* fix: double client.end() hang
fixes https://github.com/brianc/node-postgres/issues/2716
`client.end()` will resolve early if the connection is already dead,
rather than waiting for an "end" event that will never arrive.
* fix: client.end() resolves when socket is fully closed
* Enable SASL tests in GitHub actions CI
* Add SASL test to ensure that client password is a string
* Fix SASL error handling to emit and bubble up errors
* Add informative error when SASL password is empty string
This changeset enables declaring the `stream` config value as a factory
method. Providing a much more flexible control of the socket connection.
Defining a custom `stream` config value allows the postgres driver to
support a larger variety of environments/setups such as proxy servers
and secure socket connections that are used by cloud providers such as
GCP.
Currently, usage of the `stream` config value is only viable for single
connections given that it's only possible to define a single socket
stream instance per new Client/Pool instance. By adding support to a
factory function, it becomes possible to enable usage of custom socket
streams for connection pools.
For reference, see the `mysql2` driver for MySQL (linked below) for
prior art example of this pattern.
Refs: ba15fe2570/lib/connection.js (L63-L65)
Refs: https://cloud.google.com/sql/docs/postgres/connect-overview
Signed-off-by: Ruy Adorno <ruyadorno@google.com>
Signed-off-by: Ruy Adorno <ruyadorno@google.com>
This is the initial port to github actions. Still pending are the SSL and client SSL cert tests which are currently being skipped. But perfect is the enemy of the good here, and having no CI because travis-ci keeps not working is unacceptable.
Based on the suggestion from #2078. This adds ref/unref methods to the
Connection and Client classes and then uses them to allow the process to
exit if all of the connections in the pool are idle. This behavior is
controlled by the allowExitOnIdle flag to the Pool constructor; it defaults
to the old behavior.
* pg: Re-export DatabaseError from 'pg-protocol'
Before, users would have to import DatabaseError from 'pg-protocol'. If
there are multiple versions of 'pg-protocol', you might end up using the
wrong one.
Closes#2378
* Update error-handling-tests.js
* Update query-error-handling-tests.js
Co-authored-by: Brian C <brian.m.carlson@gmail.com>
* Make tests pass in github codespaces
There were a few tests which didn't specify a host or port which wasn't working well inside the codespaces docker environment. Added host & port where required. Also noticed one test wasn't actually _testing_, it was just `console.log`-ing its output, so I added proper assertions there. Finally set `PGTESTNOSSL: true` in the codespaces environment until I can get the postgres docker container configured w/ SSL...which I will do l8r.
* lint
* Add sha256 SASL helper
* Rename internal createHMAC(...) to hmacSha256(...)
* Add parseAttributePairs(...) helper for SASL
* Tighten arg checks in SASL xorBuffers(...)
* Add SASL nonce check for printable chars
* Add SASL server salt and server signature base64 validation
* Add check for non-empty SASL server nonce
* Rename SASL helper to parseServerFirstMessage(...)
* Add parameter validation to SASL continueSession(...)
* Split out SASL final message parsing into parseServerFinalMessage(...)
* Fix SCRAM tests
Removes custom assert.throws(...) so that the real one from the assert package is used and
fixes the SCRAM tests to reflect the updated error messages and actual checking of errors.
Previously the custom assert.throws(...) was ignoring the error signature validation.
Move from 3 loops (prepareValue, check for buffers, write param types, write param values) to a single loop. This speeds up the insert benchmark by around 100 queries per second. Performance improvement depends on number of parameters being bound.
* `password` already has this set, but was a little long considering we only want to override default of one property
* `ssl.key` was showing up in tracebacks
