[KentarouTakeda]

This pull request enables pdo-firebird tests within Linux containers for CI processes, resolving the current gap where tests are run on Windows but lack artifact outputs for review (see also #12443 on "strange" skips).

The approach taken here mirrors the one used for PostgreSQL and MySQL in our CI pipelines, utilizing GitHub Actions Service Containers.

I have selected the Docker image jacobalberty/firebird . Although this is not an official Firebird project distribution, its longstanding maintenance record and the substantial number of stars it has accumulated are indicative of its reliability and are believed to provide a stable and trusted testing environment.

In addition to testing this Pull Request, just to be safe, I imported #12664 #12657 , which was passed in the Pull Request for the currently open Extension: pdo_firebird , into my repository and tested it on Linux. The following has been successful without any problems:

https://github.com/KentarouTakeda/php-src/actions/runs/6877444791?pr=5
https://github.com/KentarouTakeda/php-src/actions/runs/6877441551?pr=6

[Girgias]

This lookd reasonable to me, but will wait for @iluuu1994 to have a look as he maintains the CI :)

[iluuu1994]

Great, thank you! We should add the same in nightly.yml. Can you do that too?

[KentarouTakeda]

Thank you for merging! I started modifying nightly.yml.

In that environment, the results of some of the newly added pdo-firebird tests seem to be unstable. Please take a moment to check the details.
https://github.com/KentarouTakeda/php-src/actions/runs/6890531828/job/18743805627#step:12:151
(Some of the other tests that already exist seem to be unstable, so CI may not be involved.)

[iluuu1994]

@KentarouTakeda I can't really comment on the cause of the leak as I'm not familiar with Firebird. They all originate in isc_attach_database, so maybe we're not calling some cleanup function?

If these are legitimate issues in Firebird, I'd suggest reporting upstream and marking the relevant tests with XFAIL.

[KentarouTakeda]

Note:
I'm currently working on the fix written above.

I'm still looking into it, but it's probably an issue with existing code or tests, so I think I'll have to fix it, add CI, and send two pull requests.

[SakiTakamachi]

I have reported
FirebirdSQL/firebird#7849

[SakiTakamachi]

I got a reply. It might be a good idea to put it in the ignore list.

[SakiTakamachi]

Hmm. I added leak:isc_attach_database to lsan-suppressions.txt, but the indirect leak from <unknown module> remains.

XLEAK may not be able to be removed yet...