ext/session/.../gh23043.phpt: session.save_path can be nonempty

[orlitzky]

Allow session.save_path to be nonempty in this test, as the test itself does not empty it, and in automated setups it is common to force all "temporary stuff" to a dedicated location.

[Sjord]

Looks good to me.

For reference, error is generated here.

@NickSdot could you take a look at this?

[NickSdot]

@NickSdot could you take a look at this?

This makes sense to me; gh7787.phpt already does the same.

[jorgsowa]

Test suite is not properly isolated, so such artifacts happen. I would rather to have properly isolated environment or INI setting for tests, but technically it's also correct.

[orlitzky]

Test suite is not properly isolated, so such artifacts happen. I would rather to have properly isolated environment or INI setting for tests, but technically it's also correct.

Thanks, we are actually trying quite hard to isolate the tests, but in this case we're passing -d "session.save_path=..." to the TEST_PHP_EXECUTABLE to avoid cluttering the build directory with temporary files.

If you are looking for something to do today, I have a few PRs open that fix leaks in the (crude, but usable) TEST_PHP_ isolation mechanism :)

• sapi/cli/tests/php_cli_server_ipv6_error_message.phpt: use TEST_PHP vars #22414
• ext/soap/tests/bugs/bug62900.phpt: make use of TEST_PHP_ vars #22405
• tests: add missing TEST_PHP_EXTRA_ARGS #11669

The tl;dr is that TEST_PHP_EXECUTABLE needs to be respected to avoid using the system PHP in place of the just-built one, and that TEST_PHP_EXTRA_ARGS needs to be respected so that -n can be reliably passed to that PHP. (Otherwise, it can load the system php.ini and extensions.)