Repository navigation
Conversation
1f952ed to
5a97566
Compare
|
Rebased the branch onto current Validation after rebase:
|
SamMorrowDrums
left a comment
There was a problem hiding this comment.
Targeted fix for #2524 — makes it clear that create without event produces a pending review, and refreshes the description to cover all five methods (including resolve_thread/unresolve_thread). README and toolsnap are updated consistently.
Minor nit: the user-facing Title annotation still doesn't mention thread resolve/unresolve, while the description does. Could broaden in a follow-up (e.g. "Manage pull request reviews and review threads"), not blocking. LGTM.
133347f to
cc3f9b9
Compare
|
Rebased this approved branch onto current Validation after rebase:
The PR diff remains limited to |
cc3f9b9 to
1101a48
Compare
|
Rebased this approved branch onto latest Validation after rebase: |
1101a48 to
5cb469b
Compare
|
Rebased this approved branch onto current Validation after rebase: |
a988a47 to
6ab4dad
Compare
6ab4dad to
b2bf8bb
Compare
|
Friendly nudge: this has been approved since May 30 and I've kept the branch current with main since. Happy to adjust anything if the review picture has changed. |
|
Is anything blocking this from merge? Happy to rebase or adjust the wording if the pending-review flow has changed since the approval in May. |
b2bf8bb to
c74723c
Compare
|
Rebased onto current main and resolved the conflicts from the typed-inputs migration (#3396): kept the new
|
Upstream github#3727 removed the thread_resolution_reason feature flag and deleted pkg/github/__toolsnaps__/pull_request_review_write_resolution_reason.snap, because resolution reasons are now always available outside GHES. This branch still carried a wording edit to that snapshot, which produced a modify/delete conflict. Accept the deletion. The variant it described no longer exists, so the wording fix belongs only in the surviving pull_request_review_write tool and snapshot. The pending-review clarification is unchanged there. Test Plan: ```sh go test ./pkg/github/ -run 'TestCreateAndSubmitPullRequestReview|TestCreatePendingPullRequestReview|TestSubmitPendingPullRequestReview|TestDeletePendingPullRequestReview|TestResolveReviewThread|TestUnresolveReviewThread' -count=1 go vet ./pkg/github/ gofmt -l pkg/github/pullrequests.go ``` 5 top-level tests and 23 subtests pass, including the toolsnaps comparison for the surviving snapshot. go vet is clean and gofmt reports no diff. Net diff against current main is 3 files, 6 insertions, 6 deletions: the tool description and title, the matching snapshot, and the README tool table row. Signed-off-by: Yufeng He <40085740+he-yufeng@users.noreply.github.com>
There was a problem hiding this comment.
🟢 Approval recommended
No active issues are identified, and the documentation, tool metadata, and snapshot consistently clarify pending versus submitted review creation behavior.
No open issues 🎉
What changed in this PR
Clarifies the pull_request_review_write tool's review-creation behavior, including that omitting event creates an unsubmitted pending review while preserving submission when an event is provided.
Changes:
- Updates the tool title and description.
- Refreshes the tool metadata snapshot.
- Updates the README tool listing.
| File | Description |
|---|---|
| README.md | Updates the documented tool title to clarify pending review creation. |
| pkg/github/pullrequests.go | Updates the review-write tool metadata and description to clarify its creation behavior. |
| pkg/github/__toolsnaps__/pull_request_review_write.snap | Records the updated review-write tool metadata and description. |
Review effort: Lite (adaptive)
|
@cagesellchen @SamMorrowDrums Requesting review. This is a small wording fix that was conflict-blocked, and the conflict came from your #3727, so you're the right people to confirm the resolution. New head What the PR does: clarifies that The conflict and how I resolved it: #3727 removed the Verified on the merged tree: 5 top-level tests and 23 subtests pass, including the toolsnaps comparison against the surviving snapshot. One thing worth your confirmation: with the flag gone, resolution reasons are always available outside GHES, so the single description now covers both cases. If you'd rather the description explicitly mention that GHES omits |
Summary
pull_request_review_writetitle say thatcreateopens a pending reviewmethod: "create"withouteventdoes not submit the reviewFixes #2524.
To verify
UPDATE_TOOLSNAPS=true go test ./pkg/github -run TestCreateAndSubmitPullRequestReview -count=1go test ./pkg/github -run TestCreateAndSubmitPullRequestReview -count=1go test ./pkg/github -count=1go test ./...git diff --check