can You explain please ? about python-pathspec

[cpburnz]

Hi Robert,

This library provides methods for matching files similar to how a .gitignore functions. It does not implement additional behaviors of git such as scanning descendant directories for .gitignore files.

There is certainly room for improvement with this library. I'm not currently removing blank or commented lines.

My initial thought on how to you could use nested .gitignore files is along the lines of:

import os.path
import pathspec.util

base_path = "..."
all_files = set(pathspec.util.iter_tree_files(base_path))

# Find each ".gitignore" and the files to ignore.
ignore_files = set()
for root, _dirs, files in os.walk(base_path):
    if '.gitignore' in files:
        with open(os.path.join(root, '.gitignore'), 'r', encoding='UTF-8') as fh:
            spec = pathspec.Pathspec.from_lines(fh)

        relative_root = os.path.relpath(root, base_path)
        for path in spec.match_tree_files(root):
            ignore_files.add(os.path.join(relative_root, path))

for path in all_files - ignore_files:
    print(path)
        

[bitranox]

thanks caleb -
I used another project as base for my own solution, check it out :

https://github.com/bitranox/igittigitt
mybe we can bundle our efforts, and merge our projects ?

yours sincerely

Robert

[bitranox]

Dear Caleb,
I looked deeper into the issue, and found that it might be easier to use wcmatch patterns, instead of regexp.
I made a package that supports nested .gitignore Files, provides matching function and provides a ignore function for shutil.copytree():

>>> shutil.copytree(src, dst, ignore=parser.shutil_ignore)

Since You have deep knowledge of the issue, please take a look, and take out what You might need from it.

Thanks for the inspiration !

Robert

[cpburnz]

@bitranox wcmatch looks interesting. The downside to using it is it would limit this library to only the glob expressions supported by wcmatch. With that in mind, I'm going to stick with regular expressions.

[cpburnz]

@bitranox Regarding igittigitt, pathspec was a fun proof of concept for matching files using Git's wildmatch. The goal of this project is scaffolding for using different pattern styles (even though only wildmatch is implemented). The goal was not intended to fully emulate Git with .gitignore. Unfortunately, I do not have the time available to devote on another project to any meaningful capacity.

[cpburnz]

I'm closing this because I think I've answered all of your questions.