kayg/infraternal
Watch
1
0
Fork
You've already forked infraternal
 0
Code Issues 356 Pull requests Projects Releases Packages Wiki Activity Actions

Terminal: recognise links and make them clickable (incl. wrapped URLs and OSC 8) #214

New issue
Closed
opened 2026-10-14 11:14:57 +00:00 by kayg · 114 comments
kayg commented 2026-10-14 11:14:57 +00:00
Owner
Copy link

Owner, 2026-10-14 (screenshot: a long OAuth URL wrapped across 8 lines, not clickable): "we need to recognise and make links clickable in the terminal."

Owner, 2026-10-14 (screenshot: a long OAuth URL wrapped across 8 lines, not clickable): "we need to recognise and make links clickable in the terminal."
key-after-2.log
10 KiB
cat-before-11.log
1.8 KiB
key-before-1-bun-no-tests.log
455 B
key-after-3.log
9.9 KiB
terminal-wrapped-link-hover.png
27 KiB
terminal-osc8-target-hover.png
27 KiB
terminal-wrapped-link-hover.png
28 KiB
terminal-osc8-target-hover.png
27 KiB
terminal-wrapped-link-hover.png
28 KiB
terminal-osc8-target-hover.png
27 KiB
key-before-1-bun-path-failure.log
324 B
terminal-wrapped-link-hover.png
28 KiB
terminal-osc8-target-hover.png
27 KiB
transfer-budget-final.log
6.5 KiB
baseline-static-build.log
9.5 KiB
link-unit-fuzz.log
1.1 KiB
baseline-tracer-build.log
1.2 KiB
cat-before-2.log
680 B
cat-before-4.log
680 B
key-before-1-default-300ms-env.log
10 KiB
terminal-wrapped-link-hover.png
28 KiB
key-before-1.log
1.5 KiB
terminal-osc8-target-hover.png
27 KiB
key-before-3.log
9.9 KiB
cat-before-3.log
1 KiB
cat-after-1.log
679 B
iterate.log
294 KiB
cat-before-2.log
1.6 KiB
cat-before-4.log
954 B
key-before-2.log
10 KiB
cat-before-6.log
1.5 KiB
cat-after-3.log
679 B
cat-after-4.log
679 B
iterate-after-dev.log
293 KiB
fast.log
303 KiB
cat-before-1.log
957 B
cat-before-13.log
1.9 KiB
baseline-transfer-budget.log
6.5 KiB
binary-size-finding.txt
408 B
cat-before-8.log
1.6 KiB
cat-after-2.log
2.6 KiB
key-after-1.log
9.9 KiB
cat-before-10.log
1.8 KiB
bun-check.log
36 B
terminal-osc8-target-hover.png
27 KiB
throughput-phase.log
4.8 KiB
browser-setup-failure.log
819 B
browser-setup-failure-2.log
383 B
browser-phase.log
47 KiB
throughput-phase.log
4.8 KiB
browser-phase.log
47 KiB
ab-results.json
20 KiB
verify-fast.log
291 KiB
rerun-review-input-credit.log
2.1 KiB
rerun-review-liveness.log
526 B
rerun-control-queue-remote.log
1.6 KiB
mailto-before.log
1.9 KiB
mailto-after.log
1 KiB
mailto-check.log
36 B
scrollback-before.log
1.7 KiB
scrollback-after.log
1.2 KiB
scrollback-check.log
36 B
demo-typecheck.log
36 B
terminal-links-playwright.log
5.1 KiB
error-context.md
8.7 KiB
error-context.md
8.7 KiB
real-demo-build.log
12 KiB
real-demo-rerun.log
2.1 KiB
real-demo-hostmap.log
719 B
verify-iterate.log
163 KiB
verify-fast.log
279 KiB
baseline-size-build.log
8.3 KiB
static-size-comparison.txt
603 B
demo.webm
1.4 MiB
mock-browser.log
2 KiB
terminal-links-playwright-rerun.log
2 KiB
terminal-wrapped-link-hover.png
28 KiB
terminal-wrapped-link-hover.png terminal-osc8-target-hover.png terminal-wrapped-link-hover.png terminal-osc8-target-hover.png terminal-wrapped-link-hover.png terminal-osc8-target-hover.png terminal-wrapped-link-hover.png terminal-osc8-target-hover.png terminal-wrapped-link-hover.png terminal-osc8-target-hover.png terminal-osc8-target-hover.png terminal-wrapped-link-hover.png
kayg commented 2026-10-14 11:16:42 +00:00
Author
Owner
Copy link

Starting terminal link recognition on branch terminal-links, based on origin/dev at b0a3799cf459425c3402e745f329f4ebc441c144.

Starting terminal link recognition on branch `terminal-links`, based on `origin/dev` at `b0a3799cf459425c3402e745f329f4ebc441c144`.
kayg commented 2026-10-14 11:22:45 +00:00
Author
Owner
Copy link

Finding: the pinned renderer already parses OSC 8 and exposes cell hyperlink URIs, but its built-in plain URL provider scans one row at a time and only detects HTTP(S), mailto and several other schemes; it does not join soft-wrapped text or recognize file URLs. Both built-in providers require Ctrl/Command-click, and OSC 8 activation passes its URI directly to window.open. Evidence: apps/tracer/vendor/ghostty-web/dist/ghostty-web.js:1445-1478 and :1575-1597. Terminal sessions are loaded dynamically from apps/tracer/src/canvas/sessions/live.ts:202-205. I’ll keep the new behavior in the terminal’s on-demand path and reuse the engine’s OSC 8 parser and renderer underline range.

Finding: the pinned renderer already parses OSC 8 and exposes cell hyperlink URIs, but its built-in plain URL provider scans one row at a time and only detects HTTP(S), mailto and several other schemes; it does not join soft-wrapped text or recognize file URLs. Both built-in providers require Ctrl/Command-click, and OSC 8 activation passes its URI directly to window.open. Evidence: `apps/tracer/vendor/ghostty-web/dist/ghostty-web.js:1445-1478` and `:1575-1597`. Terminal sessions are loaded dynamically from `apps/tracer/src/canvas/sessions/live.ts:202-205`. I’ll keep the new behavior in the terminal’s on-demand path and reuse the engine’s OSC 8 parser and renderer underline range.
kayg commented 2026-10-14 12:39:10 +00:00
Author
Owner
Copy link

Finding: after a second terminal write, the live Ghostty cells held OSC 8 URI https://links.example.test/actual-destination across columns 0–6, but the link cache stayed stale. The client queue held visible rows 0–37 because queueLinkRows retried controller loading after the controller already existed. apps/tracer/src/terminal.ts now flushes queued rows to the existing controller; parsing remains idle-scheduled. The production browser run passed wrapped URL click, palette activation, OSC 8 target tooltip/click; unit tests passed 7/7 and bun run check passed. The latest expanded run then timed out before either browser case connected to the test double; I’m investigating that setup failure.

wrapped link hover
OSC 8 target tooltip

Finding: after a second terminal write, the live Ghostty cells held OSC 8 URI `https://links.example.test/actual-destination` across columns 0–6, but the link cache stayed stale. The client queue held visible rows 0–37 because `queueLinkRows` retried controller loading after the controller already existed. `apps/tracer/src/terminal.ts` now flushes queued rows to the existing controller; parsing remains idle-scheduled. The production browser run passed wrapped URL click, palette activation, OSC 8 target tooltip/click; unit tests passed 7/7 and `bun run check` passed. The latest expanded run then timed out before either browser case connected to the test double; I’m investigating that setup failure. ![wrapped link hover](https://git.kayg.org/attachments/9779d240-a511-4279-95f2-2c635bebfac1) ![OSC 8 target tooltip](https://git.kayg.org/attachments/80c38999-751b-4c18-910b-5302b3e07879)
kayg commented 2026-10-14 12:52:13 +00:00
Author
Owner
Copy link

Finding: Chromium mobile emulation delivered pointerdown and pointerup with pointerType=touch but suppressed the synthesized click on the terminal surface. The controller now opens on an unmoved touch pointerup, retains the six-pixel drag guard, and consumes a later duplicate click. Production browser test passed: touch tap opens a recognized terminal link (1/1; 33.0 s test duration), with window.open(target, "_blank", "noopener,noreferrer") observed.

Finding: Chromium mobile emulation delivered `pointerdown` and `pointerup` with `pointerType=touch` but suppressed the synthesized `click` on the terminal surface. The controller now opens on an unmoved touch `pointerup`, retains the six-pixel drag guard, and consumes a later duplicate click. Production browser test passed: `touch tap opens a recognized terminal link` (1/1; 33.0 s test duration), with `window.open(target, "_blank", "noopener,noreferrer")` observed.
kayg commented 2026-10-14 13:05:47 +00:00
Author
Owner
Copy link

Finding: bun run check caught the scrollback unit fixture omitting optional wrapKnown (src/terminal-links.on-demand.test.ts:130-131). Added the optional field to the fixture type. Re-run: bun run check passed; link unit tests 8/8; production browser tests 2/2 on working tree based at b0a3799cf459425c3402e745f329f4ebc441c144.

Finding: `bun run check` caught the scrollback unit fixture omitting optional `wrapKnown` (`src/terminal-links.on-demand.test.ts:130-131`). Added the optional field to the fixture type. Re-run: `bun run check` passed; link unit tests 8/8; production browser tests 2/2 on working tree based at `b0a3799cf459425c3402e745f329f4ebc441c144`.
kayg commented 2026-10-14 13:22:55 +00:00
Author
Owner
Copy link

Finding: a URL cut by the bottom of the visible viewport was offered as a shorter, valid URL (https://scroll.example.test/a-). The new regression in apps/tracer/src/terminal-links.on-demand.test.ts failed before the fix. The controller now checks the next row's continuation flag and waits for the complete URL; the test passes with the full target after the last row becomes visible. Parser suite: 9/9.

Finding: a URL cut by the bottom of the visible viewport was offered as a shorter, valid URL (`https://scroll.example.test/a-`). The new regression in `apps/tracer/src/terminal-links.on-demand.test.ts` failed before the fix. The controller now checks the next row's continuation flag and waits for the complete URL; the test passes with the full target after the last row becomes visible. Parser suite: 9/9.
kayg commented 2026-10-14 13:25:16 +00:00
Author
Owner
Copy link

Correction to the preceding finding: the final implementation does not read the next row. To keep scanning limited to visible rows, it checks the active buffer length and defers a plain URL ending at the viewport's last cell until the continuation boundary is visible. The 9-test parser suite passes.

Correction to the preceding finding: the final implementation does not read the next row. To keep scanning limited to visible rows, it checks the active buffer length and defers a plain URL ending at the viewport's last cell until the continuation boundary is visible. The 9-test parser suite passes.
kayg commented 2026-10-14 13:34:01 +00:00
Author
Owner
Copy link

Finding: while replacement parsing was deferred, a changed row could still expose its prior cached URL to hover or the keyboard palette. apps/tracer/src/terminal-links.on-demand.test.ts reproduced the stale target before the fix. updateRows now removes links overlapping dirty rows synchronously, then parses replacements during idle; parser suite: 10/10.

Finding: while replacement parsing was deferred, a changed row could still expose its prior cached URL to hover or the keyboard palette. `apps/tracer/src/terminal-links.on-demand.test.ts` reproduced the stale target before the fix. `updateRows` now removes links overlapping dirty rows synchronously, then parses replacements during idle; parser suite: 10/10.
kayg commented 2026-10-14 13:47:28 +00:00
Author
Owner
Copy link

Production browser evidence on terminal-links: wrapped OAuth URL opens via click and palette; drag still selects; OSC 8 shows its destination and opens; touch tap opens. 2/2 Playwright tests passed. Parser suite: 11/11 passed, including 750 deterministic fuzz cases (54,601 assertions). bun run check passed. Screenshots: wrapped link OSC 8 destination

Production browser evidence on terminal-links: wrapped OAuth URL opens via click and palette; drag still selects; OSC 8 shows its destination and opens; touch tap opens. 2/2 Playwright tests passed. Parser suite: 11/11 passed, including 750 deterministic fuzz cases (54,601 assertions). `bun run check` passed. Screenshots: ![wrapped link](https://git.kayg.org/attachments/222c1ea5-d752-4e13-9587-866730b07db1) ![OSC 8 destination](https://git.kayg.org/attachments/de823174-82eb-4865-b870-b675cb62d88b)
kayg commented 2026-10-14 13:53:17 +00:00
Author
Owner
Copy link

Commit f459b8c8 adds the on-demand link parser/controller, shared point-anchored tooltip API, parser unit tests, and a deterministic fuzz target. bun run check passed; parser suite passed 11/11 with 54,601 assertions. Link parsing and cell reads run in idle scans (D188, D160, D259).

Commit f459b8c8 adds the on-demand link parser/controller, shared point-anchored tooltip API, parser unit tests, and a deterministic fuzz target. `bun run check` passed; parser suite passed 11/11 with 54,601 assertions. Link parsing and cell reads run in idle scans (D188, D160, D259).
kayg commented 2026-10-14 13:53:44 +00:00
Author
Owner
Copy link

Commit 67b98c06 wires the on-demand controller to Ghostty's visible rows, wrap flags, and OSC 8 destinations. The render path now queues row indices and invalidates stale targets; text extraction and parsing run in idle callbacks. bun run check, parser/fuzz tests (11/11), and production browser tests (2/2) pass.

Commit 67b98c06 wires the on-demand controller to Ghostty's visible rows, wrap flags, and OSC 8 destinations. The render path now queues row indices and invalidates stale targets; text extraction and parsing run in idle callbacks. `bun run check`, parser/fuzz tests (11/11), and production browser tests (2/2) pass.
kayg commented 2026-10-14 13:54:04 +00:00
Author
Owner
Copy link

Commit 1aa09e75 exposes safe visible links through the existing palette and adds production browser coverage for wrapped click, drag selection, OSC 8 destination confirmation, palette open, and touch tap. bun run check passed; Playwright passed 2/2. Final screenshots: wrapped link OSC 8 target

Commit 1aa09e75 exposes safe visible links through the existing palette and adds production browser coverage for wrapped click, drag selection, OSC 8 destination confirmation, palette open, and touch tap. `bun run check` passed; Playwright passed 2/2. Final screenshots: ![wrapped link](https://git.kayg.org/attachments/6ba536e8-d275-4f9b-ac9f-b0dd07d876c2) ![OSC 8 target](https://git.kayg.org/attachments/150de784-47bb-48ba-aa2f-ebe30c22ca4a)
kayg commented 2026-10-14 13:57:15 +00:00
Author
Owner
Copy link

Finding during review: the stale-link guard briefly fingerprinted terminal cell text synchronously on each rendered output row. Commit 67b98c06 moves all cell reads and fingerprints into the idle scan; the render path now only queues row indices and invalidates nearby cached targets. bun run check passed, parser/fuzz 11/11 (54,601 assertions), browser 2/2. Final perf repeat is running on the perf VM.

Finding during review: the stale-link guard briefly fingerprinted terminal cell text synchronously on each rendered output row. Commit 67b98c06 moves all cell reads and fingerprints into the idle scan; the render path now only queues row indices and invalidates nearby cached targets. `bun run check` passed, parser/fuzz 11/11 (54,601 assertions), browser 2/2. Final perf repeat is running on the perf VM.
kayg commented 2026-10-14 14:01:43 +00:00
Author
Owner
Copy link

D188 production bundle budget passed. Merge-base → branch: critical JS gzip 295,952 → 297,248 B, Brotli 246,125 → 247,092 B; deferred gzip/Brotli 47,859/42,454 → 48,095/42,628 B; on-demand class 78,237/69,684 → 81,183/72,398 B. The parser is in on-demand chunk chunk-gpq4hm3e.js (7,221 B identity, 2,985 gzip, 2,722 Brotli); all budgets pass and no crate or feature was added. Transfer report: transfer-budget-final.log

D188 production bundle budget passed. Merge-base → branch: critical JS gzip 295,952 → 297,248 B, Brotli 246,125 → 247,092 B; deferred gzip/Brotli 47,859/42,454 → 48,095/42,628 B; on-demand class 78,237/69,684 → 81,183/72,398 B. The parser is in on-demand chunk `chunk-gpq4hm3e.js` (7,221 B identity, 2,985 gzip, 2,722 Brotli); all budgets pass and no crate or feature was added. Transfer report: [transfer-budget-final.log](https://git.kayg.org/attachments/9df2bbca-36e8-4658-8180-f56b94b8cd87)
kayg commented 2026-10-14 14:15:58 +00:00
Author
Owner
Copy link

Finding: the 2026-10-14 15:xx production browser rerun passed touch (1/2) but the wrapped-link case hit its 90 s test timeout while waiting for the synthetic OSC 8 label. Playwright's page snapshot still showed the connected terminal and a 326 ms connection badge; the Rust iterate suite was running concurrently on the shared build host. Error context is outside the checkout at $HOME/.cache/codex-scratch/terminal-links/evidence/final-playwright/playwright/terminal-links-wrapped-and-0668d-om-the-terminal-and-palette/error-context.md. I will rerun the functional browser check after the host gate clears with zero simulated RTT.

Finding: the 2026-10-14 15:xx production browser rerun passed touch (1/2) but the wrapped-link case hit its 90 s test timeout while waiting for the synthetic OSC 8 label. Playwright's page snapshot still showed the connected terminal and a 326 ms connection badge; the Rust iterate suite was running concurrently on the shared build host. Error context is outside the checkout at `$HOME/.cache/codex-scratch/terminal-links/evidence/final-playwright/playwright/terminal-links-wrapped-and-0668d-om-the-terminal-and-palette/error-context.md`. I will rerun the functional browser check after the host gate clears with zero simulated RTT.
kayg commented 2026-10-14 14:16:07 +00:00
Author
Owner
Copy link

Correction to the previous finding: 15:xx was a placeholder, not a measured timestamp. The timeout occurred during this same-day run after commit 1aa09e75.

Correction to the previous finding: `15:xx` was a placeholder, not a measured timestamp. The timeout occurred during this same-day run after commit 1aa09e75.
kayg commented 2026-10-14 14:17:17 +00:00
Author
Owner
Copy link

Commit 1bd56ea9 adds idle fingerprint filtering for scrolled visible rows. Unchanged rows keep their cached links without reparsing; candidate rows are non-activatable until checked, and changed targets are replaced in the idle scan. bun run check passed; parser/fuzz tests passed 12/12 with 54,608 assertions. Browser rerun after the host gate remains pending.

Commit 1bd56ea9 adds idle fingerprint filtering for scrolled visible rows. Unchanged rows keep their cached links without reparsing; candidate rows are non-activatable until checked, and changed targets are replaced in the idle scan. `bun run check` passed; parser/fuzz tests passed 12/12 with 54,608 assertions. Browser rerun after the host gate remains pending.
kayg commented 2026-10-14 14:19:18 +00:00
Author
Owner
Copy link

Finding in scripts/verify iterate: auth::tests::server_setup_terminal_echo_and_resume_preserve_offsets_once reported FAILED during the shared-host Rust suite. The test covers the authenticated WebSocket terminal input/echo/resume path in crates/infraternal-server/src/auth/tests.rs:1254. I’m running this exact test on the detached dev baseline in the same host-load window per #118, then I’ll record both outputs.

Finding in `scripts/verify iterate`: `auth::tests::server_setup_terminal_echo_and_resume_preserve_offsets_once` reported FAILED during the shared-host Rust suite. The test covers the authenticated WebSocket terminal input/echo/resume path in `crates/infraternal-server/src/auth/tests.rs:1254`. I’m running this exact test on the detached `dev` baseline in the same host-load window per #118, then I’ll record both outputs.
kayg commented 2026-10-14 14:21:43 +00:00
Author
Owner
Copy link

Additional iterate finding: auth::tests::socket_tests::review_liveness::review_mesh_watch_reports_liveness_expiry_without_another_edit also reported FAILED in the same workspace test run. This is the auth socket timeout/liveness class; I’ll run it on the detached dev baseline after the first baseline comparison completes.

Additional iterate finding: `auth::tests::socket_tests::review_liveness::review_mesh_watch_reports_liveness_expiry_without_another_edit` also reported FAILED in the same workspace test run. This is the auth socket timeout/liveness class; I’ll run it on the detached `dev` baseline after the first baseline comparison completes.
kayg commented 2026-10-15 06:06:38 +00:00
Author
Owner
Copy link

Continuing after the disk-full stop. Branch terminal-links is at 1bd56ea9a2596e7d7def45282e3a06eff07cf8036, based on origin/dev at b0a3799cf459425c3402e745f329f4ebc441c144. The worktree has no tracked changes. I read all 19 comments on #214; the latest pending work is the production browser rerun and the two reported iterate failures’ #118 baseline comparisons. The parent issue comment endpoint returned InvalidContentType; I will retry it.

Continuing after the disk-full stop. Branch `terminal-links` is at `1bd56ea9a2596e7d7def45282e3a06eff07cf8036`, based on `origin/dev` at `b0a3799cf459425c3402e745f329f4ebc441c144`. The worktree has no tracked changes. I read all 19 comments on #214; the latest pending work is the production browser rerun and the two reported iterate failures’ #118 baseline comparisons. The parent issue comment endpoint returned `InvalidContentType`; I will retry it.
kayg commented 2026-10-15 06:07:45 +00:00
Author
Owner
Copy link

Production browser rerun on 1bd56ea9a2596e7d7def45282e3a06eff07cf8036: TRACER_TEST_PORT=4214, INFRATERNAL_TEST_RTT=0; production build completed; wrapped URL + OSC 8 target/tooltip + palette + drag selection and touch tap passed, 2/2 tests in 14.6 s. Screenshots: wrapped link OSC 8 destination

Production browser rerun on `1bd56ea9a2596e7d7def45282e3a06eff07cf8036`: `TRACER_TEST_PORT=4214`, `INFRATERNAL_TEST_RTT=0`; production build completed; wrapped URL + OSC 8 target/tooltip + palette + drag selection and touch tap passed, 2/2 tests in 14.6 s. Screenshots: ![wrapped link](https://git.kayg.org/attachments/763c6f3d-f482-417d-87fe-f7f67d5b81a2) ![OSC 8 destination](https://git.kayg.org/attachments/c46c2b13-b754-4f82-ba65-3ff11c6493a6)
kayg commented 2026-10-15 06:45:17 +00:00
Author
Owner
Copy link

Finding from scripts/verify iterate on 1bd56ea9: fmt, VT native fmt, CLI plugin compile-out, clippy, and workspace tests passed. cargo-deny failed because this branch locks rustls 0.23.45 (Cargo.lock:2514), newly flagged by RUSTSEC-2026-0346 through 0349; the current origin/dev has the owner-approved 0.23.46 lock update (1b23ef311). Both previously reported auth tests passed in this run, so no #118 baseline comparison is needed. The gate stopped before TypeScript checks. Raw output: iterate.log. I will sync with current origin/dev, then rerun the gate. The #5 comments endpoint still returns InvalidContentType.

Finding from `scripts/verify iterate` on `1bd56ea9`: fmt, VT native fmt, CLI plugin compile-out, clippy, and workspace tests passed. `cargo-deny` failed because this branch locks `rustls 0.23.45` (`Cargo.lock:2514`), newly flagged by RUSTSEC-2026-0346 through 0349; the current `origin/dev` has the owner-approved `0.23.46` lock update (`1b23ef311`). Both previously reported auth tests passed in this run, so no #118 baseline comparison is needed. The gate stopped before TypeScript checks. Raw output: [iterate.log](https://git.kayg.org/attachments/a685d224-8ed3-4ce0-9c4e-2215f12d3eb5). I will sync with current `origin/dev`, then rerun the gate. The #5 comments endpoint still returns `InvalidContentType`.
kayg commented 2026-10-15 07:09:27 +00:00
Author
Owner
Copy link

Dev sync gate complete. The branch still points at 1bd56ea9 while the validated origin/dev merge is staged; rustls is now 0.23.46. scripts/verify iterate finished with: ITERATE gate passed. This is not the final gate: run scripts/verify fast before reporting done. Rust and doc tests passed, including the previously load-sensitive auth checks; TypeScript protocol tests: 37 passed, 0 failed, 34,463 assertions. Tracer transfer budgets: critical 297,347/307,200 gzip and 247,318/256,000 Brotli bytes; deferred 48,089/51,200 gzip and 42,588/45,056 Brotli bytes. Gate log: iterate-after-dev.log. Link unit/fuzz rerun is queued for a managed build slot.

Dev sync gate complete. The branch still points at 1bd56ea9 while the validated origin/dev merge is staged; rustls is now 0.23.46. `scripts/verify iterate` finished with: `ITERATE gate passed. This is not the final gate: run scripts/verify fast before reporting done.` Rust and doc tests passed, including the previously load-sensitive auth checks; TypeScript protocol tests: 37 passed, 0 failed, 34,463 assertions. Tracer transfer budgets: critical 297,347/307,200 gzip and 247,318/256,000 Brotli bytes; deferred 48,089/51,200 gzip and 42,588/45,056 Brotli bytes. Gate log: [iterate-after-dev.log](https://git.kayg.org/attachments/b8f874b8-6a93-4a8f-98f9-2a8970e4f743). Link unit/fuzz rerun is queued for a managed build slot.
kayg commented 2026-10-15 07:10:16 +00:00
Author
Owner
Copy link

Validated origin/dev sync committed and pushed to terminal-links: 7eae81604. Commit records the rustls 0.23.46 deny fix and D250; merged scripts/verify iterate passed. Focused link unit/fuzz tests are waiting for a managed build slot.

Validated origin/dev sync committed and pushed to `terminal-links`: 7eae81604. Commit records the rustls 0.23.46 deny fix and D250; merged `scripts/verify iterate` passed. Focused link unit/fuzz tests are waiting for a managed build slot.
kayg commented 2026-10-15 07:19:39 +00:00
Author
Owner
Copy link

Focused link checks passed at 7eae81604: 12 tests, 0 failures, 54,608 assertions. Coverage includes eight-row soft wraps, trailing punctuation and bracket balance, IPv6/ports/query strings, OSC 8, file URL display-only behavior, deferred row invalidation, and malformed-input fuzzing. Raw log: link-unit-fuzz.log.

Focused link checks passed at 7eae81604: 12 tests, 0 failures, 54,608 assertions. Coverage includes eight-row soft wraps, trailing punctuation and bracket balance, IPv6/ports/query strings, OSC 8, file URL display-only behavior, deferred row invalidation, and malformed-input fuzzing. Raw log: [link-unit-fuzz.log](https://git.kayg.org/attachments/c5c40316-e780-437a-ac02-369ce0e99b88).
kayg commented 2026-10-15 07:22:10 +00:00
Author
Owner
Copy link

Fetched the two development-line commits that landed during verification; they add D278 and D279 only in DECISIONS.md. Merged and pushed that owner decision sync as 400acc2db. No product source changed; the final fast gate will run at this head.

Fetched the two development-line commits that landed during verification; they add D278 and D279 only in DECISIONS.md. Merged and pushed that owner decision sync as 400acc2db. No product source changed; the final fast gate will run at this head.
kayg commented 2026-10-15 07:33:15 +00:00
Author
Owner
Copy link

The production build graph classifies the link parser in the on-demand class: chunk-1mat2q84.js is 7,774 identity / 3,189 gzip / 2,896 Brotli bytes. The largest on-demand chunk remains 15,942 gzip / 14,180 Brotli, below its 20,480 / 18,432 limits; critical and deferred totals also remain within D188. The baseline comparison is still running. Evidence is in the attached iterate-after-dev.log.

The production build graph classifies the link parser in the on-demand class: `chunk-1mat2q84.js` is 7,774 identity / 3,189 gzip / 2,896 Brotli bytes. The largest on-demand chunk remains 15,942 gzip / 14,180 Brotli, below its 20,480 / 18,432 limits; critical and deferred totals also remain within D188. The baseline comparison is still running. Evidence is in the attached [iterate-after-dev.log](https://git.kayg.org/attachments/b8f874b8-6a93-4a8f-98f9-2a8970e4f743).
kayg commented 2026-10-15 07:33:43 +00:00
Author
Owner
Copy link

D188 transfer comparison, baseline b0a3799c → branch build: critical gzip/Brotli 295,956/246,132 → 297,347/247,318 bytes (+1,391/+1,186); deferred 47,862/42,429 → 48,089/42,588 (+227/+159); on-demand 78,209/69,690 → 81,375/72,563 (+3,166/+2,873). The link parser is in on-demand chunk chunk-1mat2q84.js at 7,774 identity / 3,189 gzip / 2,896 Brotli bytes. The largest chunk remains 15,942 gzip / 14,180 Brotli, within 20,480 / 18,432 limits. Baseline transfer receipt: baseline-transfer-budget.log; branch receipt: iterate-after-dev.log.

D188 transfer comparison, baseline b0a3799c → branch build: critical gzip/Brotli 295,956/246,132 → 297,347/247,318 bytes (+1,391/+1,186); deferred 47,862/42,429 → 48,089/42,588 (+227/+159); on-demand 78,209/69,690 → 81,375/72,563 (+3,166/+2,873). The link parser is in on-demand chunk `chunk-1mat2q84.js` at 7,774 identity / 3,189 gzip / 2,896 Brotli bytes. The largest chunk remains 15,942 gzip / 14,180 Brotli, within 20,480 / 18,432 limits. Baseline transfer receipt: [baseline-transfer-budget.log](https://git.kayg.org/attachments/81bb837f-4a1e-4449-b9c2-593530ac57b5); branch receipt: [iterate-after-dev.log](https://git.kayg.org/attachments/b8f874b8-6a93-4a8f-98f9-2a8970e4f743).
kayg commented 2026-10-15 07:58:15 +00:00
Author
Owner
Copy link

Synced the two new origin/dev decision commits (D280/D281); merge head is 7f702c090526d8fc55333557391dbb9c6802933e.

Perf finding: the first key-to-paint run collected no timings because Playwright was launched from /root, so it could not resolve mock-perf.config.ts (raw receipt: key-before-1.log). Updated the perf runner to start in its config directory and verified its shell syntax. I’m rerunning the measurement under the perf VM lock.

Synced the two new origin/dev decision commits (D280/D281); merge head is `7f702c090526d8fc55333557391dbb9c6802933e`. Perf finding: the first key-to-paint run collected no timings because Playwright was launched from `/root`, so it could not resolve `mock-perf.config.ts` (raw receipt: [key-before-1.log](https://git.kayg.org/attachments/c5329608-a9cb-4b5b-9c9c-628fdf16a4b6)). Updated the perf runner to start in its config directory and verified its shell syntax. I’m rerunning the measurement under the perf VM lock.
kayg commented 2026-10-15 08:16:15 +00:00
Author
Owner
Copy link

Perf harness finding: the second attempt reached Playwright but collected no samples because the perf VM has no system Node runtime; launching its CLI under Bun failed to register the tests (No tests found; receipt: key-before-1-bun-no-tests.log). I installed Node 22.23.3 under the perf job scratch and switched the runner to the pinned local Playwright CLI. --list now finds the expected three terminal tests. The corrected baseline key run is queued under the shared VM lock.

Perf harness finding: the second attempt reached Playwright but collected no samples because the perf VM has no system Node runtime; launching its CLI under Bun failed to register the tests (`No tests found`; receipt: [key-before-1-bun-no-tests.log](https://git.kayg.org/attachments/8c43ee85-cae5-4cc0-b82b-b9fb6f42e526)). I installed Node 22.23.3 under the perf job scratch and switched the runner to the pinned local Playwright CLI. `--list` now finds the expected three terminal tests. The corrected baseline key run is queued under the shared VM lock.
kayg commented 2026-10-15 08:17:07 +00:00
Author
Owner
Copy link

Perf harness follow-up: with the local Node runtime, Playwright correctly enumerates the terminal tests, but the test web server could not find Bun on its PATH (bun is required to start the tracer test double; receipt: key-before-1-bun-path-failure.log). Added /root/.bun/bin to the VM runner PATH and syntax-checked it. No timing samples were collected in either failed setup attempt; retrying the baseline run.

Perf harness follow-up: with the local Node runtime, Playwright correctly enumerates the terminal tests, but the test web server could not find Bun on its PATH (`bun is required to start the tracer test double`; receipt: [key-before-1-bun-path-failure.log](https://git.kayg.org/attachments/c04d4ac3-9cbb-44ff-bfd4-96896782644e)). Added `/root/.bun/bin` to the VM runner PATH and syntax-checked it. No timing samples were collected in either failed setup attempt; retrying the baseline run.
kayg commented 2026-10-15 08:21:07 +00:00
Author
Owner
Copy link

Perf harness follow-up: the test spec expects INFRATERNAL_TEST_RTT, while the runner passed the misspelled INFRATERAL_TEST_RTT. It therefore used the default 300 ms RTT; reconnect p99 was 1,524.40 ms against its 900 ms assertion and the run exited 1. The intended zero-RTT key-to-paint result is not available from this attempt. Receipt: key-before-1-default-300ms-env.log. Fixed the variable name and am retrying.

Perf harness follow-up: the test spec expects `INFRATERNAL_TEST_RTT`, while the runner passed the misspelled `INFRATERAL_TEST_RTT`. It therefore used the default 300 ms RTT; reconnect p99 was 1,524.40 ms against its 900 ms assertion and the run exited 1. The intended zero-RTT key-to-paint result is not available from this attempt. Receipt: [key-before-1-default-300ms-env.log](https://git.kayg.org/attachments/7f231418-740a-4c76-9052-cff8e79931f6). Fixed the variable name and am retrying.
kayg commented 2026-10-15 08:22:53 +00:00
Author
Owner
Copy link

Correction to the previous RTT note: the runner typo caused the spec and proxy to use their 300 ms default, and the key probe completed all 20 samples. Those key measurements are valid for the 300 ms synthetic RTT profile, though the same run later failed the separate reconnect assertion (p99=1,524.40 ms, budget 900 ms). The test requires RTT above 60 ms to exercise prediction, so I’m pinning 300 ms explicitly for the remaining before/after samples. Receipt: key-before-1-default-300ms-env.log.

Correction to the previous RTT note: the runner typo caused the spec and proxy to use their 300 ms default, and the key probe completed all 20 samples. Those key measurements are valid for the 300 ms synthetic RTT profile, though the same run later failed the separate reconnect assertion (`p99=1,524.40 ms`, budget `900 ms`). The test requires RTT above 60 ms to exercise prediction, so I’m pinning 300 ms explicitly for the remaining before/after samples. Receipt: [key-before-1-default-300ms-env.log](https://git.kayg.org/attachments/7f231418-740a-4c76-9052-cff8e79931f6).
kayg commented 2026-10-15 08:26:15 +00:00
Author
Owner
Copy link

Baseline key trial 2 on infraternal-perf-test (load before 0.26 0.57 1.05): 20 samples; predictive glyph visible p50/p99 12.90/13.80 ms (17 ms budget), prediction-off echo p50/p99 314.50/315.50 ms at 300 ms configured RTT. The probe metrics passed; the enclosing test later failed only its reconnect check (p99=1,532.30 ms, budget=900 ms). Raw receipt: key-before-2.log.

Baseline key trial 2 on `infraternal-perf-test` (load before `0.26 0.57 1.05`): 20 samples; predictive glyph visible p50/p99 `12.90/13.80 ms` (17 ms budget), prediction-off echo p50/p99 `314.50/315.50 ms` at 300 ms configured RTT. The probe metrics passed; the enclosing test later failed only its reconnect check (`p99=1,532.30 ms`, `budget=900 ms`). Raw receipt: [key-before-2.log](https://git.kayg.org/attachments/9fc6dc30-f9b6-459e-94e9-5c722b1cd2fe).
kayg commented 2026-10-15 08:31:20 +00:00
Author
Owner
Copy link

Baseline key trial 3 on infraternal-perf-test (load before 1.86 1.32 1.23): RTT measured 306.60 ms; predictive glyph p50/p99 12.70/28.60 ms (17 ms budget), prediction-off echo 314.30/328.90 ms. The 28.60 ms p99 and reconnect p99 1,556.30 ms exceeded their test budgets, so this raw trial is retained as a loaded baseline sample, not a passing test. Receipt: key-before-3.log.

Baseline key trial 3 on `infraternal-perf-test` (load before `1.86 1.32 1.23`): RTT measured `306.60 ms`; predictive glyph p50/p99 `12.70/28.60 ms` (17 ms budget), prediction-off echo `314.30/328.90 ms`. The 28.60 ms p99 and reconnect p99 `1,556.30 ms` exceeded their test budgets, so this raw trial is retained as a loaded baseline sample, not a passing test. Receipt: [key-before-3.log](https://git.kayg.org/attachments/d01cfb1c-2ebd-45cb-a54e-e4d4458fc497).
kayg commented 2026-10-15 08:34:31 +00:00
Author
Owner
Copy link

Branch key trial 1 on infraternal-perf-test (load before 0.37 0.89 1.08): 20 samples at 300 ms RTT; predictive glyph visible p50/p99 12.60/13.80 ms (17 ms budget), prediction-off echo 314.90/327.10 ms. The key probe passed; the enclosing test again failed the separate reconnect assertion (p99=1,525.50 ms, budget 900 ms). Raw receipt: key-after-1.log.

Branch key trial 1 on `infraternal-perf-test` (load before `0.37 0.89 1.08`): 20 samples at 300 ms RTT; predictive glyph visible p50/p99 `12.60/13.80 ms` (17 ms budget), prediction-off echo `314.90/327.10 ms`. The key probe passed; the enclosing test again failed the separate reconnect assertion (`p99=1,525.50 ms`, budget `900 ms`). Raw receipt: [key-after-1.log](https://git.kayg.org/attachments/03f36bac-4c14-45dc-919f-d17c8f208769).
kayg commented 2026-10-15 08:37:37 +00:00
Author
Owner
Copy link

Branch key trial 2 on infraternal-perf-test (load before 0.44 0.87 1.04): 20 samples at 300 ms RTT; predictive glyph p50/p99 13.20/13.90 ms, prediction-off echo 298.20/301.60 ms. Predictive p99 remains under the 17 ms budget. The enclosing test again failed the separate reconnect assertion (p99=1,526.70 ms, budget 900 ms). Raw receipt: key-after-2.log.

Branch key trial 2 on `infraternal-perf-test` (load before `0.44 0.87 1.04`): 20 samples at 300 ms RTT; predictive glyph p50/p99 `13.20/13.90 ms`, prediction-off echo `298.20/301.60 ms`. Predictive p99 remains under the 17 ms budget. The enclosing test again failed the separate reconnect assertion (`p99=1,526.70 ms`, budget `900 ms`). Raw receipt: [key-after-2.log](https://git.kayg.org/attachments/PLACEHOLDER).
kayg commented 2026-10-15 08:37:41 +00:00
Author
Owner
Copy link

Correction: trial 2 receipt link is key-after-2.log.

Correction: trial 2 receipt link is [key-after-2.log](https://git.kayg.org/attachments/99fec9d4-08a0-46c2-9386-f3ae3648b465).
kayg commented 2026-10-15 08:40:44 +00:00
Author
Owner
Copy link

Branch key trial 3 on infraternal-perf-test (load before 0.10 0.55 0.88): 20 samples at 300 ms RTT; predictive glyph p50/p99 13.10/14.30 ms, prediction-off echo 312.20/314.90 ms. Predictive p99 remained under 17 ms. The separate reconnect assertion again failed (p99=1,537.10 ms, budget 900 ms). Raw receipt: key-after-3.log.

Branch key trial 3 on `infraternal-perf-test` (load before `0.10 0.55 0.88`): 20 samples at 300 ms RTT; predictive glyph p50/p99 `13.10/14.30 ms`, prediction-off echo `312.20/314.90 ms`. Predictive p99 remained under 17 ms. The separate reconnect assertion again failed (`p99=1,537.10 ms`, budget `900 ms`). Raw receipt: [key-after-3.log](https://git.kayg.org/attachments/65cf8b0e-a1f8-4721-a1ec-00d90fc05d25).
kayg commented 2026-10-15 08:58:08 +00:00
Author
Owner
Copy link

Static binary size check: branch 7f702c0905 is 22,277,392 bytes against the 22,204,304-byte cap (+73,088). The same recipe measured merge-base b0a3799c at 22,275,888 bytes, so branch growth is 1,504 bytes, below D243's 16 KiB review threshold. The fast gate is continuing through remaining checks. Raw gate log: /home/kayg/.cache/codex-scratch/terminal-links/evidence/fast.log.

Static binary size check: branch 7f702c090526d8fc55333557391dbb9c6802933e is 22,277,392 bytes against the 22,204,304-byte cap (+73,088). The same recipe measured merge-base b0a3799c at 22,275,888 bytes, so branch growth is 1,504 bytes, below D243's 16 KiB review threshold. The fast gate is continuing through remaining checks. Raw gate log: /home/kayg/.cache/codex-scratch/terminal-links/evidence/fast.log.
kayg commented 2026-10-15 09:39:46 +00:00
Author
Owner
Copy link

Fast gate completed at 7f702c0905. fmt, VT native fmt, CLI plugin compile-out, clippy, workspace tests and cargo-deny passed. Static size checks failed: x86_64 22,277,392 bytes (+73,088 over cap; +1,504 from measured merge-base 22,275,888), aarch64 19,705,112 bytes (+66,368 over cap). Static gate also reports server/agent subcommands exit 2 after the oversized fixture is rejected. Full log: https://git.kayg.org/attachments/cc89d9bd-9dfc-48f8-8799-2a35be5b0833

+---------------------------+--------------+
| Rust gate | Result |
+---------------------------+--------------+
| fmt | PASS |
| VT native fmt | PASS |
| CLI plugin compile-out | PASS |
| clippy | PASS |
| test | PASS |
| cargo-deny | PASS |
| static builds | FAIL (1) |
+---------------------------+--------------+

Fast gate completed at 7f702c090526d8fc55333557391dbb9c6802933e. fmt, VT native fmt, CLI plugin compile-out, clippy, workspace tests and cargo-deny passed. Static size checks failed: x86_64 22,277,392 bytes (+73,088 over cap; +1,504 from measured merge-base 22,275,888), aarch64 19,705,112 bytes (+66,368 over cap). Static gate also reports server/agent subcommands exit 2 after the oversized fixture is rejected. Full log: https://git.kayg.org/attachments/cc89d9bd-9dfc-48f8-8799-2a35be5b0833 +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | test | PASS | | cargo-deny | PASS | | static builds | FAIL (1) | +---------------------------+--------------+
kayg commented 2026-10-15 09:45:15 +00:00
Author
Owner
Copy link

The perf-VM 10 MB cat harness acquired the lock at load 0.18 / 0.32 / 0.58 but exited before timing: Playwright loaded apps/tracer/real/server.ts as CommonJS and failed on import.meta (Error: Cannot use 'import.meta' outside a module; no tests found). No throughput number was produced. Raw receipt: https://git.kayg.org/attachments/ec761b2c-36a8-4a8b-9d77-ab96ee8ed898. The remote scratch harness is being corrected to include the app's ESM package metadata before rerunning.

The perf-VM 10 MB cat harness acquired the lock at load 0.18 / 0.32 / 0.58 but exited before timing: Playwright loaded apps/tracer/real/server.ts as CommonJS and failed on import.meta (Error: Cannot use 'import.meta' outside a module; no tests found). No throughput number was produced. Raw receipt: https://git.kayg.org/attachments/ec761b2c-36a8-4a8b-9d77-ab96ee8ed898. The remote scratch harness is being corrected to include the app's ESM package metadata before rerunning.
kayg commented 2026-10-15 09:47:06 +00:00
Author
Owner
Copy link

After fixing ESM loading, perf-VM cat attempt 2 reached the test but stopped before timing: the long INFRATERAL_TEST_STATE_ROOT made the Unix session socket exceed the 108-byte path limit. No throughput number was produced. Raw receipt: https://git.kayg.org/attachments/ef1ae430-341f-4687-9290-405cf7e7ea96. I will shorten only the scratch runner's state root and rerun.

After fixing ESM loading, perf-VM cat attempt 2 reached the test but stopped before timing: the long INFRATERAL_TEST_STATE_ROOT made the Unix session socket exceed the 108-byte path limit. No throughput number was produced. Raw receipt: https://git.kayg.org/attachments/ef1ae430-341f-4687-9290-405cf7e7ea96. I will shorten only the scratch runner's state root and rerun.
kayg commented 2026-10-15 09:49:15 +00:00
Author
Owner
Copy link

Perf-VM cat attempt 3 passed the short-socket-path setup but stopped before timing because the runner used INFRATERAL_BIN; the real-server harness expects INFRATERNAL_BIN, so it fell back to a non-existent target path. No throughput number was produced. Raw receipt: https://git.kayg.org/attachments/85999b80-fb1d-4d55-9ce1-203a49f2b0a4. The scratch runner now uses the required binary variable and short state root.

Perf-VM cat attempt 3 passed the short-socket-path setup but stopped before timing because the runner used `INFRATERAL_BIN`; the real-server harness expects `INFRATERNAL_BIN`, so it fell back to a non-existent target path. No throughput number was produced. Raw receipt: https://git.kayg.org/attachments/85999b80-fb1d-4d55-9ce1-203a49f2b0a4. The scratch runner now uses the required binary variable and short state root.
kayg commented 2026-10-15 09:50:44 +00:00
Author
Owner
Copy link

Perf-VM cat attempt 4 acquired the lock, started the Playwright test at load average 0.50 / 0.33 / 0.50, then hit the runner's 30-second default timeout before producing a throughput result. Raw receipt: https://git.kayg.org/attachments/144c1558-bbd6-4919-a7de-2afe04d997fc. I am checking the stalled test step and will adjust only the scratch test timeout if needed.

Perf-VM cat attempt 4 acquired the lock, started the Playwright test at load average 0.50 / 0.33 / 0.50, then hit the runner's 30-second default timeout before producing a throughput result. Raw receipt: https://git.kayg.org/attachments/144c1558-bbd6-4919-a7de-2afe04d997fc. I am checking the stalled test step and will adjust only the scratch test timeout if needed.
kayg commented 2026-10-15 09:57:54 +00:00
Author
Owner
Copy link

Perf-VM attempt 6 completed the stream setup but the measurement helper timed out because the newly bounded 64-character diagnostic tail was searched using a cursor from the old cumulative-buffer implementation. The marker can shift before that cursor. No throughput number was produced. Raw receipt: https://git.kayg.org/attachments/8b6d78a7-ff07-488e-8586-1c54eee34d81. I am updating the scratch test to search the full bounded tail.

Perf-VM attempt 6 completed the stream setup but the measurement helper timed out because the newly bounded 64-character diagnostic tail was searched using a cursor from the old cumulative-buffer implementation. The marker can shift before that cursor. No throughput number was produced. Raw receipt: https://git.kayg.org/attachments/8b6d78a7-ff07-488e-8586-1c54eee34d81. I am updating the scratch test to search the full bounded tail.
kayg commented 2026-10-15 10:03:23 +00:00
Author
Owner
Copy link

Perf-VM attempt 8 reached the test but the 5-second preflight marker failed before cat started; the server bridge reported output_bytes=none for every stream. The harness waited for terminal attach only, so I am adding a wait for replay_complete before sending input. No throughput result was produced. Raw receipt: https://git.kayg.org/attachments/0ac31707-149b-422d-9edc-44dbf09e1f49.

Perf-VM attempt 8 reached the test but the 5-second preflight marker failed before cat started; the server bridge reported output_bytes=none for every stream. The harness waited for terminal attach only, so I am adding a wait for replay_complete before sending input. No throughput result was produced. Raw receipt: https://git.kayg.org/attachments/0ac31707-149b-422d-9edc-44dbf09e1f49.
kayg commented 2026-10-15 10:08:58 +00:00
Author
Owner
Copy link

Perf-VM attempt 10 passed the terminal-ready check but the keyboard preflight still produced no protocol frames: attached stream 1, replay count 1, input bytes none, output bytes none. Raw receipt: https://git.kayg.org/attachments/e0ff6540-6216-4fb9-b4d6-2b99df9382ce. The scratch harness will use the existing real-protocol probe to issue the fixed cat command, while measuring its output reaching and painting in the production browser terminal.

Perf-VM attempt 10 passed the terminal-ready check but the keyboard preflight still produced no protocol frames: attached stream 1, replay count 1, input bytes none, output bytes none. Raw receipt: https://git.kayg.org/attachments/e0ff6540-6216-4fb9-b4d6-2b99df9382ce. The scratch harness will use the existing real-protocol probe to issue the fixed cat command, while measuring its output reaching and painting in the production browser terminal.
kayg commented 2026-10-15 10:11:23 +00:00
Author
Owner
Copy link

Perf-VM attempt 11 used the existing protocol probe to inject a command and received its input acknowledgement, but the browser bridge still recorded no output frames. Raw receipt: https://git.kayg.org/attachments/8398f033-cb96-482f-b15e-8203dd76fc72. I am changing the cat measurement to consume the real PTY output on that protocol probe; this isolates production PTY/protocol throughput, while the separate key-to-paint comparison covers browser input and rendering latency.

Perf-VM attempt 11 used the existing protocol probe to inject a command and received its input acknowledgement, but the browser bridge still recorded no output frames. Raw receipt: https://git.kayg.org/attachments/8398f033-cb96-482f-b15e-8203dd76fc72. I am changing the cat measurement to consume the real PTY output on that protocol probe; this isolates production PTY/protocol throughput, while the separate key-to-paint comparison covers browser input and rendering latency.
kayg commented 2026-10-15 10:16:50 +00:00
Author
Owner
Copy link

Perf-VM real-protocol attempt 13 produced its first valid cat sample: 10,000,000 bytes in 1,163.43 ms (8.60 MB/s) at load average 0.38 / 0.67 / 0.75. The PTY session closed after that command, so the follow-up reset/second sample failed; the harness now runs one sample in a fresh PTY per locked trial. Receipt: https://git.kayg.org/attachments/de99e8d0-9151-484d-a267-c86dd827b185.

Perf-VM real-protocol attempt 13 produced its first valid cat sample: 10,000,000 bytes in 1,163.43 ms (8.60 MB/s) at load average 0.38 / 0.67 / 0.75. The PTY session closed after that command, so the follow-up reset/second sample failed; the harness now runs one sample in a fresh PTY per locked trial. Receipt: https://git.kayg.org/attachments/de99e8d0-9151-484d-a267-c86dd827b185.
kayg commented 2026-10-15 10:20:20 +00:00
Author
Owner
Copy link

Branch PTY throughput sample 2 closed before completion: protocol Close 1013 at 2,035,850 output bytes, no throughput metric. This is an invalid sample; sample 1 completed at 14.15 MB/s. Receipt: cat-after-2.log. Continuing with fresh one-sample runs.

Branch PTY throughput sample 2 closed before completion: protocol Close 1013 at 2,035,850 output bytes, no throughput metric. This is an invalid sample; sample 1 completed at 14.15 MB/s. Receipt: [cat-after-2.log](https://git.kayg.org/attachments/28095d67-62b7-4a11-b320-70528567ca2e). Continuing with fresh one-sample runs.
kayg commented 2026-10-15 10:22:13 +00:00
Author
Owner
Copy link

Perf step complete on the 4-vCPU/7-GB perf VM. Real PTY protocol 10 MB samples: baseline 8.60, 16.65, 33.88 MB/s (median 16.65); branch 14.15, 19.81, 15.44 MB/s (median 15.44). The ranges overlap materially (baseline 8.60–33.88; branch 14.15–19.81), so these samples do not establish a throughput regression or improvement. Load averages are in the attached receipts. This probes PTY protocol output, not browser rendering; the separate key-to-paint runs cover browser input/rendering. The baseline 8.60 MB/s sample was valid, though that invocation later failed during its reset step (recorded earlier).

Baseline receipts: sample 1, sample 2, sample 3.
Branch receipts: sample 1, sample 2, sample 3.

Perf step complete on the 4-vCPU/7-GB perf VM. Real PTY protocol 10 MB samples: baseline 8.60, 16.65, 33.88 MB/s (median 16.65); branch 14.15, 19.81, 15.44 MB/s (median 15.44). The ranges overlap materially (baseline 8.60–33.88; branch 14.15–19.81), so these samples do not establish a throughput regression or improvement. Load averages are in the attached receipts. This probes PTY protocol output, not browser rendering; the separate key-to-paint runs cover browser input/rendering. The baseline 8.60 MB/s sample was valid, though that invocation later failed during its reset step (recorded earlier). Baseline receipts: [sample 1](https://git.kayg.org/attachments/de99e8d0-9151-484d-a267-c86dd827b185), [sample 2](https://git.kayg.org/attachments/6c46067b-a455-40a1-82cd-383213a3a0b7), [sample 3](https://git.kayg.org/attachments/0bea3056-d3bb-41ec-99f2-ad6f3e2c33fc). Branch receipts: [sample 1](https://git.kayg.org/attachments/c151f2c4-ef0f-49d6-962e-f09e0d7aace0), [sample 2](https://git.kayg.org/attachments/2b06e223-1d0d-43f1-8edc-ac4286585f23), [sample 3](https://git.kayg.org/attachments/0f61b7fa-505d-424f-89f2-a41c1761ca72).
kayg commented 2026-10-15 10:26:29 +00:00
Author
Owner
Copy link

TypeScript gate status: apps/tracer bun run check has been queued in build-slot since 11:45 local. At 12:26 local all four host build slots remained occupied by other jobs; this check has not started and has produced no output. It remains queued under the required scheduler.

TypeScript gate status: `apps/tracer` `bun run check` has been queued in build-slot since 11:45 local. At 12:26 local all four host build slots remained occupied by other jobs; this check has not started and has produced no output. It remains queued under the required scheduler.
kayg commented 2026-10-15 10:37:40 +00:00
Author
Owner
Copy link

TypeScript gate complete: apps/tracer bun run check passed (exit 0; it ran bun x tsc --project tsconfig.json). Receipt: bun-check.log.

TypeScript gate complete: `apps/tracer` `bun run check` passed (exit 0; it ran `bun x tsc --project tsconfig.json`). Receipt: [bun-check.log](https://git.kayg.org/attachments/43c4681e-3898-4887-bf2b-7c69b7cf7940).
kayg commented 2026-10-15 10:38:17 +00:00
Author
Owner
Copy link

Finished this lane on terminal-links, HEAD 7f702c090526d8fc55333557391dbb9c6802933e (already pushed; no new source changes during this continuation). apps/tracer bun run check passed (exit 0; receipt). scripts/verify iterate passed. Link unit/fuzz tests: 12 tests, 0 failures, 54,608 assertions. Production browser tests passed 2/2; screenshots: wrapped link, OSC 8 and affordances.

Performance receipts and comparisons are in the preceding comments. Key-to-paint median p50/p99 was 12.90/14.00 ms baseline and 13.10/13.90 ms branch; worst p99 was 28.60 ms baseline (loaded trial) and 14.30 ms branch. The reconnect assertion failed in both variants. Real PTY protocol 10 MB throughput medians were 16.65 MB/s baseline and 15.44 MB/s branch, with ranges 8.60–33.88 and 14.15–19.81 MB/s; this measures PTY protocol output rather than browser rendering and does not establish a throughput change. D188 transfer sizes baseline→branch: critical gzip/Brotli 295,956/246,132→297,347/247,318 (+1,391/+1,186); deferred 47,862/42,429→48,089/42,588 (+227/+159); on-demand 78,209/69,690→81,375/72,563 (+3,166/+2,873).

Stripped x86_64 binary: 22,275,888 bytes at baseline and 22,277,392 on branch, delta +1,504 bytes. The fast gate’s absolute caps still fail: x86_64 is +73,088 bytes over cap and aarch64 is +66,368 bytes over cap. This is recorded with the size evidence and full fast log.

scripts/verify fast output:

+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| test                      | PASS         |
| cargo-deny                | PASS         |
| static builds             | FAIL (1)     |
+---------------------------+--------------+

Stopping for review. The static-size failure is the remaining gate issue; no deploy was performed.

Finished this lane on `terminal-links`, HEAD `7f702c090526d8fc55333557391dbb9c6802933e` (already pushed; no new source changes during this continuation). `apps/tracer` `bun run check` passed (exit 0; [receipt](https://git.kayg.org/attachments/43c4681e-3898-4887-bf2b-7c69b7cf7940)). `scripts/verify iterate` passed. Link unit/fuzz tests: 12 tests, 0 failures, 54,608 assertions. Production browser tests passed 2/2; screenshots: [wrapped link](https://git.kayg.org/attachments/763c6f3d-f482-417d-87fe-f7f67d5b81a2), [OSC 8 and affordances](https://git.kayg.org/attachments/c46c2b13-b754-4f82-ba65-3ff11c6493a6). Performance receipts and comparisons are in the preceding comments. Key-to-paint median p50/p99 was 12.90/14.00 ms baseline and 13.10/13.90 ms branch; worst p99 was 28.60 ms baseline (loaded trial) and 14.30 ms branch. The reconnect assertion failed in both variants. Real PTY protocol 10 MB throughput medians were 16.65 MB/s baseline and 15.44 MB/s branch, with ranges 8.60–33.88 and 14.15–19.81 MB/s; this measures PTY protocol output rather than browser rendering and does not establish a throughput change. D188 transfer sizes baseline→branch: critical gzip/Brotli 295,956/246,132→297,347/247,318 (+1,391/+1,186); deferred 47,862/42,429→48,089/42,588 (+227/+159); on-demand 78,209/69,690→81,375/72,563 (+3,166/+2,873). Stripped x86_64 binary: 22,275,888 bytes at baseline and 22,277,392 on branch, delta +1,504 bytes. The fast gate’s absolute caps still fail: x86_64 is +73,088 bytes over cap and aarch64 is +66,368 bytes over cap. This is recorded with the [size evidence](https://git.kayg.org/attachments/db751512-39e8-47b1-b4c0-f5582067b506) and [full fast log](https://git.kayg.org/attachments/cc89d9bd-9dfc-48f8-8799-2a35be5b0833). `scripts/verify fast` output: ```text +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | test | PASS | | cargo-deny | PASS | | static builds | FAIL (1) | +---------------------------+--------------+ ``` Stopping for review. The static-size failure is the remaining gate issue; no deploy was performed.
kayg commented 2026-10-15 10:40:24 +00:00
Author
Owner
Copy link

Independent review started on branch review-terminal-links, target head 7f702c090526d8fc55333557391dbb9c6802933e, diff base fb9830291d416499c531650cd48b97a3415937ec. I read the issue and parent history, the project rules, and the job report. Review will focus on link parsing, stale-cache invalidation, activation safety, production browser behavior, and the claimed hot-path isolation.

Independent review started on branch `review-terminal-links`, target head `7f702c090526d8fc55333557391dbb9c6802933e`, diff base `fb9830291d416499c531650cd48b97a3415937ec`. I read the issue and parent history, the project rules, and the job report. Review will focus on link parsing, stale-cache invalidation, activation safety, production browser behavior, and the claimed hot-path isolation.
kayg commented 2026-10-15 10:46:47 +00:00
Author
Owner
Copy link

Review finding F1 (P2), committed on review-terminal-links at 39d9695ec074cfc9c26b24f7b1133529d4f39641: plain mailto:person@example.test is not recognized. apps/tracer/src/terminal-links.on-demand.ts:41 omits mailto: while apps/tracer/src/terminal.ts:319 disables Ghostty's provider; the pinned provider regex at apps/tracer/vendor/ghostty-web/dist/ghostty-web.js:1621 includes it. The new recognizes plain mailto links as openable terminal links test fails; suite result: 11 pass, 1 fail. Details are in /home/kayg/.local/state/codex-jobs/infraternal/review-terminal-links/findings.md.

Review finding F1 (P2), committed on `review-terminal-links` at `39d9695ec074cfc9c26b24f7b1133529d4f39641`: plain `mailto:person@example.test` is not recognized. `apps/tracer/src/terminal-links.on-demand.ts:41` omits `mailto:` while `apps/tracer/src/terminal.ts:319` disables Ghostty's provider; the pinned provider regex at `apps/tracer/vendor/ghostty-web/dist/ghostty-web.js:1621` includes it. The new `recognizes plain mailto links as openable terminal links` test fails; suite result: 11 pass, 1 fail. Details are in `/home/kayg/.local/state/codex-jobs/infraternal/review-terminal-links/findings.md`.
kayg commented 2026-10-15 10:47:54 +00:00
Author
Owner
Copy link

Review finding F2 (P2), committed on review-terminal-links at 1695d824b2cf85d09eedbb72ddfcf49c608d8f6b: a previously unseen scrollback viewport can expose a valid prefix of a wrapped URL. apps/tracer/src/terminal.ts:1993 marks scrollback wrap state unknown; apps/tracer/src/terminal-links.on-demand.ts:232-236 then treats uncached rows as unwrapped. The new does not expose a valid prefix when scrollback wrap history is unknown test expected https://scroll.example.test/a-long-path but got https://scroll.example.t (link suite: 11 pass, 2 fail). Evidence is recorded in /home/kayg/.local/state/codex-jobs/infraternal/review-terminal-links/findings.md.

Review finding F2 (P2), committed on `review-terminal-links` at `1695d824b2cf85d09eedbb72ddfcf49c608d8f6b`: a previously unseen scrollback viewport can expose a valid prefix of a wrapped URL. `apps/tracer/src/terminal.ts:1993` marks scrollback wrap state unknown; `apps/tracer/src/terminal-links.on-demand.ts:232-236` then treats uncached rows as unwrapped. The new `does not expose a valid prefix when scrollback wrap history is unknown` test expected `https://scroll.example.test/a-long-path` but got `https://scroll.example.t` (link suite: 11 pass, 2 fail). Evidence is recorded in `/home/kayg/.local/state/codex-jobs/infraternal/review-terminal-links/findings.md`.
kayg commented 2026-10-15 10:54:21 +00:00
Author
Owner
Copy link

Focused parser/fuzz verification completed at review head 1695d824b2cf85d09eedbb72ddfcf49c608d8f6b: 14 tests, 12 pass, 2 fail, 54,610 assertions. The failures are the committed F1 mailto-detection and F2 uncached-scrollback partial-link regressions. Raw output: $HOME/.cache/codex-scratch/terminal-links-review/link-tests.log.

Focused parser/fuzz verification completed at review head `1695d824b2cf85d09eedbb72ddfcf49c608d8f6b`: 14 tests, 12 pass, 2 fail, 54,610 assertions. The failures are the committed F1 mailto-detection and F2 uncached-scrollback partial-link regressions. Raw output: `$HOME/.cache/codex-scratch/terminal-links-review/link-tests.log`.
kayg commented 2026-10-15 10:59:45 +00:00
Author
Owner
Copy link

TypeScript review check completed on review-terminal-links at 1695d824b2cf85d09eedbb72ddfcf49c608d8f6b: after installing the locked tracer dependencies, apps/tracer bun run check passed (exit 0). The initial attempt had stopped before source analysis because the worktree lacked installed Bun/Node type definitions.

TypeScript review check completed on `review-terminal-links` at `1695d824b2cf85d09eedbb72ddfcf49c608d8f6b`: after installing the locked tracer dependencies, `apps/tracer` `bun run check` passed (exit 0). The initial attempt had stopped before source analysis because the worktree lacked installed Bun/Node type definitions.
kayg commented 2026-10-15 11:06:13 +00:00
Author
Owner
Copy link

Verification update: the production browser cases (TRACER_TEST_PORT=4246) and scripts/verify fast on detached target head 7f702c090526d8fc55333557391dbb9c6802933e are queued through the required build-slot scheduler. All four host slots remain occupied; neither command has produced output. I have left the other jobs untouched and will continue when capacity opens.

Verification update: the production browser cases (`TRACER_TEST_PORT=4246`) and `scripts/verify fast` on detached target head `7f702c090526d8fc55333557391dbb9c6802933e` are queued through the required build-slot scheduler. All four host slots remain occupied; neither command has produced output. I have left the other jobs untouched and will continue when capacity opens.
kayg commented 2026-10-15 11:59:15 +00:00
Author
Owner
Copy link

Independent scripts/verify fast on exact job head 7f702c090526d8fc55333557391dbb9c6802933e completed, but its test row hit a missing Cargo output path (target/debug/deps/...rcgu.o: No such file or directory, exit 101); I am rerunning to distinguish this from a test failure. fmt, VT native fmt, CLI plugin compile-out, clippy, and cargo-deny passed. Static size checks reproduced the reported overruns: x86_64 22,277,392 / 22,204,304 bytes (+73,088); aarch64 19,705,112 / 19,638,744 bytes (+66,368). Gate table: fmt PASS; VT native fmt PASS; CLI plugin compile-out PASS; clippy PASS; test FAIL (101); cargo-deny PASS; static builds FAIL (1). Raw log: $HOME/.cache/codex-scratch/terminal-links-review/fast.log.

Independent `scripts/verify fast` on exact job head `7f702c090526d8fc55333557391dbb9c6802933e` completed, but its test row hit a missing Cargo output path (`target/debug/deps/...rcgu.o: No such file or directory`, exit 101); I am rerunning to distinguish this from a test failure. fmt, VT native fmt, CLI plugin compile-out, clippy, and cargo-deny passed. Static size checks reproduced the reported overruns: x86_64 22,277,392 / 22,204,304 bytes (+73,088); aarch64 19,705,112 / 19,638,744 bytes (+66,368). Gate table: fmt PASS; VT native fmt PASS; CLI plugin compile-out PASS; clippy PASS; test FAIL (101); cargo-deny PASS; static builds FAIL (1). Raw log: `$HOME/.cache/codex-scratch/terminal-links-review/fast.log`.
kayg commented 2026-10-15 12:22:00 +00:00
Author
Owner
Copy link

Second scripts/verify fast run on job head 7f702c090526d8fc55333557391dbb9c6802933e completed at review SHA 1695d824b2cf85d09eedbb72ddfcf49c608d8f6b. Workspace tests: 422 passed, 1 failed, 46 ignored; the only failure was auth socket review_full_input_writer_preserves_browser_control_progress, whose inputs.recv() timed out at 5 seconds. The matching test is running on a detached merge-base dev checkout under the shared build-slot gate. Size checks again reported x86_64 22,277,392 / 22,204,304 (+73,088) and aarch64 19,705,112 / 19,638,744 (+66,368). Full raw gate output: $HOME/.cache/codex-scratch/terminal-links-review/fast-rerun.log.

Second `scripts/verify fast` run on job head `7f702c090526d8fc55333557391dbb9c6802933e` completed at review SHA `1695d824b2cf85d09eedbb72ddfcf49c608d8f6b`. Workspace tests: 422 passed, 1 failed, 46 ignored; the only failure was auth socket `review_full_input_writer_preserves_browser_control_progress`, whose `inputs.recv()` timed out at 5 seconds. The matching test is running on a detached merge-base `dev` checkout under the shared build-slot gate. Size checks again reported x86_64 22,277,392 / 22,204,304 (+73,088) and aarch64 19,705,112 / 19,638,744 (+66,368). Full raw gate output: `$HOME/.cache/codex-scratch/terminal-links-review/fast-rerun.log`.
kayg commented 2026-10-15 12:30:45 +00:00
Author
Owner
Copy link

The exact auth socket timeout test passed on detached merge-base fb9830291d416499c531650cd48b97a3415937ec: 1 passed; 0 failed; finished in 1.26s under the managed build-slot run. The target full suite had timed out waiting for input credit after 5 seconds. I am repeating the exact test in isolation on the target head to check whether suite concurrency explains the timeout. Logs: $HOME/.cache/codex-scratch/terminal-links-review/baseline-auth-timeout.log and fast-rerun.log.

The exact auth socket timeout test passed on detached merge-base `fb9830291d416499c531650cd48b97a3415937ec`: `1 passed; 0 failed; finished in 1.26s` under the managed build-slot run. The target full suite had timed out waiting for input credit after 5 seconds. I am repeating the exact test in isolation on the target head to check whether suite concurrency explains the timeout. Logs: `$HOME/.cache/codex-scratch/terminal-links-review/baseline-auth-timeout.log` and `fast-rerun.log`.
kayg commented 2026-10-15 12:47:50 +00:00
Author
Owner
Copy link

The target-only auth socket test also passed in isolation: 1 passed; 0 failed; finished in 0.68s on 7f702c090526d8fc55333557391dbb9c6802933e. The merge-base dev run passed in 1.26s. This was the only workspace test failure in the full target run (422 passed, 1 failed, 46 ignored); it is a full-suite timeout under host load, not a repeatable branch failure. Logs: $HOME/.cache/codex-scratch/terminal-links-review/job-head-auth-timeout.log and baseline-auth-timeout.log.

The target-only auth socket test also passed in isolation: `1 passed; 0 failed; finished in 0.68s` on `7f702c090526d8fc55333557391dbb9c6802933e`. The merge-base `dev` run passed in 1.26s. This was the only workspace test failure in the full target run (422 passed, 1 failed, 46 ignored); it is a full-suite timeout under host load, not a repeatable branch failure. Logs: `$HOME/.cache/codex-scratch/terminal-links-review/job-head-auth-timeout.log` and `baseline-auth-timeout.log`.
kayg commented 2026-10-15 13:59:17 +00:00
Author
Owner
Copy link

Independent review finished. Review branch review-terminal-links is at 1695d824b2cf85d09eedbb72ddfcf49c608d8f6b; target head reviewed: 7f702c090526d8fc55333557391dbb9c6802933e (merge base fb9830291d416499c531650cd48b97a3415937ec).

Verdict: MERGE (no P0/P1). Findings: P0 0, P1 0, P2 2, P3 0. The two P2 regressions are committed as 39d9695ec074cfc9c26b24f7b1133529d4f39641 (plain mailto: URLs are no longer detected) and 1695d824b2cf85d09eedbb72ddfcf49c608d8f6b (an unseen wrapped scrollback URL can be exposed as a valid partial destination). Both failing tests and evidence are in findings.md and on this issue.

Focused parser/fuzz checks: 14 tests, 12 pass, 2 fail, 54,610 assertions. apps/tracer TypeScript check passed. Production tracer browser tests passed 2/2 in 25.3s on unique port 4246. I reviewed and attached the wrapped URL and OSC 8 tooltip screenshots:

terminal-wrapped-link-hover.png
terminal-osc8-target-hover.png

The link index loads on demand, queues visible or changed rows, and parses in an idle slice; no per-byte parsing was found by source inspection. The requested paired PTY/key-to-paint A/B and confidence intervals were not independently measured because this brief explicitly forbids using the perf VM, and no host timing was substituted. The author's timing samples therefore remain unverified.

scripts/verify fast on the job head reproduced the size-cap failures: x86_64 22,277,392 / 22,204,304 bytes (+73,088); aarch64 19,705,112 / 19,638,744 bytes (+66,368). Its test row had one full-suite auth input-credit timeout (422 passed, 1 failed, 46 ignored); the exact test passed isolated on the merge-base in 1.26s and on the target head in 0.68s. The first run's missing Cargo object path did not recur. Exact final gate table:

+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| test                      | FAIL (101)   |
| cargo-deny                | PASS         |
| static builds             | FAIL (1)     |
+---------------------------+--------------+
Independent review finished. Review branch `review-terminal-links` is at `1695d824b2cf85d09eedbb72ddfcf49c608d8f6b`; target head reviewed: `7f702c090526d8fc55333557391dbb9c6802933e` (merge base `fb9830291d416499c531650cd48b97a3415937ec`). Verdict: **MERGE (no P0/P1)**. Findings: P0 0, P1 0, P2 2, P3 0. The two P2 regressions are committed as `39d9695ec074cfc9c26b24f7b1133529d4f39641` (plain `mailto:` URLs are no longer detected) and `1695d824b2cf85d09eedbb72ddfcf49c608d8f6b` (an unseen wrapped scrollback URL can be exposed as a valid partial destination). Both failing tests and evidence are in `findings.md` and on this issue. Focused parser/fuzz checks: 14 tests, 12 pass, 2 fail, 54,610 assertions. `apps/tracer` TypeScript check passed. Production tracer browser tests passed 2/2 in 25.3s on unique port 4246. I reviewed and attached the wrapped URL and OSC 8 tooltip screenshots: ![terminal-wrapped-link-hover.png](https://git.kayg.org/attachments/afe1c63d-66e6-4325-a0f5-b45c0d895ae6) ![terminal-osc8-target-hover.png](https://git.kayg.org/attachments/ca059638-734b-43c3-865d-57c8d0e853d1) The link index loads on demand, queues visible or changed rows, and parses in an idle slice; no per-byte parsing was found by source inspection. The requested paired PTY/key-to-paint A/B and confidence intervals were not independently measured because this brief explicitly forbids using the perf VM, and no host timing was substituted. The author's timing samples therefore remain unverified. `scripts/verify fast` on the job head reproduced the size-cap failures: x86_64 22,277,392 / 22,204,304 bytes (+73,088); aarch64 19,705,112 / 19,638,744 bytes (+66,368). Its test row had one full-suite auth input-credit timeout (422 passed, 1 failed, 46 ignored); the exact test passed isolated on the merge-base in 1.26s and on the target head in 0.68s. The first run's missing Cargo object path did not recur. Exact final gate table: ```text +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | test | FAIL (101) | | cargo-deny | PASS | | static builds | FAIL (1) | +---------------------------+--------------+ ```
kayg commented 2026-10-15 14:03:52 +00:00
Author
Owner
Copy link

Started the terminal-links paired A/B on branch ab-checks, based at origin/dev 65d6421492. The worktree is clean. Next I’ll build both revisions with matching settings, then run at least 10 interleaved pairs for PTY output throughput and browser key-to-paint latency on the locked perf VM, reporting paired deltas with bootstrap 95% CIs.

Started the terminal-links paired A/B on branch `ab-checks`, based at `origin/dev` 65d64214926e4b799e53a0b442bf97a4756d8c7b. The worktree is clean. Next I’ll build both revisions with matching settings, then run at least 10 interleaved pairs for PTY output throughput and browser key-to-paint latency on the locked perf VM, reporting paired deltas with bootstrap 95% CIs.
kayg commented 2026-10-15 14:51:26 +00:00
Author
Owner
Copy link

Built the PTY/browser comparison artifacts on the build host at unchanged worktree HEAD 65d6421492. Both binaries use Rust 1.99.0, Zig 0.15.2, --release --locked, target x86_64-unknown-linux-musl, and the same cargo target directory. Both tracer bundles use Bun 1.4.2, INFRATERNAL_CLUSTERS_PREVIEW=1, and bun install --frozen-lockfile.

Side Revision Stripped binary bytes SHA-256 Tracer dist bytes
dev 65d64214926e4b799e53a0b442bf97a4756d8c7b 22,277,392 27bd848e85eb552b7c7915b157472a2be4278e6beda3017619785fcee7556745 3,521,700
terminal-links 7f702c090526d8fc55333557391dbb9c6802933e 22,277,392 16d9aaa5c17372109e74bdd74ef490bbeb16065d3c7814bb137a4a4ff3dee2c7 3,543,718

Artifacts are staged under /root/infraternal-bench/ab-checks/ on the perf VM. No timing data collected yet.

Built the PTY/browser comparison artifacts on the build host at unchanged worktree HEAD 65d64214926e4b799e53a0b442bf97a4756d8c7b. Both binaries use Rust 1.99.0, Zig 0.15.2, `--release --locked`, target `x86_64-unknown-linux-musl`, and the same cargo target directory. Both tracer bundles use Bun 1.4.2, `INFRATERNAL_CLUSTERS_PREVIEW=1`, and `bun install --frozen-lockfile`. | Side | Revision | Stripped binary bytes | SHA-256 | Tracer dist bytes | |---|---|---:|---|---:| | dev | `65d64214926e4b799e53a0b442bf97a4756d8c7b` | 22,277,392 | `27bd848e85eb552b7c7915b157472a2be4278e6beda3017619785fcee7556745` | 3,521,700 | | terminal-links | `7f702c090526d8fc55333557391dbb9c6802933e` | 22,277,392 | `16d9aaa5c17372109e74bdd74ef490bbeb16065d3c7814bb137a4a4ff3dee2c7` | 3,543,718 | Artifacts are staged under `/root/infraternal-bench/ab-checks/` on the perf VM. No timing data collected yet.
kayg commented 2026-10-15 15:01:22 +00:00
Author
Owner
Copy link

The first PTY invocation stopped before producing a sample: the copied harness expected work/target/, which was absent on the perf VM. Added that job-local directory and moved the setup-only log to throughput-setup-failure.log; no measurements from that attempt are included. The 10-pair run is restarted and waiting for the perf lock. Worktree HEAD remains 65d6421492.

The first PTY invocation stopped before producing a sample: the copied harness expected `work/target/`, which was absent on the perf VM. Added that job-local directory and moved the setup-only log to `throughput-setup-failure.log`; no measurements from that attempt are included. The 10-pair run is restarted and waiting for the perf lock. Worktree HEAD remains 65d64214926e4b799e53a0b442bf97a4756d8c7b.
kayg commented 2026-10-15 15:13:36 +00:00
Author
Owner
Copy link

PTY lane complete on the 4-vCPU perf VM under CPUWeight=10000: 10 interleaved pairs, one existing bench-agent-output.py LF run per side per pair (50 MiB), INFRATERNAL_TEST_RTT unset so the harness used its 300 ms default. Baseline and candidate medians of the per-run rates were 3.3195 and 3.261 MiB/s. The median paired candidate-minus-baseline delta was -0.104 MiB/s; paired bootstrap 95% CI [-0.163, +0.165] MiB/s. Since the interval includes zero, verdict: no measurable regression. Pair load averages and raw trial outputs: throughput-phase.log. Worktree HEAD remains 65d6421492.

PTY lane complete on the 4-vCPU perf VM under `CPUWeight=10000`: 10 interleaved pairs, one existing `bench-agent-output.py` LF run per side per pair (50 MiB), `INFRATERNAL_TEST_RTT` unset so the harness used its 300 ms default. Baseline and candidate medians of the per-run rates were 3.3195 and 3.261 MiB/s. The median paired candidate-minus-baseline delta was -0.104 MiB/s; paired bootstrap 95% CI [-0.163, +0.165] MiB/s. Since the interval includes zero, verdict: **no measurable regression**. Pair load averages and raw trial outputs: [throughput-phase.log](https://git.kayg.org/attachments/744f3012-4839-4f2f-a19f-35d1c560c236). Worktree HEAD remains 65d64214926e4b799e53a0b442bf97a4756d8c7b.
kayg commented 2026-10-15 15:24:11 +00:00
Author
Owner
Copy link

The first browser-lane invocation stopped before launch because Node rejected the unopened server log stream passed as child stdout; no timing sample was produced. The runner now passes an open file descriptor and closes it in finally. Setup-only log: browser-setup-failure.log. The 10-pair browser run is restarting; worktree HEAD remains 65d6421492.

The first browser-lane invocation stopped before launch because Node rejected the unopened server log stream passed as child stdout; no timing sample was produced. The runner now passes an open file descriptor and closes it in `finally`. Setup-only log: [browser-setup-failure.log](https://git.kayg.org/attachments/a32c3402-969e-425b-808b-6f767cde0142). The 10-pair browser run is restarting; worktree HEAD remains 65d64214926e4b799e53a0b442bf97a4756d8c7b.
kayg commented 2026-10-15 15:25:33 +00:00
Author
Owner
Copy link

The next browser setup reached the page but timed out waiting for the terminal test double because the fresh context lacked a mock owner session. The existing latency spec provisions one with POST /__mock/reset and POST /__mock/open-terminal; both calls are now part of the per-trial setup. No timing sample was produced. Setup log: browser-setup-failure-2.log. Retrying the 10-pair run at HEAD 65d6421492.

The next browser setup reached the page but timed out waiting for the terminal test double because the fresh context lacked a mock owner session. The existing latency spec provisions one with `POST /__mock/reset` and `POST /__mock/open-terminal`; both calls are now part of the per-trial setup. No timing sample was produced. Setup log: [browser-setup-failure-2.log](https://git.kayg.org/attachments/423c8d2a-2e84-4f19-81b6-670510963e0c). Retrying the 10-pair run at HEAD 65d64214926e4b799e53a0b442bf97a4756d8c7b.
kayg commented 2026-10-15 15:33:32 +00:00
Author
Owner
Copy link

Browser lane complete on the 4-vCPU perf VM under CPUWeight=10000: 10 interleaved pairs, 20 keys per trial, configured RTT 300 ms (mean measured RTT: dev 303.4 ms, terminal-links 297.7 ms). All 20 runs painted 20/20 keys with zero page errors; trial ports were unique (24400–24419). Pair load averages and raw per-key stages: browser-phase.log.

Metric (links minus dev) dev median links median Paired delta median Bootstrap 95% CI
Predictive glyph visible p50 13.60 ms 13.50 ms -0.10 ms [-0.35, +0.10] ms
Predictive glyph visible p99 14.40 ms 14.30 ms 0.00 ms [-0.60, +0.15] ms
Echo-confirmed paint p50 314.30 ms 314.10 ms 0.00 ms [-32.80, +16.50] ms
Echo-confirmed paint p99 317.40 ms 315.15 ms +0.60 ms [-24.55, +9.80] ms

Verdict: no measurable regression; all intervals include zero. One predicted-glyph interval in dev pair 9 was -5.0 ms, an impossible low-tail timestamp. I excluded only that sample from the predictive percentiles (19 valid samples for that trial); the raw value remains in the attached log. Worktree HEAD remains 65d6421492.

Browser lane complete on the 4-vCPU perf VM under `CPUWeight=10000`: 10 interleaved pairs, 20 keys per trial, configured RTT 300 ms (mean measured RTT: dev 303.4 ms, terminal-links 297.7 ms). All 20 runs painted 20/20 keys with zero page errors; trial ports were unique (24400–24419). Pair load averages and raw per-key stages: [browser-phase.log](https://git.kayg.org/attachments/03efa24c-5f82-4e7c-93ad-39c31cae60c6). | Metric (links minus dev) | dev median | links median | Paired delta median | Bootstrap 95% CI | |---|---:|---:|---:|---:| | Predictive glyph visible p50 | 13.60 ms | 13.50 ms | -0.10 ms | [-0.35, +0.10] ms | | Predictive glyph visible p99 | 14.40 ms | 14.30 ms | 0.00 ms | [-0.60, +0.15] ms | | Echo-confirmed paint p50 | 314.30 ms | 314.10 ms | 0.00 ms | [-32.80, +16.50] ms | | Echo-confirmed paint p99 | 317.40 ms | 315.15 ms | +0.60 ms | [-24.55, +9.80] ms | Verdict: **no measurable regression**; all intervals include zero. One predicted-glyph interval in dev pair 9 was -5.0 ms, an impossible low-tail timestamp. I excluded only that sample from the predictive percentiles (19 valid samples for that trial); the raw value remains in the attached log. Worktree HEAD remains 65d64214926e4b799e53a0b442bf97a4756d8c7b.
kayg commented 2026-10-15 16:25:39 +00:00
Author
Owner
Copy link

Paired A/B complete: dev vs origin/terminal-links (7f702c09), 10 interleaved pairs for each lane. PTY used the existing LF workload (50 MiB credited output); browser used the existing latency harness at configured RTT 300 ms with 20 keys per trial. Same build settings; all timings are from the 4-vCPU perf VM. Baseline/candidate columns are medians across pairs; delta is the median paired candidate-minus-dev difference. CIs are paired bootstrap percentile 95% intervals (50,000 resamples).

Metric dev median links median paired delta 95% CI Verdict
PTY throughput (MiB/s) 3.320 3.261 -0.104 [-0.163, +0.165] MiB/s no measurable regression
Predictive glyph p50 (ms) 13.60 13.50 -0.10 [-0.35, +0.10] ms no measurable regression
Predictive glyph p99 (ms) 14.40 14.30 0.00 [-0.60, +0.15] ms no measurable regression
Echo-confirmed paint p50 (ms) 314.30 314.10 0.00 [-32.80, +16.50] ms no measurable regression
Echo-confirmed paint p99 (ms) 317.40 315.15 +0.60 [-24.55, +9.80] ms no measurable regression

Every interval includes zero. One impossible -5 ms predicted-glyph sample (dev pair 9) was excluded from glyph percentile summaries and remains in the raw log; all trials painted 20/20 keys with no page errors. Pair load averages are recorded at start/end in ab-results.json. Raw PTY and browser measurements: throughput-phase.log, browser-phase.log.

HEAD: 65d64214926e4b799e53a0b442bf97a4756d8c7b; no repository files changed.

Paired A/B complete: dev vs `origin/terminal-links` (`7f702c09`), 10 interleaved pairs for each lane. PTY used the existing LF workload (50 MiB credited output); browser used the existing latency harness at configured RTT 300 ms with 20 keys per trial. Same build settings; all timings are from the 4-vCPU perf VM. Baseline/candidate columns are medians across pairs; delta is the median paired candidate-minus-dev difference. CIs are paired bootstrap percentile 95% intervals (50,000 resamples). | Metric | dev median | links median | paired delta | 95% CI | Verdict | |---|---:|---:|---:|---:|---| | PTY throughput (MiB/s) | 3.320 | 3.261 | -0.104 | [-0.163, +0.165] MiB/s | no measurable regression | | Predictive glyph p50 (ms) | 13.60 | 13.50 | -0.10 | [-0.35, +0.10] ms | no measurable regression | | Predictive glyph p99 (ms) | 14.40 | 14.30 | 0.00 | [-0.60, +0.15] ms | no measurable regression | | Echo-confirmed paint p50 (ms) | 314.30 | 314.10 | 0.00 | [-32.80, +16.50] ms | no measurable regression | | Echo-confirmed paint p99 (ms) | 317.40 | 315.15 | +0.60 | [-24.55, +9.80] ms | no measurable regression | Every interval includes zero. One impossible -5 ms predicted-glyph sample (dev pair 9) was excluded from glyph percentile summaries and remains in the raw log; all trials painted 20/20 keys with no page errors. Pair load averages are recorded at start/end in [ab-results.json](https://git.kayg.org/attachments/cc0b4c10-5758-44f3-97a7-8a7d7d2da565). Raw PTY and browser measurements: [throughput-phase.log](https://git.kayg.org/attachments/032a151b-ea46-4775-b61e-7e8a89567325), [browser-phase.log](https://git.kayg.org/attachments/dacb66a4-04bf-4ebd-8d59-96de87044009). HEAD: `65d64214926e4b799e53a0b442bf97a4756d8c7b`; no repository files changed.
kayg commented 2026-10-15 18:53:30 +00:00
Author
Owner
Copy link

Full scripts/verify fast on unchanged HEAD 65d64214926e4b799e53a0b442bf97a4756d8c7b exited 1 after 3,323 seconds. Transcript: verify-fast.log

Gate summary (verbatim):

+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| test                      | FAIL (101)   |
| cargo-deny                | PASS         |
| static builds             | FAIL (1)     |
+---------------------------+--------------+

Workspace tests: 420 passed, 3 failed, 46 ignored. Failures were browser_input_overload_retires_only_the_terminal_stream, review_mesh_watch_reports_liveness_expiry_without_another_edit, and saturated_native_control_queue_does_not_hide_primary_terminal_output. The first two are auth socket timing tests; the third is remote control queue delivery.

Static binary caps: x86_64 was 22,277,392 bytes against 22,204,304 (+73,088); aarch64 was 19,705,112 bytes against 19,638,744 (+66,368). The isolated test rerun is in progress and will be added as a follow-up.

Full `scripts/verify fast` on unchanged HEAD `65d64214926e4b799e53a0b442bf97a4756d8c7b` exited 1 after 3,323 seconds. Transcript: [verify-fast.log](https://git.kayg.org/attachments/aa1200d4-d87a-411f-853e-2363611c05d2) Gate summary (verbatim): ``` +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | test | FAIL (101) | | cargo-deny | PASS | | static builds | FAIL (1) | +---------------------------+--------------+ ``` Workspace tests: 420 passed, 3 failed, 46 ignored. Failures were `browser_input_overload_retires_only_the_terminal_stream`, `review_mesh_watch_reports_liveness_expiry_without_another_edit`, and `saturated_native_control_queue_does_not_hide_primary_terminal_output`. The first two are auth socket timing tests; the third is remote control queue delivery. Static binary caps: x86_64 was 22,277,392 bytes against 22,204,304 (+73,088); aarch64 was 19,705,112 bytes against 19,638,744 (+66,368). The isolated test rerun is in progress and will be added as a follow-up.
kayg commented 2026-10-15 19:09:44 +00:00
Author
Owner
Copy link

Finished the #214 A/B measurements. No source changes or commits; HEAD 65d64214926e4b799e53a0b442bf97a4756d8c7b equals the merge base with origin/dev.

The PTY throughput and browser key-to-paint verdicts are all no measurable regression. The three workspace test failures each passed when rerun individually (1 passed, 0 failed); logs: credit test, liveness test, remote control queue test. The #214 result table and per-pair load averages are in the earlier measurement comment and attached raw results.

The full gate still exits 1 because of those three workspace failures and both static-size caps. Isolated reruns passing does not change the full-gate result. x86_64 stripped static size is 22,277,392 bytes on HEAD and merge base (delta 0 bytes); it exceeds the 22,204,304-byte cap by 73,088 bytes. aarch64 is 19,705,112 bytes, 66,368 above its 19,638,744-byte cap.

Final scripts/verify fast summary (verbatim):

+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| test                      | FAIL (101)   |
| cargo-deny                | PASS         |
| static builds             | FAIL (1)     |
+---------------------------+--------------+

Transcript: verify-fast.log

Finished the #214 A/B measurements. No source changes or commits; HEAD `65d64214926e4b799e53a0b442bf97a4756d8c7b` equals the merge base with `origin/dev`. The PTY throughput and browser key-to-paint verdicts are all **no measurable regression**. The three workspace test failures each passed when rerun individually (1 passed, 0 failed); logs: [credit test](https://git.kayg.org/attachments/84462348-4c7b-49e7-bd47-1e74dd808eae), [liveness test](https://git.kayg.org/attachments/9f188d08-a23c-43b9-96df-66f86e5c4e84), [remote control queue test](https://git.kayg.org/attachments/55570f74-f52e-4f36-9d8d-3513af92bd58). The #214 result table and per-pair load averages are in the earlier measurement comment and attached raw results. The full gate still exits 1 because of those three workspace failures and both static-size caps. Isolated reruns passing does not change the full-gate result. x86_64 stripped static size is 22,277,392 bytes on HEAD and merge base (delta 0 bytes); it exceeds the 22,204,304-byte cap by 73,088 bytes. aarch64 is 19,705,112 bytes, 66,368 above its 19,638,744-byte cap. Final `scripts/verify fast` summary (verbatim): ``` +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | test | FAIL (101) | | cargo-deny | PASS | | static builds | FAIL (1) | +---------------------------+--------------+ ``` Transcript: [verify-fast.log](https://git.kayg.org/attachments/aa1200d4-d87a-411f-853e-2363611c05d2)
kayg commented 2026-10-15 19:13:56 +00:00
Author
Owner
Copy link

Starting continuation on branch terminal-links, HEAD 7f702c090526d8fc55333557391dbb9c6802933e; base is origin/dev at 40cc9c5be9b863ab23a92dfeca9e46cc726737aa. I read the 76 comments on #214 and the one comment on parent #5. Review commits 39d9695e and 1695d824 add failing tests for plain mailto: detection and partial URLs from unknown wrapped scrollback; I will rebase, apply both regressions, and fix them before the production demo.

Starting continuation on branch `terminal-links`, HEAD `7f702c090526d8fc55333557391dbb9c6802933e`; base is `origin/dev` at `40cc9c5be9b863ab23a92dfeca9e46cc726737aa`. I read the 76 comments on #214 and the one comment on parent #5. Review commits `39d9695e` and `1695d824` add failing tests for plain `mailto:` detection and partial URLs from unknown wrapped scrollback; I will rebase, apply both regressions, and fix them before the production demo.
kayg commented 2026-10-15 19:22:59 +00:00
Author
Owner
Copy link

The mailto regression test is queued through build-slot, but all four shared build slots and the reserved slot are occupied. It has not started and has produced no test output; waiting for a slot before changing the implementation.

The mailto regression test is queued through build-slot, but all four shared build slots and the reserved slot are occupied. It has not started and has produced no test output; waiting for a slot before changing the implementation.
kayg commented 2026-10-15 19:33:40 +00:00
Author
Owner
Copy link

Reproduced review finding F1 before the fix: focused link tests were 11 pass / 1 fail (27 assertions); detectTerminalLinks returned [] for mailto:person@example.test. The failure is at apps/tracer/src/terminal-links.on-demand.test.ts:82. Raw receipt: mailto-before.log.

Reproduced review finding F1 before the fix: focused link tests were 11 pass / 1 fail (27 assertions); `detectTerminalLinks` returned `[]` for `mailto:person@example.test`. The failure is at `apps/tracer/src/terminal-links.on-demand.test.ts:82`. Raw receipt: [mailto-before.log](https://git.kayg.org/attachments/506df473-5efb-408d-99be-b45531202abd).
kayg commented 2026-10-15 19:43:20 +00:00
Author
Owner
Copy link

The mailto fix passes the focused link suite (12 pass, 0 fail, 27 assertions). apps/tracer bun run check is queued through build-slot and has not started because all four standard slots remain occupied. I have not committed the fix before its TypeScript gate.

The mailto fix passes the focused link suite (12 pass, 0 fail, 27 assertions). `apps/tracer` `bun run check` is queued through build-slot and has not started because all four standard slots remain occupied. I have not committed the fix before its TypeScript gate.
kayg commented 2026-10-15 20:08:50 +00:00
Author
Owner
Copy link

F1 fixed and pushed as 4640436a3b8b32f1204f323b7fb01344b47cdf78. The scanner now detects plain mailto: links; the focused suite passes 12/12 (27 assertions), and apps/tracer bun run check passed with output $ bun x tsc --project tsconfig.json. Receipts: mailto-after.log, mailto-check.log.

F1 fixed and pushed as `4640436a3b8b32f1204f323b7fb01344b47cdf78`. The scanner now detects plain `mailto:` links; the focused suite passes 12/12 (27 assertions), and `apps/tracer` `bun run check` passed with output `$ bun x tsc --project tsconfig.json`. Receipts: [mailto-after.log](https://git.kayg.org/attachments/5e595d66-ecb0-4d1b-a0a9-07d24ca5455a), [mailto-check.log](https://git.kayg.org/attachments/2cb30b33-2960-4f37-b830-17dcdbb08c02).
kayg commented 2026-10-15 20:09:03 +00:00
Author
Owner
Copy link

Reproduced review finding F2 before the fix: the focused suite is 12 pass / 1 fail (28 assertions). For a previously unseen wrapped scrollback viewport, the controller offers https://scroll.example.t instead of the full https://scroll.example.test/a-long-path. Failure: apps/tracer/src/terminal-links.on-demand.test.ts:178. Raw receipt: scrollback-before.log.

Reproduced review finding F2 before the fix: the focused suite is 12 pass / 1 fail (28 assertions). For a previously unseen wrapped scrollback viewport, the controller offers `https://scroll.example.t` instead of the full `https://scroll.example.test/a-long-path`. Failure: `apps/tracer/src/terminal-links.on-demand.test.ts:178`. Raw receipt: [scrollback-before.log](https://git.kayg.org/attachments/21847c35-a1a5-4f57-9712-8720a4422305).
kayg commented 2026-10-15 20:09:21 +00:00
Author
Owner
Copy link

F2 fix is in the working tree: uncached scrollback rows use the buffer row's wrap flag for the current scan, while matching cached fingerprints still preserve their earlier boundary. The link and fuzz suites pass 14/14 with 54,610 assertions; the regression now yields the full https://scroll.example.test/a-long-path. Receipt: scrollback-after.log. The tracer TypeScript check is next.

F2 fix is in the working tree: uncached scrollback rows use the buffer row's wrap flag for the current scan, while matching cached fingerprints still preserve their earlier boundary. The link and fuzz suites pass 14/14 with 54,610 assertions; the regression now yields the full `https://scroll.example.test/a-long-path`. Receipt: [scrollback-after.log](https://git.kayg.org/attachments/662549b3-3980-4256-8b03-c312af9eaab2). The tracer TypeScript check is next.
kayg commented 2026-10-15 20:28:11 +00:00
Author
Owner
Copy link

F2 focused tests and fuzz are green (14/14, 54,610 assertions). The required apps/tracer bun run check remains queued: all four standard build slots and the reserved priority slot are occupied, so the command has not started. I will commit after the TypeScript gate completes.

F2 focused tests and fuzz are green (14/14, 54,610 assertions). The required `apps/tracer` `bun run check` remains queued: all four standard build slots and the reserved priority slot are occupied, so the command has not started. I will commit after the TypeScript gate completes.
kayg commented 2026-10-15 21:04:22 +00:00
Author
Owner
Copy link

F2 fixed and pushed as bf7bb64412492ad85fe5800c93dae3d208626f23. Uncached scrollback now preserves the terminal buffer's wrap flag for the current scan; matching cached fingerprints still preserve prior wrap history. Parser/fuzz tests: 14 passed, 0 failed, 54,610 assertions. apps/tracer bun run check passed with output $ bun x tsc --project tsconfig.json. Receipts: scrollback-after.log, scrollback-check.log.

F2 fixed and pushed as `bf7bb64412492ad85fe5800c93dae3d208626f23`. Uncached scrollback now preserves the terminal buffer's wrap flag for the current scan; matching cached fingerprints still preserve prior wrap history. Parser/fuzz tests: 14 passed, 0 failed, 54,610 assertions. `apps/tracer` `bun run check` passed with output `$ bun x tsc --project tsconfig.json`. Receipts: [scrollback-after.log](https://git.kayg.org/attachments/662549b3-3980-4256-8b03-c312af9eaab2), [scrollback-check.log](https://git.kayg.org/attachments/0717c9da-cb3c-46f1-9e2c-3c89a9c4cf01).
kayg commented 2026-10-15 21:04:48 +00:00
Author
Owner
Copy link

Rebased the six terminal-link feature/fix commits onto current origin/dev at 40cc9c5be9b863ab23a92dfeca9e46cc726737aa; no conflicts. Rewritten branch head cd2638358 was pushed to terminal-links with --force-with-lease.

Rebased the six terminal-link feature/fix commits onto current `origin/dev` at `40cc9c5be9b863ab23a92dfeca9e46cc726737aa`; no conflicts. Rewritten branch head `cd2638358` was pushed to `terminal-links` with `--force-with-lease`.
kayg commented 2026-10-15 21:18:41 +00:00
Author
Owner
Copy link

Demo recording spec added at apps/tracer/real/terminal-links-demo.spec.ts (uncommitted). bun run check and strict standalone TypeScript check both passed at cd2638358835f3ce4b3093ff378d91109e8681ed; demo-typecheck.log. The real-browser run is queued through build-slot; all five slots are occupied by other jobs.

Demo recording spec added at `apps/tracer/real/terminal-links-demo.spec.ts` (uncommitted). `bun run check` and strict standalone TypeScript check both passed at `cd2638358835f3ce4b3093ff378d91109e8681ed`; [demo-typecheck.log](https://git.kayg.org/attachments/57a8db03-d70f-4f7d-a7db-4a4083db8ce4). The real-browser run is queued through build-slot; all five slots are occupied by other jobs.
kayg commented 2026-10-15 21:27:41 +00:00
Author
Owner
Copy link

The queued terminal-links Playwright run has waited over 12 minutes for a build slot and has not started. All five managed slots remain occupied. I am waiting for a slot release; no other job was changed or stopped.

The queued terminal-links Playwright run has waited over 12 minutes for a build slot and has not started. All five managed slots remain occupied. I am waiting for a slot release; no other job was changed or stopped.
kayg commented 2026-10-15 21:29:14 +00:00
Author
Owner
Copy link

Real-server fixture path check: INFRATERAL_TEST_STATE_ROOT=/home/kayg/.cache/codex-scratch/terminal-links/runtime would produce a 121-byte Unix session socket path, exceeding the fixture's 107-byte limit in apps/tracer/real/server.ts. Using the dedicated scratch root /home/kayg/.cache/codex-scratch/tl produces a 101-byte path. No source change is needed.

Real-server fixture path check: `INFRATERAL_TEST_STATE_ROOT=/home/kayg/.cache/codex-scratch/terminal-links/runtime` would produce a 121-byte Unix session socket path, exceeding the fixture's 107-byte limit in `apps/tracer/real/server.ts`. Using the dedicated scratch root `/home/kayg/.cache/codex-scratch/tl` produces a 101-byte path. No source change is needed.
kayg commented 2026-10-15 21:29:18 +00:00
Author
Owner
Copy link

Correction to the previous comment: the variable name is INFRATERNAL_TEST_STATE_ROOT (the prior comment omitted an N). The 121-byte and 101-byte path measurements are unchanged.

Correction to the previous comment: the variable name is `INFRATERNAL_TEST_STATE_ROOT` (the prior comment omitted an `N`). The 121-byte and 101-byte path measurements are unchanged.
kayg commented 2026-10-15 21:37:23 +00:00
Author
Owner
Copy link

The required link browser run exposed a harness setup failure: both specs timed out after 30 s in openTerminal() waiting for window.__tracerProductTestDouble.isConnected(). No link assertions ran (0/2); the test double served at port 4295 behind the proxy on 4294. Receipts: run log, wrapped/OSC 8 context, touch context.

The required link browser run exposed a harness setup failure: both specs timed out after 30 s in `openTerminal()` waiting for `window.__tracerProductTestDouble.isConnected()`. No link assertions ran (0/2); the test double served at port 4295 behind the proxy on 4294. Receipts: [run log](https://git.kayg.org/attachments/77d2c989-8012-47f1-a054-0c48c490102d), [wrapped/OSC 8 context](https://git.kayg.org/attachments/14508cf3-e387-40b0-98fa-e3afb279df4b), [touch context](https://git.kayg.org/attachments/54bcfbec-76a8-472f-8a0d-66f54b3f2042).
kayg commented 2026-10-15 21:39:03 +00:00
Author
Owner
Copy link

Root cause of the browser setup failure: apps/tracer/test/mock-server.ts hard-coded apps/tracer/dist, while bun run build wrote to the configured external TRACER_DIST; the double served the stale checkout-local build and never loaded __tracerProductTestDouble. The server now honors TRACER_DIST, with its existing local path as the default. I’m rerunning the focused browser suite.

Root cause of the browser setup failure: `apps/tracer/test/mock-server.ts` hard-coded `apps/tracer/dist`, while `bun run build` wrote to the configured external `TRACER_DIST`; the double served the stale checkout-local build and never loaded `__tracerProductTestDouble`. The server now honors `TRACER_DIST`, with its existing local path as the default. I’m rerunning the focused browser suite.
kayg commented 2026-10-15 21:47:08 +00:00
Author
Owner
Copy link

After the test-double dist-root fix, the focused bun run check && bun x playwright test test/terminal-links.spec.ts rerun has waited about eight minutes for build-slot. All five slots remain occupied; there is no rerun output yet. No further source changes were made during this wait.

After the test-double dist-root fix, the focused `bun run check && bun x playwright test test/terminal-links.spec.ts` rerun has waited about eight minutes for build-slot. All five slots remain occupied; there is no rerun output yet. No further source changes were made during this wait.
kayg commented 2026-10-15 22:01:58 +00:00
Author
Owner
Copy link

At 2026-10-16 00:02 CEST, the corrected focused browser rerun had waited about 20 minutes for a standard build slot and still had no output. All four standard slots are occupied; the command remains queued.

At 2026-10-16 00:02 CEST, the corrected focused browser rerun had waited about 20 minutes for a standard build slot and still had no output. All four standard slots are occupied; the command remains queued.
kayg commented 2026-10-15 22:10:41 +00:00
Author
Owner
Copy link

At 2026-10-16 00:10 CEST, the focused rerun remained queued for about 30 minutes with no output. All four standard build slots were occupied. The corrected command remains pending; no run was made outside build-slot.

At 2026-10-16 00:10 CEST, the focused rerun remained queued for about 30 minutes with no output. All four standard build slots were occupied. The corrected command remains pending; no run was made outside build-slot.
kayg commented 2026-10-15 22:21:05 +00:00
Author
Owner
Copy link

At 2026-10-16 00:21 CEST, the corrected focused rerun remained queued for roughly 43 minutes with no output. New managed jobs continue to take released standard slots; no slot limits were changed and no build/test ran outside the wrapper.

At 2026-10-16 00:21 CEST, the corrected focused rerun remained queued for roughly 43 minutes with no output. New managed jobs continue to take released standard slots; no slot limits were changed and no build/test ran outside the wrapper.
kayg commented 2026-10-15 22:35:22 +00:00
Author
Owner
Copy link

At 2026-10-16 00:35 CEST, the corrected focused rerun remained queued for about 55 minutes with no output. All four standard slots remain occupied. No code changes were made while waiting.

At 2026-10-16 00:35 CEST, the corrected focused rerun remained queued for about 55 minutes with no output. All four standard slots remain occupied. No code changes were made while waiting.
kayg commented 2026-10-15 22:44:02 +00:00
Author
Owner
Copy link

At 2026-10-16 00:45 CEST, the corrected focused rerun had waited about 65 minutes with no output. Standard slots remain occupied, and a recently released slot was acquired by another managed job. The command remains queued; I did not alter the slot pool or other jobs.

At 2026-10-16 00:45 CEST, the corrected focused rerun had waited about 65 minutes with no output. Standard slots remain occupied, and a recently released slot was acquired by another managed job. The command remains queued; I did not alter the slot pool or other jobs.
kayg commented 2026-10-15 23:04:41 +00:00
Author
Owner
Copy link

Committed and pushed e293cf0cf for the test-server dist path fix. bun run check passed; browser terminal-link suite passed 2/2 (wrapped link/OSC 8/palette, touch) on port 4297. Receipt: terminal-links-playwright-rerun.log.

Committed and pushed `e293cf0cf` for the test-server dist path fix. `bun run check` passed; browser terminal-link suite passed 2/2 (wrapped link/OSC 8/palette, touch) on port 4297. Receipt: [terminal-links-playwright-rerun.log](https://git.kayg.org/attachments/be6e6e30-c6e9-419c-b3b8-42eadf376191).
kayg commented 2026-10-15 23:24:30 +00:00
Author
Owner
Copy link

Real-demo build outcome: release x86_64 musl server build, package check, strict demo-spec typecheck and production tracer build passed. The real-server fixture then timed out waiting for /healthz after 20 s (server startup failed); receipt: real-demo-build.log. I found no remaining process or listener for the fixture. I’m enabling its server log on the next uniquely ported run to identify the startup error.

Real-demo build outcome: release x86_64 musl server build, package check, strict demo-spec typecheck and production tracer build passed. The real-server fixture then timed out waiting for `/healthz` after 20 s (`server startup failed`); receipt: [real-demo-build.log](https://git.kayg.org/attachments/26ba58f6-32d5-43ba-8fbf-42223bde20dd). I found no remaining process or listener for the fixture. I’m enabling its server log on the next uniquely ported run to identify the startup error.
kayg commented 2026-10-15 23:33:33 +00:00
Author
Owner
Copy link

At 2026-10-16 01:33 CEST, the instrumented real-server demo retry had waited about 10 minutes for a standard slot with no output. All four standard slots remain occupied; the server log capture is ready and the command remains queued.

At 2026-10-16 01:33 CEST, the instrumented real-server demo retry had waited about 10 minutes for a standard slot with no output. All four standard slots remain occupied; the server log capture is ready and the command remains queued.
kayg commented 2026-10-15 23:35:43 +00:00
Author
Owner
Copy link

The server log identified the startup failure: rp_id is not an effective_domain of rp_origin. Binding the HTTPS fixture on 10.69.69.125 while also using that IP as its WebAuthn origin is invalid. I’ll keep the server bound to the requested interface, use a DNS RP hostname, and map that hostname to the interface inside the demo browser. Receipt: real-demo-rerun.log.

The server log identified the startup failure: `rp_id is not an effective_domain of rp_origin`. Binding the HTTPS fixture on `10.69.69.125` while also using that IP as its WebAuthn origin is invalid. I’ll keep the server bound to the requested interface, use a DNS RP hostname, and map that hostname to the interface inside the demo browser. Receipt: [real-demo-rerun.log](https://git.kayg.org/attachments/0238c943-896b-4028-9ea9-9c1ac06a5525).
kayg commented 2026-10-15 23:45:40 +00:00
Author
Owner
Copy link

The production real-PTY demo passed 1/1 on port 4397; bun run check and strict demo-spec typecheck passed. The server bound to 10.69.69.125:4397 and used terminal-links.test as its RP origin via the browser-local host map. The video is 42.6 s (1,430,558 bytes): demo.webm. Run receipt: real-demo-hostmap.log. It shows the normal and wrapped URL control-clicks, OSC 8 target tooltip, display-only file URL, mailto and palette entry. Ready for the owner’s visual review.

The production real-PTY demo passed 1/1 on port 4397; `bun run check` and strict demo-spec typecheck passed. The server bound to `10.69.69.125:4397` and used `terminal-links.test` as its RP origin via the browser-local host map. The video is 42.6 s (1,430,558 bytes): [demo.webm](https://git.kayg.org/attachments/bf6dcb92-23b0-43c8-9018-d263b0eda97e). Run receipt: [real-demo-hostmap.log](https://git.kayg.org/attachments/06e75eb4-adfa-4e8d-b07a-bee1a3e6daf8). It shows the normal and wrapped URL control-clicks, OSC 8 target tooltip, display-only file URL, mailto and palette entry. Ready for the owner’s visual review.
kayg commented 2026-10-15 23:46:02 +00:00
Author
Owner
Copy link

Committed and pushed 983299bd6 for the opt-in demo host mapping. The package check, strict demo-spec typecheck and real-PTY demo passed; video duration 42.6 s. The browser maps terminal-links.test to 10.69.69.125 while the fixture binds there. Receipt and video are attached above.

Committed and pushed `983299bd6` for the opt-in demo host mapping. The package check, strict demo-spec typecheck and real-PTY demo passed; video duration 42.6 s. The browser maps `terminal-links.test` to `10.69.69.125` while the fixture binds there. Receipt and video are attached above.
kayg commented 2026-10-15 23:46:17 +00:00
Author
Owner
Copy link

Committed and pushed 65cec91fe, adding the opt-in real-PTY demo. bun run check, strict demo-spec typecheck, and the real browser test passed (1/1; 43.2 s test, 42.6 s video). The attached video and raw receipt are linked above. Remaining: scripts/verify iterate, final scripts/verify fast, static size comparison against the merge base, then hand off for owner review.

Committed and pushed `65cec91fe`, adding the opt-in real-PTY demo. `bun run check`, strict demo-spec typecheck, and the real browser test passed (1/1; 43.2 s test, 42.6 s video). The attached video and raw receipt are linked above. Remaining: `scripts/verify iterate`, final `scripts/verify fast`, static size comparison against the merge base, then hand off for owner review.
kayg commented 2026-10-16 00:09:39 +00:00
Author
Owner
Copy link

Finding: first scripts/verify iterate run failed in unrelated real_binary_preserves_input_and_output_across_24_abrupt_stops (crates/infraternal/tests/agent_daemon.rs). Iterations 01–21 replayed exactly; iteration 22 stalled at 316,927/468,224 output bytes after 4,193,649 input bytes were sent and acknowledged, then the 1-second socket read timed out. The focused feature suites and preceding Rust tests passed. Full scripts/verify fast is still running; I will report its result separately.
Evidence: $HOME/.cache/codex-scratch/terminal-links/evidence/verify-iterate.log.

Finding: first `scripts/verify iterate` run failed in unrelated `real_binary_preserves_input_and_output_across_24_abrupt_stops` (`crates/infraternal/tests/agent_daemon.rs`). Iterations 01–21 replayed exactly; iteration 22 stalled at 316,927/468,224 output bytes after 4,193,649 input bytes were sent and acknowledged, then the 1-second socket read timed out. The focused feature suites and preceding Rust tests passed. Full `scripts/verify fast` is still running; I will report its result separately. Evidence: `$HOME/.cache/codex-scratch/terminal-links/evidence/verify-iterate.log`.
kayg commented 2026-10-16 00:23:34 +00:00
Author
Owner
Copy link

Finding: scripts/verify fast passed fmt, VT native fmt, CLI compile-out, clippy, workspace tests, and cargo-deny, but its static-build size row failed. Stripped x86_64 was 22,329,104 bytes against 22,204,304 (+124,800); aarch64 was 19,752,920 against 19,638,744 (+114,176). The diff from merge base 40cc9c5be9 contains 12 TypeScript files and no Rust/Cargo source or manifest changes. I am measuring the merge-base binary with the same recipe before assigning the size delta. iterate had one separate unrelated restart-oracle timeout recorded earlier.

Gate logs: iterate, fast.

Finding: `scripts/verify fast` passed fmt, VT native fmt, CLI compile-out, clippy, workspace tests, and cargo-deny, but its static-build size row failed. Stripped x86_64 was 22,329,104 bytes against 22,204,304 (+124,800); aarch64 was 19,752,920 against 19,638,744 (+114,176). The diff from merge base 40cc9c5be9b863ab23a92dfeca9e46cc726737aa contains 12 TypeScript files and no Rust/Cargo source or manifest changes. I am measuring the merge-base binary with the same recipe before assigning the size delta. `iterate` had one separate unrelated restart-oracle timeout recorded earlier. Gate logs: [iterate](https://git.kayg.org/attachments/ef51c4cd-82ea-4971-bfd9-5449c603182b), [fast](https://git.kayg.org/attachments/1317954a-3edd-4b37-bb11-bd506dea4838).
kayg commented 2026-10-16 00:36:34 +00:00
Author
Owner
Copy link

Terminal links work is ready for owner review.

Head: 65cec91fe5de9ad46483546f439782f8e213a2c0 (test(tracer): add real terminal link demo (#214)), pushed to terminal-links. The requested base at gate time was 40cc9c5be9b863ab23a92dfeca9e46cc726737aa; origin/dev has since advanced to 492e1505ab6822078268344382f4d246d2c3d65e (#225).

The real production demo is 42.6 seconds / 1,430,558 bytes and is attached: demo.webm. It covers normal and wrapped links, OSC 8 target hover/open, file URL display-only, mailto and the palette. Production browser run: 1 passed; mock browser suite: 2 passed. bun run check and the strict demo spec TypeScript check passed.

Gate outputs:

scripts/verify iterate (exit 1; only failing row was the unrelated restart oracle at iteration 22; detailed counters and timeout are in the attached log):

+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| test                      | FAIL (101)   |
| cargo-deny                | PASS         |
+---------------------------+--------------+

scripts/verify fast (exit 1):

PT_INTERP: absent; size: 22329104 bytes; cap: 22204304 bytes; delta: +124800 bytes
binary exceeds the 22204304-byte size budget: 22329104 bytes
PT_INTERP: absent; size: 19752920 bytes; cap: 19638744 bytes; delta: +114176 bytes
binary exceeds the 19638744-byte size budget: 19752920 bytes
+---------------------------+--------------+
| Rust gate                 | Result       |
+---------------------------+--------------+
| fmt                       | PASS         |
| VT native fmt             | PASS         |
| CLI plugin compile-out    | PASS         |
| clippy                    | PASS         |
| test                      | PASS         |
| cargo-deny                | PASS         |
| static builds             | FAIL (1)     |
+---------------------------+--------------+

D243 comparison used the same locked x86_64 static release recipe and one target directory: merge base 40cc9c5be9b863ab23a92dfeca9e46cc726737aa = 22,329,104 bytes; branch = 22,329,104 bytes; delta = 0 bytes. Both are 124,800 bytes above the configured x86_64 ceiling. This lane adds no Rust/Cargo changes, so the fast static-size failure is pre-existing. The fast Rust test row passed; the static-build failure stops the dispatcher before its TypeScript row. Evidence: iterate log, fast log, baseline build log, size comparison.

The earlier perf-VM A/B recorded no measurable regression: PTY throughput 3.320 → 3.261 MiB/s (paired median delta -0.104; 95% CI [-0.163, +0.165]); glyph paint p50 13.60 → 13.50 ms and p99 14.40 → 14.30 ms. Transfer deltas and detailed paint intervals are in the earlier issue evidence.

Issue remains open. No merge or deploy.

Terminal links work is ready for owner review. Head: `65cec91fe5de9ad46483546f439782f8e213a2c0` (`test(tracer): add real terminal link demo (#214)`), pushed to `terminal-links`. The requested base at gate time was `40cc9c5be9b863ab23a92dfeca9e46cc726737aa`; `origin/dev` has since advanced to `492e1505ab6822078268344382f4d246d2c3d65e` (#225). The real production demo is 42.6 seconds / 1,430,558 bytes and is attached: [demo.webm](https://git.kayg.org/attachments/bf6dcb92-23b0-43c8-9018-d263b0eda97e). It covers normal and wrapped links, OSC 8 target hover/open, file URL display-only, mailto and the palette. Production browser run: 1 passed; mock browser suite: 2 passed. `bun run check` and the strict demo spec TypeScript check passed. Gate outputs: ` scripts/verify iterate ` (exit 1; only failing row was the unrelated restart oracle at iteration 22; detailed counters and timeout are in the attached log): ``` +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | test | FAIL (101) | | cargo-deny | PASS | +---------------------------+--------------+ ``` ` scripts/verify fast ` (exit 1): ``` PT_INTERP: absent; size: 22329104 bytes; cap: 22204304 bytes; delta: +124800 bytes binary exceeds the 22204304-byte size budget: 22329104 bytes PT_INTERP: absent; size: 19752920 bytes; cap: 19638744 bytes; delta: +114176 bytes binary exceeds the 19638744-byte size budget: 19752920 bytes +---------------------------+--------------+ | Rust gate | Result | +---------------------------+--------------+ | fmt | PASS | | VT native fmt | PASS | | CLI plugin compile-out | PASS | | clippy | PASS | | test | PASS | | cargo-deny | PASS | | static builds | FAIL (1) | +---------------------------+--------------+ ``` D243 comparison used the same locked x86_64 static release recipe and one target directory: merge base `40cc9c5be9b863ab23a92dfeca9e46cc726737aa` = 22,329,104 bytes; branch = 22,329,104 bytes; delta = 0 bytes. Both are 124,800 bytes above the configured x86_64 ceiling. This lane adds no Rust/Cargo changes, so the fast static-size failure is pre-existing. The fast Rust test row passed; the static-build failure stops the dispatcher before its TypeScript row. Evidence: [iterate log](https://git.kayg.org/attachments/ef51c4cd-82ea-4971-bfd9-5449c603182b), [fast log](https://git.kayg.org/attachments/1317954a-3edd-4b37-bb11-bd506dea4838), [baseline build log](https://git.kayg.org/attachments/482cd5cf-bb7e-42ee-9bc4-261ba5d47032), [size comparison](https://git.kayg.org/attachments/0ad2f8b5-c3bb-4e9c-89a5-e71f4d71bfd4). The earlier perf-VM A/B recorded no measurable regression: PTY throughput 3.320 → 3.261 MiB/s (paired median delta -0.104; 95% CI [-0.163, +0.165]); glyph paint p50 13.60 → 13.50 ms and p99 14.40 → 14.30 ms. Transfer deltas and detailed paint intervals are in the earlier issue evidence. Issue remains open. No merge or deploy.
kayg commented 2026-10-16 04:24:16 +00:00
Author
Owner
Copy link

Starting continuation on terminal-links at 65cec91fe5de9ad46483546f439782f8e213a2c0; fetched origin/dev at 9e84426c48f3818cebaa99acede14fec1d24ab4e. I’m rebasing, preserving dev behavior on conflicts, then will run the mock browser suite and scripts/verify fast through build-slot and push the updated branch with --force-with-lease.

Starting continuation on `terminal-links` at `65cec91fe5de9ad46483546f439782f8e213a2c0`; fetched `origin/dev` at `9e84426c48f3818cebaa99acede14fec1d24ab4e`. I’m rebasing, preserving `dev` behavior on conflicts, then will run the mock browser suite and `scripts/verify fast` through build-slot and push the updated branch with `--force-with-lease`.
kayg commented 2026-10-16 04:24:32 +00:00
Author
Owner
Copy link

Rebased the nine terminal-links commits onto origin/dev at 9e84426c48f3818cebaa99acede14fec1d24ab4e; no conflicts. New head: 44435146697b29bc081e6adb66ffe208be47fe6a. The 12-file feature diff is unchanged; next are the focused mock browser suite and the full fast gate through build-slot.

Rebased the nine `terminal-links` commits onto `origin/dev` at `9e84426c48f3818cebaa99acede14fec1d24ab4e`; no conflicts. New head: `44435146697b29bc081e6adb66ffe208be47fe6a`. The 12-file feature diff is unchanged; next are the focused mock browser suite and the full fast gate through build-slot.
kayg commented 2026-10-16 04:31:37 +00:00
Author
Owner
Copy link

The focused mock browser rerun (TRACER_TEST_PORT=4314, zero simulated RTT) remains queued through build-slot after about five minutes. It has produced no test output; when checked, all four standard slots and the reserved slot were occupied. No command has run outside the scheduler.

The focused mock browser rerun (`TRACER_TEST_PORT=4314`, zero simulated RTT) remains queued through build-slot after about five minutes. It has produced no test output; when checked, all four standard slots and the reserved slot were occupied. No command has run outside the scheduler.
kayg commented 2026-10-16 04:32:42 +00:00
Author
Owner
Copy link

Rebased-head mock browser suite passed at 44435146697b29bc081e6adb66ffe208be47fe6a: TRACER_TEST_PORT=4314 INFRATERNAL_TEST_RTT=0 bun x playwright test test/terminal-links.spec.ts ran 2/2 tests in 12.7 s. Wrapped URL/OSC 8/palette and touch tap both passed. Receipt: mock-browser.log. Next: scripts/verify fast through build-slot.

Rebased-head mock browser suite passed at `44435146697b29bc081e6adb66ffe208be47fe6a`: `TRACER_TEST_PORT=4314 INFRATERNAL_TEST_RTT=0 bun x playwright test test/terminal-links.spec.ts` ran 2/2 tests in 12.7 s. Wrapped URL/OSC 8/palette and touch tap both passed. Receipt: [mock-browser.log](https://git.kayg.org/attachments/efd36224-47c8-4eaf-b6e2-9b75b2e333ea). Next: `scripts/verify fast` through build-slot.
kayg commented 2026-10-16 05:32:19 +00:00
Author
Owner
Copy link

Orchestrator: merged to dev as f210a4f25 (owner approved the demo). Deployed to production web-only at ~11:30 IST (no server restart, MainPID 20425 unchanged; previous web kept in web.prev). Trigger: owner's video of agy's wrapped sign-in link on production opening only one row.

Orchestrator: merged to dev as f210a4f25 (owner approved the demo). Deployed to production web-only at ~11:30 IST (no server restart, MainPID 20425 unchanged; previous web kept in web.prev). Trigger: owner's video of agy's wrapped sign-in link on production opening only one row.
kayg commented 2026-10-28 17:53:18 +00:00
Author
Owner
Copy link

Closed by the orchestrator during the 2026-10-28 handover (#254), on the owner's request to close finished issues. Evidence: Terminal links merged as f210a4f25 (ancestor of origin/dev and release-prod-live), deployed web-only and owner-approved; underline/CJK follow-ups tracked in #232. Reopen if anything is still missing.

Closed by the orchestrator during the 2026-10-28 handover (#254), on the owner's request to close finished issues. Evidence: Terminal links merged as f210a4f25 (ancestor of origin/dev and release-prod-live), deployed web-only and owner-approved; underline/CJK follow-ups tracked in #232. Reopen if anything is still missing.
kayg closed this issue 2026-10-28 17:53:18 +00:00
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
j470b-server-mesh-rejoin
mesh-compat-382
j444b-d381-key-wins-flood
j461b-macos-utun-reader-thread
j475b-remove-host-motion
fix-reconnect-new-link
j481-d380
dev
j479-mesh-soak
fix/j476-host-facts
j475-motion-pill-split
j470-mac-node-join
root-upgrade-d376
j474-review-merge
latency-targets-120hz
nightly-suite-shards
gotatun-pin-j469
macos-utun-reader-thread
fix-flood-echo-stall
wip/j416
motion-fast-ease-out
mesh-addhost-j259
takeover-close-terminals-d368
kbd-instant-followups
fix-trust-key-double-click
nightly-suite-partial-report
appwin-icons-j207
rusqlite-slice1
swipe-content-first-d374
mesh-mtu-381
addhost-instant-open
glass-contrast-d365-mock
glass-contrast-d365
signout-followups-442
fix-startup-toast-red
research-gotatun-matrix-j317
themes-sync-j304
fix-load-flakes
fix-nix-zig-patch
fix-compute01-browser-baseline
ratchet-size-baseline-419
predict-auto-d130
revoke-coverage-d367
nightly-suite-queue-timeout
nightly-browser-suite
fix-rustsec-2026-0459
fonts-ioskeley-obel
log-retention-d375
fix-kbd-no-anim-427
fix-dev-red-specs-2
takeover-proof-305-j205
wip/j410-server-pump
ui-consistency-d372
fix-osc52-clipboard
libghostty-vt-1.0
forget-retired-host-d369
j404-merge-426
fix-setup-lock-race
fix-touch-select-426
revoke-access-report-ui
fix-predictive-echo-specs
fix-selinux-agent-exec
wip/j395-checkpoint2
wip/j395
j391-review-scratch
ui-consistency-j195-deferred-rebased
fix-rustc-paths-421
decisions-wording-341
fix-install-suite-red
fix-dev-rtt-specs
fix-pending-removal-align
fix-privileged-skip
fix-install-suite-404
review-j368
mesh-dns-j254
revoke-access-report-d356b
revoke-privilege-round5
fix-removal-confirm
review-j360
gotatun-rotation-367
deps-refresh-340
review-j354
review-J351
review-j350
addhost-takeover-d359
revoke-report-items-347
revoke-end-sessions-d362
review-j348
size-mesh-ssh-346
size-trim-337
revoke-access-diff-d356
j259-pre-rebase-b06c5a6
j305-ended-at-start-spec
addhost-sudo-consent
fix-layout-real-specs-411
fix-shell-open-411
fix-console-hosts-373
fix-shell-exit
fuzz-host-name-353
fix-host-details-410
macos-filter-j197
research-gotatun-tricks
revoke-p3-j216l
stale-state-pill
revoke-followups-j216k
revoke-progress-j216e
backup/macos01-20261106/lume/j197
backup/macos01-20261106/lume/appwin-icons-j207-stash0
backup/macos01-20261106/claude/stash-0
backup/macos01-20261106/claude/stash-1
backup/macos01-20261106/claude/J80-macos-s1
backup/macos01-20261106/claude/J80
backup/macos01-20261106/claude/J80-slice1-20261102
backup/macos01-20261106/claude/J16b-appwin-on-dev
backup/macos01-20261106/claude/appwin-launch-idle-146
backup/macos01-20261106/claude/J80-s1-run
wip/j259
trust-285-p3c
runtime-lock-j216f
mesh-ssh-p3-j175
tun-proof-p3b
trust-285-p3
terminal-polish-j271
tun-proof-p3
ssh-install-flakes-j216i
fuzz-fixes-j264
trust-285-a
trust-285-b
restart-oversized-fixture-j198
backup/macos01-20261106/lume/appwin-icons-j207
restart-bound-j216h
up-preflight-p3-j198
backup/macos01-20261106/lume/j251
up-preflight-j198
wip/up-preflight-j198-j245-p2
wip/up-preflight-j198-j245-p1
agent-fsync-offloop-j196
wip/up-preflight-j198-j245
settings-offscreen-j231
wip/up-preflight-j198-j226-inspection
wip/j103f
themes-j206
wip/up-preflight-j198-j226
flakes-j216c
backup/macos01-20261106/claude/j197
flakes-j216b
backup/macos01-20261106/claude/j197-red
wip/up-preflight-j198-review
flakes-j216
backup/macos01-20261106/claude/appwin-icons-j207
engineering-j208b
appwin-mock-j209
toast-stack-j165
engineering-j208
decisions-d351
decisions-d350
decisions-d349
appearance-presets-mock-j129
decisions-d348
docs-audit-spec
doc-coverage-baseline-raise
decisions-d347
decisions-d346b
decisions-d346
j88h-profile
ui-consistency-j195-deferred
ui-consistency-j195
engineering-practices-j194
preferred-login-ui-j166
signin-pools-263-j174
macos-node-265-s1
decisions-d345
decisions-d344
proxy-cgroup-j164
verify-quick-fuzz-329
proxy-start-io-j168
timing-flakes-j163
up-263-s1
mcp-events-j147
verify-quick-fuzz-j158
proxy-p3-j159
agent-resource-j157-detect
wip/j103d-final
remote-cancel-hang-j155
proxy-traefik-281-j62d
j149-verify-quick
proxy-traefik-281-j62b
j150-drop-unnamed
wip/j79d-red-r4
wip/j103d-d332
rootless-toast-j142
wip/j103d-head
wip/j103d
j125c-mcp
macos-build-j136
backup/macos01-20261106/claude/j136
wip/j79d-red-cli
wip/j79d-red-server
j121-link-liveness
j118c-tracer-failures
wip/j79c-head
release-desktop-j131
j124c-signed-macos
wip/j79c-red-p3
wip/j79c-red-p23
wip/j79c-red-p22
merge-j128-check
perf-deferred-j128
sidebar-toggle-260
j127-shortcut-changed-dot
j125-mcp
j118-tracer-failures
j126-account-summary
settings-sections-mock-j92
j120-pinch-lock
sidebar-292
backup/macos01-20261106/claude/j120
taste-2026-11-03
row-actions-242-215
backup/macos01-20261106/claude/stash-0-untracked
backup/macos01-20261106/claude/J80-265-s1-run
backup/macos01-20261106/claude/J80-macos-s1-20261102
j112-canvas-commands
ui-no-overshoot-scroll
fix-recovery-restart
backup/macos01-20261106/claude/stash-1-untracked
taste-2026-11-02
size-baseline-ratchet
machine-cards-structure-j108
j109-boot-copy-mockups
fix-dev-red
wip/up-263-s1
j95b-settings-rows
proxy-traefik-281
j95-option-rows
docs-human-j91
docs-keys-palette
fix-flood-1013-270-j38d
mesh-ssh-s1
mesh-ssh-s0
addhost-d291-tunnel
test-stability-load
mesh-relay-220
repro-296
shim-thp-off-24
mesh-pins-258-v2
mesh-dns
signin-d311-merged
docs-pages
ui-size-283
fix-ui-batch-specs
wip/j12b-edge-update
gate-on-arm
fix-rename-sync-293
trust-285-j60b
fix-focus-ring-clip
integ-ui-1101
edge-update-v2
signin-d311
taste-2026-11-01
layout-per-tab-250
ui-approved-splitopt-224
ui-approved-splitopt-224-on-dev
icon-text-centre-282
ui-approved-189-212-216
research-reverse-proxy
transfer-budget-320
review-mesh-pins-258-r2
sidebar-pill-259
review-mesh-relay-220
wip/j12-edge-update-v2
nix-package
status-crash-loop-237
fix-dev-test-failures
fix-fixture-socket-path
wip/mesh-relay-220-pre-dev-advance-rebase
fix-canvas-side-clip
research-head-to-head
frosted-real-256
review-mesh-pins-258
wip/mesh-relay-220-pre-final-rebase
review-layout-per-tab-250-r2
broad-suite-triage
nix-devshell
review-status-crash-loop-237-r3
wip/nix-devshell-pre-p3
taste-frosted-real
wip/code-serve-demo-2026-10-30
wip/code-serve-glass-2026-10-30
wip/code-serve-clusters-2026-10-30
wip/code-merge-release94-dev-2026-10-30
wip/rproxy-delta-scratch-2026-10-30
wip/stale-note-257-staged-2026-10-30
wip/d241-size-staged-2026-10-30
review-broad-suite-207
wip/mesh-pins-258-pre-split
wip/mesh-relay-220-pre-rebase
mesh-measure-220
plan-211-update
review-edge-autoupdate-r3
perf24-reconnect
review-stale-note-r2
release-prod-live
d241-shell-marker
review-mesh-pins-server-258
review-revocation-r3
review-setup-token-r2
review-sync-journal-r3
wip/review-mesh-pins-258-prune-review-1995fee46
frosted-scroll-256
sync-journal-r3
revocation-per-account-wip
review-tokio-asyncfd-0386
review-sync-p3s
wip/stash-8by64-stash0-2026-10-30
mesh-drag-r2
setup-token-data-wip
release-9-4
auth-busy-1013
mesh-view-overlap
sync-p3s
review-auth-busy-wake
review-readd-r8-wip
mesh-drag-bounds
audit-auth-93
app-windows-stream-start
review-authority-cap-r3
app-windows-r7-measure
readd-minimal-r6
readd-minimal-r5
app-windows-r7-soak
readd-minimal-r4
app-windows-r7
app-windows-r6
readd-minimal-r3
readd-minimal-r2
readd-minimal
merge-release8-dev
review-busy-page-liveness-r2
review-busy-page-liveness
mesh-view-trim
review-authority-read-retry-r2
review-remove-uninstall-r2
release-retry-reattach
release-retry
review-remove-uninstall
review-authority-read-retry
app-windows-ship
mesh-autoswitch-p2
mesh-ssh
mesh-relay
mesh-macos-node
screen-and-control
session-labels
review-real-logins-r5
review-real-logins-r4
review-real-logins-r3
shim-memory-r2
review-real-logins-r2
shim-memory
shadow-addhost-instant
shadow-real-logins
shadow-same-session-flood
shadow-prod-perf-fix
hosts-from-server
app-windows-real
app-windows-transport
release-prod-6w
mesh-policy-verify
quiet-close-toast
dev-merge-r4
review-mesh-agent-transport
review-app-windows-transport
settings-roundtrip
review-release-pipeline-r4
mesh-agent-transport
ended-strip-fix
review-audit-ca-update-d244
settings-rows
phone-width
signout-clear
fix-restart-tests
dev-merge-rp
signer-hardening
review-audit-ca-update
terminal-fidelity
review-audit-ca-update-wip
mesh-integrate
review-mesh-agent-transport-515d3712
review-mesh-dataplane-feec1988
mesh-dataplane
signout
research-canvas-node
wip/review-renderer-cost
wip/review-terminal-fidelity
padlock-glyph
review-release-pipeline-2e95b689
phone-polish
sidebar-polish
inbox-generic
review-mesh-dataplane-378d4055
prod-batch
ui-inbox
ui-ca-unlock
ui-machine-sections
review-mesh-dataplane-a5f744fb
tls-session-prod
session-bounds-prod
fix-dead-session
session-bounds
tls-session
palette-polish
shortcuts-d222
research-appmeta
phone-gutter
ui-centre-button
wip/ui-centre-button
wip/tls-session
review-addhost-breakit
ui-platform
renderer-cost
remote-agents
mesh-api
ssh-bootstrap
prod-hotfix2
prod-fixes
review-remote-agents
review-mesh-dataplane
session-life
research-browser
review-mesh-api
research-capture
review-mesh-api-round3
review-mesh-api-round1
gate-stability
review-gate-stability
review-relay-direct
review-mesh-api-findings
proto-sessions
font-picker
proto-addhost
proto-windows
mesh-core
latency-gate
research-bsd
research-discovery
keytray-mods
mock-core
ui-tweaks
fix-motion-sel
fix-motion
integrate2
fix-robust
t-web
t-server
vt-ghostty
fix-snap-echo
t-web-auth
review-vt2
t-agent
lane-protocol
lane-shim
lane-auth
s1-base
s1-contracts
s1-workspace
No results found.
Labels
Clear labels   
launch-qa
Findings from launch QA.

No labels
launch-qa
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
kayg/infraternal#214
No description provided.