@ impl Term {
         term
     }
 
+    pub(crate) fn from_field_bytes(field: Field, bytes: &[u8]) -> Term {
+        let mut term = Term::for_field(field);
+        term.set_bytes(bytes);
+        term
+    }
+
     /// Creates a new Term for a given field.
     pub(crate) fn for_field(field: Field) -> Term {
         let mut term = Term(Vec::with_capacity(100));
@@ -100,12 +109,10 @@ impl Term {
         term
     }
 
-    /// Returns the field.
+    pub(crate) fn set_field(&mut self, field: Field) {
-    pub fn set_field(&mut self, field: Field) {
+        self.0.clear();
-        if self.0.len() < 4 {
+        self.0
-            self.0.resize(4, 0u8);
+            .extend_from_slice(&field.field_id().to_be_bytes()[..]);
-        }
-        BigEndian::write_u32(&mut self.0[0..4], field.field_id());
     }
 
     /// Sets a u64 value in the term.
@@ -116,7 +123,7 @@ impl Term {
     /// the natural order of the values.
     pub fn set_u64(&mut self, val: u64) {
         self.0.resize(INT_TERM_LEN, 0u8);
-        BigEndian::write_u64(&mut self.0[4..], val);
+        self.0[4..12].copy_from_slice(val.to_be_bytes().as_ref());
     }
 
     /// Sets a `i64` value in the term.
@@ -134,12 +141,6 @@ impl Term {
         self.0.extend(bytes);
     }
 
-    pub(crate) fn from_field_bytes(field: Field, bytes: &[u8]) -> Term {
-        let mut term = Term::for_field(field);
-        term.set_bytes(bytes);
-        term
-    }
-
     /// Set the texts only, keeping the field untouched.
     pub fn set_text(&mut self, text: &str) {
         self.set_bytes(text.as_bytes());
@@ -157,7 +158,9 @@ where
 
     /// Returns the field.
     pub fn field(&self) -> Field {
-        Field::from_field_id(BigEndian::read_u32(&self.0.as_ref()[..4]))
+        let mut field_id_bytes = [0u8; 4];
+        field_id_bytes.copy_from_slice(&self.0.as_ref()[..4]);
+        Field::from_field_id(u32::from_be_bytes(field_id_bytes))
     }
 
     /// Returns the `u64` value stored in a term.
@@ -166,7 +169,9 @@ where
     /// ... or returns an invalid value
     /// if the term is not a `u64` field.
     pub fn get_u64(&self) -> u64 {
-        BigEndian::read_u64(&self.0.as_ref()[4..])
+        let mut field_id_bytes = [0u8; 8];
+        field_id_bytes.copy_from_slice(self.value_bytes());
+        u64::from_be_bytes(field_id_bytes)
     }
 
     /// Returns the `i64` value stored in a term.
@@ -175,7 +180,7 @@ where
     /// ... or returns an invalid value
     /// if the term is not a `i64` field.
     pub fn get_i64(&self) -> i64 {
-        common::u64_to_i64(BigEndian::read_u64(&self.0.as_ref()[4..]))
+        common::u64_to_i64(self.get_u64())
     }
 
     /// Returns the `f64` value stored in a term.
@@ -184,7 +189,7 @@ where
     /// ... or returns an invalid value
     /// if the term is not a `f64` field.
     pub fn get_f64(&self) -> f64 {
-        common::u64_to_f64(BigEndian::read_u64(&self.0.as_ref()[4..]))
+        common::u64_to_f64(self.get_u64())
     }
 
     /// Returns the text associated with the term.

src/schema/value.rs

@@ -221,6 +221,12 @@ impl<'a> From<&'a str> for Value {
     }
 }
 
+impl<'a> From<&'a [u8]> for Value {
+    fn from(bytes: &'a [u8]) -> Value {
+        Value::Bytes(bytes.to_vec())
+    }
+}
+
 impl<'a> From<Facet> for Value {
     fn from(facet: Facet) -> Value {
         Value::Facet(facet)

src/snippet/mod.rs

@@ -221,7 +221,7 @@ fn select_best_fragment_combination(fragments: &[FragmentCandidate], text: &str)
 /// #    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, 30_000_000)?;
+/// #    let mut index_writer = index.writer_with_num_threads(1, 10_000_000)?;
 /// #    let doc = doc!(text_field => r#"Comme je descendais des Fleuves impassibles,
 /// #   Je ne me sentis plus guidé par les haleurs :
 /// #  Des Peaux-Rouges criards les avaient pris pour cibles,
@@ -506,7 +506,7 @@ Survey in 2016, 2017, and 2018."#;
         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_for_tests().unwrap();
             index_writer.add_document(doc!(text_field => "a"));
             index_writer.add_document(doc!(text_field => "a"));
             index_writer.add_document(doc!(text_field => "a b"));
@@ -562,7 +562,7 @@ Survey in 2016, 2017, and 2018."#;
         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_for_tests().unwrap();
             {
                 let doc = doc ! (text_field => TEST_TEXT);
                 index_writer.add_document(doc);

src/space_usage/mod.rs

@@ -25,6 +25,8 @@ pub enum ComponentSpaceUsage {
     Store(StoreSpaceUsage),
     /// Some sort of raw byte count
     Basic(ByteCount),
+    ///
+    Unimplemented,
 }
 
 /// Represents combined space usage of an entire searcher and its component segments.
@@ -119,7 +121,7 @@ impl SegmentSpaceUsage {
     /// Clones the underlying data.
     /// Use the components directly if this is somehow in performance critical code.
     pub fn component(&self, component: SegmentComponent) -> ComponentSpaceUsage {
-        use self::ComponentSpaceUsage::*;
+        use self::ComponentSpaceUsage::{Basic, PerField, Store, Unimplemented};
         use crate::SegmentComponent::*;
         match component {
             POSTINGS => PerField(self.postings().clone()),
@@ -130,6 +132,7 @@ impl SegmentSpaceUsage {
             TERMS => PerField(self.termdict().clone()),
             STORE => Store(self.store().clone()),
             DELETE => Basic(self.deletes()),
+            FIELDSTATS => Unimplemented,
         }
     }
 
@@ -336,7 +339,7 @@ mod test {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(name => 1u64));
             index_writer.add_document(doc!(name => 2u64));
             index_writer.add_document(doc!(name => 10u64));
@@ -374,7 +377,7 @@ mod test {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(name => "hi"));
             index_writer.add_document(doc!(name => "this is a test"));
             index_writer.add_document(
@@ -414,7 +417,7 @@ mod test {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(name => "hi"));
             index_writer.add_document(doc!(name => "this is a test"));
             index_writer.add_document(
@@ -453,7 +456,7 @@ mod test {
         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_for_tests().unwrap();
             index_writer.add_document(doc!(name => 1u64));
             index_writer.add_document(doc!(name => 2u64));
             index_writer.add_document(doc!(name => 3u64));

src/termdict/mod.rs

@@ -44,11 +44,13 @@ mod tests {
 
     const BLOCK_SIZE: usize = 1_500;
 
-    fn make_term_info(val: u64) -> TermInfo {
+    fn make_term_info(term_ord: u64) -> TermInfo {
+        let offset = |term_ord: u64| term_ord * 100 + term_ord * term_ord;
         TermInfo {
-            doc_freq: val as u32,
+            doc_freq: term_ord as u32,
-            positions_idx: val * 2u64,
+            postings_start_offset: offset(term_ord),
-            postings_offset: val * 3u64,
+            postings_end_offset: offset(term_ord + 1),
+            positions_idx: offset(term_ord) * 2u64,
         }
     }
 
@@ -138,7 +140,7 @@ mod tests {
         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_for_tests().unwrap();
             {
                 {
                     let mut doc = Document::default();
@@ -208,20 +210,14 @@ mod tests {
     }
 
     #[test]
-    fn test_stream_high_range_prefix_suffix() {
+    fn test_stream_high_range_prefix_suffix() -> std::io::Result<()> {
         let buffer: Vec<u8> = {
             let mut term_dictionary_builder = TermDictionaryBuilder::create(vec![]).unwrap();
             // term requires more than 16bits
-            term_dictionary_builder
+            term_dictionary_builder.insert("abcdefghijklmnopqrstuvwxy", &make_term_info(1))?;
-                .insert("abcdefghijklmnopqrstuvwxy", &make_term_info(1))
+            term_dictionary_builder.insert("abcdefghijklmnopqrstuvwxyz", &make_term_info(2))?;
-                .unwrap();
+            term_dictionary_builder.insert("abr", &make_term_info(3))?;
-            term_dictionary_builder
+            term_dictionary_builder.finish()?
-                .insert("abcdefghijklmnopqrstuvwxyz", &make_term_info(2))
-                .unwrap();
-            term_dictionary_builder
-                .insert("abr", &make_term_info(2))
-                .unwrap();
-            term_dictionary_builder.finish().unwrap()
         };
         let source = ReadOnlySource::from(buffer);
         let term_dictionary: TermDictionary = TermDictionary::from_source(&source);
@@ -229,12 +225,15 @@ mod tests {
         assert!(kv_stream.advance());
         assert_eq!(kv_stream.key(), "abcdefghijklmnopqrstuvwxy".as_bytes());
         assert_eq!(kv_stream.value(), &make_term_info(1));
+        dbg!(make_term_info(1));
         assert!(kv_stream.advance());
         assert_eq!(kv_stream.key(), "abcdefghijklmnopqrstuvwxyz".as_bytes());
         assert_eq!(kv_stream.value(), &make_term_info(2));
         assert!(kv_stream.advance());
         assert_eq!(kv_stream.key(), "abr".as_bytes());
+        assert_eq!(kv_stream.value(), &make_term_info(3));
         assert!(!kv_stream.advance());
+        Ok(())
     }
 
     #[test]

src/termdict/term_info_store.rs

@@ -57,21 +57,28 @@ impl TermInfoBlockMeta {
         self.doc_freq_nbits + self.postings_offset_nbits + self.positions_idx_nbits
     }
 
+    // Here inner_offset is the offset within the block, WITHOUT the first term_info.
+    // In other word, term_info #1,#2,#3 gets inner_offset 0,1,2... While term_info #0
+    // is encoded without bitpacking.
     fn deserialize_term_info(&self, data: &[u8], inner_offset: usize) -> TermInfo {
+        assert!(inner_offset < BLOCK_LEN - 1);
         let num_bits = self.num_bits() as usize;
         let mut cursor = num_bits * inner_offset;
 
+        let postings_start_offset = extract_bits(data, cursor, self.postings_offset_nbits);
+        let postings_end_offset = self.ref_term_info.postings_start_offset
+            + extract_bits(data, cursor + num_bits, self.postings_offset_nbits);
+        cursor += self.postings_offset_nbits as usize;
+
         let doc_freq = extract_bits(data, cursor, self.doc_freq_nbits) as u32;
         cursor += self.doc_freq_nbits as usize;
 
-        let postings_offset = extract_bits(data, cursor, self.postings_offset_nbits);
-        cursor += self.postings_offset_nbits as usize;
-
         let positions_idx = extract_bits(data, cursor, self.positions_idx_nbits);
 
         TermInfo {
             doc_freq,
-            postings_offset: postings_offset + self.ref_term_info.postings_offset,
+            postings_start_offset: postings_start_offset + self.ref_term_info.postings_start_offset,
+            postings_end_offset,
             positions_idx: positions_idx + self.ref_term_info.positions_idx,
         }
     }
@@ -126,14 +133,13 @@ impl TermInfoStore {
             .expect("Failed to deserialize terminfoblockmeta");
         let inner_offset = (term_ord as usize) % BLOCK_LEN;
         if inner_offset == 0 {
-            term_info_block_data.ref_term_info
+            return term_info_block_data.ref_term_info;
-        } else {
-            let term_info_data = self.term_info_source.as_slice();
-            term_info_block_data.deserialize_term_info(
-                &term_info_data[term_info_block_data.offset as usize..],
-                inner_offset - 1,
-            )
         }
+        let term_info_data = self.term_info_source.as_slice();
+        term_info_block_data.deserialize_term_info(
+            &term_info_data[term_info_block_data.offset as usize..],
+            inner_offset - 1,
+        )
     }
 
     pub fn num_terms(&self) -> usize {
@@ -154,16 +160,17 @@ fn bitpack_serialize<W: Write>(
     term_info_block_meta: &TermInfoBlockMeta,
     term_info: &TermInfo,
 ) -> io::Result<()> {
+    bit_packer.write(
+        term_info.postings_start_offset,
+        term_info_block_meta.postings_offset_nbits,
+        write,
+    )?;
     bit_packer.write(
         u64::from(term_info.doc_freq),
         term_info_block_meta.doc_freq_nbits,
         write,
     )?;
-    bit_packer.write(
+
-        term_info.postings_offset,
-        term_info_block_meta.postings_offset_nbits,
-        write,
-    )?;
     bit_packer.write(
         term_info.positions_idx,
         term_info_block_meta.positions_idx_nbits,
@@ -183,23 +190,27 @@ impl TermInfoStoreWriter {
     }
 
     fn flush_block(&mut self) -> io::Result<()> {
-        if self.term_infos.is_empty() {
-            return Ok(());
-        }
         let mut bit_packer = BitPacker::new();
         let ref_term_info = self.term_infos[0].clone();
+
+        let last_term_info = if let Some(last_term_info) = self.term_infos.last().cloned() {
+            last_term_info
+        } else {
+            return Ok(());
+        };
+        let postings_end_offset =
+            last_term_info.postings_end_offset - ref_term_info.postings_start_offset;
         for term_info in &mut self.term_infos[1..] {
-            term_info.postings_offset -= ref_term_info.postings_offset;
+            term_info.postings_start_offset -= ref_term_info.postings_start_offset;
             term_info.positions_idx -= ref_term_info.positions_idx;
         }
 
         let mut max_doc_freq: u32 = 0u32;
-        let mut max_postings_offset: u64 = 0u64;
+        let max_postings_offset: u64 = postings_end_offset;
-        let mut max_positions_idx: u64 = 0u64;
+        let max_positions_idx: u64 = last_term_info.positions_idx;
+
         for term_info in &self.term_infos[1..] {
             max_doc_freq = cmp::max(max_doc_freq, term_info.doc_freq);
-            max_postings_offset = cmp::max(max_postings_offset, term_info.postings_offset);
-            max_positions_idx = cmp::max(max_positions_idx, term_info.positions_idx);
         }
 
         let max_doc_freq_nbits: u8 = compute_num_bits(u64::from(max_doc_freq));
@@ -224,6 +235,12 @@ impl TermInfoStoreWriter {
             )?;
         }
 
+        bit_packer.write(
+            postings_end_offset,
+            term_info_block_meta.postings_offset_nbits,
+            &mut self.buffer_term_infos,
+        )?;
+
         // Block need end up at the end of a byte.
         bit_packer.flush(&mut self.buffer_term_infos)?;
         self.term_infos.clear();
@@ -232,6 +249,7 @@ impl TermInfoStoreWriter {
     }
 
     pub fn write_term_info(&mut self, term_info: &TermInfo) -> io::Result<()> {
+        assert!(term_info.postings_end_offset >= term_info.postings_start_offset);
         self.num_terms += 1u64;
         self.term_infos.push(term_info.clone());
         if self.term_infos.len() >= BLOCK_LEN {
@@ -291,10 +309,11 @@ mod tests {
     #[test]
     fn test_term_info_block_meta_serialization() {
         let term_info_block_meta = TermInfoBlockMeta {
-            offset: 2009,
+            offset: 2009u64,
             ref_term_info: TermInfo {
                 doc_freq: 512,
-                postings_offset: 51,
+                postings_start_offset: 51,
+                postings_end_offset: 57u64,
                 positions_idx: 3584,
             },
             doc_freq_nbits: 10,
@@ -312,10 +331,12 @@ mod tests {
     fn test_pack() {
         let mut store_writer = TermInfoStoreWriter::new();
         let mut term_infos = vec![];
+        let offset = |i| (i * 13 + i * i) as u64;
         for i in 0..1000 {
             let term_info = TermInfo {
                 doc_freq: i as u32,
-                postings_offset: (i / 10) as u64,
+                postings_start_offset: offset(i),
+                postings_end_offset: offset(i + 1),
                 positions_idx: (i * 7) as u64,
             };
             store_writer.write_term_info(&term_info).unwrap();
@@ -325,7 +346,12 @@ mod tests {
         store_writer.serialize(&mut buffer).unwrap();
         let term_info_store = TermInfoStore::open(&ReadOnlySource::from(buffer));
         for i in 0..1000 {
-            assert_eq!(term_info_store.get(i as u64), term_infos[i]);
+            assert_eq!(
+                term_info_store.get(i as u64),
+                term_infos[i],
+                "term info {}",
+                i
+            );
         }
     }
 }