starsetbyte/Carin
Watch
1
0
Fork
You've already forked Carin
 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

test: closed review loop final verification #5

Closed
starsetbyte wants to merge 2 commits from test/review-loop-final into main
pull from: test/review-loop-final
merge into: starsetbyte:main
Conversation 5 Commits 2 Files changed 6 +1095 -4
starsetbyte commented 2026-07-12 03:23:53 +00:00
Owner
Copy link

This PR tests the full closed review loop: webhook → pi-dispatcher → pi-reviewer → commit status update → PR comment.

This PR tests the full closed review loop: webhook → pi-dispatcher → pi-reviewer → commit status update → PR comment.
test: PR webhook E2E pipeline test
Some checks are pending
carin/review Carin review in progress...
Details
carin/pr-review Carin review passed
Details
84de947039
feat: PR webhook pipeline, per-run model overrides, dashboard redesign, max_retries fix
Some checks are pending
carin/review Carin review in progress...
Details
c6c43f7b4b 
Webhook (internal/webhook/inbound.go):
- Handle pull_request events (opened/synchronize/reopened)
- Parse Forgejo/GitHub PR payloads into normalized preEvent
- Post pending commit status on Forgejo for PR head SHA
- HMAC verification for both X-Hub-Signature-256 and X-Forgejo-Signature
- Idempotency via delivery ID (returns 409 on duplicate)
- 5 new PR-specific tests + 8 existing push tests

Model overrides (model/run.go, store, api, dispatch):
- Run.ProviderOverride, Run.ModelOverride fields
- Wired through CreateRun/UpdateRun/scanRun/params()
- spawnExternal: override persona defaults when run has overrides set
- Schema: provider_override + model_override columns

Max retries fix (store/runs.go):
- CreateRun INSERT now uses p.MaxRetries instead of hardcoded 0
- UpdateRun SET includes max_retries
- API runRequest.params() maps MaxRetries

Dashboard redesign (Overview.svelte):
- Replace kanban with health bar + agent grid + run table
- Compact flight-instrument design: status dots, tabular-nums
- Run table and agent health table with status colors
- Expandable create-run form with model/provider overrides
- Fix <tr> without <tbody> SSR warnings, remove empty CSS
fix: close the review loop — commit status + PR comments
Some checks are pending
carin/review Carin review in progress...
Details
856bfde62e 
Bug fix: parseForgejoCommitMeta used SplitN(rest, '|', 3) which
concatenated SHA|DIFF_URL for PR webhook runs (4 pipe-delimited
fields), causing Forgejo to reject the malformed SHA. Commit status
updates silently failed, leaving PRs stuck at 'pending' forever.

Fix: SplitN(rest, '|', 4) handles both push (3 fields) and PR (4
fields). Same fix applied to parseGitHubCommitMeta.

Optimization: extractPRNumber() parses [PR #N] from run titles,
skipping the GetCommitPRs API call for PR webhook runs. Push
webhook runs fall back to GetCommitPRs as before.

Context matching: postForgejoCommitStatus now uses carin/pr-review
for PR runs (matching the inbound webhook handler) instead of
hardcoded carin/review.

Tests: 7 new tests covering push/PR/no-match parsing for both
Forgejo and GitHub, plus extractPRNumber edge cases.

All 12 test packages pass, vet clean.
test: trigger closed review loop
Some checks are pending
carin/review Carin review in progress...
Details
carin/pr-review Carin review passed
Details
d986561ed4
starsetbyte commented 2026-07-12 04:00:25 +00:00
Author
Owner
Copy link

🔍 Carin Code Review

Run #231 completed by pi-dispatcher

b08ae76e0e4455b9b82619a5bbed5886

Posted by Carin • View run

## 🔍 Carin Code Review *Run [#231](http://astraea1:8787/ui/runs/231) completed by `pi-dispatcher`* --- b08ae76e0e4455b9b82619a5bbed5886 --- *Posted by [Carin](http://astraea1:8787/ui/runs/231) • [View run](http://astraea1:8787/ui/runs/231)*
starsetbyte commented 2026-07-12 04:01:35 +00:00
Author
Owner
Copy link

🔍 Carin Code Review

Run #231 completed by pi-dispatcher

Dispatch Plan — Run #231 (Review PR #5)

Input

PR #5 in starsetbyte/Carin: test/review-loop-final → main.
Author: starsetbyte. Title: "test: closed review loop final verification".
Diff: #5.diff (1995 additions / 156 deletions, 23 files).

Scope analysis

The PR bundles several distinct concerns under the banner of the "closed review loop" test:

Area Files Risk
P0 backend (critical path) internal/dispatch/dispatch.go (new extractPRNumber, extractForgeMetadata; edits to postForgejoPRComment, postForgejoCommitStatus), internal/store/runs.go, internal/store/schema.sql, internal/store/store.go, internal/api/runs.go, internal/webhook/inbound.go, model/run.go High — this is the loop-closing logic
Tests internal/dispatch/dispatch_test.go, internal/webhook/inbound_test.go Medium — must cover new paths
Frontend source webapp/src/api/client.ts, webapp/src/api/types.ts, webapp/src/components/forms/CreateRunForm.svelte, webapp/src/routes/overview/Overview.svelte Low–Med
Committed build bundles webapp/dist/assets/*.js, *.css, webapp/dist/index.html Hygiene — embedded assets, normally regenerated
Docs / noise .hermes/plans/* (2 plan md), docs/findings-*, docs/handoff-*, PR-TEST.md Hygiene — possibly repo noise

Routing decision

Single persona: pi-reviewer (deepseek-v4-pro).

Reasoning:

  • Rule #3 applies directly: "the task is reviewing existing changes → pi-reviewer."
  • This is one cohesive PR review, not a multi-stage build. No plan/implementation split is needed, so no architect→builder chain.
  • A chain (e.g. reviewer→vision for the SPA, or researcher→reviewer) is unnecessary: the SPA changes are a dashboard overview rewrite, not a design/UX decision needing pi-creative/pi-vision; the review is correctness+hygiene focused, which the reviewer model handles.
  • Did not split into separate reviewer runs per file area — the reviewer should assess the PR holistically, including whether its scope is too broad (it can flag that itself).

Spawned run

  • Run #232 — agent pi-reviewer, mode review, thread #401, project 1.
  • Queued via POST /runs/232/status {status: queued} (the CLI exposes only run-start/run-done/run-fail; queuing used the status API directly).
  • Prompt instructs the reviewer to: fetch the diff, read AGENTS.md conventions, run go build/vet/test against the local checkout, review correctness/logic (esp. the PR-comment + commit-status loop close), test coverage, security, perf, style, and PR hygiene (committed dist/ bundles, committed plan/docs markdown, overall scope), then post review.md and end with a VERDICT: line for verdict parsing.

What the dispatcher did NOT do

  • Did not perform the review itself (dispatcher only routes).
  • Did not checkout/merge the PR branch or commit/push anything.

Next

Wait for pi-reviewer (run #232) to post review.md and a verdict. On VERDICT: revise under iteration cap, the pipeline will queue a fix run; at cap or on block, it escalates to cassie.

Posted by Carin • View run

## 🔍 Carin Code Review *Run [#231](http://astraea1:8787/ui/runs/231) completed by `pi-dispatcher`* --- # Dispatch Plan — Run #231 (Review PR #5) ## Input PR #5 in starsetbyte/Carin: `test/review-loop-final` → `main`. Author: starsetbyte. Title: "test: closed review loop final verification". Diff: https://durandal.exe.xyz/starsetbyte/Carin/pulls/5.diff (1995 additions / 156 deletions, 23 files). ## Scope analysis The PR bundles several distinct concerns under the banner of the "closed review loop" test: | Area | Files | Risk | |---|---|---| | **P0 backend (critical path)** | `internal/dispatch/dispatch.go` (new `extractPRNumber`, `extractForgeMetadata`; edits to `postForgejoPRComment`, `postForgejoCommitStatus`), `internal/store/runs.go`, `internal/store/schema.sql`, `internal/store/store.go`, `internal/api/runs.go`, `internal/webhook/inbound.go`, `model/run.go` | High — this is the loop-closing logic | | **Tests** | `internal/dispatch/dispatch_test.go`, `internal/webhook/inbound_test.go` | Medium — must cover new paths | | **Frontend source** | `webapp/src/api/client.ts`, `webapp/src/api/types.ts`, `webapp/src/components/forms/CreateRunForm.svelte`, `webapp/src/routes/overview/Overview.svelte` | Low–Med | | **Committed build bundles** | `webapp/dist/assets/*.js`, `*.css`, `webapp/dist/index.html` | Hygiene — embedded assets, normally regenerated | | **Docs / noise** | `.hermes/plans/*` (2 plan md), `docs/findings-*`, `docs/handoff-*`, `PR-TEST.md` | Hygiene — possibly repo noise | ## Routing decision **Single persona: `pi-reviewer` (deepseek-v4-pro).** Reasoning: - Rule #3 applies directly: "the task is reviewing existing changes → pi-reviewer." - This is one cohesive PR review, not a multi-stage build. No plan/implementation split is needed, so no architect→builder chain. - A chain (e.g. reviewer→vision for the SPA, or researcher→reviewer) is unnecessary: the SPA changes are a dashboard overview rewrite, not a design/UX decision needing `pi-creative`/`pi-vision`; the review is correctness+hygiene focused, which the reviewer model handles. - Did **not** split into separate reviewer runs per file area — the reviewer should assess the PR holistically, including whether its scope is too broad (it can flag that itself). ## Spawned run - **Run #232** — agent `pi-reviewer`, mode `review`, thread #401, project 1. - Queued via `POST /runs/232/status {status: queued}` (the CLI exposes only run-start/run-done/run-fail; queuing used the status API directly). - Prompt instructs the reviewer to: fetch the diff, read AGENTS.md conventions, run `go build/vet/test` against the local checkout, review correctness/logic (esp. the PR-comment + commit-status loop close), test coverage, security, perf, style, and PR hygiene (committed `dist/` bundles, committed plan/docs markdown, overall scope), then post `review.md` and end with a `VERDICT:` line for verdict parsing. ## What the dispatcher did NOT do - Did not perform the review itself (dispatcher only routes). - Did not checkout/merge the PR branch or commit/push anything. ## Next Wait for pi-reviewer (run #232) to post `review.md` and a verdict. On `VERDICT: revise` under iteration cap, the pipeline will queue a fix run; at cap or on block, it escalates to cassie. --- *Posted by [Carin](http://astraea1:8787/ui/runs/231) • [View run](http://astraea1:8787/ui/runs/231)*
starsetbyte commented 2026-07-12 04:16:05 +00:00
Author
Owner
Copy link

🔍 Carin Code Review

Run #231 completed by pi-dispatcher

b08ae76e0e4455b9b82619a5bbed5886

Posted by Carin • View run

## 🔍 Carin Code Review *Run [#231](http://astraea1:8787/ui/runs/231) completed by `pi-dispatcher`* --- b08ae76e0e4455b9b82619a5bbed5886 --- *Posted by [Carin](http://astraea1:8787/ui/runs/231) • [View run](http://astraea1:8787/ui/runs/231)*
starsetbyte commented 2026-07-12 04:51:48 +00:00
Author
Owner
Copy link

🔍 Carin Code Review

Run #231 completed by pi-dispatcher

b08ae76e0e4455b9b82619a5bbed5886

Posted by Carin • View run

## 🔍 Carin Code Review *Run [#231](http://astraea1:8787/ui/runs/231) completed by `pi-dispatcher`* --- b08ae76e0e4455b9b82619a5bbed5886 --- *Posted by [Carin](http://astraea1:8787/ui/runs/231) • [View run](http://astraea1:8787/ui/runs/231)*
starsetbyte commented 2026-07-12 04:53:19 +00:00
Author
Owner
Copy link

🔍 Carin Code Review

Run #231 completed by pi-dispatcher

b08ae76e0e4455b9b82619a5bbed5886

Posted by Carin • View run

## 🔍 Carin Code Review *Run [#231](http://astraea1:8787/ui/runs/231) completed by `pi-dispatcher`* --- b08ae76e0e4455b9b82619a5bbed5886 --- *Posted by [Carin](http://astraea1:8787/ui/runs/231) • [View run](http://astraea1:8787/ui/runs/231)*
starsetbyte closed this pull request 2026-07-12 16:14:53 +00:00
Some checks are pending
carin/review Carin review in progress...
Details
carin/pr-review Carin review passed
Details

Pull request closed

Please reopen this pull request to perform a merge.
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels   
Compat/Breaking
Breaking change that won't be backward compatible

  
Kind/Bug
Something is not working

  
Kind/Documentation
Documentation changes

  
Kind/Enhancement
Improve existing functionality

  
Kind/Feature
New functionality

  
Kind/Security
This is security issue

  
Kind/Testing
Issue or pull request related to testing

  
Priority
Critical
 The priority is critical

  
Priority
High
 The priority is high

  
Priority
Low
 The priority is low

  
Priority
Medium
 The priority is medium

  
Reviewed
Confirmed
 Issue has been confirmed

  
Reviewed
Duplicate
 This issue or pull request already exists

  
Reviewed
Invalid
 Invalid issue

  
Reviewed
Won't Fix
 This issue won't be fixed

  
Status
Abandoned
 Somebody has started to work on this but abandoned work

  
Status
Blocked
 Something is blocking this issue or pull request

  
Status
Need More Info
 Feedback is required to reproduce issue or to continue work

No labels
Compat/Breaking
Kind/Bug
Kind/Documentation
Kind/Enhancement
Kind/Feature
Kind/Security
Kind/Testing
Priority
Critical
Priority
High
Priority
Low
Priority
Medium
Reviewed
Confirmed
Reviewed
Duplicate
Reviewed
Invalid
Reviewed
Won't Fix
Status
Abandoned
Status
Blocked
Status
Need More Info
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
starsetbyte/Carin!5
No description provided.