test: Fix PODMAN_BATS_LEAK_CHECK

[ricardobranco777]

This variable is set by hack/bats and it fails because PODMAN_CMD is unset in basic_teardown. I could manually reproduce as well when calling bats directly with PODMAN_BATS_LEAK_CHECK=1.

Otherwise you get errors like this:
https://openqa-assets.opensuse.org/tests/5555887/file/podman-bats-user-local.tap.txt

# [teardown] $ podman pod rm -t 0 --all --force --ignore
# timeout: failed to run command ‘pod’: No such file or directory
# [teardown] $ podman rm -t 0 --all --force --ignore
# rm: invalid option -- 't' Try 'rm --help' for more information.
# [teardown] $ podman network prune --force
# timeout: failed to run command ‘network’: No such file or directory
# [teardown] $ podman volume rm -a -f
# timeout: failed to run command ‘volume’: No such file or directory
# setup(): removing stray external container error: (unknown)
# setup(): removing stray external container Usage: ()
# setup(): removing stray external container ps ([options])
# setup(): removing stray external container Try ('ps)
# setup(): removing stray external container or ('ps)
# setup(): removing stray external container for (additional)
# setup(): removing stray external container For (more)
# setup(): removing stray image timeout:
# setup(): removing stray image failed

Checklist

Ensure you have completed the following checklist for your pull request to be reviewed:

• Certify you wrote the patch or otherwise have the right to pass it on as an open-source patch by signing all
  commits. (git commit -s). (If needed, use git commit -s --amend). The author email must match
  the sign-off email address. See CONTRIBUTING.md
  for more information.
• Referenced issues using Fixes: #00000 in commit message (if applicable)
• Tests have been added/updated (or no tests are needed)
• Documentation has been updated (or no documentation changes are needed)
• All commits pass make validatepr (format/lint checks)
• Release note entered in the section below (or None if no user-facing changes)

Does this PR introduce a user-facing change?

None

[lsm5]

LGTM as best I can tell. Clearly an improvement over main when run with PODMAN_BATS_LEAK_CHECK set.

@containers/podman-maintainers PTAL.

[mheon]

LGTM once comment addressed

[ricardobranco777]

LGTM as best I can tell. Clearly an improvement over main when run with PODMAN_BATS_LEAK_CHECK set.

I was wondering why you didn't catch it earlier in your CI.

[ricardobranco777]

Full verification run for latest change: https://openqa.opensuse.org/tests/5580232

[Luap99]

I was wondering why you didn't catch it earlier in your CI.

Because we don't run with this option set in CI, which is the reason why we have this option to begin with. The leak check is considered to "slow" so we don't do it. WE rather have the few minutes of CI speed up

LGTM

[ricardobranco777]

I was wondering why you didn't catch it earlier in your CI.

Because we don't run with this option set in CI, which is the reason why we have this option to begin with. The leak check is considered to "slow" so we don't do it. WE rather have the few minutes of CI speed up

LGTM

Wow. Our openQA jobs run for like ~2 hours on x86_64. Will check how much they take without this option and maybe unset it in slower arches like aarch64 & ppc64le.

[Luap99]

see 81c90f5 for the history, I noted 400s back then (on x86_64) and of course with the more test cases it is likely more now. However that is only them not run in parallel, in parallel we cannot do the check anyway.

It is certainly good that we have someone testing with this option on regular to avoid broken or buggy tests. I agree that turning off on certain arches is fine.