fixer: Regal fix command

[charlieegan3]

At the moment it works like this:

regal $ cat test.rego
package wow

#comment

allow = 1 {
        input.foo == true
}
regal $ ./regal fix test.rego
3 fixes applied:
test.rego:
- no-whitespace-comment
- use-assignment-operator
- use-rego-v1
regal $ cat test.rego
package wow

import rego.v1

# comment

allow := 1 if {
        input.foo == true
}

Related to #92

[charlieegan3]

Also, interested to get thoughts on --fix vs regal fix. Having a dedicated command gives us more options for flags etc. but --fix is what might feel more familiar.

[anderseknert]

Nice work! I'm fine with the direction taken here 👍 Although I'm now leaning more and more towards providing a regal fix command for a cleaner separation of linting and transformations. I've left some comments here and there for further discussion.

[charlieegan3]

@anderseknert this is ready for review again. I am more happy with the implementation here now. If you agree on the direction & interfaces etc, then I think we can get this in.

We might want to add code actions for the whitespace comment and assignment operator too, but these could happen as a follow on PR since this is already kinda large.

[anderseknert]

Amazing work on this! The code is really clean, and I had no problems whatsoever following along exactly in what it does. I did have a few thoughts and questions, but this is definitely getting there 😍 Next release is gonna be 💥

[charlieegan3]

This is ready for another review. I think we are getting there. Main changes:

• The fixer will now run repeatedly until there are no fixes applied fixer: Regal fix command  #653 (comment)
• Interfaces for fixing are now more generic and should better support other fix operations such as renaming.

[anderseknert]

Terrific! Only some nits, but none of them blocking, so go on and merge when you feel done! 👏

This will be an awesome addition to Regal, and will save a lot of time both for us and others.