proxy(tokio-postgres): refactor typeinfo query to occur earlier (#11993)

## Problem

For #11992 I realised we need to get the type info before executing the
query. This is important to know how to decode rows with custom types,
eg the following query:

```sql
CREATE TYPE foo AS ENUM ('foo','bar','baz');
SELECT ARRAY['foo'::foo, 'bar'::foo, 'baz'::foo] AS data;
```

Getting that to work was harder that it seems. The original
tokio-postgres setup has a split between `Client` and `Connection`,
where messages are passed between. Because multiple clients were
supported, each client message included a dedicated response channel.
Each request would be terminated by the `ReadyForQuery` message.

The flow I opted to use for parsing types early would not trigger a
`ReadyForQuery`. The flow is as follows:

```
PARSE ""    // parse the user provided query
DESCRIBE "" // describe the query, returning param/result type oids
FLUSH       // force postgres to flush the responses early

// wait for descriptions

  // check if we know the types, if we don't then
  // setup the typeinfo query and execute it against each OID:

  PARSE typeinfo    // prepare our typeinfo query
  DESCRIBE typeinfo
  FLUSH // force postgres to flush the responses early

  // wait for typeinfo statement

    // for each OID we don't know:
    BIND typeinfo
    EXECUTE
    FLUSH

    // wait for type info, might reveal more OIDs to inspect

  // close the typeinfo query, we cache the OID->type map and this is kinder to pgbouncer.
  CLOSE typeinfo 

// finally once we know all the OIDs:
BIND ""   // bind the user provided query - already parsed - to the user provided params
EXECUTE   // run the user provided query
SYNC      // commit the transaction
```

## Summary of changes

Please review commit by commit. The main challenge was allowing one
query to issue multiple sub-queries. To do this I first made sure that
the client could fully own the connection, which required removing any
shared client state. I then had to replace the way responses are sent to
the client, by using only a single permanent channel. This required some
additional effort to track which query is being processed. Lastly I had
to modify the query/typeinfo functions to not issue `sync` commands, so
it would fit into the desired flow above.

To note: the flow above does force an extra roundtrip into each query. I
don't know yet if this has a measurable latency overhead.
This commit is contained in:
Conrad Ludgate
2025-05-23 20:41:12 +01:00
committed by GitHub
parent 87fc0a0374
commit 6768a71c86
15 changed files with 500 additions and 745 deletions

View File

@@ -1,35 +1,15 @@
use std::fmt;
use std::sync::{Arc, Weak};
use std::sync::Arc;
use crate::types::Type;
use postgres_protocol2::Oid;
use postgres_protocol2::message::backend::Field;
use postgres_protocol2::message::frontend;
use crate::client::InnerClient;
use crate::codec::FrontendMessage;
use crate::connection::RequestMessages;
use crate::types::Type;
struct StatementInner {
client: Weak<InnerClient>,
name: &'static str,
params: Vec<Type>,
columns: Vec<Column>,
}
impl Drop for StatementInner {
fn drop(&mut self) {
if let Some(client) = self.client.upgrade() {
let buf = client.with_buf(|buf| {
frontend::close(b'S', self.name, buf).unwrap();
frontend::sync(buf);
buf.split().freeze()
});
let _ = client.send(RequestMessages::Single(FrontendMessage::Raw(buf)));
}
}
}
/// A prepared statement.
///
/// Prepared statements can only be used with the connection that created them.
@@ -37,38 +17,14 @@ impl Drop for StatementInner {
pub struct Statement(Arc<StatementInner>);
impl Statement {
pub(crate) fn new(
inner: &Arc<InnerClient>,
name: &'static str,
params: Vec<Type>,
columns: Vec<Column>,
) -> Statement {
Statement(Arc::new(StatementInner {
client: Arc::downgrade(inner),
name,
params,
columns,
}))
}
pub(crate) fn new_anonymous(params: Vec<Type>, columns: Vec<Column>) -> Statement {
Statement(Arc::new(StatementInner {
client: Weak::new(),
name: "<anonymous>",
params,
columns,
}))
pub(crate) fn new(name: &'static str, columns: Vec<Column>) -> Statement {
Statement(Arc::new(StatementInner { name, columns }))
}
pub(crate) fn name(&self) -> &str {
self.0.name
}
/// Returns the expected types of the statement's parameters.
pub fn params(&self) -> &[Type] {
&self.0.params
}
/// Returns information about the columns returned when the statement is queried.
pub fn columns(&self) -> &[Column] {
&self.0.columns
@@ -78,7 +34,7 @@ impl Statement {
/// Information about a column of a query.
pub struct Column {
name: String,
type_: Type,
pub(crate) type_: Type,
// raw fields from RowDescription
table_oid: Oid,