Remove salt as a dependency

[mew1033]

Salt, even when installed via pip, is huge and has a ton of dependencies. All it was being used for was six, and the output coloring. This PR removes salt as a dependency and brings in the relevant code from the salt repo. Thanks to them for using Apache 2!

[roaldnefs]

Thanks @mew1033 for your PR and investing your time to make salt-lint a better tool!

Even though this a great optimization for the current version of salt-lint, we are including salt as a dependency to be able to (partially) render states (in order to fix #19 for example) in future releases. It also allows (custom) rules to use the internals of salt (we could for example use the existing Jinja tokenizer to fix #51).

I'm closing this PR for now, we can always reopen it if we change our mind.

[mew1033]

@roaldnefs That makes a lot of sense. I was looking into ansible-lint and saw how it was using ansible directly.
Do you think it would make sense to create multiple methods of installing salt-lint? Maybe using pip install extras? So if you want a lighter install you can pip install salt-lint, and if you want a full-featured install, you can do pip install salt-lint[salt]?

I just know that, at least for now, I'd much rather have a smaller, lightweight version of salt-lint.

[roaldnefs]

Do you think it would make sense to create multiple methods of installing salt-lint?

Yes, it would make sense but also adds some complexity to the source code and on the user side during the installation.

I just know that, at least for now, I'd much rather have a smaller, lightweight version of salt-lint.

Would it be an idea to create an issue for this request? If there are more users that are interested in a lightweight variant of salt-lint we can change our priority accordingly. Feel free to refer back to this issue!

[mew1033]

@roaldnefs Sounds good, I made an issue for this. I'll try to maintain my fork for a while because I personally would much prefer to use salt-lint without salt for now.

Thanks!