fix: make issue-thread backpressure tracker-aware #1214

Merged

coilysiren merged 1 commit from issue-1213 into main

2026-07-13 23:43:21 +00:00

coilyco-ops commented

2026-07-13 23:34:31 +00:00

Member

What changed: repo backpressure now routes through the carried ref and tracker-aware issue scanning.
Why: Forgejo was incorrectly treated as the implicit authority for GitHub, GitLab, and Shortcut refs.
Tests: ward exec test.
closes #1213
ward.workflow: pull-request-and-merge

tracker-aware repo backpressure now carries the issue ref instead of rebuilding authority from owner/repo, so issue-thread capacity checks stay aligned with the tracked tracker and explicit unsupported trackers fall back safely. - What changed: repo backpressure now routes through the carried ref and tracker-aware issue scanning. - Why: Forgejo was incorrectly treated as the implicit authority for GitHub, GitLab, and Shortcut refs. - Tests: ward exec test. closes #1213 ward.workflow: pull-request-and-merge

coilyco-ops added 1 commit

2026-07-13 23:34:31 +00:00

fix: make repo backpressure tracker-aware

test / test (pull_request) Successful in 32s

Details

af9fa61524

coilyco-ops commented

2026-07-13 23:36:51 +00:00

Author

Member

WARDED_WORKFLOW review-note

verdict: block
confidence: high
summary: The implementation carries agentIssueRef into repo backpressure and avoids the old direct Forgejo client for the tested Shortcut path, but it does not satisfy the explicit cross-tracker test contract from ward#1213. The PR should not merge until GitHub and GitLab refs are covered too.
findings:

severity: medium
file: cmd/ward/agent_capacity_test.go
line: 416
problem: The new non-Forgejo backpressure coverage only exercises trackerShortcut. The issue acceptance criteria requires proving the issue-thread capacity path does not call the Forgejo client for GitHub, GitLab, and Shortcut refs.
impact: A later change could regress GitHub or GitLab dispatch backpressure back to Forgejo authority while this test suite still stays green, preserving the exact multi-node issue-authority bug this issue was filed to remove.
conclusion: Review blocks PR 1214 pending explicit GitHub and GitLab backpressure tests alongside the Shortcut case. The code shape looks close, but the safety claim is cross-tracker and the current tests only prove one non-Forgejo tracker.

WARDED_WORKFLOW review-note verdict: block confidence: high summary: The implementation carries `agentIssueRef` into repo backpressure and avoids the old direct Forgejo client for the tested Shortcut path, but it does not satisfy the explicit cross-tracker test contract from ward#1213. The PR should not merge until GitHub and GitLab refs are covered too. findings: - severity: medium file: cmd/ward/agent_capacity_test.go line: 416 problem: The new non-Forgejo backpressure coverage only exercises `trackerShortcut`. The issue acceptance criteria requires proving the issue-thread capacity path does not call the Forgejo client for GitHub, GitLab, and Shortcut refs. impact: A later change could regress GitHub or GitLab dispatch backpressure back to Forgejo authority while this test suite still stays green, preserving the exact multi-node issue-authority bug this issue was filed to remove. conclusion: Review blocks PR 1214 pending explicit GitHub and GitLab backpressure tests alongside the Shortcut case. The code shape looks close, but the safety claim is cross-tracker and the current tests only prove one non-Forgejo tracker.

~~coilyco-ops referenced this pull request 2026-07-13 23:37:49 +00:00~~

Make issue-thread backpressure tracker-aware #1213

coilyco-ops referenced this pull request

2026-07-13 23:38:18 +00:00

Make issue-thread backpressure tracker-aware #1213