I'd like support for ** as well, is there any plan to add that before merging?

When I said limited to glob patterns involving just *, I meant foo/**/* and foo/*..
See test for reference..

okay, great 😄.

I just tried switching to this branch and I'm still having issues with emacs lock files causing the entire site to be rebuilt. That was the original problem I wanted to get addressed.

Whenever I edit a file, emacs creates a lock file in the same directory with a similair format. I.E. if I start editing src/foo/bar.md emacs makes a src/foo/.#bar.md file. I've added this **/.#* pattern to my exclude section so that the jekyll doesn't place these files in the output directory, but jekyll-watch didn't respect exclude so the entire site was rebuilt anyways (when using jekyll serve).

I've added this **/.#* pattern to my exclude section so that the jekyll doesn't place these files in the output directory

@mohkale That is not the correct glob pattern for the given filename. The correct pattern would be **/*.#*.* or **/.#*.*

exclude:
  - "**/*.#*.*"  # matches src/foo/lorem.#bar.md
  - "**/.#*.*"   # matches src/foo/.#bar.md

but... File.fnmatch("**/.#*", "src/foo/.#bar.md") evaluates to true. Besides I'm quite sure the issue here isn't with my patterns, jekyll isn't outputting any emacs temp files to my build directory anyways.

To clarify.

"**/.#.*" # matches src/foo/lorem.#bar.md

I don't need to account for such files, but thank you anyway.

"**/.#." # matches src/foo/.#bar.md

The only difference between this pattern and the one in my config is that this one demands the file has an extension. I can't switch to this one because I've also got some shell scripts on my site for which I've ommited the extension and simply used a shebang line.

# new pattern
File.fnmatch?("**/.#*.*", "src/foo/.#bar.md") #=> true
File.fnmatch?("**/.#*.*", "src/foo/.#bar") #=> false

# old pattern
File.fnmatch?("**/.#*", "src/foo/.#bar.md") #=> true
File.fnmatch?("**/.#*", "src/foo/.#bar") #=> true

That said, I've added the two patterns you've shared to my config and I've still got the same issue using your branch. Here is the log output

Configuration file: mohkale.io/_config.yml
            Source: ./src
       Destination: ./bin
 Incremental build: disabled. Enable with --incremental
      Generating...
         AutoPages: Generating pages for tags & categories
        Pagination: Complete, processed 9 pagination page(s)
       Jekyll Feed: Generating feed for posts
                    done in 12.622 seconds.
 Auto-regeneration: enabled for './src'
    Server address: http://127.0.0.1:4567
  Server running... press ctrl-c to stop.
# here I started editing FAQ.md, leading to the creation of .#FAQ.md
      Regenerating: 1 file(s) changed at 2020-01-04 17:35:57
                    .#FAQ.md
         AutoPages: Generating pages for tags & categories
        Pagination: Complete, processed 9 pagination page(s)
       Jekyll Feed: Generating feed for posts
                    ...done in 2.4888566 seconds.

to reiterate, my build directory does not contain a .#FAQ.md file, but jekyll still decides to rebuild the entire site when emacs made it, even though the file should've been excluded.

Thank you for the clarifications. I stand corrected..

I also see the problem with my implementation — it can only ignore existing paths that match the wildcard pattern. A new path matching the pattern will pass through undetected.. ☹️

This needs more work

well I look forward to the imminent fix. in the mean time I'll just keep using my own fork. 😄.