feat: do not update/commit files in .gitignore

[bcoe]

I've bumped into this issue a few times, and noticed a similar conversation in #213.

How does this solution look? @nexdrew, @satazor, @arobson?

[satazor]

Looks good, though tests are failing

[stevemao]

LGTM!

Side note: Looks like lib/gitignore.js could be a separate npm package :)

[nexdrew]

For Node 4 support, this file needs 'use strict' at the top. This fixes the failing tests for me locally.

[bcoe]

@nexdrew I guess I'll need to spin up a Windows VM ... grumble.

[bcoe]

@nexdrew smart 👍

[coveralls]

Coverage decreased (-0.1%) to 99.548% when pulling dc21162 on respect-gitignore into 45fcad5 on master.

[nexdrew]

My attempt to fix the Windows build failed. Anyone else have ideas?

[bcoe]

@nexdrew I noticed the underlying library has some logic that uses / rather than path.sep; I might take @stevemao's advice and write our own little library that uses find-up and will patch the issue.

[nexdrew]

@bcoe Are you talking about this line (and maybe line 9 as well)? I don't really think this would affect the failing test, but we could try setting matcher.delimiter = path.sep.

[nexdrew]

@bcoe I was able to spin up a Windows VM, and it looks like my suggestion of matcher.delimiter = path.sep actually works/fixes the failing test (which I find a bit surprising). Not sure my suggestion is the best long-term solution, but it might be enough to get this PR over the finish line.

[coveralls]

Coverage decreased (-0.1%) to 99.548% when pulling 8f85f7e on respect-gitignore into 45fcad5 on master.

[coveralls]

Coverage increased (+0.3%) to 100.0% when pulling 9efe36f on respect-gitignore into 45fcad5 on master.

[nexdrew]

@bcoe This should be ready to go, assuming you're still ok using the dotignore package.

[coveralls]

Coverage increased (+0.3%) to 100.0% when pulling 9b8dc26 on respect-gitignore into 45fcad5 on master.

[coveralls]

Coverage increased (+0.3%) to 100.0% when pulling 93d2624 on respect-gitignore into 45fcad5 on master.

[coveralls]

Coverage increased (+0.3%) to 100.0% when pulling 2e9268b on respect-gitignore into 45fcad5 on master.

[Aigeec]

@bcoe just wanted to let you know that I raised a bug on dotgitignore. It means that lines that are comments are included in the matches. So for example someone puts their directory name is as a comment the dotgitignore starts matching everything.

[chrisgalvan]

@Aigeec Any update on when that will be fixed I just bump into that situation where my package.json and package-lock.json where ignored because some weird issue with https://github.com/bcoe/dotgitignore where it does match some ignored lines withe the full path of the package.json.

For example my package.json is at:
c:/user/myuser/dev/mypackage/package.json

and in my .gitignore I have the following

# ignore user files
.user*

because the path contains user package.json is being ignored.

[Aigeec]

I've open a PR bcoe/dotgitignore#3 on https://github.com/bcoe/dotgitignore to resolve the issue but I believe @bcoe is triaging it at the moment.