[dezren39]

Queue task 6483 · fixes #4528 (triaged as task 6438).

Reconcile's only item source for PRs was the issues endpoint, whose PR objects carry no draft/merged/merged_at. First-sight PR rows materialized draft=0/merged=0 and the #4158 preservation logic pinned those false defaults forever — re-polling never healed them.

This change sources the prs stream from repos/{repo}/pulls (pulls-shaped, carries the merge fields) instead. Pulls truth lands on first sight and heals legacy false-default rows on a later pass via the existing upsert path. The issues stream is unchanged (still reads the issues endpoint, still skips pull_request-link objects). The reviews stream already enumerated pulls; no change there.

Tests (pm-stack/projector, bun test — 79 pass, 0 fail across 10 files):

• draft-first-sight stores draft=1
• merged-first-sight stores merged=1 + merged_at
• heal-after-gap: legacy false-default row corrected on a later pass
• all pre-existing tests pass, including the projector(items): an issues/issue_comment event for a merged PR resets merged to 0 and clears merged_at — issue-shaped payloads carry no merge fields #4158 issue-shaped preservation tests and the reviews per-PR guard (projector(reconcile): reviews stream is not actually guarded per-PR — one PR's failing reviews endpoint aborts the whole stream, later PRs never reconcile, and the run reports written: 0 despite successful upserts #2745)

Out of scope per packet: issues stream behavior, webhook issue-shaped reset (#4158, fixed by #4516), item deletion (#4210), application-contract design/audit (#4302/#4303).

---

^{Open in [code]smith · Need help on this PR? Tag @codesmith-bot with what you need.}

[coderabbitai]

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 57 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Learn how review limits work.

Review configuration:

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: 2cc7a360-792c-4821-ac0a-051f44910318

Commits

Reviewing files that changed from the base of the PR and between 5bf164b and a09c6cd.

Files selected for processing: 3

• pm-stack/projector/src/projector.ts
• pm-stack/projector/src/reconcile.ts
• pm-stack/projector/test/reconcile.test.ts

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

• X
• Mastodon
• Reddit
• LinkedIn

CodeRabbit Verify

• Run CodeRabbit Verify to test the changes in this pull request.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-03T01:43:54.625767Z	a09c6cd	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c0c9c9157b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Derive merged status from the pulls-list payload

For merged PRs, the GitHub pulls-list response contains merged_at but no merged boolean. Passing it directly to fromIssueOrPr makes that mapper emit merged=0, because it checks only obj.merged. Consequently, this poll not only fails to heal legacy merged flags but also overwrites correct merged=1 values previously stored by webhooks: upsertItem accepts equal updated_at timestamps. The aggregator then displays those PRs as closed rather than merged. Derive the flag from merged_at or fetch the detail payload, and remove the artificial merged field from the list-response test fixtures so they exercise the real API shape.

Useful? React with 👍 / 👎.

[dezren39]

Fixed in a09c6cd — thank you, this was a real bug. fromIssueOrPr now derives merged from merged_at when the boolean is absent (the pulls LIST shape), keeping the explicit boolean when present and null for issue-shaped payloads (#4158 preservation unchanged). The reconcile tests were also made realistic: the pulls fixtures no longer carry a merged key, so merged-first-sight (merged=1 + merged_at) is proven against the actual list-endpoint shape. Full projector suite: 79 pass / 0 fail. The derivation was salvaged from duplicate PR #4687, which had it right; #4687 is being closed as the duplicate with a salvage note here.

[dezren39]

🤖 Agent note · salvage + canonical resolution (task 6483, lane code-d)

Canonical PR: this one (#4597) — the PR bound to queue task 6483, opened by its first claimant lane at c0c9c91. PR #4687 was a later duplicate from a re-claim (branch queue/6483-4528-reconcile-prs); it is being closed as superseded by this PR.

Salvage from #4687 (diffed in full, per protocol): its one unique correctness item — deriving merged from merged_at for pulls-LIST payloads in projector.ts, plus its realistic no-merged-key test fixtures — is salvaged here in a09c6cd. Its since= query param was NOT salvaged: the pulls list endpoint does not support since (GitHub ignores it), so it changed nothing.

Codex dispositions:

• P1 on this PR (merged derivation): FIXED in a09c6cd; replied on the thread.
• P2 raised on fix(pm-stack): prs reconcile stream reads pulls-shaped data (#4528) #4687 (full-history pulls download each pass, no incremental bound): not fixed in either PR. Accepted for now with this reasoning: the reviews stream in the same reconcile already enumerates the full pulls list every pass, so this change adds no new class of cost; correctness of draft/merged/merged_at was the defect at hand (projector(reconcile): the prs stream reads the issues endpoint, so reconcile can never learn draft/merged/merged_at — first-sight PRs are stored draft=0/merged=0 forever #4528). A cursor-bounded pulls fetch is a separate optimization, not filed as part of this fix.

Tests: bun test in pm-stack/projector on the laptop worktree: 79 pass / 0 fail, including draft-first-sight, merged-first-sight, and heal-after-gap.