Summary
RefactronPipeline.apply processes each file independently: some files may apply fixes while others are blocked. Backup session is created once, but there is no single orchestration policy (transactional batch vs best-effort) documented for users or for rollback UX.
Suggested direction
- Define supported behavior: e.g. best-effort (current) vs
--fail-fast stop on first blocked file.
- Ensure
backup_session_id and rollback story matches partial applies (document which files are included in backup).
- Optionally add session summary:
files_attempted, files_succeeded, files_failed.
Acceptance
- README or docstring for pipeline behavior; tests for fail-fast if implemented.
- No silent data loss: blocked files always recorded in
blocked_fixes.
Code
refactron/core/pipeline.py — apply
refactron/cli/run.py
Summary
RefactronPipeline.applyprocesses each file independently: some files may apply fixes while others are blocked. Backup session is created once, but there is no single orchestration policy (transactional batch vs best-effort) documented for users or for rollback UX.Suggested direction
--fail-faststop on first blocked file.backup_session_idandrollbackstory matches partial applies (document which files are included in backup).files_attempted,files_succeeded,files_failed.Acceptance
blocked_fixes.Code
refactron/core/pipeline.py—applyrefactron/cli/run.py