docs: clarify directory-pattern semantics and update tests

[RyanL2004]

While working on enhancing the test coverage for directory pattern matching in GitIngest, I encountered discrepancies between the assumptions in the TODO comments and the actual behavior of the pattern matching. Specifically, patterns like src/*, /src/*, and src/ do not behave as expected. Below is a summary of the observations and suggested updates for clarity and next steps.

---

Key Findings

1. Pattern Mismatches

• src/*: Matches all files recursively under src/, instead of just direct children.
• /src/*: Does not match any files. Leading slashes are not supported.
• src/: Does not match recursively. Correct pattern for recursive matching is src/**.

2. Cross-Platform Path Separator Issues

• Problem: Path separators were inconsistent across Windows and Unix systems, causing failures in test comparisons.
• Solution: Implemented path normalization using Path() to ensure compatibility.

3. Improved Tests

• test_include_src_star_pattern: Correctly verifies that src/* matches all files recursively under src/.
• test_include_src_recursive: Uses src/** for explicit recursive matching.
• Path normalization fixes ensure the tests work across different platforms.

4. Test Documentation & Structure

• Added descriptive docstrings and improved error messages.
• Utilized set comparisons for more reliable and readable file path verification.
• Comments added to clarify test expectations and behavior.

---

Conclusion

• Through this testing process, I discovered several differences between the expected and actual behavior of the pattern matching system.
• I proceeded with updating the test cases to reflect the real functionality, so it would help clarify the intended behavior.

---

[cyclotruc]

This is a very good point, I did in fact cheat a little when I implemented pattern matching (for instance, src/ is modified into src/*)

I think a good starting point is to follow unix-style Glob pattern matching as described here:
https://en.wikipedia.org/wiki/Glob_(programming)

But maybe not implement all the features right away, first step would be to make them consistent with Unix pattern matching

[github-actions]

Hi there! We haven’t seen activity here for 45 days, so I’m marking this issue as stale.
If you’d like to keep it open, please leave a comment within 10 days. Thanks!