Scandir improvements

[jose-elias-alvarez]

This PR combines a handful of small fixes and features for the scandir module as well as a refactor of the respect_gitignore functionality. Instead of attempting to interpret the .gitignore file, which is quite difficult when the file contains more advanced patterns, it instead checks relative file paths against the output of git ls-tree -rt HEAD --name-only. This should close #76.

I've split up the changes into separate commits for easier review but will squash them whenever this PR is ready to merge.

[jose-elias-alvarez]

I also wanted to take care of #75 as part of this PR but noted that changing search_pattern as described in the issue breaks the relevant test, so to clarify: do we want search_pattern to match only against the file / directory name and not the full path?

[Conni2461]

What? Why? What? If we want to get the output of git ls-tree -rt HEAD --name-only sync why would we even want to call scan_dir and iterate the filesystem ourselves? We then could just return git ls-tree -rt HEAD --name-only and be done. No need to compare anything.

I just done get it right now. I know its hard. So i would look at already existing implementations like https://github.com/ggreer/the_silver_searcher/blob/master/src/ignore.c or https://github.com/BurntSushi/ripgrep/tree/master/crates/ignore

[jose-elias-alvarez]

Understood. Frankly, my Lua skills aren't up to the task of creating an implementation on the level of rg's or ag's, but the implementation I just pushed should bring gitignore handling closer to what's specified in the documentation. Let me know how it looks.

[jose-elias-alvarez]

@Conni2461 Let me know if there's anything I can add / change on this to get it merged!

[Conni2461]

LGTM, sorry i hadn't have much time to review it and now tami knows about it and forces me to merge it 😆

I was thinking, maybe we could also expose the git ignore things as scandir.__ functions and then test these things isolated. Thoughts

[jose-elias-alvarez]

Thanks for taking a look at this again! I exposed make_gitignore and added tests asserting against the new functionality. Let me know how everything looks.

[tjdevries]

Sorry, I haven't been reviewing much lately. If you fix conflicts I will take a look again and we can merge after I review

[jose-elias-alvarez]

Thanks! I just fixed the conflicts (and formatted) so let me know. Not sure what's going on with the test failures, but the scandir tests seem fine on my end, and I'm getting the same issues on the master branch.

[Conni2461]

Everything look good here thanks :) can you do another rebase to make sure that the tests are happy. I think all of those issues are solved in master but just in case 😆 I keep this tab open. I intend to merge this today :)

[Conni2461]

I just run it locally. Everything seems to work. Thanks you very much :)

Sorry for neglecting the PR