mktemp: treat empty TMPDIR as unset and fallback to /tmp

[reubenwong97]

This should resolve #10189 . Apologies for the duplicate PR, I had made a mistake in the earlier one and closed it.

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tail/inotify-dir-recreate (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/basenc/bounded-memory is now passing!

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tail/inotify-dir-recreate (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/basenc/bounded-memory is now passing!

[sylvestre]

Could you please add a test to make sure we don't regress in the future? Thanks

[reubenwong97]

Hi @sylvestre , thanks for taking the time to take a look. I added tests for this behavior, but was only able to test the successful creation locally. Based on my understanding of the thread, usage on windows should still attempt the creation, but fail due to non-existence of /tmp. I've added tests based on the expected behavior, but am unable to run that as I don't have a Windows machine... How do other contributors manage that?

Edit: I got a hint based on the windows build from the CI.

[github-actions]

GNU testsuite comparison:

Skip an intermittent issue tests/shuf/shuf-reservoir (fails in this run but passes in the 'main' branch)
Skip an intermittent issue tests/sort/sort-stale-thread-mem (fails in this run but passes in the 'main' branch)

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tail/inotify-dir-recreate (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/basenc/bounded-memory is now passing!

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tail/inotify-dir-recreate (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/basenc/bounded-memory is now passing!

[reubenwong97]

Hi @sylvestre , I've managed to rectify the issue with windows builds. Android builds were failing as well, as based on some searches, I believe android doesn't have /tmp, and therefore should also be failing like windows builds. I've updated the tests to indicate as such.

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tail/inotify-dir-recreate (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/basenc/bounded-memory is now passing!

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/shuf/shuf-reservoir (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/sort/sort-stale-thread-mem (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tail/inotify-dir-recreate (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/basenc/bounded-memory is now passing!

[ChrisDryden]

The change looks good, the only thing I could think causing issues is that on windows typically the default /tmp folder will not exist but that can be something that can be addressed separately since there is no existing behaviour that we are trying to replicate for windows.