[adambaratz]

No description provided.

[Girgias]

Can you add a README at the same time in the tests folder for PDO DBLIB so that one can reproduce it locally in some way?

[adambaratz]

@nikic, could you let me know if this looks good to merge? Thanks!

[nikic]

Looks good to me, thank you! Could you please squash the commits when merging?

[adambaratz]

@nikic, yes, I'll definitely squash them. I just tried to push and I don't have karma for the azure directory. Is that something you can grant me?

[nikic]

@adambaratz Done, might take a while to propagate.

[adambaratz]

Merged with b569698.

[nikic]

Looks like there's some test failures in nightly builds: https://dev.azure.com/phpazuredevops/PHP/_build/results?buildId=14085&view=ms.vss-test-web.build-test-results-tab

There's two failing tests:

• ext/pdo_dblib/tests/types.phpt fails on COVERAGE_DEBUG_ZTS and DEBUG_NTS_FILE_CACHE builders.
• ext/pdo_dblib/tests/timeout.phpt fails on DEBUG_NTS_REPEAT builder

[adambaratz]

Thanks for flagging. I'll take a look later today.

[nikic]

Looks like the ext/pdo_dblib/tests/bug_50755.phpt test still occasionally fails with the following error: https://dev.azure.com/phpazuredevops/PHP/_build/results?buildId=14292&view=ms.vss-test-web.build-test-results-tab&runId=323892&resultId=107600&paneView=debug

This happens on the REPEAT job, which means that the same test is run two times in the same process (--repeat 2 in run-tests).

[nikic]

There's still common failures in ext/pdo_dblib/tests/bug_50755.phpt with --repeat 2 testing with the following diff:

001+ Fatal error: Uncaught PDOException: SQLSTATE[HY000]: General error: 20018 Transaction (Process ID 60) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction. [20018] (severity 13) [select *
001- OK
002+ from information_schema.columns ic1
003+ cross join information_schema.columns ic2
004+ cross join information_schema.columns ic3] in /home/vsts/work/1/s/ext/pdo_dblib/tests/bug_50755.php:12
005+ Stack trace:
006+ #0 /home/vsts/work/1/s/ext/pdo_dblib/tests/bug_50755.php(12): PDOStatement->fetch()
007+ #1 {main}
008+   thrown in /home/vsts/work/1/s/ext/pdo_dblib/tests/bug_50755.php on line 12

@adambaratz Do you have any idea what is causing this? Otherwise I'll just mark the test as skipped on this job. Though it does look like it could be a legitimate issue.

[adambaratz]

I'm okay skipping this test. The test is trying to run a query that uses too much memory. I'm guessing the test worked in a different environment, but it's dependent enough on how MSSQL handles the query that it's exercising MSSQL just as much as it's exercising pdo_dblib.