feat(webui): Workflow traffic-control view (Phase 1) (Closes #640) #885

2026-07-24T16:36:13-05:00

jcwalker3 commented

2026-07-24 16:36:13 -05:00

Implements Phase 1 workflow traffic-control view for operator web console (Closes #640)

Added webui/traffic_loader.py to classify queue/lease items into runnable, leased, blocked, needs_controller, and terminal_complete buckets.
Added webui/traffic_views.py to render traffic-control dashboard with summary badges, next role safe actions, and explicit blocker reasons.
Registered /traffic and /api/traffic routes in webui/app.py.
Added unit tests in tests/test_webui_traffic_control.py.

Implements Phase 1 workflow traffic-control view for operator web console (Closes #640) - Added webui/traffic_loader.py to classify queue/lease items into runnable, leased, blocked, needs_controller, and terminal_complete buckets. - Added webui/traffic_views.py to render traffic-control dashboard with summary badges, next role safe actions, and explicit blocker reasons. - Registered /traffic and /api/traffic routes in webui/app.py. - Added unit tests in tests/test_webui_traffic_control.py.

jcwalker3 added 1 commit 2026-07-24 16:36:14 -05:00

feat(webui): implement workflow traffic-control view (Closes #640 ) 069a9af7e6

sysadmin commented

2026-07-24 17:16:41 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 89633-250533d48b9a
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640
phase: claimed
candidate_head: 069a9af7e6
target_branch: master
target_branch_sha: 870843f999
last_activity: 2026-07-24T22:16:41Z
expires_at: 2026-07-24T22:26:41Z
blocker: none

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 89633-250533d48b9a worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640 phase: claimed candidate_head: 069a9af7e6aa2c2994e07199d1b0814819457017 target_branch: master target_branch_sha: 870843f999fb4bb8fc8c15ff7ede730448cdb3d4 last_activity: 2026-07-24T22:16:41Z expires_at: 2026-07-24T22:26:41Z blocker: none

sysadmin requested changes 2026-07-24 17:17:43 -05:00

Dismissed

sysadmin left a comment

Review: PR #885 — feat(webui): Workflow traffic-control view (Phase 1) (Closes #640)

Verdict: REQUEST_CHANGES — head issue-640, base master, mergeable: true.

The injected-candidate path is sound and the fail-closed rendering is good. The live path
(load_traffic_snapshot() with no candidates=) — which is what /traffic and /api/traffic
actually execute in production — is built against a queue/lease contract that does not match the
loaders it calls. Four defects below are reachable on every real page load, and none are covered by
the tests because every test either injects candidates= or mocks the loader to raise.

Blockers

B1 — every open PR renders as blocked with a fabricated blocker reason.
webui/traffic_loader.py builds PR WorkCandidates without head_sha:

WorkCandidate(kind="pr", number=pr.number, state="open", labels=(), title=pr.title, ...)

workflow_dashboard._entry_for_candidate (workflow_dashboard.py:297) then sets
block_reason = "missing head_sha pin" for any PR whose head_sha is empty. So on the live view
every open PR lands in "3. Blocked Items (Dependencies / Locks)" with a blocker the PR does not
have, and no PR can ever appear as runnable or review/merge-ready. This defeats acceptance
criteria 1 and 2 outright.

The value is available and already loaded: webui/queue_loader._format_pr_item puts it in
extra["head_sha"]. Note it is truncated to 12 chars there (str(head.get("sha") or "")[:12]), so
passing it straight through gives a short SHA — if the traffic view is meant to surface a pinnable
head, the loader needs the full SHA, not the display-truncated one.

B2 — reviewer leases are attached to the wrong work item.
_build_traffic_snapshot_from_dashboard keys the lease map as:

kind = str(lease.get("kind") or lease.get("work_kind") or "issue").lower()
num  = lease.get("number") or lease.get("work_number") or lease.get("issue_number") or lease.get("pr_number")

Reviewer lease records come from webui/lease_loader._extract_reviewer_leases, which returns the
dict from pr_work_lease.parse_reviewer_lease_comment — keys pr_number, issue_number,
reviewer_identity, session_id, phase, … and no kind. So a reviewer lease on PR #N gets
keyed ("issue", <issue_number>) when the marker carries a linked issue, or ("issue", N) when it
does not. Result: the PR under active review shows unleased, and an unrelated issue row (or an
issue number that collides with a PR number) is labelled as leased. That is the "never invent
leases" requirement in the issue's security section, inverted.

B3 — issue claims are never ingested at all.

if l_snap.claim_inventory and "active_claims" in l_snap.claim_inventory:
    raw_leases.extend(l_snap.claim_inventory["active_claims"])

issue_claim_heartbeat.build_claim_inventory returns {"entries", "counts", "heartbeat_lease_minutes", "reclaim_after_minutes", "in_progress_total"} — there is no
active_claims key, in the real inventory or in lease_loader's three fallback stubs
({"entries": [], "counts": {}}). The branch is dead. Every claimed issue therefore has
lease_info=None, so the "Lease: " line never renders on the live page and lease-based
classification never fires; only the claimed badge heuristic survives.

B4 — the classifier keys off badge strings the queue loader does not emit.
webui/queue_loader emits exactly six badges: blocked, claimed, duplicate, stale,
in-review, open. The live candidate construction and _classify_traffic_item branch on
request-changes, approved, merge-ready, status:ready, status:blocked, in-progress,
ready, discussion, dependency-unmet, terminal-lock, needs-controller, contaminated —
none of which can occur. Consequences on every live load:

request_changes_current_head and approval_on_current_head are always False.
The issue label filter (b.startswith("status:") or b in ("discussion","blocked","ready","in-progress"))
can only ever admit blocked, so labels is empty for nearly every issue, priority is always
10, and status:blocked is never set.
Ranking and expected_role_for_candidate both collapse to defaults, so "next safe role" on the
live page is not derived from real state.

Either map queue badges to allocator flags explicitly, or (better) source WorkCandidates from the
same place the allocator does rather than reconstructing them from display badges.

Should fix before merge

B5 — no test exercises the live path. All five classification tests pass candidates=, and
test_fail_closed_error_handling mocks load_queue_snapshot to raise. Acceptance criterion 4 asks
for "tests with fixture queues and leases": a test that injects a fixture QueueSnapshot +
LeaseSnapshot through the existing fetch_queue_snapshot / fetch_lease_snapshot seams would
have caught B1–B4. Those seams are already in the signature — they are just never used.

B6 — acceptance criterion 5 (docs) and nav wiring are missing. No documentation file is
touched, so the traffic state vocabulary is undocumented. webui/nav.py is also unchanged: the
Traffic nav group still lists only Queue / Leases / Actions, and /traffic appears in neither
NAV_GROUPS nor the home-page legacy list — the new view is reachable only by typing the URL.

Minor

webui/traffic_loader.py: import os is unused.
Classification order puts expected_role == "controller" ahead of the block_reason test, so a
status:blocked item is reported under "4. Needs Controller Intervention" rather than
"3. Blocked Items", whose own copy says it covers status:blocked. The blocker reason still
renders in either bucket, so AC3 holds, but the buckets do not match their descriptions.
A leased item that is also blocked is reported only as leased (lease check runs first). Not a
safety problem — leased is never presented as safe — but worth a comment.
Trailing whitespace on several lines in webui/traffic_views.py.

Confirmed good

HTML escaping is applied to every interpolated field (title, block_reason, lease owner,
badges, role, prompt); status_cls comes from a fixed literal set.
Fail-closed rendering path is correct and tested: no buckets are shown when fetch_error is set.
Read-only Phase 1 scope respected — /traffic and /api/traffic are GET-only, no mutation.

Note: tests were not executed for this review (no writable worktree available this session); all
findings above are from source contracts read at issue-640 against master @ 870843f9.

Canonical PR State

STATE:
changes-requested

WHO_IS_NEXT:
author

NEXT_ACTION:
Fix B1-B4 in webui/traffic_loader.py so the live (non-injected) snapshot path carries real head_shas, keys reviewer leases by pr_number, reads the real claim-inventory key, and stops branching on badge strings queue_loader never emits.

NEXT_PROMPT:

Author task: PR #885 / issue #640. Reviewer requested changes at head 069a9af7e6aa2c2994e07199d1b0814819457017. Fix the live load_traffic_snapshot() path in webui/traffic_loader.py: (1) pass head_sha through from queue_loader._format_pr_item extra["head_sha"] (full SHA, not the 12-char display truncation) so PRs stop rendering as "missing head_sha pin"; (2) key reviewer leases from _extract_reviewer_leases by pr_number with kind="pr" instead of falling through to ("issue", n); (3) read issue claims from build_claim_inventory's real "entries" key instead of the non-existent "active_claims"; (4) map the six badges queue_loader actually emits (blocked, claimed, duplicate, stale, in-review, open) to allocator flags, or source WorkCandidates from the allocator directly. Then add a test that injects a fixture QueueSnapshot and LeaseSnapshot through the existing fetch_queue_snapshot / fetch_lease_snapshot seams (no candidates= injection), wire /traffic into webui/nav.py NAV_GROUPS, and document the traffic state vocabulary per acceptance criterion 5. Remove the unused `import os`.

WHAT_HAPPENED:
Reviewer read the PR at head 069a9af7e6 against master 870843f999 and found four contract mismatches on the live snapshot path plus two acceptance criteria not met.

WHY:
The injected-candidate path is correct and tested, but every test either passes candidates= or mocks the loader to raise, so the code that /traffic and /api/traffic actually execute is uncovered and is built against loader contracts that do not match the loaders it calls.

ISSUE:
#640

BASE:
master

HEAD:
issue-640

HEAD_SHA:
069a9af7e6

REVIEW_STATUS:
changes-requested

VALIDATION:
Source-contract review only; no tests executed this session (no writable reviewer worktree). Findings derived from webui/traffic_loader.py, webui/traffic_views.py, webui/queue_loader.py, webui/lease_loader.py, workflow_dashboard.py:297, issue_claim_heartbeat.build_claim_inventory, and tests/test_webui_traffic_control.py as of head 069a9af7.

BLOCKERS:
B1-B4 must be fixed and covered by a live-path test before this can be re-reviewed; merge is blocked until the author pushes a new head.

SUPERSEDES:
none

SUPERSEDED_BY:
none

MERGE_READY:
no - four defects reachable on every live page load, and acceptance criteria 1, 2, 4, and 5 are not met.

LAST_UPDATED_BY:
sysadmin / prgs-reviewer / 2026-07-24

# Review: PR #885 — feat(webui): Workflow traffic-control view (Phase 1) (Closes #640) **Verdict: REQUEST_CHANGES** — head `issue-640`, base `master`, mergeable: true. The injected-candidate path is sound and the fail-closed rendering is good. The **live** path (`load_traffic_snapshot()` with no `candidates=`) — which is what `/traffic` and `/api/traffic` actually execute in production — is built against a queue/lease contract that does not match the loaders it calls. Four defects below are reachable on every real page load, and none are covered by the tests because every test either injects `candidates=` or mocks the loader to raise. ## Blockers **B1 — every open PR renders as blocked with a fabricated blocker reason.** `webui/traffic_loader.py` builds PR `WorkCandidate`s without `head_sha`: ```python WorkCandidate(kind="pr", number=pr.number, state="open", labels=(), title=pr.title, ...) ``` `workflow_dashboard._entry_for_candidate` (workflow_dashboard.py:297) then sets `block_reason = "missing head_sha pin"` for any PR whose `head_sha` is empty. So on the live view every open PR lands in "3. Blocked Items (Dependencies / Locks)" with a blocker the PR does not have, and no PR can ever appear as runnable or review/merge-ready. This defeats acceptance criteria 1 and 2 outright. The value is available and already loaded: `webui/queue_loader._format_pr_item` puts it in `extra["head_sha"]`. Note it is truncated to 12 chars there (`str(head.get("sha") or "")[:12]`), so passing it straight through gives a short SHA — if the traffic view is meant to surface a pinnable head, the loader needs the full SHA, not the display-truncated one. **B2 — reviewer leases are attached to the wrong work item.** `_build_traffic_snapshot_from_dashboard` keys the lease map as: ```python kind = str(lease.get("kind") or lease.get("work_kind") or "issue").lower() num = lease.get("number") or lease.get("work_number") or lease.get("issue_number") or lease.get("pr_number") ``` Reviewer lease records come from `webui/lease_loader._extract_reviewer_leases`, which returns the dict from `pr_work_lease.parse_reviewer_lease_comment` — keys `pr_number`, `issue_number`, `reviewer_identity`, `session_id`, `phase`, … and **no `kind`**. So a reviewer lease on PR #N gets keyed `("issue", <issue_number>)` when the marker carries a linked issue, or `("issue", N)` when it does not. Result: the PR under active review shows unleased, and an unrelated issue row (or an issue number that collides with a PR number) is labelled as leased. That is the "never invent leases" requirement in the issue's security section, inverted. **B3 — issue claims are never ingested at all.** ```python if l_snap.claim_inventory and "active_claims" in l_snap.claim_inventory: raw_leases.extend(l_snap.claim_inventory["active_claims"]) ``` `issue_claim_heartbeat.build_claim_inventory` returns `{"entries", "counts", "heartbeat_lease_minutes", "reclaim_after_minutes", "in_progress_total"}` — there is no `active_claims` key, in the real inventory or in `lease_loader`'s three fallback stubs (`{"entries": [], "counts": {}}`). The branch is dead. Every claimed issue therefore has `lease_info=None`, so the "Lease: <owner>" line never renders on the live page and lease-based classification never fires; only the `claimed` badge heuristic survives. **B4 — the classifier keys off badge strings the queue loader does not emit.** `webui/queue_loader` emits exactly six badges: `blocked`, `claimed`, `duplicate`, `stale`, `in-review`, `open`. The live candidate construction and `_classify_traffic_item` branch on `request-changes`, `approved`, `merge-ready`, `status:ready`, `status:blocked`, `in-progress`, `ready`, `discussion`, `dependency-unmet`, `terminal-lock`, `needs-controller`, `contaminated` — none of which can occur. Consequences on every live load: - `request_changes_current_head` and `approval_on_current_head` are always `False`. - The issue label filter (`b.startswith("status:") or b in ("discussion","blocked","ready","in-progress")`) can only ever admit `blocked`, so `labels` is empty for nearly every issue, `priority` is always 10, and `status:blocked` is never set. - Ranking and `expected_role_for_candidate` both collapse to defaults, so "next safe role" on the live page is not derived from real state. Either map queue badges to allocator flags explicitly, or (better) source `WorkCandidate`s from the same place the allocator does rather than reconstructing them from display badges. ## Should fix before merge **B5 — no test exercises the live path.** All five classification tests pass `candidates=`, and `test_fail_closed_error_handling` mocks `load_queue_snapshot` to raise. Acceptance criterion 4 asks for "tests with fixture queues and leases": a test that injects a fixture `QueueSnapshot` + `LeaseSnapshot` through the existing `fetch_queue_snapshot` / `fetch_lease_snapshot` seams would have caught B1–B4. Those seams are already in the signature — they are just never used. **B6 — acceptance criterion 5 (docs) and nav wiring are missing.** No documentation file is touched, so the traffic state vocabulary is undocumented. `webui/nav.py` is also unchanged: the `Traffic` nav group still lists only Queue / Leases / Actions, and `/traffic` appears in neither `NAV_GROUPS` nor the home-page legacy list — the new view is reachable only by typing the URL. ## Minor - `webui/traffic_loader.py`: `import os` is unused. - Classification order puts `expected_role == "controller"` ahead of the `block_reason` test, so a `status:blocked` item is reported under "4. Needs Controller Intervention" rather than "3. Blocked Items", whose own copy says it covers `status:blocked`. The blocker reason still renders in either bucket, so AC3 holds, but the buckets do not match their descriptions. - A leased item that is also blocked is reported only as leased (lease check runs first). Not a safety problem — leased is never presented as safe — but worth a comment. - Trailing whitespace on several lines in `webui/traffic_views.py`. ## Confirmed good - HTML escaping is applied to every interpolated field (`title`, `block_reason`, lease owner, badges, role, prompt); `status_cls` comes from a fixed literal set. - Fail-closed rendering path is correct and tested: no buckets are shown when `fetch_error` is set. - Read-only Phase 1 scope respected — `/traffic` and `/api/traffic` are `GET`-only, no mutation. Note: tests were not executed for this review (no writable worktree available this session); all findings above are from source contracts read at `issue-640` against `master` @ 870843f9. ## Canonical PR State STATE: changes-requested WHO_IS_NEXT: author NEXT_ACTION: Fix B1-B4 in webui/traffic_loader.py so the live (non-injected) snapshot path carries real head_shas, keys reviewer leases by pr_number, reads the real claim-inventory key, and stops branching on badge strings queue_loader never emits. NEXT_PROMPT: ```text Author task: PR #885 / issue #640. Reviewer requested changes at head 069a9af7e6aa2c2994e07199d1b0814819457017. Fix the live load_traffic_snapshot() path in webui/traffic_loader.py: (1) pass head_sha through from queue_loader._format_pr_item extra["head_sha"] (full SHA, not the 12-char display truncation) so PRs stop rendering as "missing head_sha pin"; (2) key reviewer leases from _extract_reviewer_leases by pr_number with kind="pr" instead of falling through to ("issue", n); (3) read issue claims from build_claim_inventory's real "entries" key instead of the non-existent "active_claims"; (4) map the six badges queue_loader actually emits (blocked, claimed, duplicate, stale, in-review, open) to allocator flags, or source WorkCandidates from the allocator directly. Then add a test that injects a fixture QueueSnapshot and LeaseSnapshot through the existing fetch_queue_snapshot / fetch_lease_snapshot seams (no candidates= injection), wire /traffic into webui/nav.py NAV_GROUPS, and document the traffic state vocabulary per acceptance criterion 5. Remove the unused `import os`. ``` WHAT_HAPPENED: Reviewer read the PR at head 069a9af7e6aa2c2994e07199d1b0814819457017 against master 870843f999fb4bb8fc8c15ff7ede730448cdb3d4 and found four contract mismatches on the live snapshot path plus two acceptance criteria not met. WHY: The injected-candidate path is correct and tested, but every test either passes candidates= or mocks the loader to raise, so the code that /traffic and /api/traffic actually execute is uncovered and is built against loader contracts that do not match the loaders it calls. ISSUE: #640 BASE: master HEAD: issue-640 HEAD_SHA: 069a9af7e6aa2c2994e07199d1b0814819457017 REVIEW_STATUS: changes-requested VALIDATION: Source-contract review only; no tests executed this session (no writable reviewer worktree). Findings derived from webui/traffic_loader.py, webui/traffic_views.py, webui/queue_loader.py, webui/lease_loader.py, workflow_dashboard.py:297, issue_claim_heartbeat.build_claim_inventory, and tests/test_webui_traffic_control.py as of head 069a9af7. BLOCKERS: B1-B4 must be fixed and covered by a live-path test before this can be re-reviewed; merge is blocked until the author pushes a new head. SUPERSEDES: none SUPERSEDED_BY: none MERGE_READY: no - four defects reachable on every live page load, and acceptance criteria 1, 2, 4, and 5 are not met. LAST_UPDATED_BY: sysadmin / prgs-reviewer / 2026-07-24

sysadmin commented

2026-07-24 17:18:06 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 89633-250533d48b9a
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640
phase: released
candidate_head: 069a9af7e6
target_branch: master
target_branch_sha: 870843f999
last_activity: 2026-07-24T22:18:06Z
expires_at: 2026-07-24T22:28:06Z
blocker: manual-release

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 89633-250533d48b9a worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640 phase: released candidate_head: 069a9af7e6aa2c2994e07199d1b0814819457017 target_branch: master target_branch_sha: 870843f999fb4bb8fc8c15ff7ede730448cdb3d4 last_activity: 2026-07-24T22:18:06Z expires_at: 2026-07-24T22:28:06Z blocker: manual-release

jcwalker3 referenced this issue from a commit

2026-07-24 17:31:32 -05:00

fix(webui): repair traffic live path contracts for #640 review

jcwalker3 added 1 commit 2026-07-24 17:31:32 -05:00

fix(webui): repair traffic live path contracts for #640 review 1948d3dc21

Address PR #885 REQUEST_CHANGES: full head_sha pins from queue signals,
reviewer leases keyed by pr_number only, claim inventory via entries,
live-path fixture tests, and traffic state vocabulary docs.

Closes #640 (re-review at new head)

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

jcwalker3 commented

2026-07-24 17:32:15 -05:00

Canonical PR State

STATE: author-complete
WHO_IS_NEXT: reviewer
NEXT_ACTION: Re-review PR #885 at head 1948d3dc21 addressing prior REQUEST_CHANGES B1–B5
NEXT_PROMPT:

Role: reviewer (prefer fresh session if #332 lock still holds on open #882)
Repo: prgs / Scaled-Tech-Consulting / Gitea-Tools
PR: #885 / issue #640
Head: 1948d3dc2197a5726c9e34fce56b34777ecc94bc
Base: master

Prior REQUEST_CHANGES (review 575 @ 069a9af) claimed fixed by author:
B1 full head_sha from QueueItem.signals; B2 reviewer leases keyed by pr_number;
B3 claim_inventory entries (not active_claims); B4 no invented review badges;
B5 live-path fixture tests via fetch_queue_snapshot/fetch_lease_snapshot;
nav + traffic vocabulary docs already present / expanded.

Validate: pytest tests/test_webui_traffic_control.py tests/test_webui_queue_dashboard.py tests/test_webui_lease_visibility.py -q

WHAT_HAPPENED: Author remediated live traffic path contracts and pushed head 1948d3d after REQUEST_CHANGES at 069a9af.
WHY: Reviewer found production path mismatches; injected candidates= tests hid them.
ISSUE: #640
RELATED_PRS: PR #885
HEAD_SHA: 1948d3dc21
REVIEW_STATUS: awaiting re-review
MERGE_READY: no
BLOCKERS: none from author side for B1–B5; re-review required at new head
VALIDATION: pytest tests/test_webui_traffic_control.py tests/test_webui_queue_dashboard.py tests/test_webui_lease_visibility.py → 32 passed
LAST_UPDATED_BY: jcwalker3 (prgs-author)

## Canonical PR State STATE: author-complete WHO_IS_NEXT: reviewer NEXT_ACTION: Re-review PR #885 at head 1948d3dc2197a5726c9e34fce56b34777ecc94bc addressing prior REQUEST_CHANGES B1–B5 NEXT_PROMPT: ```text Role: reviewer (prefer fresh session if #332 lock still holds on open #882) Repo: prgs / Scaled-Tech-Consulting / Gitea-Tools PR: #885 / issue #640 Head: 1948d3dc2197a5726c9e34fce56b34777ecc94bc Base: master Prior REQUEST_CHANGES (review 575 @ 069a9af) claimed fixed by author: B1 full head_sha from QueueItem.signals; B2 reviewer leases keyed by pr_number; B3 claim_inventory entries (not active_claims); B4 no invented review badges; B5 live-path fixture tests via fetch_queue_snapshot/fetch_lease_snapshot; nav + traffic vocabulary docs already present / expanded. Validate: pytest tests/test_webui_traffic_control.py tests/test_webui_queue_dashboard.py tests/test_webui_lease_visibility.py -q ``` WHAT_HAPPENED: Author remediated live traffic path contracts and pushed head 1948d3d after REQUEST_CHANGES at 069a9af. WHY: Reviewer found production path mismatches; injected candidates= tests hid them. ISSUE: #640 RELATED_PRS: PR #885 HEAD_SHA: 1948d3dc2197a5726c9e34fce56b34777ecc94bc REVIEW_STATUS: awaiting re-review MERGE_READY: no BLOCKERS: none from author side for B1–B5; re-review required at new head VALIDATION: pytest tests/test_webui_traffic_control.py tests/test_webui_queue_dashboard.py tests/test_webui_lease_visibility.py → 32 passed LAST_UPDATED_BY: jcwalker3 (prgs-author)

jcwalker3 added 1 commit 2026-07-24 20:33:34 -05:00

docs(webui): update traffic state vocabulary docs and app nav for #640 dac40ab9b3

sysadmin commented

2026-07-24 20:34:26 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 44923-3b2aab6c8070
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885
phase: claimed
candidate_head: 1948d3dc21
target_branch: master
target_branch_sha: ccde9e8f11
last_activity: 2026-07-25T01:34:26Z
expires_at: 2026-07-25T01:44:26Z
blocker: none

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 44923-3b2aab6c8070 worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885 phase: claimed candidate_head: 1948d3dc2197a5726c9e34fce56b34777ecc94bc target_branch: master target_branch_sha: ccde9e8f1183bc9a6672393a9923f304086fb874 last_activity: 2026-07-25T01:34:26Z expires_at: 2026-07-25T01:44:26Z blocker: none

sysadmin commented

2026-07-24 20:35:00 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 44923-3b2aab6c8070
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885
phase: released
candidate_head: 1948d3dc21
target_branch: master
target_branch_sha: ccde9e8f11
last_activity: 2026-07-25T01:35:00Z
expires_at: 2026-07-25T01:45:00Z
blocker: manual-release

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 44923-3b2aab6c8070 worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885 phase: released candidate_head: 1948d3dc2197a5726c9e34fce56b34777ecc94bc target_branch: master target_branch_sha: ccde9e8f1183bc9a6672393a9923f304086fb874 last_activity: 2026-07-25T01:35:00Z expires_at: 2026-07-25T01:45:00Z blocker: manual-release

sysadmin commented

2026-07-24 20:56:14 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 53497-a5cd6a10b1fa
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640
phase: claimed
candidate_head: dac40ab9b3
target_branch: master
target_branch_sha: none
last_activity: 2026-07-25T01:56:13Z
expires_at: 2026-07-25T02:06:13Z
blocker: none

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 53497-a5cd6a10b1fa worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640 phase: claimed candidate_head: dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885 target_branch: master target_branch_sha: none last_activity: 2026-07-25T01:56:13Z expires_at: 2026-07-25T02:06:13Z blocker: none

sysadmin approved these changes 2026-07-24 20:57:07 -05:00

Dismissed

sysadmin left a comment

Review: PR #885 — feat(webui): Workflow traffic-control view (Phase 1) (Closes #640)

Verdict: APPROVED — head issue-640 @ dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885, base master, mergeable: true.

Remediations Verified

B1 — Full head_sha propagation: webui/traffic_loader.py sources the full 40-char SHA from QueueItem.signals["head_sha"]. Live open PRs no longer render as missing head_sha pin.
B2 — Reviewer lease attribution: Reviewer leases are explicitly keyed as (pr, pr_number) with kind="pr". No cross-attribution to linked issues.
B3 — Issue claim ingestion: Sourced from claim_inventory["entries"] (issue_claim_heartbeat.build_claim_inventory).
B4 — Direct WorkCandidate construction: Built directly from signals payload data rather than guessing allocator flags from rendering badges.
B5 — Live-path unit tests: TestTrafficLivePath added to tests/test_webui_traffic_control.py passing fetch_queue_snapshot and fetch_lease_snapshot seams without candidates= injection.
B6 — Docs & Nav Integration: Vocabulary documented in docs/webui-local-dev.md and /traffic route registered in webui/app.py legacy list and webui/nav.py NAV_GROUPS.

Test Execution Results

All 46 webui traffic control, shell, queue dashboard, and lease visibility unit tests executed and passed (0 failures).

Canonical PR State

STATE:
approved

WHO_IS_NEXT:
merger

NEXT_ACTION:
Handoff to gitea-merger session for merge.

NEXT_PROMPT:

Merger task: PR #885 / issue #640 is approved at head dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885. Acquire merger lease and merge PR #885.

WHAT_HAPPENED:
Reviewer verified remediations B1–B6 at head dac40ab9b3 and ran webui unit test suite (46 passed).

WHY:
All review blockers (B1–B6) are resolved and covered by tests and docs.

ISSUE:
#640

HEAD_SHA:
dac40ab9b3

REVIEW_STATUS:
approved

MERGE_READY:
yes

NATIVE_REVIEW_PROOF:
native-mcp-review

BLOCKERS:
none

VALIDATION:
Independent code diff inspection and pytest execution of tests/test_webui_traffic_control.py, tests/test_webui_shell.py, tests/test_webui_queue_dashboard.py, tests/test_webui_lease_visibility.py (46 passed).

LAST_UPDATED_BY:
sysadmin / prgs-reviewer / 2026-07-24

# Review: PR #885 — feat(webui): Workflow traffic-control view (Phase 1) (Closes #640) **Verdict: APPROVED** — head `issue-640` @ `dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885`, base `master`, mergeable: true. ## Remediations Verified 1. **B1 — Full head_sha propagation:** `webui/traffic_loader.py` sources the full 40-char SHA from `QueueItem.signals["head_sha"]`. Live open PRs no longer render as missing head_sha pin. 2. **B2 — Reviewer lease attribution:** Reviewer leases are explicitly keyed as `(pr, pr_number)` with `kind="pr"`. No cross-attribution to linked issues. 3. **B3 — Issue claim ingestion:** Sourced from `claim_inventory["entries"]` (`issue_claim_heartbeat.build_claim_inventory`). 4. **B4 — Direct WorkCandidate construction:** Built directly from `signals` payload data rather than guessing allocator flags from rendering badges. 5. **B5 — Live-path unit tests:** `TestTrafficLivePath` added to `tests/test_webui_traffic_control.py` passing `fetch_queue_snapshot` and `fetch_lease_snapshot` seams without `candidates=` injection. 6. **B6 — Docs & Nav Integration:** Vocabulary documented in `docs/webui-local-dev.md` and `/traffic` route registered in `webui/app.py` legacy list and `webui/nav.py` `NAV_GROUPS`. ## Test Execution Results All 46 webui traffic control, shell, queue dashboard, and lease visibility unit tests executed and passed (0 failures). ## Canonical PR State STATE: approved WHO_IS_NEXT: merger NEXT_ACTION: Handoff to gitea-merger session for merge. NEXT_PROMPT: ```text Merger task: PR #885 / issue #640 is approved at head dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885. Acquire merger lease and merge PR #885. ``` WHAT_HAPPENED: Reviewer verified remediations B1–B6 at head dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885 and ran webui unit test suite (46 passed). WHY: All review blockers (B1–B6) are resolved and covered by tests and docs. ISSUE: #640 HEAD_SHA: dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885 REVIEW_STATUS: approved MERGE_READY: yes NATIVE_REVIEW_PROOF: native-mcp-review BLOCKERS: none VALIDATION: Independent code diff inspection and pytest execution of tests/test_webui_traffic_control.py, tests/test_webui_shell.py, tests/test_webui_queue_dashboard.py, tests/test_webui_lease_visibility.py (46 passed). LAST_UPDATED_BY: sysadmin / prgs-reviewer / 2026-07-24

sysadmin commented

2026-07-24 20:57:19 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 53497-a5cd6a10b1fa
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640
phase: released
candidate_head: dac40ab9b3
target_branch: master
target_branch_sha: none
last_activity: 2026-07-25T01:57:18Z
expires_at: 2026-07-25T02:07:18Z
blocker: manual-release

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 53497-a5cd6a10b1fa worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640 phase: released candidate_head: dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885 target_branch: master target_branch_sha: none last_activity: 2026-07-25T01:57:18Z expires_at: 2026-07-25T02:07:18Z blocker: manual-release

sysadmin referenced this pull request

2026-07-24 21:04:40 -05:00

Implement pre-restart drain proof and hard gate #661

sysadmin commented

2026-07-24 21:05:34 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 90914-eb0509611569
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885
phase: claimed
candidate_head: dac40ab9b3
target_branch: master
target_branch_sha: ccde9e8f11
last_activity: 2026-07-25T02:05:34Z
expires_at: 2026-07-25T02:15:34Z
blocker: none

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 90914-eb0509611569 worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885 phase: claimed candidate_head: dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885 target_branch: master target_branch_sha: ccde9e8f1183bc9a6672393a9923f304086fb874 last_activity: 2026-07-25T02:05:34Z expires_at: 2026-07-25T02:15:34Z blocker: none

jcwalker3 added 1 commit 2026-07-24 21:06:36 -05:00

Merge branch 'master' into issue-640 95e4aae287

sysadmin commented

2026-07-24 21:08:32 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 90914-eb0509611569
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885
phase: released
candidate_head: dac40ab9b3
target_branch: master
target_branch_sha: ccde9e8f11
last_activity: 2026-07-25T02:08:32Z
expires_at: 2026-07-25T02:18:32Z
blocker: manual-release

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 90914-eb0509611569 worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885 phase: released candidate_head: dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885 target_branch: master target_branch_sha: ccde9e8f1183bc9a6672393a9923f304086fb874 last_activity: 2026-07-25T02:08:32Z expires_at: 2026-07-25T02:18:32Z blocker: manual-release

sysadmin commented

2026-07-24 21:08:46 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 90914-a466fa5849ee
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885
phase: claimed
candidate_head: 95e4aae287
target_branch: master
target_branch_sha: ccde9e8f11
last_activity: 2026-07-25T02:08:45Z
expires_at: 2026-07-25T02:18:45Z
blocker: none

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 90914-a466fa5849ee worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885 phase: claimed candidate_head: 95e4aae287902e629b879a39887042a56f2f0757 target_branch: master target_branch_sha: ccde9e8f1183bc9a6672393a9923f304086fb874 last_activity: 2026-07-25T02:08:45Z expires_at: 2026-07-25T02:18:45Z blocker: none

sysadmin approved these changes 2026-07-24 21:09:49 -05:00

sysadmin left a comment

Review: PR #885 — feat(webui): Workflow traffic-control view (Phase 1) (Closes #640)

Verdict: APPROVE at head 95e4aae287902e629b879a39887042a56f2f0757, base master @ ccde9e8f1183bc9a6672393a9923f304086fb874.

This is a fresh review at the current live head. The head moved twice since the B1–B6 remediation was
first examined: 1948d3dc… → dac40ab9… → 95e4aae2…. The approval at dac40ab9… (review 580) is
stale against the live head and is not being relied on; every claim below was re-established at
95e4aae2….

Delta inspected

1948d3dc… → dac40ab9… — one commit, docs(webui): update traffic state vocabulary docs and app nav for #640:

docs/webui-local-dev.md (+state vocabulary section)
webui/app.py (+1 line: /traffic added to _LEGACY_PAGES)
webui/traffic_views.py (trailing-whitespace cleanup, plus one copy correction to the "Blocked Items" description so it matches the actual routing of status:blocked to section 4)

dac40ab9… → 95e4aae2… — one commit, Merge branch 'master' into issue-640, bringing in
master through ccde9e8f1183… (#884 / #682 httpx2 TestClient migration). git diff dac40ab9 95e4aae -- webui/ docs/
is empty: this merge changed no file owned by this PR. It touches requirements.txt,
tests/webui_testclient.py, and the httpx2 migration of unrelated tests/test_webui_*.py files — all
already on master.

B1–B6 re-verified at the current head

B1 — full head_sha. webui/traffic_loader.py:165 reads signals.get("head_sha") and passes it
at line 176 with the comment "Full 40-char SHA from head.sha — never the 12-char display value."
webui/queue_loader.py:163 puts the untruncated SHA in signals, while the 12-char value at line 158
is explicitly marked display-only. PRs no longer render with a fabricated "missing head_sha pin".
B2 — reviewer-lease mapping. Candidates are built with kind="pr" (line 170), and reviewer
leases are normalized to pr_number (lines 357–372) with number = pr_number if kind == "pr"
(line 265). No cross-attribution to a linked issue.
B3 — real claim ingestion. _claim_lease_records reads (inventory or {}).get("entries")
(line 211) and documents the real build_claim_inventory contract. The dead active_claims branch
is gone.
B4 — signal-based classification and routing. WorkCandidates are constructed from signals
(labels, mergeable, head_sha) rather than from display badges.
request_changes_current_head / approval_on_current_head are deliberately left at their fail-safe
False, with an explicit comment that the queue loader never fetches review verdicts and that an
unproven approval must never route a PR to the merger. That is the correct fail-closed choice.
B5 — live-path regression tests. TestTrafficLivePath drives fetch_queue_snapshot /
fetch_lease_snapshot with no candidates= injection, and covers each earlier defect directly:
test_live_pr_uses_full_head_sha_and_is_runnable, test_live_pr_without_head_sha_is_blocked,
test_reviewer_lease_keys_by_pr_not_linked_issue, test_claim_inventory_entries_key_marks_issue_leased,
and test_active_claims_key_is_ignored (a canary that the bogus key stays unused).
B6 — state-vocabulary documentation. docs/webui-local-dev.md carries a
"Traffic-control state vocabulary (#640)" section defining all five buckets, and states that
status:blocked routes to needs_controller, matching the code.

Navigation was already wired and is confirmed present at this head: webui/nav.py contains
NavGroup("Traffic", (NavItem("/traffic", "Traffic control"), …)), and webui/app.py lists /traffic
in _LEGACY_PAGES. The earlier navigation blocker does not apply and is not restated.

Validation at the exact head

Focused on the moved delta and affected behavior; the full suite was not re-run.

pytest tests/test_webui_traffic_control.py tests/test_webui_shell.py -q at 95e4aae2… →
28 passed, 54 subtests passed, exit 0.
Traffic-control suite alone at dac40ab9… → 14 passed; shell/nav suite → 14 passed, 54 subtests.
No existing worktree was modified: validation at 95e4aae2… ran from a read-only export of that
commit. branches/issue-640 was left at dac40ab9… with git status --porcelain empty before and
after every command.

Merge-state note for the merger

Gitea currently reports mergeable: false / has_conflicts: true for this PR, and
assess_pr_sync_status routes author_conflict_remediation on that basis. That server-side flag
disagrees with the repository: git merge-tree --write-tree ccde9e8f1183… 95e4aae2879… completes
cleanly (exit 0, tree 4fc9650d889f…), and commits_behind is now 0 because the author already
merged master in. This looks like the known stale post-sync conflict cache, not a real conflict.
Merger must re-check mergeable at merge time and must not force anything — if Gitea still reports
a conflict once its cache refreshes, route the author, do not merge.

Author-safety: PR author jcwalker3 != reviewer sysadmin.

Canonical PR State

STATE:
approved

WHO_IS_NEXT:
merger

NEXT_ACTION:
Acquire the merger lease and merge PR #885 at head 95e4aae287 after merger preflight re-confirms approval_at_current_head and a live mergeable check.

NEXT_PROMPT:

Merger task: PR #885 (issue-640) / issue #640 is approved at head 95e4aae287902e629b879a39887042a56f2f0757, base master ccde9e8f1183bc9a6672393a9923f304086fb874. remote=prgs org=Scaled-Tech-Consulting repo=Gitea-Tools. Run full merger gates, pin expected_head_sha to 95e4aae287902e629b879a39887042a56f2f0757, and re-check mergeable at merge time: Gitea reported has_conflicts=true while local git merge-tree against master was clean (tree 4fc9650d889f), so confirm the live flag before acting. If a real conflict is present, route author remediation in the existing branches/ worktree and do not rebase or force-push. Do not re-review unless the head moved.

WHAT_HAPPENED:
Reviewer re-reviewed PR #885 at the current live head 95e4aae287 after the head moved twice, re-verified remediations B1-B6 against the source at that head, and ran the focused webui traffic-control and shell/nav suites (28 passed, 54 subtests).

WHY:
All six previously recorded blockers are resolved at the current head, the only change since the verified content is a clean merge of master that touches no file owned by this PR, and the focused suites pass at the exact head.

ISSUE:
#640

BASE:
master

HEAD:
issue-640

HEAD_SHA:
95e4aae287

REVIEW_STATUS:
approved

MERGE_READY:
yes - subject to the merger re-confirming the live mergeable flag noted above.

BLOCKERS:
none

SUPERSEDES:
Review 580 (approve at former head dac40ab9b3, now stale against the live head).

SUPERSEDED_BY:
none

VALIDATION:
pytest tests/test_webui_traffic_control.py tests/test_webui_shell.py at head 95e4aae287 -> 28 passed, 54 subtests passed, exit 0. Source re-verification of B1-B6 in webui/traffic_loader.py, webui/queue_loader.py, webui/nav.py, webui/app.py, docs/webui-local-dev.md, tests/test_webui_traffic_control.py. git diff dac40ab9..95e4aae -- webui/ docs/ empty. git merge-tree against master clean.

NATIVE_REVIEW_PROOF:
transport=native_mcp; entrypoint=mcp_server; pid=90914

LAST_UPDATED_BY:
sysadmin / prgs-reviewer / 2026-07-25

# Review: PR #885 — feat(webui): Workflow traffic-control view (Phase 1) (Closes #640) **Verdict: APPROVE** at head `95e4aae287902e629b879a39887042a56f2f0757`, base `master` @ `ccde9e8f1183bc9a6672393a9923f304086fb874`. This is a fresh review at the current live head. The head moved twice since the B1–B6 remediation was first examined: `1948d3dc…` → `dac40ab9…` → `95e4aae2…`. The approval at `dac40ab9…` (review `580`) is stale against the live head and is not being relied on; every claim below was re-established at `95e4aae2…`. ## Delta inspected **`1948d3dc…` → `dac40ab9…`** — one commit, `docs(webui): update traffic state vocabulary docs and app nav for #640`: - `docs/webui-local-dev.md` (+state vocabulary section) - `webui/app.py` (+1 line: `/traffic` added to `_LEGACY_PAGES`) - `webui/traffic_views.py` (trailing-whitespace cleanup, plus one copy correction to the "Blocked Items" description so it matches the actual routing of `status:blocked` to section 4) **`dac40ab9…` → `95e4aae2…`** — one commit, `Merge branch 'master' into issue-640`, bringing in master through `ccde9e8f1183…` (#884 / #682 httpx2 TestClient migration). `git diff dac40ab9 95e4aae -- webui/ docs/` is **empty**: this merge changed no file owned by this PR. It touches `requirements.txt`, `tests/webui_testclient.py`, and the httpx2 migration of unrelated `tests/test_webui_*.py` files — all already on master. ## B1–B6 re-verified at the current head 1. **B1 — full `head_sha`.** `webui/traffic_loader.py:165` reads `signals.get("head_sha")` and passes it at line 176 with the comment "Full 40-char SHA from head.sha — never the 12-char display value." `webui/queue_loader.py:163` puts the untruncated SHA in `signals`, while the 12-char value at line 158 is explicitly marked display-only. PRs no longer render with a fabricated "missing head_sha pin". 2. **B2 — reviewer-lease mapping.** Candidates are built with `kind="pr"` (line 170), and reviewer leases are normalized to `pr_number` (lines 357–372) with `number = pr_number if kind == "pr"` (line 265). No cross-attribution to a linked issue. 3. **B3 — real claim ingestion.** `_claim_lease_records` reads `(inventory or {}).get("entries")` (line 211) and documents the real `build_claim_inventory` contract. The dead `active_claims` branch is gone. 4. **B4 — signal-based classification and routing.** `WorkCandidate`s are constructed from `signals` (labels, `mergeable`, `head_sha`) rather than from display badges. `request_changes_current_head` / `approval_on_current_head` are deliberately left at their fail-safe `False`, with an explicit comment that the queue loader never fetches review verdicts and that an unproven approval must never route a PR to the merger. That is the correct fail-closed choice. 5. **B5 — live-path regression tests.** `TestTrafficLivePath` drives `fetch_queue_snapshot` / `fetch_lease_snapshot` with no `candidates=` injection, and covers each earlier defect directly: `test_live_pr_uses_full_head_sha_and_is_runnable`, `test_live_pr_without_head_sha_is_blocked`, `test_reviewer_lease_keys_by_pr_not_linked_issue`, `test_claim_inventory_entries_key_marks_issue_leased`, and `test_active_claims_key_is_ignored` (a canary that the bogus key stays unused). 6. **B6 — state-vocabulary documentation.** `docs/webui-local-dev.md` carries a "Traffic-control state vocabulary (#640)" section defining all five buckets, and states that `status:blocked` routes to `needs_controller`, matching the code. **Navigation was already wired** and is confirmed present at this head: `webui/nav.py` contains `NavGroup("Traffic", (NavItem("/traffic", "Traffic control"), …))`, and `webui/app.py` lists `/traffic` in `_LEGACY_PAGES`. The earlier navigation blocker does not apply and is not restated. ## Validation at the exact head Focused on the moved delta and affected behavior; the full suite was not re-run. - `pytest tests/test_webui_traffic_control.py tests/test_webui_shell.py -q` at `95e4aae2…` → **28 passed, 54 subtests passed**, exit 0. - Traffic-control suite alone at `dac40ab9…` → **14 passed**; shell/nav suite → **14 passed, 54 subtests**. - No existing worktree was modified: validation at `95e4aae2…` ran from a read-only export of that commit. `branches/issue-640` was left at `dac40ab9…` with `git status --porcelain` empty before and after every command. ## Merge-state note for the merger Gitea currently reports `mergeable: false` / `has_conflicts: true` for this PR, and `assess_pr_sync_status` routes `author_conflict_remediation` on that basis. That server-side flag disagrees with the repository: `git merge-tree --write-tree ccde9e8f1183… 95e4aae2879…` completes cleanly (exit 0, tree `4fc9650d889f…`), and `commits_behind` is now `0` because the author already merged master in. This looks like the known stale post-sync conflict cache, not a real conflict. **Merger must re-check `mergeable` at merge time and must not force anything** — if Gitea still reports a conflict once its cache refreshes, route the author, do not merge. Author-safety: PR author `jcwalker3` != reviewer `sysadmin`. ## Canonical PR State STATE: approved WHO_IS_NEXT: merger NEXT_ACTION: Acquire the merger lease and merge PR #885 at head 95e4aae287902e629b879a39887042a56f2f0757 after merger preflight re-confirms approval_at_current_head and a live mergeable check. NEXT_PROMPT: ```text Merger task: PR #885 (issue-640) / issue #640 is approved at head 95e4aae287902e629b879a39887042a56f2f0757, base master ccde9e8f1183bc9a6672393a9923f304086fb874. remote=prgs org=Scaled-Tech-Consulting repo=Gitea-Tools. Run full merger gates, pin expected_head_sha to 95e4aae287902e629b879a39887042a56f2f0757, and re-check mergeable at merge time: Gitea reported has_conflicts=true while local git merge-tree against master was clean (tree 4fc9650d889f), so confirm the live flag before acting. If a real conflict is present, route author remediation in the existing branches/ worktree and do not rebase or force-push. Do not re-review unless the head moved. ``` WHAT_HAPPENED: Reviewer re-reviewed PR #885 at the current live head 95e4aae287902e629b879a39887042a56f2f0757 after the head moved twice, re-verified remediations B1-B6 against the source at that head, and ran the focused webui traffic-control and shell/nav suites (28 passed, 54 subtests). WHY: All six previously recorded blockers are resolved at the current head, the only change since the verified content is a clean merge of master that touches no file owned by this PR, and the focused suites pass at the exact head. ISSUE: #640 BASE: master HEAD: issue-640 HEAD_SHA: 95e4aae287902e629b879a39887042a56f2f0757 REVIEW_STATUS: approved MERGE_READY: yes - subject to the merger re-confirming the live mergeable flag noted above. BLOCKERS: none SUPERSEDES: Review 580 (approve at former head dac40ab9b3ef13255c4fe0f6353e6d93e0d0d885, now stale against the live head). SUPERSEDED_BY: none VALIDATION: pytest tests/test_webui_traffic_control.py tests/test_webui_shell.py at head 95e4aae287902e629b879a39887042a56f2f0757 -> 28 passed, 54 subtests passed, exit 0. Source re-verification of B1-B6 in webui/traffic_loader.py, webui/queue_loader.py, webui/nav.py, webui/app.py, docs/webui-local-dev.md, tests/test_webui_traffic_control.py. git diff dac40ab9..95e4aae -- webui/ docs/ empty. git merge-tree against master clean. NATIVE_REVIEW_PROOF: transport=native_mcp; entrypoint=mcp_server; pid=90914 LAST_UPDATED_BY: sysadmin / prgs-reviewer / 2026-07-25

sysadmin commented

2026-07-24 21:10:03 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-reviewer
session_id: 90914-a466fa5849ee
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885
phase: released
candidate_head: 95e4aae287
target_branch: master
target_branch_sha: ccde9e8f11
last_activity: 2026-07-25T02:10:03Z
expires_at: 2026-07-25T02:20:03Z
blocker: manual-release

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-reviewer session_id: 90914-a466fa5849ee worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/review-pr885 phase: released candidate_head: 95e4aae287902e629b879a39887042a56f2f0757 target_branch: master target_branch_sha: ccde9e8f1183bc9a6672393a9923f304086fb874 last_activity: 2026-07-25T02:10:03Z expires_at: 2026-07-25T02:20:03Z blocker: manual-release

sysadmin commented

2026-07-24 21:10:40 -05:00

repo: Scaled-Tech-Consulting/Gitea-Tools
pr: #885
issue: #640
reviewer_identity: sysadmin
profile: prgs-merger
session_id: 53504-2faa0a10eb5e
worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640
phase: claimed
candidate_head: 95e4aae287
target_branch: master
target_branch_sha: none
last_activity: 2026-07-25T02:10:40Z
expires_at: 2026-07-25T02:20:40Z
blocker: none

repo: Scaled-Tech-Consulting/Gitea-Tools pr: #885 issue: #640 reviewer_identity: sysadmin profile: prgs-merger session_id: 53504-2faa0a10eb5e worktree: /Users/jasonwalker/Development/Gitea-Tools/branches/issue-640 phase: claimed candidate_head: 95e4aae287902e629b879a39887042a56f2f0757 target_branch: master target_branch_sha: none last_activity: 2026-07-25T02:10:40Z expires_at: 2026-07-25T02:20:40Z blocker: none

sysadmin merged commit a4c73766f4 into master

2026-07-24 21:10:52 -05:00

sysadmin commented

2026-07-24 21:10:58 -05:00

Stale #332 review-decision lock cleanup (#594)

Status: APPLIED

actor: sysadmin
profile: prgs-merger
timestamp: 2026-07-25T02:10:58.268020+00:00
last terminal: approve on PR feat(webui): Workflow traffic-control view (Phase 1) (Closes #640) (#885)
PR state: closed (merged=True)
merge_commit_sha: a4c73766f4b0cc32f7c3808688eceeb6fee74335
prior live_mutations_count: 4
prior profile_identity: prgs-reviewer

Manual deletion of session-state files is not the workflow.
This path only clears a lock when the referenced PR is merged/closed.

## Stale #332 review-decision lock cleanup (#594) Status: **APPLIED** - actor: `sysadmin` - profile: `prgs-merger` - timestamp: `2026-07-25T02:10:58.268020+00:00` - last terminal: `approve` on PR #885 - PR state: `closed` (merged=True) - merge_commit_sha: `a4c73766f4b0cc32f7c3808688eceeb6fee74335` - prior live_mutations_count: `4` - prior profile_identity: `prgs-reviewer` Manual deletion of session-state files is **not** the workflow. This path only clears a lock when the referenced PR is merged/closed.

jcwalker3 referenced this pull request

2026-07-24 21:39:26 -05:00

fix(reviewer): isolate final-decision locks by PR and exact head #896

Sign in to join this conversation.

2 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: Scaled-Tech-Consulting/Gitea-Tools#885