queue: "ready" means buildable, not merely form-complete - five of five are blocked #1900

Merged
jaylouisw merged 1 commit from catalogue/1117-2055 into main 2026-11-17 21:31:23 +00:00
Owner

The queue's own front page said open: 5 ready: 5 this evening, for a queue in which not one of the five can be written without inventing a port census. node scripts/request-queue.js --skeleton 810 writes OptiXtrans E9624 as a one-port eth0 box; the real subrack is 24 service slots with no ports until a board is named. That gap is step 3 of docs/plans/stated-configurations.md, and this PR lands it.

What changed

  • scripts/forgejo-api.js - new listIssueComments(number), the reader the plan said was missing. It is only called for a request that already carries the label, so a healthy queue makes no extra calls.
  • scripts/request-queue.js - blocked-on-build is its own state, never counted ready. Its group prints the newest comment that names the block, verbatim, capped at 220 characters, plus the issue URL. Markdown emphasis is stripped so the line reads as a sentence. If the comments cannot be read, the request still shows as blocked and says so - a failed read must not turn a visible blocker into an invisible "ready".
  • --skeleton on a blocked request still emits the stub, but warns on stderr that it is an empty chassis and that filling it from datasheet maxima is the error docs/MAINTAINING.md warns about.
  • test/request-queue.test.js - nine assertions: the verdict line is picked over a newer off-topic comment, emphasis and > quotes are stripped, a long comment is capped, an empty or failed read is null rather than an empty string.

Why a label and not a regex over maintainer prose

The plan proposed matching "blocked on" in the last comment. That re-classifies a request the moment someone writes "blocked on the bench", and it cannot be un-done by the person who knows the answer. The label is one click, appears in the issue list without opening it, and the comment still carries the reason - so the queue quotes the evidence instead of deriving a verdict from it.

Measured, live queue, 2026-11-17

before:  open: 5   ready: 5   waiting on requester: 0
after:   open: 5   ready: 0   blocked-on-build: 5   waiting on requester: 0

The label was applied to #130 (UDW-PRO - no vendor page at all), #803 (E6616X - E6600 board census hedex-only), #807 (E9605), #808 (E9612), #810 (E9624) - the last three on one client-board chapter each.

Two corrections that fell out of it

  1. research/open-device-requests.json still listed #742 and #743, both closed. Dropped, so docs/DEVICE_REGISTRY.md stops claiming seven open requests; it now says five.
  2. That page's count was typed into scripts/regen-registry-doc.js as the word "Seven", so a generated page went stale about its own length the moment two of the seven closed. The count is now derived from the JSON (0-10 spelled out, digits above), with a queue-empty branch, because a page that exists to catch staleness should not be the one that goes stale.

Gates

  • node scripts/syntax-check.js - 394 files, all OK
  • node scripts/audit-port-types.js - "No port contradicts its declared type."
  • node scripts/regen-registry-doc.js --check - current: 2,348 models, 258 vendors, 29,235 sockets, 966 port shapes
  • npm test - 322 suites, 11,202 assertions passed, 0 failed
  • node test/request-queue.test.js - 32 passed, 0 failed (9 new)

No device spec is added here and no issue is closed by this PR: the five remain blocked on their named fields, and docs/plans/stated-configurations.md step 6 (an OTN port type, needing Jay's approval) is what unblocks three of them.

The queue's own front page said `open: 5 ready: 5` this evening, for a queue in which not one of the five can be written without inventing a port census. `node scripts/request-queue.js --skeleton 810` writes `OptiXtrans E9624` as a one-port `eth0` box; the real subrack is 24 service slots with no ports until a board is named. That gap is step 3 of `docs/plans/stated-configurations.md`, and this PR lands it. **What changed** - `scripts/forgejo-api.js` - new `listIssueComments(number)`, the reader the plan said was missing. It is only called for a request that already carries the label, so a healthy queue makes no extra calls. - `scripts/request-queue.js` - `blocked-on-build` is its own state, never counted ready. Its group prints the newest comment that names the block, verbatim, capped at 220 characters, plus the issue URL. Markdown emphasis is stripped so the line reads as a sentence. If the comments cannot be read, the request still shows as blocked and says so - a failed read must not turn a visible blocker into an invisible "ready". - `--skeleton` on a blocked request still emits the stub, but warns on stderr that it is an empty chassis and that filling it from datasheet maxima is the error `docs/MAINTAINING.md` warns about. - `test/request-queue.test.js` - nine assertions: the verdict line is picked over a newer off-topic comment, emphasis and `>` quotes are stripped, a long comment is capped, an empty or failed read is `null` rather than an empty string. **Why a label and not a regex over maintainer prose** The plan proposed matching "blocked on" in the last comment. That re-classifies a request the moment someone writes "blocked on the bench", and it cannot be un-done by the person who knows the answer. The label is one click, appears in the issue list without opening it, and the comment still carries the reason - so the queue quotes the evidence instead of deriving a verdict from it. **Measured, live queue, 2026-11-17** ``` before: open: 5 ready: 5 waiting on requester: 0 after: open: 5 ready: 0 blocked-on-build: 5 waiting on requester: 0 ``` The label was applied to #130 (UDW-PRO - no vendor page at all), #803 (E6616X - E6600 board census hedex-only), #807 (E9605), #808 (E9612), #810 (E9624) - the last three on one client-board chapter each. **Two corrections that fell out of it** 1. `research/open-device-requests.json` still listed #742 and #743, both closed. Dropped, so `docs/DEVICE_REGISTRY.md` stops claiming seven open requests; it now says five. 2. That page's count was typed into `scripts/regen-registry-doc.js` as the word "Seven", so a generated page went stale about its own length the moment two of the seven closed. The count is now derived from the JSON (0-10 spelled out, digits above), with a queue-empty branch, because a page that exists to catch staleness should not be the one that goes stale. **Gates** - `node scripts/syntax-check.js` - 394 files, all OK - `node scripts/audit-port-types.js` - "No port contradicts its declared type." - `node scripts/regen-registry-doc.js --check` - current: 2,348 models, 258 vendors, 29,235 sockets, 966 port shapes - `npm test` - 322 suites, 11,202 assertions passed, 0 failed - `node test/request-queue.test.js` - 32 passed, 0 failed (9 new) No device spec is added here and no issue is closed by this PR: the five remain blocked on their named fields, and `docs/plans/stated-configurations.md` step 6 (an OTN port type, needing Jay's approval) is what unblocks three of them.
queue: "ready" means buildable, not merely form-complete - five of five are blocked
All checks were successful
CI / Classify the change (pull_request) Successful in 8s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 14s
CI / Server smoke test (pull_request) Successful in 8s
CI / Docker build (pull_request) Successful in 24s
CI / Test (Node 20) (pull_request) Successful in 50s
CI / Test (Node 22) (pull_request) Successful in 52s
CI / Test (Node 18) (pull_request) Successful in 57s
CI / Arm auto-merge (pull_request) Successful in 7s
964f6a137b
``node scripts/request-queue.js`` counted a request ready when the four form fields
were filled, so today's whole queue read ``ready: 5`` and ``--skeleton 810`` offered a
24-slot OTN subrack as a one-port box. That is the failure mode
``docs/plans/stated-configurations.md`` ranked fifth: the queue calls a request ready
that a worker cannot write without inventing a port census.

- ``scripts/forgejo-api.js`` gains ``listIssueComments`` (one call, made only for a
  labelled request).
- ``scripts/request-queue.js`` reports ``blocked-on-build`` as its own group and
  quotes the newest comment that names the block, so the reason is the thread's
  sentence and not this script's paraphrase. ``--skeleton`` still runs on a blocked
  request but warns first. Nine new assertions in ``test/request-queue.test.js``.
- The label was applied to #130, #803, #807, #808 and #810. Measured on the live
  queue: ``open: 5  ready: 5`` became ``open: 5  ready: 0  blocked-on-build: 5``.
- ``research/open-device-requests.json`` drops #742 and #743, both closed, and
  ``scripts/regen-registry-doc.js`` now derives the count instead of typing
  "Seven": ``docs/DEVICE_REGISTRY.md`` says five open requests, which is the truth
  after those two closes.

Gates: ``node scripts/syntax-check.js`` 394 files OK; ``node scripts/audit-port-types.js``
no port contradicts its declared type; ``node scripts/regen-registry-doc.js --check``
current at 2,348 models; ``npm test`` 322 suites / 11,202 assertions / 0 failed.
jaylouisw force-pushed catalogue/1117-2055 from 964f6a137b
All checks were successful
CI / Classify the change (pull_request) Successful in 8s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 14s
CI / Server smoke test (pull_request) Successful in 8s
CI / Docker build (pull_request) Successful in 24s
CI / Test (Node 20) (pull_request) Successful in 50s
CI / Test (Node 22) (pull_request) Successful in 52s
CI / Test (Node 18) (pull_request) Successful in 57s
CI / Arm auto-merge (pull_request) Successful in 7s
to 6a0b3b8299
All checks were successful
CI / Classify the change (pull_request) Successful in 9s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 12s
CI / Server smoke test (pull_request) Successful in 9s
CI / Docker build (pull_request) Successful in 24s
CI / Test (Node 20) (pull_request) Successful in 49s
CI / Test (Node 18) (pull_request) Successful in 58s
CI / Test (Node 22) (pull_request) Successful in 56s
CI / Arm auto-merge (pull_request) Successful in 7s
2026-11-17 21:26:05 +00:00
Compare
jaylouisw force-pushed catalogue/1117-2055 from 6a0b3b8299
All checks were successful
CI / Classify the change (pull_request) Successful in 9s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 12s
CI / Server smoke test (pull_request) Successful in 9s
CI / Docker build (pull_request) Successful in 24s
CI / Test (Node 20) (pull_request) Successful in 49s
CI / Test (Node 18) (pull_request) Successful in 58s
CI / Test (Node 22) (pull_request) Successful in 56s
CI / Arm auto-merge (pull_request) Successful in 7s
to 6ed3e9e2d3
All checks were successful
CI / Classify the change (pull_request) Successful in 11s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 9s
CI / Test (Node 20) (pull_request) Successful in 49s
CI / Test (Node 18) (pull_request) Successful in 57s
CI / Test (Node 22) (pull_request) Successful in 52s
CI / Server smoke test (pull_request) Successful in 8s
CI / Docker build (pull_request) Successful in 17s
CI / Arm auto-merge (pull_request) Successful in 16s
2026-11-17 21:27:10 +00:00
Compare
Sign in to join this conversation.
No reviewers
No milestone
No project
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
jaylouisw/packetforge!1900
No description provided.