The PostgreSQL server will reject messages greater than ~1 GB anyway.
However, worse than that is that a message that is larger than 4 GB
could wrap the 32-bit integer message size and be interpreted by the
server as multiple messages. This could allow a malicious client to
inject arbitrary protocol messages.
https://github.com/jackc/pgx/security/advisories/GHSA-mrww-27vc-gghv
Otherwise, it might be possible to panic when closing the pipeline if it
tries to read a connection that should be closed but still has a fatal
error on the wire.
https://github.com/jackc/pgx/issues/1920
OnPGError is called on every error response received from Postgres and can
be used to close connections on specific errors. Defaults to closing on
FATAL-severity errors.
Fixes#1803
stdlib_test.TestConnConcurrency had been flickering on CI deadline /
timeout errors. This was extremely confusing because the test deadline
was set for 2 minutes and the errors would occur much quicker.
The problem only manifested in an extremely specific and timing
sensitive situation.
1. The watchdog timer for deadlocked writes starts the goroutine to
start the background reader
2. The background reader is stopped
3. The next operation is a read without a preceding write (AFAIK only
CheckConn does this)
4. The deadline is set to interrupt the read
5. The goroutine from 1 actually starts the background reader
6. The background reader gets an error reading the connection with the
deadline
7. The deadline is cleared
8. The next read on the connection will get the timeout error
connect_timeout given in conn string was not obeyed if sslmode is not specified (default is prefer) or equals sslmode=allow|prefer. It took twice the amount of time specified by connect_timeout in conn string. While this behavior is correct if multi-host is provided in conn string, it doesn't look correct in case of single host. This behavior was also not matching with libpq.
The root cause was to implement sslmode=allow|prefer conn are tried twice. First with TLSConfig and if that doesn't work then without TLSConfig. The fix for this issue now uses the same context if same host is being tried out. This change won't affect the existing multi-host behavior.
This PR goal is to close issue [jackc/pgx/issues/1672](https://github.com/jackc/pgx/issues/1672)
The context timeouts for tests are designed to give a better error
message when something hangs rather than the test just timing out.
Unfortunately, the potato CI frequently has some test or another
randomly take a long time. While the increased times are somewhat less
than optimal on a real computer, hopefully this will solve the
flickering CI.
on Windows connecting on a closed port takes about 2 seconds.
You can test with something like this
start := time.Now()
_, err := d.DialContext(context.Background(), "tcp", "127.0.0.1:1")
fmt.Printf("finished, time %s, err: %v\n", time.Since(start), err)
This seems by design
https://groups.google.com/g/comp.os.ms-windows.programmer.win32/c/jV6kRVY3BqM
Generally TestConnectWithFallback takes about 8-9 seconds on Windows.
Increase timeout to avoid random failures under load
This commit adds a background reader that can optionally buffer reads.
It is used whenever a potentially blocking write is made to the server.
The background reader is started on a slight delay so there should be no
meaningful performance impact as it doesn't run for quick queries and
its overhead is minimal relative to slower queries.
Jack Christensen
Oleksandr Redko
vahid-sohrabloo
James Hartig
Ryan Fowler
Samuel Stauffer
Simon Paredes
Ivan Posazhennikov
Anton Levakin
Ville Skyttä
Alexey Palazhchenko
smaher-edb
Adrian-Stefan Mares
Brandon Kauffman
Nicola Murino