[juliusmarminge]

No new items to fix this round, so the branch is unchanged at ef1c048. I watched PR #36910 until the checks and bot reviews finished. All 37 checks passed on ef1c048. Macroscope approved it with no line comments. CodeRabbit reported 'No actionable comments were generated'. The change itself is as before: PullRequestService.readTogether passes each repository's seenAt as {repository, number} pairs through the new ProviderAcrossListCursor. GitHubPullRequestApi.searchPullRequests and listPullRequests drop rows at the boundary that were already delivered before applying the limit, using the shared collect pager. Truncation counts only the rows actually dropped. CodeRabbit noted that the PR description's 'Not in this change' bullet still says the readTogether multi-repository path is left alone. That bullet came from the first round's outOfScope and has been out of date since the second round fixed that path, so whoever edits the PR body should remove it.

Fixes #36885

Reproduction on main

Commit baf2558 (origin/main). Ran a fake GitHub Enterprise server (GraphQL search honouring updated:<= inclusively and first/after paging) on 127.0.0.1:443 with a self-signed cert, and started vp run dev with T3CODE_DEV_SEED=0, GH_HOST=127.0.0.1, GH_ENTERPRISE_TOKEN=, NODE_EXTRA_CA_CERTS=. Fixture repository acme/web with 230 open PRs: PR1-PR20 newer with distinct times, PR21-PR220 sharing one update time (2026-07-02T00:00:00Z), PR221-PR230 older. Added a git checkout with that remote as a project, paired the web client, opened /pull-requests?involvement=all&state=open&projectId=. First page: PR1-PR99 (shipped limit 99) plus 'Load more pull requests'. First Load more: server sent is:pr is:open updated:<=2026-07-02T00:00:00Z sort:updated-desc repo:acme/web (first 100) and the fake returned PR21-PR120; only PR100-PR119 were added. Second Load more: identical search, same PR21-PR120, nothing added. Third Load more: no request at all. The list stays at PR119 with Load more still shown; PR120-PR220 (tied) and PR221-PR230 (older) cannot be reached. Captures: https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/7acd918a69bdcfdb/01-first-page.png (first page) and https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/21eee16e2653a489/02-stuck-after-load-more.png (stuck at PR119). Fake server driver: https://gh-file-drop-api-prod-mi5fy3sowv63ufte.pinglabs.workers.dev/f/854ecf0ad7488d98/server.py. Also added a unit test in packages/source-control-github/src/server/GitHubPullRequestApi.test.ts ('skips the rows already delivered at a tied boundary before applying the limit'): with 20 tied rows plus 5 older and a cursor covering PR1-PR10 (limit 10), listPullRequests returns PR1-PR10 again instead of PR11-PR20; it fails on main.

Evidence

Real web app, Chromium, 1280×800, light theme, fresh database on each commit. The same GitHub Enterprise fixture contains 230 open PRs, with PR21–PR220 sharing one update time. Open the pull request list with All involvement and Open state, then repeatedly click Load more.

Proven: base stops at PR119; head reaches all 230 PRs and removes Load more. scenarios:prove exited 0.

Step	Before — baf2558017	After — ef1c048917
First page
First Load more
Second Load more
Third Load more		List exhausted after two clicks; button absent.
Final result
Runner outcome	119 rows; failure capture	230 rows; passed
Recording	Before interaction	After interaction

Scenario source · Synthetic GitHub server · Fixture request transcript

Tests

This round: CI on ef1c048, all 37 checks passed (Lint, Build and Typecheck, Test Server 1-8, Test Web 1-2, Test, Macroscope checks, CodeRabbit). The local runs from the previous round still apply: GitHub package vp test run src/server/ 384/384, server vp test run src/pullRequest/ 162/162, vp run check:changed passing.

Not in this change

• PullRequestService readTogether (the listChangeRequestsAcross / multi-repository searchPullRequests path) passes only updatedBefore and delivered in its cursor, never seenAt, so a full page of equal update times across several repositories in one search may stall the same way; fixing it needs per-repository seenAt across a shared boundary, which is a separate change.
• searchPullRequests now reads further pages when a full first page is needed to get past seen rows. Without a cursor its request count is unchanged for limits up to 99.
• I did not repeat the end-to-end run against a fake GitHub Enterprise server from the issue; the unit tests cover both the batched search and the per-repository path.
• The PR description's 'Not in this change' bullet about readTogether is out of date (the second round fixed that path); it needs a PR body edit, which this step does not make.

Who did the work

• Written by Claude Opus 5.5, in Claude Code
• Reviewed by GPT-6.1 Sol, in Codex, and by Claude Opus 5.5, each on its own
• Proof captured by GPT-6.1 Sol and checked by Claude Opus 5.5

Run by the t3code factory in T3 Code.

[macroscopeapp]

Approvability

Verdict: Approved at ef1c048

Macroscope's review found this PR approvable — GitHub “Load more” now skips previously delivered rows sharing an update timestamp, including across repositories. The change is confined to read-only pagination, with limited result sizes and cursor stopping safeguards.

^{You can add or adjust custom eligibility rules. Learn more.}

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	21.5 KiB	21.5 KiB	0 B (0.0%)	24.4 KiB	✅
Codex	Live turn messages	1	1	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.9 KiB	21.9 KiB	0 B (0.0%)	24.4 KiB	✅
Claude	Live turn messages	1	1	0 (0.0%)	8	✅

Other scenarios

All 64 scenario measurements match main.

All 64 scenario measurements

Provider	Scenario	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	short turn resume	WebSocket decoded	10.5 KiB	10.5 KiB	0 B (0.0%)	12.7 KiB	✅
Codex	short turn resume	WebSocket messages	2	2	0 (0.0%)	2	✅
Codex	long turn running, cold open	HTTP decoded	281.7 KiB	281.7 KiB	0 B (0.0%)	302.7 KiB	✅
Codex	long turn running, cold open	rows	75	75	0 (0.0%)	75	✅
Codex	long turn running, history page	HTTP decoded	92.4 KiB	92.4 KiB	0 B (0.0%)	107.4 KiB	✅
Codex	long turn running, history page	rows	75	75	0 (0.0%)	75	✅
Codex	long turn finished, cold open	HTTP decoded	381.5 KiB	381.5 KiB	0 B (0.0%)	410.2 KiB	✅
Codex	long turn finished, cold open	rows	151	151	0 (0.0%)	151	✅
Codex	long turn finished, socket cold open	WebSocket decoded	381.6 KiB	381.6 KiB	0 B (0.0%)	410.2 KiB	✅
Codex	long turn finished, socket cold open	WebSocket messages	2	2	0 (0.0%)	2	✅
Codex	long turn finished, socket cold open	rows	151	151	0 (0.0%)	151	✅
Codex	long turn finished, work page	HTTP decoded	186.2 KiB	186.2 KiB	0 B (0.0%)	200.2 KiB	✅
Codex	long turn finished, work page	rows	150	150	0 (0.0%)	150	✅
Codex	idle thread resume	WebSocket decoded	65 B	65 B	0 B (0.0%)	200 B	✅
Codex	idle thread resume	WebSocket messages	1	1	0 (0.0%)	1	✅
Codex	fleet shell snapshot	HTTP decoded	42.4 KiB	42.4 KiB	0 B (0.0%)	53.7 KiB	✅
Codex	fleet shell snapshot	rows	27	27	0 (0.0%)	27	✅
Codex	fleet shell subscribe	WebSocket decoded	65 B	65 B	0 B (0.0%)	200 B	✅
Codex	fleet shell subscribe	WebSocket messages	1	1	0 (0.0%)	1	✅
Codex	fleet shell resume check	HTTP decoded	18 B	18 B	0 B (0.0%)	100 B	✅
Codex	fleet shell resume check	rows	0	0	0	0	✅
Codex	stale shell reload	HTTP decoded	42.4 KiB	42.4 KiB	0 B (0.0%)	53.7 KiB	✅
Codex	stale shell reload	rows	27	27	0 (0.0%)	27	✅
Codex	stale shell resume	WebSocket decoded	140 B	140 B	0 B (0.0%)	400 B	✅
Codex	stale shell resume	WebSocket messages	2	2	0 (0.0%)	2	✅
Codex	stale todo list cold open	HTTP decoded	72.4 KiB	72.4 KiB	0 B (0.0%)	83.0 KiB	✅
Codex	stale todo list cold open	rows	75	75	0 (0.0%)	75	✅
Codex	stale todo list cold open	nodes	2	2	0 (0.0%)	4	✅
Codex	stale todo list cold open	subagents	2	2	0 (0.0%)	4	✅
Codex	stale todo list cold open	plans	1	1	0 (0.0%)	2	✅
Codex	stale todo list cold open	providerTurns	1	1	0 (0.0%)	2	✅
Codex	stale todo list cold open	checkpoints	1	1	0 (0.0%)	2	✅
Claude	short turn resume	WebSocket decoded	10.6 KiB	10.6 KiB	0 B (0.0%)	12.7 KiB	✅
Claude	short turn resume	WebSocket messages	2	2	0 (0.0%)	2	✅
Claude	long turn running, cold open	HTTP decoded	283.3 KiB	283.3 KiB	0 B (0.0%)	302.7 KiB	✅
Claude	long turn running, cold open	rows	75	75	0 (0.0%)	75	✅
Claude	long turn running, history page	HTTP decoded	93.4 KiB	93.4 KiB	0 B (0.0%)	107.4 KiB	✅
Claude	long turn running, history page	rows	75	75	0 (0.0%)	75	✅
Claude	long turn finished, cold open	HTTP decoded	384.2 KiB	384.2 KiB	0 B (0.0%)	410.2 KiB	✅
Claude	long turn finished, cold open	rows	151	151	0 (0.0%)	151	✅
Claude	long turn finished, socket cold open	WebSocket decoded	384.3 KiB	384.3 KiB	0 B (0.0%)	410.2 KiB	✅
Claude	long turn finished, socket cold open	WebSocket messages	2	2	0 (0.0%)	2	✅
Claude	long turn finished, socket cold open	rows	151	151	0 (0.0%)	151	✅
Claude	long turn finished, work page	HTTP decoded	188.1 KiB	188.1 KiB	0 B (0.0%)	200.2 KiB	✅
Claude	long turn finished, work page	rows	150	150	0 (0.0%)	150	✅
Claude	idle thread resume	WebSocket decoded	66 B	66 B	0 B (0.0%)	200 B	✅
Claude	idle thread resume	WebSocket messages	1	1	0 (0.0%)	1	✅
Claude	fleet shell snapshot	HTTP decoded	43.0 KiB	43.0 KiB	0 B (0.0%)	53.7 KiB	✅
Claude	fleet shell snapshot	rows	27	27	0 (0.0%)	27	✅
Claude	fleet shell subscribe	WebSocket decoded	66 B	66 B	0 B (0.0%)	200 B	✅
Claude	fleet shell subscribe	WebSocket messages	1	1	0 (0.0%)	1	✅
Claude	fleet shell resume check	HTTP decoded	18 B	18 B	0 B (0.0%)	100 B	✅
Claude	fleet shell resume check	rows	0	0	0	0	✅
Claude	stale shell reload	HTTP decoded	43.0 KiB	43.0 KiB	0 B (0.0%)	53.7 KiB	✅
Claude	stale shell reload	rows	27	27	0 (0.0%)	27	✅
Claude	stale shell resume	WebSocket decoded	140 B	140 B	0 B (0.0%)	400 B	✅
Claude	stale shell resume	WebSocket messages	2	2	0 (0.0%)	2	✅
Claude	stale todo list cold open	HTTP decoded	73.4 KiB	73.4 KiB	0 B (0.0%)	83.0 KiB	✅
Claude	stale todo list cold open	rows	75	75	0 (0.0%)	75	✅
Claude	stale todo list cold open	nodes	2	2	0 (0.0%)	4	✅
Claude	stale todo list cold open	subagents	2	2	0 (0.0%)	4	✅
Claude	stale todo list cold open	plans	1	1	0 (0.0%)	2	✅
Claude	stale todo list cold open	providerTurns	1	1	0 (0.0%)	2	✅
Claude	stale todo list cold open	checkpoints	1	1	0 (0.0%)	2	✅

Baseline: 2c607ad · PR result: ef1c048 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[coderabbitai]

No actionable comments were generated in the recent review.

Walkthrough

GitHub pull request pagination now carries repository-and-number cursor exclusions across repositories. List and search requests account for previously delivered rows at tied update-time boundaries when collecting results and determining truncation.

Changes: Tied-timestamp pagination

Layer / File(s)	Summary
Cursor contract and propagation packages/source-control-core/src/server/PullRequestProvider.ts, apps/server/src/pullRequest/PullRequestService.ts, apps/server/src/pullRequest/PullRequestService.test.ts	The across-repository cursor accepts seen pull requests as repository-and-number pairs. Cross-repository reads pass those entries with the timestamp boundary and delivered count. The service test expects the cursor to include a seen entry.
GitHub pagination and boundary filtering packages/source-control-github/src/server/GitHubPullRequestApi.ts, packages/source-control-github/src/server/GitHubPullRequestApi.test.ts	List and search pagination account for previously delivered rows at the timestamp boundary before applying result limits. The shared page collector handles raw-row limits and repeated cursors. Tests cover tied timestamps and moved boundaries.

Estimated code review effort: 3 (Moderate) | ~25 minutes

Change: Bug fix · Severity of issue fixed: Low

Sequence Diagram(s)

sequenceDiagram
  participant PullRequestService
  participant GitHubPullRequestApi
  participant GitHubListAndSearch
  PullRequestService->>GitHubPullRequestApi: Request rows with updatedBefore and seenAt
  GitHubPullRequestApi->>GitHubListAndSearch: Fetch list or search pages
  GitHubListAndSearch-->>GitHubPullRequestApi: Return rows and page cursors
  GitHubPullRequestApi->>GitHubPullRequestApi: Exclude seen boundary rows and apply limit
  GitHubPullRequestApi-->>PullRequestService: Return rows and truncation status

Loading

Suggested reviewers: maria-rcks, t3dotgg, bil0000

Merge Risk: Minimal · up to ef1c0

This change lets "load more" continue past pull requests that share the same update time, and it carries repository-and-number exclusions across repositories. No concrete merge-blocking defect remains after review. The multi-repository search path in readTogether is deliberately left for a separate change.

Pre-merge checks: 4

Check name	Status	Explanation
Title check		The title names the fix (GitHub load-more pagination stalling on a full page of rows with equal update times). The raw summary shows the changes address this: the cursor now carries seenAt, and the se…
Linked Issues check		Issue #36885 requires GitHub pull-request lists to continue past a full group of equal updatedAt values, then reach older rows, without duplicates or an unchanged continuation cursor. Implementation …
Out of Scope Changes check		All changed areas serve the cursor and pagination fix for #36885: - PullRequestService.ts and PullRequestService.test.ts: pass seenAt through the readTogether cursor and update the cursor assertion. -…
Description check		The description follows the repository template closely. It has a problem statement, a change summary, a scope section that links issue #36885, and a verification section with before/after evidence an…

Full details: Title check

Explanation

The title names the fix (GitHub load-more pagination stalling on a full page of rows with equal update times). The raw summary shows the changes address this: the cursor now carries seenAt, and the search and list reads skip already-delivered tied rows. The title is specific and matches the main change. It uses the conventional commit form. It does not mention the readTogether multi-repository path, which the author lists as out of scope, but that is a detail and not the main change.

Full details: Linked Issues check

Explanation

Issue #36885 requires GitHub pull-request lists to continue past a full group of equal updatedAt values, then reach older rows, without duplicates or an unchanged continuation cursor.

Implementation at head ef1c048:

• PullRequestService.readTogether (apps/server/src/pullRequest/PullRequestService.ts) now sends seenAt as {repository, number} pairs for each project in the chunk. The diff confirms this. The PR description section 'Not in this change' says the opposite; that bullet is stale and conflicts with the code.
• PullRequestProvider.ts adds ProviderAcrossListCursor, which carries the repository-scoped seenAt, and changes listChangeRequestsAcross to accept it.
• GitHubPullRequestApi.ts searchPullRequests reads additional pages through the shared collect pager. It drops rows at updatedBefore that were already delivered, then applies the limit. Truncation counts only rows actually dropped.
• listPullRequests applies the same skip-before-limit logic.
• GitHubPullRequestApi.test.ts adds a tied-boundary regression test for both methods. The author reports it fails on the previous head and passes at head.

The author reports src/server (384/384) and src/pullRequest (162/162) pass. I did not run these tests. The end-to-end fake GitHub Enterprise run is not captured, and the Evidence section is still pending. The issue's acceptance criteria are coding requirements covered by unit tests, so the missing end-to-end capture does not block this verdict.

Full details: Out of Scope Changes check

Explanation

All changed areas serve the cursor and pagination fix for #36885:

• PullRequestService.ts and PullRequestService.test.ts: pass seenAt through the readTogether cursor and update the cursor assertion.
• PullRequestProvider.ts: add ProviderAcrossListCursor and ProviderRepositoryNumber, and update the listChangeRequestsAcross signature.
• GitHubPullRequestApi.ts: make the list page type generic, share the collect pager at module level, add seenAt skipping and truncation accounting to list and search, and type searchPullRequests with the new cursor.
• GitHubPullRequestApi.test.ts: regression coverage for tied boundaries.

The summary shows no unrelated feature or refactoring changes. The PR description states that no new regex or parser was added. Formatting-only changes, if any, apply to changed files.

Full details: Description check

Explanation

The description follows the repository template closely. It has a problem statement, a change summary, a scope section that links issue #36885, and a verification section with before/after evidence and test results. It also names the models and harness used. Three gaps remain. The 'Not in this change' section says the readTogether multi-repository path is left alone, but the author says that bullet is out of date. The author also says the readTogether path still passes only updatedBefore and delivered, so the cursor gap is open. The Problem section has no separate reproduction steps in template form, although the Reproduction section covers them. Overall the description is mostly complete and gives enough detail to review the change.

Finishing Touches

Generate unit tests (beta)

• Commit to this branch
• Create a new PR

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

Recent review info

Run configuration

• Source mode: Auto
• Repository: pingdotgg/t3code
• Branch: factory/issue-36885-a36886
• File: .coderabbit.config.ts
• Commit: ef1c048
• Layers: Repository TS config
• Review profile: CHILL
• Plan: Team
• Run ID: 0fbabfe2-df9b-4892-8411-ca321c537c9c

Commits

Reviewing files that changed from the base of the PR and between 2c607ad and ef1c048.

Files selected for processing: 5

• apps/server/src/pullRequest/PullRequestService.test.ts
• apps/server/src/pullRequest/PullRequestService.ts
• packages/source-control-core/src/server/PullRequestProvider.ts
• packages/source-control-github/src/server/GitHubPullRequestApi.test.ts
• packages/source-control-github/src/server/GitHubPullRequestApi.ts

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 3 remain after this review.