fix(packets): one pktEsc listener per page and #268/#260/#212 follow-ups (#282) by dborup-agent · Pull Request #313 · dborup/CoreScope

dborup-agent · 2026-10-06T09:45:12Z

Relates to #282

Follow-up nits from the reviews of #268, #260 and #212. All 8 points handled, in order. No behaviour change beyond the listed points; cmd/server stays read-only (only a new *_test.go added there).

Plan (as built)

Tests first, then the code (two commits). Every point has a test and at least one mutant; where a point is a pure comment/text fix the guard is on the source text, and the behaviour it describes is covered by existing tests (named below).

#	Point	Fix	Test	Mutant
1	`pktEsc` leaks a listener	`public/packets.js`: `_pktEsc` is a stable module-level handler, added once in `renderLeft()` (repeat add = DOM no-op) and removed in `destroy()` — same shape as #260's `_nodesPanelEsc`	`test-issue-282-pktesc-listener.js` drives init/destroy over several #/packets visits and counts document `keydown` listeners	master stacks listeners / `destroy()` leaves one → red
2	stale `#866` deep-link comment	`public/packets.js`: `#/packets/<hash>?obs=<id>`	`test-issue-282-comment-guards.js` (point 2)	re-introduce `/<obs>` form → red
3	untested `_nodesPanelEsc` guard (mutant ND)	n/a (guard already correct)	`test-issue-259-nodes-esc-listener.js`: Escape while the panel is still `.empty` must do nothing	dropping the `.empty` guard → red
4	wrong `#258` E2E header comment	`test-issue-258-column-widths-e2e.js`: describes dragging the Time handle right	`test-issue-282-comment-guards.js` (point 4)	re-introduce "Path handle … left" → red
5	wrong PR claim "analytics never create an observer"	`test-issue-258-column-widths.js` header: the trigger is the first render's usable-row count, any table (analytics included) with < 5 does create the one-shot observer	`test-issue-282-comment-guards.js` (point 5); behaviour already pinned by the "empty"/"near-empty first body" cases in `test-issue-258-column-widths.js`	drop the sparse-body observer branch → those cases red
6	untested transport-route frame through `GetChannelMessages`	n/a (read-only)	`cmd/server/channel_transport_hash_size_282_test.go`: TRANSPORT frames (v2+v3, db+store) so `substr(t.raw_hex, 1, 12)` reaches the byte-5 path length	shorten to `substr(…, 1, 11)` → red (FLOOD test stays green)
7	two offset sources in the Path Length row	`public/packets.js`: the encoded hash size is read from the same path-length byte the row prints (offset from `pkt.route_type`), not re-derived inside `senderPathHashSize(buf)` from the header byte	`test-packets.js`: a transport frame whose stored route_type and on-wire header route bits disagree — the printed byte and its `hash_size` label must describe the same byte	revert to `senderPathHashSize(buf)` → red
8	exports used only by tests	`public/channels.js`: removed `renderObservedPathHashBadge` + `OBSERVED_PATH_HASH_TOOLTIP` (no shipped reader)	`test-channels-observed-path-hash-size.js`: badge-only cases dropped; normalize/union coverage retained	— (dead-code removal)

Verification

sh test-all.sh: 224/224 files pass (incl. the two new node suites + the meta-registration test).
cd cmd/server && go build . && go vet ./... && go test ./...: green; channel-message suite also green under -race.
scripts/check-xss-sinks.sh --diff origin/master: clean.
Fork-guards unchanged: 9 in deploy.yml, 1 in release-fast-path.yml. No new map[string]interface{} outside tests. No hardcoded colours.
Browser check against a local Go server on the CI-prepared e2e-fixture.db:
- test-issue-147-packets-url-obs-e2e.js (deep link + Esc): 12/12.
- test-issue-180-packets-url-modal-e2e.js (Esc / detail sheet): 12/12.
- test-issue-259-mobile-expanded-group-e2e.js (group expansion): 10/10.
- transport-route packet detail: the Byte Breakdown renders Path Length 0x40 → hash_size=2 bytes, hash_count=0 with Transport Codes / Next Hop / Last Hop (point 7 live).

Draft until review.

…212 Tests first, red on master where there is a behaviour/offset change to pin: - #282 (1) test-issue-282-pktesc-listener.js: drives packets.js init/destroy over several #/packets visits and counts document keydown listeners. Red on master (listeners stack; destroy() leaves one behind), green once _pktEsc is a stable module-level reference removed in destroy(). - #282 (3) test-issue-259-nodes-esc-listener.js: adds the missing _nodesPanelEsc "only act while a detail panel is open" guard case (kills mutant ND). - #282 (6) cmd/server/channel_transport_hash_size_282_test.go: sends TRANSPORT frames through GetChannelMessages so the substr(t.raw_hex, 1, 12) slice is exercised to byte 5; shortening it to 11 turns this red while the FLOOD test stays green. - #282 (7) test-packets.js: buildFieldTable must read the Path Length width from the byte it prints (one offset source), not a second header-derived offset. - #282 (8) test-channels-observed-path-hash-size.js: drops the badge-only cases for the removed renderObservedPathHashBadge export. - #282 (2,4,5) test-issue-282-comment-guards.js: source-text guards for the comment/text corrections, red against the pre-#282 wording. - test-all.sh: register the two new node suites. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

…/text fixes - #282 (1) public/packets.js: _pktEsc is now a stable module-level handler, registered once in renderLeft() (a repeat add is a DOM no-op) and removed in destroy(), so #/packets no longer stacks a document keydown listener per visit. Same fix shape as nodes.js' #259 nodesPanelEsc. - #282 (7) public/packets.js: the Path Length row in the hex breakdown derives the encoded hash size from the same path-length byte it prints (offset from pkt.route_type) instead of re-deriving the offset inside senderPathHashSize() from the raw_hex header byte. One offset source; for a well-formed frame the result is unchanged, including transport-route frames. - #282 (2) public/packets.js: the Kpa-clawbot#866 deep-link comment now reads #/packets/<hash>?obs=<id>, the actual form. - #282 (8) public/channels.js: remove renderObservedPathHashBadge and OBSERVED_PATH_HASH_TOOLTIP -- the only reader was the test export; no shipped code rendered the "Observed path hash" badge. - #282 (4) test-issue-258-column-widths-e2e.js: header now describes dragging the Time handle right (what the test does), not the Path handle left. - #282 (5) test-issue-258-column-widths.js: correct the claim that analytics tables never create an observer -- any table whose first render has < 5 usable rows creates the one-shot MutationObserver. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

dborup-agent · 2026-10-06T10:00:50Z

Rapport — CS-MacBook PR#313 #282 — head `275296a`

Status: All 8 points handled in order on a draft PR; every point has a test and a mutant, all local suites are green, the browser checks pass, and the single (fork-scoped) CI job is still running — results below will be updated when it finishes.

Evidence tags: [T] test/run output, [A] assessment/inference, [K] checked in code/diff/CI.

Branch

codex/issue-282-pktesc-leak-followups from origin/master 298609e0 (clean ancestor; origin/master has since advanced to 03f9a9d6 on unrelated work — no rebase). Two commits, tests before code, author+committer dborup <kontakt@meshview.dk>. Only explicit git add; no rebase/amend/force-push. No closing keywords in title/body/commits. [K]

59fd2437 tests (red on master where a behaviour/offset changes)
275296a1 the fixes + comment/text corrections

Requirements

#	Point	Fix	Test	Mutant	Result
1	`pktEsc` listener leak	`public/packets.js`: `_pktEsc` is module-level, added once in `renderLeft()` (repeat add = DOM no-op), removed in `destroy()`	`test-issue-282-pktesc-listener.js` (7 cases): counts document `keydown` listeners over several init/destroy visits	master stacks / `destroy()` leaves one → 3 cases red [T]	✅ [T]
2	stale Kpa-clawbot#866 deep-link comment	`public/packets.js`: `#/packets/<hash>?obs=<id>`	`test-issue-282-comment-guards.js` p2	re-add `/<obs>` → red [T]	✅ [T]
3	untested `_nodesPanelEsc` guard (ND)	— (guard already correct)	`test-issue-259-nodes-esc-listener.js`: Escape while panel `.empty` does nothing	drop `.empty` guard → red (verified) [T]	✅ [T]
4	wrong #258 E2E header	`test-issue-258-column-widths-e2e.js`: Time handle dragged right	`test-issue-282-comment-guards.js` p4	re-add "Path … left" → red [T]	✅ [T]
5	wrong "analytics never create an observer"	`test-issue-258-column-widths.js` header: row-count rule, analytics not exempt	`test-issue-282-comment-guards.js` p5; behaviour already pinned by the "empty"/"near-empty first body" cases	drop sparse-body observer branch → those cases red [A]	✅ [T]
6	untested transport frame via `GetChannelMessages`	— (cmd/server read-only)	`cmd/server/channel_transport_hash_size_282_test.go`: TRANSPORT frames, v2+v3, db+store	`substr(…,1,12)`→`,1,11)` → transport test red, FLOOD test green (verified) [T]	✅ [T]
7	two offset sources in Path Length row	`public/packets.js`: width read from the same byte the row prints (offset from `pkt.route_type`), not re-derived in `senderPathHashSize(buf)`	`test-packets.js`: transport frame where stored route_type ≠ header route bits	revert to `senderPathHashSize(buf)` → red (master prints `0x80` but labels `hash_size=2`) [T]	✅ [T]
8	exports only used by tests	`public/channels.js`: removed `renderObservedPathHashBadge` + `OBSERVED_PATH_HASH_TOOLTIP` (no shipped reader; `renderSenderPathHashBadge` is the live one)	`test-channels-observed-path-hash-size.js`: badge-only cases dropped, normalize/union retained	— dead-code removal [K]	✅ [T]

Local checks [T]

sh test-all.sh: 224/224 files pass (incl. the two new node suites + the meta registration test).
cd cmd/server && go build . && go vet ./... && go test ./...: green; -race on the channel-message suite green.
scripts/check-xss-sinks.sh --diff origin/master: clean.
Fork-guards unchanged: 9 in deploy.yml, 1 in release-fast-path.yml. [K]
No new map[string]interface{} outside tests; no hardcoded colours. [K]
cmd/server production code untouched (only a new *_test.go). [K]

Browser check [T] — local Go server (my build) on CI-prepared `e2e-fixture.db`, port 13900

test-issue-147-packets-url-obs-e2e.js (deep link + Esc): 12/12.
test-issue-180-packets-url-modal-e2e.js (Esc, detail sheet): 12/12.
test-issue-259-mobile-expanded-group-e2e.js (group expansion): 10/10.
Transport-route packet detail (point 7): Byte Breakdown renders Path Length 0x40 → hash_size=2 bytes, hash_count=0 with Transport Codes / Next Hop / Last Hop. Only console noise was GET /api/nodes/<fake-pubkey> 404 from the synthetic packet's dummy key — unrelated.

CI [K]

Single fork-scoped job Go Build & Test (deploy/E2E jobs are fork-guarded off on this fork). In progress at report time; will update with the per-step outcome. Known-flaky #271 noted — if it is the only failure I will re-run the job once.

Rest / notes

Points 2/4/5 are comment/text fixes, so their new tests are source-text guards (red against the pre-Follow-ups to #268, #260 and #212: pktEsc listener leak, test gaps and comment fixes #282 wording); the behaviours they describe are covered by the existing tests named above. [A]

dborup-agent · 2026-10-06T10:36:03Z

CI update — green [T]

Run 37445133914 concluded success (head 275296a1), per job:

✅ Go Build & Test: success (23m36s) — Go server + ingestor (with -race), test-all.sh JS unit tests, Preflight XSS --diff gate, frontend eslint no-undef, CSS-var lint, proto syntax — all green.
🎭 Playwright E2E Tests: success — full E2E suite + the hard-fix: slide-over close path must not transition through focused-but-hidden state (supersedes #1172) Kpa-clawbot/CoreScope#1616 slide-over flake-gate.
🏗️ Build & Publish Docker Image: success.
📦 Release Artifacts / 🚀 Deploy Staging / 📝 Publish Badges: skipped (branch/fork-guarded, expected).

No failures; #271 did not need a re-run. PR stays a draft.

dborup-agent · 2026-10-06T10:36:21Z

Review — CS-pve-agent3 PR#313 — head `275296a`

Dom: APPROVE med nits

Independent, read-only review. Evidence tags: [T] test/run output I produced, [A] assessment/inference, [K] checked in code/diff/CI.

Trees used: git archive of the head, and of the merged tree git merge-tree --write-tree origin/master 275296a1… = 09797cdd against origin/master 03f9a9d6. The merge is clean; the only master change since the merge base 298609e0 is test-issue-180-packets-url-modal-e2e.js, which this PR does not touch. [K] git ls-remote on the head gave 275296a1 both before and after the review. [K]

Findings

#	Severity	Where	Finding	Evidence
1	Nit (DRY)	`public/packets.js` `buildFieldTable` (point 7)	The fix inlines a second copy of `senderPathHashSize()`'s width rules: TRACE gives no width, `0b11` is invalid, and `0x00` on route 2/3 is the direct marker. AGENTS.md asks for one implementation. A small shared helper in `app.js`, e.g. `pathHashSizeFromByte(pathByte, routeType, headerByte)`, could be called by both `senderPathHashSize` (with the header-derived offset) and the Path Length row (with the `pkt.route_type` offset). That would still give one offset source per caller.	[K] diff; [A]
2	Nit (test gap)	same lines	Mutant `(pkt.route_type === 2 \|\| pkt.route_type === 3)` → `(pkt.route_type === 2)` survives `test-packets.js` and `test-frontend-helpers.js`, so no test covers the TRANSPORT_DIRECT `0x00` marker in the hex breakdown. The same mutant in `app.js` `senderPathHashSize` also survives on master, so the gap is pre-existing, not new. But the PR now has two untested copies of the rule. One `buildFieldTable` case with `raw_hex: '17aabbccdd00…', route_type: 3` would kill both mutants if the helper in finding 1 is shared.	[T]
3	Nit (comment/test)	`test-issue-282-pktesc-listener.js` header and case "a second render inside the same visit…"	The header says the old closure stacked "on every filter/region change within a visit". It did not. `renderLeft()` returns at the `filtersBuilt` guard before reaching `addEventListener`, so the leak was once per visit. As a result, this case is also green on master; it documents behaviour but is not a regression guard. The other three leak cases are red on master and carry the fix. Suggest correcting the header or relabelling the case.	[T] red run; [K] `renderLeft()` `filtersBuilt`
4	Nit (comment)	`public/channels.js` #282 (8) comment	The comment says "`normalizeObservedPathHashSizes()` stays: the union/merge path and the sender badge still use it". `renderSenderPathHashBadge()` does not call it; it reads `message.senderPathHashSize`. The real callers are union, the cached-message merge, the packet → message mapping and the dedup key.	[K]
5	Info	author report	The report says deploy/E2E jobs are fork-guarded off. On this run, `🎭 Playwright E2E Tests` does run (see CI below).	[K]

None of these blocks the PR. 1+2 are best fixed together; 3+4 are one-line comment edits.

Answers to the review points

pktEsc leak — fixed. _pktEsc is now a module-level function. It is added in renderLeft() and removed in destroy(), the same shape as _nodesEsc/_nodesPanelEsc in nodes.js. [K]
- Unit test. test-issue-282-pktesc-listener.js against master's packets.js gives 3/7 red: 4 listeners after 4 visits, destroy() leaves 1, and the reopened page has 2. On the PR it is 7/7 green. [T]
- Real browser (Chromium, CDP getEventListeners(document), local server). On the PR, document keydown listeners stay at 5 on the first visit to #/packets, after 5 more #/nodes ↔ #/packets round trips, and on #/nodes. On master's frontend they grow 8 → 13 → 14. [T]
- Esc still works. Esc closes an open detail panel (#pktRight goes .empty). There were no page errors. [T]
- Order (fix(packets): keep ?obs=/?viewPath= on packet detail URLs (#147) #167). The handler body is unchanged. It is a bubble-phase document listener, as before. Modal Esc handlers (packet-path-map.js, channel-color-picker.js) listen in the capture phase. A stable reference only removes the duplicates; it does not move the first one. [K]/[A]
- E2E. test-issue-180-packets-url-modal-e2e.js (12/12) and test-issue-147-packets-url-obs-e2e.js (12/12) are green. [T]
Points 3 and 6 — the new tests.
- Point 3. The new _nodesPanelEsc .empty guard case passes. Mutant ND (if (!panel) return;) turns it red: replaceState ["#/nodes"]. [T]
- Point 6. TestChannelMessagesSenderPathHashSize_TransportRoute_282 (v2+v3, db+store) passes. With mutant substr(t.raw_hex, 1, 11) at both SQL sites, the transport test fails (v2 and v3) and the FLOOD test stays green, exactly as the PR claims. With substr(…, 1, 2), both fail. [T]
Point 7 — offsets are consistent. The printed byte and the hash_size label now come from the same off. [K]
- Unit test. The new test-packets.js case is red on master (hash_size=2 printed for 0x80) and green on the PR. [T]
- Mutants. "revert to senderPathHashSize(buf)" and "read path byte at the header offset" are both killed. "drop the 0b11 guard" is killed by an existing case. [T]
- Browser, PR vs master frontend. Same local server and DB, with synthetic rows added to a scratch copy of the fixture. Both frontends render identical rows:
  - TRANSPORT_FLOOD 0x14…41…: 5 Path Length 0x41 hash_size=2 bytes, hash_count=1, with Next Hop/Last Hop at 1/3.
  - TRANSPORT_DIRECT 0x17…00: 5 Path Length 0x00 hash_count=0 (no encoded hash size).
  - FLOOD 0x15 81…: 1 Path Length 0x81 hash_size=3 bytes, hash_count=1.
  So there is no visible change for well-formed frames, as intended. [T]
- E2E. test-packet-detail-sender-hash-size-obs-e2e.js passes 3/3. [T]
Point 8 — removed exports. grep over the merged tree for renderObservedPathHashBadge, OBSERVED_PATH_HASH_TOOLTIP and _channelsRenderObservedPathHashBadgeForTest (*.js, *.html, *.go, *.md) finds only the two new explanatory comments; there are no code readers in public/. The .ch-path-hash-badge class is still used by the sender badge. test-channels-observed-path-hash-size-e2e.js passes 8/8. The injection-shaped rejection is still covered by the normalize case "rejects unknown, empty and injection-shaped values". [K]/[T]
Points 2, 4 and 5 — comments and text.
- Point 2. #/packets/<hash>?obs=<id> matches the ?obs= handling in packets.js. [K]
- Point 4. The test drags th.col-time .col-resize-handle to the right (+60 px) and asserts that Time widens, which matches the corrected header. [K]
- Point 5. In app.js, the observer is created when rows.length < COL_MEASURE_MIN_ROWS (5), with no page check, and analytics tables call makeColumnsResizable (analytics.js:521). The corrected header is right. [K]
- All three guards in test-issue-282-comment-guards.js are red on the old wording and green on the new. [T]
Scope and rules.
- Behaviour. There is no behaviour change beyond points 1 and 7. Point 7 only differs when pkt.route_type and the header's route bits disagree, which is the point of the fix. [K]/[A]
- XSS gate. scripts/check-xss-sinks.sh --diff origin/master passes (exit 0). [T]
- Colours. No hardcoded colours were added; the only #… matches in added lines are #282 in comments. [K]
- cmd/server and map[string]interface{}. cmd/server has no non-test changes, so it stays read-only, and the diff adds 0 new map[string]interface{} outside _test.go. [K]
- Fork guards and keywords. Fork guards are unchanged: 9 in deploy.yml and 1 in release-fast-path.yml. There are no closing keywords in the title, body or commits (closingIssuesReferences = 0). [K]
- Commits. Both commits have author and committer dborup <kontakt@meshview.dk>. [K]

Tests and mutants (my runs, merged tree `09797cdd`)

sh test-all.sh: 224 passed, 0 failed (224 files). [T]
node test-frontend-helpers.js: 709 passed, 0 failed. [T]
cd cmd/server && go build . && go vet ./...: clean. go test ./... with Go's default 10 m timeout panicked in Test1690_BackgroundLoadHonesty because of the timeout, while E2E runs were loading the machine. Re-run with -count=1 -timeout 25m, as CI uses -timeout 20m: ok github.com/corescope/server 743.7s. [T]
Red-before. I ran the new tests against master's packets.js, channels.js and the two makeColumnsResizable measures once and credits colspan rows to column 0 (packets Details column too narrow) #258 test files:
- test-issue-282-pktesc-listener.js: 3 red.
- test-issue-282-comment-guards.js: 3 red.
- test-packets.js: 1 red (the Follow-ups to #268, #260 and #212: pktEsc listener leak, test gaps and comment fixes #282 case).
- The ND case in test-issue-259-nodes-esc-listener.js is red only under the mutant, because the guard already exists. [T]

Mutants:

Mutant	Result
M1 (re-register per init)	killed
M2 (drop the `destroy()` removal)	killed
M3 (handler no-op)	killed
ND	killed
P7 revert to `senderPathHashSize(buf)`	killed
P7 header offset	killed
P7 drop `0b11` guard	killed
P7 drop the whole direct-marker clause	killed
Go `substr 1,11`	killed (transport test only)
Go `substr 1,2`	killed
P7 drop `route_type === 3`	survives (finding 2)

[T]

E2E against a local Go server (merged build) on e2e-fixture.db, prepared as in CI: freshen, then the Packets page collapse button in the left column of the table opens the dialog. Kpa-clawbot/CoreScope#1486/"Group Data" message type missing from packet view window's "message type" filter Kpa-clawbot/CoreScope#1791 seed SQL, then corescope-migrate, then seeds 2073, 199 and 245. [T]
- test-issue-147-packets-url-obs-e2e.js 12/12
- test-issue-180-packets-url-modal-e2e.js 12/12
- test-issue-258-column-widths-e2e.js 22/22
- test-issue-259-mobile-expanded-group-e2e.js 10/10
- test-packet-detail-sender-hash-size-obs-e2e.js 3/3
- test-channels-observed-path-hash-size-e2e.js 8/8
- test-issue-1486-collapse-reopens-detail-e2e.js 4/4
- test-issue-1692-packets-init-parallel-e2e.js 1/1
- test-slideover-1056-e2e.js 27/27
- test-e2e-playwright.js stops fail-fast at test 7 ("Version info lives on Perf dashboard", #navStats wait timeout). It fails identically against a pure origin/master tree in my environment, so it is local and not caused by this PR. I rely on CI's Playwright job for that suite.
CI (run 37445133914, attempt 1, no re-runs; per job):
- ✅ Go Build & Test: success.
- 🎭 Playwright E2E Tests: success. This includes the full test-e2e-playwright.js; "Version info lives on Perf dashboard" passes there. Only "Collect frontend coverage" was skipped.
- 🏗️ Build & Publish Docker Image: success.
- Release Artifacts, Deploy Staging and Publish Badges: skipped (fork-scoped).
Neither known flake, flaky: #226 Hash Stats adopters sort E2E — server sorts multiByteNodes on packets only, ties come out in random order #256 nor flaky: TestResolvedPathBackfill_WriteHoldUnderBudget_188 — wall-clock write-hold budget fails under CI load #267, showed up. [K]

Not verified

The full test-e2e-playwright.js locally, past test 7 (see above); covered by CI's E2E job instead.
cmd/ingestor suites: the PR does not touch the ingestor, so I did not run them.
The real #167 stack (a path-map modal open over the detail pane, then Esc twice) by hand. I relied on the handler body being unchanged and on the green fix(packets): updatePacketsUrl drops ?obs= and ?viewPath= on load and on filter changes #147/fix(packets): leaving with the SlideOver open rewrites the new page's URL; Clear Filters on a detail URL re-filters on reload #180 E2Es.
The ≤640 px mobile detail sheet closing on Esc in a browser; only closeDetailPanel(), which is unchanged, handles it.
No staging or prod access; everything ran on a local server with the fixture DB.

dborup and others added 2 commits October 6, 2026 11:37

dborup marked this pull request as ready for review October 6, 2026 10:53

dborup merged commit 30c7de4 into master Oct 6, 2026
6 checks passed

dborup mentioned this pull request Oct 6, 2026

Follow-ups to #268, #260 and #212: pktEsc listener leak, test gaps and comment fixes #282

Closed

dborup-agent mentioned this pull request Oct 6, 2026

test(server): lock log capture in the remaining cmd/server tests (#310) #311

Merged

dborup mentioned this pull request Oct 6, 2026

Follow-ups to #313: one Path Length width helper, TRANSPORT_DIRECT test, two comment fixes #322

Closed

This was referenced Oct 6, 2026

fix(packets): one path-hash width helper for the hex breakdown (#322) #327

Merged

My Mesh: accurate 24h node activity and compact repeater observer cards #351

Merged

Navigation Menu

Repository navigation

fix(packets): one pktEsc listener per page and #268/#260/#212 follow-ups (#282) - #313

fix(packets): one pktEsc listener per page and #268/#260/#212 follow-ups (#282)#313
dborup merged 2 commits into
masterfrom
codex/issue-282-pktesc-leak-followups

dborup-agent commented Oct 6, 2026

Uh oh!

dborup-agent commented Oct 6, 2026

Uh oh!

dborup-agent commented Oct 6, 2026

Uh oh!

dborup-agent commented Oct 6, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

dborup-agent commented Oct 6, 2026

Plan (as built)

Verification

Uh oh!

dborup-agent commented Oct 6, 2026

Rapport — CS-MacBook PR#313 #282 — head 275296a

Branch

Requirements

Local checks [T]

Browser check [T] — local Go server (my build) on CI-prepared e2e-fixture.db, port 13900

CI [K]

Rest / notes

Uh oh!

dborup-agent commented Oct 6, 2026

CI update — green [T]

Uh oh!

dborup-agent commented Oct 6, 2026

Review — CS-pve-agent3 PR#313 — head 275296a

Findings

Answers to the review points

Tests and mutants (my runs, merged tree 09797cdd)

Not verified

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Rapport — CS-MacBook PR#313 #282 — head `275296a`

Browser check [T] — local Go server (my build) on CI-prepared `e2e-fixture.db`, port 13900

Review — CS-pve-agent3 PR#313 — head `275296a`

Tests and mutants (my runs, merged tree `09797cdd`)