Add verification policy for bug fixes and refactoring#57
Merged
Conversation
Bug fixes require demonstrated red->green with recorded pre-fix failure evidence; refactors require pinned before/after results in both test presets; OS-dependent behavior is verified with the platform unit-test targets or example flows, with Classic-target test executables and Flow-driven scenario automation as the long-term harness. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Lihi-AI
force-pushed
the
docs/verification-policy
branch
from
c136dd9 to
a18bd93
Compare
Contributor
There was a problem hiding this comment.
Pull request overview
Codifies a repository-wide verification bar for bug fixes and refactoring by documenting it in AGENTS.md, so contributors follow consistent evidence and platform-validation expectations.
Changes:
- Added an explicit “Verification Policy For Bug Fixes And Refactoring” section covering red→green expectations, behavior-preserving refactors, and platform/Classic validation.
- Documented what to do when headless tests cannot discriminate a change and when tests are infeasible (require explicit PR note + tracked follow-up).
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| - If a design or implementation path looks fragile, hard to reason about, or likely to cause intermittent bugs, prefer a small refactor toward a simpler structure first. If no clear low-risk refactor is apparent, stop and call it out to the user before building further on top of it. | ||
|
|
||
| ## Verification Policy For Bug Fixes And Refactoring | ||
| - Bug fixes must demonstrate red->green: add or extend a test that fails (or hard-fails under ASan) on the pre-fix code, confirm that failure, then land the fix that turns it green. Record the pre-fix failure evidence in the PR. |
|
|
||
| ## Verification Policy For Bug Fixes And Refactoring | ||
| - Bug fixes must demonstrate red->green: add or extend a test that fails (or hard-fails under ASan) on the pre-fix code, confirm that failure, then land the fix that turns it green. Record the pre-fix failure evidence in the PR. | ||
| - Refactoring must be behavior-preserving under pinned results: add characterization tests first where coverage is missing, then show identical results before and after the change in both the `testing` and `testing-asan` presets. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Codifies the maintainer's verification policy in AGENTS.md so every agent and contributor works to the same bar:
testingandtesting-asan.LokaTestsMacOS/LokaTestsWin32) or by driving the affected example flow on that platform; for Classic targets, test executables in Mini vMac / real hardware, with Flow-driven scenario automation as the intended long-term harness.Note: this commit is unsigned (ssh-agent not reloaded after restart); squash-merge covers it per the signing policy.
🤖 Generated with Claude Code