Problem
code-review review-loop stalls indefinitely when the diff between --base and -b contains many files. In our case, MR !340 on agents-python had 108 changed Python files and 377 auto-gathered function bodies. The tool completed the function-gathering phase but never progressed to the actual review — it appeared to hang after printing the function list.
Steps to Reproduce
- Have a feature branch that diverges significantly from main (e.g., includes merge commits from main that add many files to the diff)
- Run
code-review review-loop -b feature-branch --base main --beliefs beliefs.md --repo /path/to/repo
- Observe: function gathering completes (prints 377 functions), auto-discovers test files, then hangs
Expected Behavior
Either:
- Chunk the review into smaller pieces and process sequentially
- Warn and abort if the diff exceeds a reasonable threshold (e.g., 50 files)
- Allow
--max-files or --paths filter to scope the review
Actual Behavior
Process hangs after function gathering with no output, no error, and no timeout. Had to review manually.
Environment
code-review installed via uv tool
- Target repo had 108 changed files / ~41k insertions (most from merge noise)
- The actual feature was only 9 files / 750 lines
Problem
code-review review-loopstalls indefinitely when the diff between--baseand-bcontains many files. In our case, MR !340 on agents-python had 108 changed Python files and 377 auto-gathered function bodies. The tool completed the function-gathering phase but never progressed to the actual review — it appeared to hang after printing the function list.Steps to Reproduce
code-review review-loop -b feature-branch --base main --beliefs beliefs.md --repo /path/to/repoExpected Behavior
Either:
--max-filesor--pathsfilter to scope the reviewActual Behavior
Process hangs after function gathering with no output, no error, and no timeout. Had to review manually.
Environment
code-reviewinstalled viauv tool