problem with re-include, excluded path about micromatch HOT 11 CLOSED

or say we have this one

var patterns = [
  'foo/**',               // include all filepaths (files and folders)
  '!foo/',                // we dont want `foo/` in the list
  '!foo/bar/',            // and also dont want `foo/bar/`
  '!foo/bar/*',           // we dont want all files
  'foo/bar/*.j*',         // except that are .js and .json
];

results

[charlike@localhost glob-fs]$ node playing.js multimatch
[ 'foo/bar.js',
  'foo/bar.md',
  'foo/bar.txt',
  'foo/bar.hbs',
  'foo/bar.json',
  'foo/bar.coffee',
  'foo/bar/bar.js',
  'foo/bar/bar.json' ] 8
[charlike@localhost glob-fs]$ node playing.js micromatch
[ 'foo/bar.js',
  'foo/bar.md',
  'foo/bar.txt',
  'foo/bar.hbs',
  'foo/bar.json',
  'foo/bar.coffee' ] 6
[charlike@localhost glob-fs]$

from micromatch.

note there foo/bar/bar.json missing

This is exactly what I was talking about with "re-inclusion" patterns. You said: micromatch/glob-fs#1 (comment).

Now you know what I mean? This isn't a bug, I did the PR to multimatch that allows it to re-include patterns like that, I can do the same here. IMO, this is the expected result since you are removing it from the set. But maybe it makes more sense to re-include.

talking about that, can you answer me why is needed to ignore negate glob if it is first in given patterns?!

I have no idea, I agree. it makes no sense

edit: actually, I think the reason they did that is because they didn't have the correct logic to do inclusion/exclusion patterns, like what was implemented in multimatch and globby

from micromatch.

Initially I was thinking maybe we should change to what multimatch does, but I'm not sure that's correct. You can get the exact result you want with foo/bar/*.j*, right? Why exclude everything first, then include that?

from micromatch.

also I'm almost ready to push up a first pass at glob-fs. I focused 100% on the correctness of resolving paths, so it's all sync atm. It will be easy to do async, promises, streams whatever once the basic path logic is down

from micromatch.

Hmm, right, clear. Yea, maybe make more sense to save this behaving and allow users to un-negate patterns in options, like not-documented ignore option.

You can get the exact result you want with foo/bar/.j, right?

Yep, but you know ... testing, complexity, possible use cases and etc.

also I'm almost ready to push up a first pass at glob-fs.

Great! From you sync, from me streams, haha. Almost pass all gulp/glob-stream tests. lol
Benchmarks shows x22,000 faster than glob-stream and its normal. :) glob-stream is slowest thing that I ever seen, srsly. And tools from this chain glob-stream -> vinyl-fs -> gulp, verb, template, assemble are very slow natively, because the glob-stream and glob + minimatch.

But yea, you know that - for this reason we work on micromatch/glob-fs :)

from micromatch.

actually, I think the reason they did that is because they didn't have the correct logic to do inclusion/exclusion patterns, like what was implemented in multimatch and globby

mm, yea and for this reason we'll never pass 90-100% of gulp/glob-stream/vinyl-fs tests and they are very very weird and low quality. They not deserves 100%-coverage-badge LOL

Cant wait for you first shot 🎉

from micromatch.

@tunnckoCore was this resolved?

from micromatch.

I'll look after 10-20mins

from micromatch.

no prob, I don't know how long I'll last anyway, trying to keep my eyes open!

from micromatch.

@jonschlinkert okey, we can close and wait your push.

from micromatch.

k thx

from micromatch.