add a plist serializer

[weswhet]

What does this PR do?

Adds a plist serializer, allowing you to manage plist files with file.serialize.

Previous Behavior

Before the only way to manage and manipulate plists, a common file type on macOS, was to use file.manage with a source file.

New Behavior

you can now use a formatter of plist in a file.serialize state to manage xml and binary plist files.

test_plist_serial:
  file.serialize:
    - name: /tmp/test_plist.plist
    - serializer_opts:
      - fmt: FMT_BINARY
    - formatter: plist
    - user: root
    - group: wheel
    - mode: 0644
    - makedirs: True
    - merge_if_exists: False
    - dataset:
        Label: com.saltstack.salt.master
        RunAtLoad: True
        KeepAlive: True
        ProgramArguments:
          - /opt/salt/bin/start-salt-master.sh
        SoftResourceLimits:
          NumberOfFiles: 100000
        HardResourceLimits:
          NumberOfFiles: 100000
        StandardOutPath: /tmp/salt-minion.out
        StandardErrorPath: /tmp/salt-minion.err

Merge requirements satisfied?

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

• Docs
• Changelog
• Tests written/updated

Commits signed with GPG?

Yes

[weswhet]

cc @sheagcraig @garethgreenaway members of the macOS working group.

[weswhet]

the previous implementation would attempt to add a new line to the end of a binary plist. which would cause sadness.

[waynew]

My guess is that was written using Py2, where binary == str == text. In Py3, binary == bytes and text == str. Definitely a good move here - unless we should have an

else:
   contents += b"\n"

[weswhet]

@waynew Would it make sense to check to see if a new line already exists? I know plistlib automatically adds a newline.

[waynew]

I'm not familiar with other tools out there, but I would assume we should do whatever the "correct" thing is.

[weswhet]

I've updated with the recommendation as well as skipping new line concatenation on plists as this will create an invalid plist file.

[weswhet]

trying to read a binary plist without the "b" would cause an exception. Also appears all the other implementations of salt.utils.files.fopen in this file use "rb" as arguments.

[weswhet]

I moved this back to its previous behavior and only add a "b" if the formatter is a plist. this change was causing the integration.states.test_file.FileTest.test_serializer_deserializer_opts test to fail.

[weswhet]

Would this be 3001?

[waynew]

Should be, assuming all is 🥇

[weswhet]

added

[sheagcraig]

@wes and I wargamed this one and I think it looks ✨🤘🏻✨

[waynew]

The module at salt/serializers/plist.py does not have a sphinx stub at doc/ref/serializers/all/salt.serializers.plist.rst

Looks like you're missing some docs. I think you can autogen - just take a look at the other docs around there.

[weswhet]

The module at salt/serializers/plist.py does not have a sphinx stub at doc/ref/serializers/all/salt.serializers.plist.rst

Looks like you're missing some docs. I think you can autogen - just take a look at the other docs around there.

Should be added. I think I did that right :)

[weswhet]

wooo, thanks all!