Add +X capability to linux_acl

[leifliddy]

What does this PR do?

It adds the +X setfacl capability feature.
"The character X stands for the execute permission if the file is a directory or already has execute permission for some user."

What issues does this PR fix or reference?

References: #33921

Previous Behavior

The setfacl characterX was not supported

New Behavior

The setfacl characterX is now supported

Merge requirements satisfied?

[NOTICE] Bug fixes or features added to Salt require tests.

• Docs
• Changelog - https://docs.saltproject.io/en/master/topics/development/changelog.html
• Tests written/updated

Commits signed with GPG?

Yes

[leifliddy]

Is there any update on this PR?

[dwoz]

Can we add a test or X rather than changing the test which uses x.. Seems like by changing this we're removing coverage or x

[leifliddy]

Agreed. I wasn't sure quite how to handle that. I'll see what I can do -- but this might be beyond my current abilities. Thanks.

[thecosmicfrog]

Hi @leifliddy, @dwoz. Has any progress been made on this? Thanks.

[leifliddy]

Yes, progress has been made on this. The proposed PR has added support for the setfacl X character
All that's needed now is a functional unit test to be drafted.
TBH, I just don't have the will/desire/time to write + test that out at the moment.
If you look at the current test_present linux_acl.py unit test -- it's nearly 300 lines long.
https://github.com/saltstack/salt/blob/master/tests/unit/states/test_linux_acl.py

So what needs to be done is to create a similar test that tests for the rwX acl

At that topic -- there's likely hundreds of PRs that are stuck in limbo due to the fact that not everyone is familiar with mock object library. The salt team must have known that imposing a hard unit test requirement would lead to this -- so it shouldn't be a total surprise that this has stymied progress on some level. Sorry, but I'm not spending several hours of my time to try and satisfy the unit test requirement when I can just run a cmd.run setfacl function instead to accomplish the same thing.

Feel free the submit your own proposal for that though.

[leifliddy]

@twangboy Can you please review this change and see if that satisfies the requirement of having a separate test?
Feel free to phrase, reword, modify...etc anything in the PR.

[leifliddy]

It looks like this is ready to be merged.

[leifliddy]

@twangboy What was the issue with this? Why did it never get merged?

[leifliddy]

@dwoz Are you still tracking this? I just re-requested a review. I can easily rebase my changes no problem -- but I need at least some sort of acknowledgement from the reviewers that they're tracking this.

[leifliddy]

@twangboy @dwoz Is anyone going to respond to this? If not -- please assign this PR a reviewer that will respond.
I shouldn't have to wait months after making the requested changes -- for a reviewer to respond.

[OrangeDog]

Apparently this is being ignored because there are failing tests, even though they have nothing to do with your changes.
https://saltstackcommunity.slack.com/archives/C8832QN4U/p1700852406066979?thread_ts=1700828180.622949&cid=C8832QN4U

[leifliddy]

Oh really!? I can rebase the changes and have the tests re-run no problem. So the reviewers weren't being notified?

[lkubb]

I would assume they were still notified, but two failing checks was the reaction of a core dev - might be an indication of the general core team workflow. (I have also been ignored on passing CI, but maybe less so)

Better wait with a rebase until the master branch CI is fixed, currently it's crashing hard.

[dwoz]

@leifliddy Can you please add a changelog entry?

[leifliddy]

@dwoz damn, forgot about that. Ok, I just pushed it.

[leifliddy]

@dwoz Could you please remove the needs-changelog-file label?

[leifliddy]

Can this get merged please?
It's been approved for over a month....

[twangboy]

These tests probably won't pass until we get 3006.x and 3007.x merged in to master

[dwoz]

@twangboy I think this should be back-ported to 3006.x and then merged forward as it's a long standing bug.

[twangboy]

@leifliddy Could you rebase this PR against the 3006.x branch please?

[twangboy]

Please rebase and address conflicts