Master-port #49860 selinux support for file.manage

[mchugh19]

Master port of #49860
This looks good to me, but happy to get more eyes on it. See conversation in #49860 for more information

What does this PR do?

Recursion not supported for persist=True on file.get_selinux_context execution module

Add selinux functionality to file.managed

/tmp/selinux.test
  file.managed:
    - user: root
    - selinux:
        seuser:  system_u
        serole: object_r
        setype: system_conf_t
        seranage: s0

# salt-call --local state.sls selinux test=True
local:
----------
          ID: check_selinux
    Function: file.managed
        Name: /tmp/selinux.test
      Result: None
     Comment: File /tmp/selinux.test not updated
     Started: 10:54:39.418112
    Duration: 17.911 ms
     Changes:
              ----------
              selinux context:
                  ----------
                  New:
                      User: system_u Type: system_conf_t
                  Old:
                      User: unconfined_u Type: user_tmp_t

Summary for local
------------
Succeeded: 1 (unchanged=1, changed=1)
Failed:    0
------------
Total states run:     1
Total run time:  17.911 ms

What issues does this PR fix or reference?

#40703 and possibly #1349

New Behavior

Uses file.get_selinux_context and file.set_selinux_context to determine if selinux updates are needed. As mentioned in a ticket #1349, set_selinux_context uses the chcon command, which might not be persistent. Feedback welcome.

Tests written?

Yes - but could always use more eyes

[twangboy]

Hey, @mchugh19

Looks like this needs to be rebased and have pre-commit run on it. Do you mind doing that?

https://docs.saltstack.com/en/latest/topics/development/contributing.html#quickstart

[mchugh19]

@twangboy

Looks like this needs to be rebased and have pre-commit run on it. Do you mind doing that?

I'm not sure what you mean. These files have been run through isort and black already. I tried manually running pre-commit run --files filenames but there weren't any changes. and there don't appear to be any conflicts with master. What needs rebasing?

[waynew]

I'm not sure either 😂

Pre-commit looks fine and is passing, but there are some other failures that look to be related to this PR. Can you take a look?

[mchugh19]

tests were missing a cmd.run_all mapping. Not sure how it ever worked before, or why it only failed on centos (maybe selinux is only enabled on redhat?). Should be good to go now though.

[mchugh19]

re-run pr-windows2019-py3

[mchugh19]

re-run pr-windows2016-py3

[mchugh19]

re-run pr-centos7-py3

[mchugh19]

Ugh. Centos is stuck on

File "/tmp/kitchen/testing/tests/integration/states/test_docker_network.py", line 133, in tearDownClass
    if process.returncode != 0 and "No such image" not in output:
TypeError: a bytes-like object is required, not 'str'