🔧 Pin changelog generation to GPT-5.6 Luna - #353
Closed
Robdel12 wants to merge 11 commits into
Closed
Conversation
Resolve review scope from the server, preserve annotations atomically, honor review versions and command idempotency, and retire raw API access to the old SDK mutations.
Delete hotspot and confirmed-region cache APIs, persistence, automatic pass logic, context projections, and reporter controls. Stale cache data can no longer affect comparison results or local review output.
Emit the canonical visual_review contract across commands, upload results, local context, types, tests, and JSON documentation. Remove status fallbacks that kept the retired review response shape alive.
Keep the local baseline reporter on its own review state vocabulary and delete obsolete tests that only described retired cache inputs. The retained context workflow still proves stale cache files cannot affect evidence.
Use direct names like review state, analysis details, and status response throughout the CLI. Remove inflated compatibility terminology from code, tests, and repo-owned docs.
Keep the current output contract tests focused on supported review data instead of preserving retired field names as negative fixtures.
Require the current status and context response shapes, remove flat and cached v1 adapters, and use plain local formatting names.
Run the real CLI process against a stateful HTTP boundary to prove decisions, annotations, stale-version errors, auth refresh, and idempotent retries.
Keep the shipped Vizzly skill aligned with the current review and diff behavior after removing hotspot and confirmed-region support.
Submit comparison decisions directly after resolving their build and organization. Keep command IDs for safe auth retries without fetching or sending a review counter.
Contributor
Author
|
Superseded by #354, which is based on the latest main and contains only the seven changelog model-setting changes. |
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
1 participant
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.
Why
Release changelog generation currently leaves model selection to Codex’s default. Pinning the model keeps CLI and client release notes on the intended model instead of letting a future default change alter release output.
Approach
Set
model: gpt-5.6-lunaon everyopenai/codex-actioninvocation in the seven release workflows. This is limited to workflow configuration; runtime code and generated changelogs are unchanged.Evidence
All seven release workflows now use the same explicit model setting, and the branch diff is limited to those workflow configuration lines.