devices(ubiquiti): one UDM-Pro row with both 10G cages, and the UDW-PRO document-slug controls (#130) #2049

Closed
jaylouisw wants to merge 1 commit from catalogue/1118-2252 into main
Owner

Refs #130

Two things this pass: the last naming-convention route for the Dream Wall Pro is now measured shut with controls, and the catalogue row a requester settles for instead was modelled twice, one copy of which could not express its own second 10G cage.

New controls for issue 130. Every earlier "no datasheet" reading was taken from a _ds URL. https://dl.ui.com/ds/udm-pro.pdf answers 200, 353,553 B, application/pdf while its _ds twin answers 302 to the techspecs front page - so _ds is one of two conventions, not the convention, and a 302 on dl.ui.com/ds/udw-pro_ds.pdf never proved there was no file. The bare-slug path was then probed under five spellings (udw-pro.pdf, udwpro.pdf, udw_pro.pdf, udw-pro_ds.pdf, udw-pro-1.pdf): all 302 to https://techspecs.ui.com. The quick-install-guide tree, which is server-rendered and slug-addressed (dl.ui.com/qig/udw/ 200/6,310 B, dl.ui.com/qig/udm-pro/ 200/6,185 B), has no entry for qig/udw-pro/ (302/313 B) or qig/udwpro/ (302/312 B). Table and reasoning: research/ubiquiti-udw-pro-doc-slug-controls.md; research/open-device-requests.json carries the same with the four new sources.

Catalogue fix, one device. devices/ubiquiti/ described the UDM-Pro twice - UDM-Pro.json and unifi-dream-machine-pro.json - and the two disagreed. The file holding the model string modelled the second 10G cage as a bare ethernet 10G port with no connector: one cage, plus a 10G port nothing can seat in. The other copy had the right count under model "UniFi Dream Machine Pro" (a product name), category host, and the engine's host capability, which means an end device that terminates packets (engine/device.js), not a console that hosts apps, on a gateway.

Vendor census, read today: datasheet https://dl.ui.com/ds/udm-pro.pdf - "Networking interface (8) LAN: GbE RJ45 ports / (1) WAN: GbE RJ45 port", "SFP+ interface (1) LAN: 10G SFP+ / (1) WAN: 10G SFP+", IDS/IPS 3.5 Gbps, 33 W max. Tech-spec page https://techspecs.ui.com/unifi/cloud-gateways/udm-pro prints the same count as a table - "1 GbE RJ45 = 9 (1G/100M/10M)", "10G SFP+ = 2 (10G/1G)" - and points at the same PDF. So: 9x 1G copper (1 WAN + 8 LAN, speedCaps 10/100/1000), 2x 10G SFP+ (connector sfp+, speedCaps 1000/10000), no PoE. One file, devices/ubiquiti/UDM-Pro.json, with aliases for both spellings so search does not lose the deleted name; capability set follows UDM-Pro-Max and UDM-SE. scripts/merge-research.js's inline entry, which is what produced the connector-less port, is corrected to match. Registry regenerated: 2,356 models, Ubiquiti 369 -> 368.

Gates: syntax-check.js 428 files OK; audit-port-types.js no port contradicts its declared type; test/hardware-invariants.test.js 26/26; test/device-aliases.test.js 13/13; npm test 352 suites, 12,179 assertions, 0 failed.

No close here: nothing in this pull request names a UDW-PRO port or PoE wattage, and issue 130 stays open with research/open-device-requests.json as its record.

Refs #130 Two things this pass: the last naming-convention route for the Dream Wall Pro is now measured shut with controls, and the catalogue row a requester settles for instead was modelled twice, one copy of which could not express its own second 10G cage. **New controls for issue 130.** Every earlier "no datasheet" reading was taken from a `_ds` URL. `https://dl.ui.com/ds/udm-pro.pdf` answers **200, 353,553 B, application/pdf** while its `_ds` twin answers 302 to the techspecs front page - so `_ds` is one of two conventions, not the convention, and a 302 on `dl.ui.com/ds/udw-pro_ds.pdf` never proved there was no file. The bare-slug path was then probed under five spellings (`udw-pro.pdf`, `udwpro.pdf`, `udw_pro.pdf`, `udw-pro_ds.pdf`, `udw-pro-1.pdf`): all 302 to `https://techspecs.ui.com`. The quick-install-guide tree, which is server-rendered and slug-addressed (`dl.ui.com/qig/udw/` 200/6,310 B, `dl.ui.com/qig/udm-pro/` 200/6,185 B), has no entry for `qig/udw-pro/` (302/313 B) or `qig/udwpro/` (302/312 B). Table and reasoning: `research/ubiquiti-udw-pro-doc-slug-controls.md`; `research/open-device-requests.json` carries the same with the four new sources. **Catalogue fix, one device.** `devices/ubiquiti/` described the UDM-Pro twice - `UDM-Pro.json` and `unifi-dream-machine-pro.json` - and the two disagreed. The file holding the model string modelled the second 10G cage as a bare `ethernet 10G` port with no connector: one cage, plus a 10G port nothing can seat in. The other copy had the right count under model `"UniFi Dream Machine Pro"` (a product name), category `host`, and the engine's `host` capability, which means an end device that terminates packets (`engine/device.js`), not a console that hosts apps, on a gateway. Vendor census, read today: datasheet `https://dl.ui.com/ds/udm-pro.pdf` - "Networking interface (8) LAN: GbE RJ45 ports / (1) WAN: GbE RJ45 port", "SFP+ interface (1) LAN: 10G SFP+ / (1) WAN: 10G SFP+", IDS/IPS 3.5 Gbps, 33 W max. Tech-spec page `https://techspecs.ui.com/unifi/cloud-gateways/udm-pro` prints the same count as a table - "1 GbE RJ45 = 9 (1G/100M/10M)", "10G SFP+ = 2 (10G/1G)" - and points at the same PDF. So: 9x 1G copper (1 WAN + 8 LAN, speedCaps 10/100/1000), 2x 10G SFP+ (connector `sfp+`, speedCaps 1000/10000), no PoE. One file, `devices/ubiquiti/UDM-Pro.json`, with `aliases` for both spellings so search does not lose the deleted name; capability set follows UDM-Pro-Max and UDM-SE. `scripts/merge-research.js`'s inline entry, which is what produced the connector-less port, is corrected to match. Registry regenerated: 2,356 models, Ubiquiti 369 -> 368. Gates: `syntax-check.js` 428 files OK; `audit-port-types.js` no port contradicts its declared type; `test/hardware-invariants.test.js` 26/26; `test/device-aliases.test.js` 13/13; `npm test` 352 suites, 12,179 assertions, 0 failed. No close here: nothing in this pull request names a UDW-PRO port or PoE wattage, and issue 130 stays open with `research/open-device-requests.json` as its record.
devices(ubiquiti): one UDM-Pro row, with both 10G SFP+ cages (#130)
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 8s
CI / Server smoke test (pull_request) Successful in 10s
CI / Test (Node 20) (pull_request) Successful in 1m0s
CI / Docker build (pull_request) Successful in 19s
CI / Test (Node 18) (pull_request) Successful in 1m5s
CI / Test (Node 22) (pull_request) Successful in 1m0s
CI / Arm auto-merge (pull_request) Successful in 2s
eeeccb21e5
The catalogue described the UDM-Pro twice. The file holding the model string
modelled the second 10G cage as a bare ethernet 10G port with no connector, so
the box had one cage and a 10G port nothing can seat in; the other copy carried
the right count under model 'UniFi Dream Machine Pro', category 'host' and the
engine's 'host' capability (an end device that terminates packets) on a gateway.

Census from the vendor datasheet dl.ui.com/ds/udm-pro.pdf (200, 353,553 B):
(8) LAN GbE RJ45, (1) WAN GbE RJ45, (1) LAN + (1) WAN 10G SFP+, 33 W max. The
tech-spec page prints the same count (1 GbE RJ45 = 9, 10G SFP+ = 2).

Also records why the Dream Wall Pro's document tree is empty under every slug
spelling, with controls: dl.ui.com/ds/udm-pro.pdf is real while its _ds twin
302s, so _ds was never the convention; the bare-slug and qig trees have no Pro
entry. Issue 130 stays open.
forgejo-actions force-pushed catalogue/1118-2252 from eeeccb21e5
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 8s
CI / Server smoke test (pull_request) Successful in 10s
CI / Test (Node 20) (pull_request) Successful in 1m0s
CI / Docker build (pull_request) Successful in 19s
CI / Test (Node 18) (pull_request) Successful in 1m5s
CI / Test (Node 22) (pull_request) Successful in 1m0s
CI / Arm auto-merge (pull_request) Successful in 2s
to d5e15d71a9 2026-11-18 23:09:38 +00:00
Compare
Author
Owner

Conflict is not mechanical — needs Jay's decision before this can land.

Rebasing onto current main gives 2 conflicted files, both over issue #130 (UDW-PRO):

  • research/open-device-requests.json — the missing string and the sources array for entry 130. main now carries #2051's grant-census text (SWX grantee enumeration, the EU store tree, "no Pro hides inside the parent grant", sources fccid.io/SWX pages 1–5 + the Wayback uacc-psu-udw capture). This branch carries a different investigation of the same field (dl.ui.com naming conventions: udm-pro.pdf 200/353,553 B proves _ds is not a convention, the five bare-slug probings that all 302 to techspecs, qig/udw-pro/ and qig/udwpro/ absent, sources dl.ui.com/ds/udm-pro.pdf, udw-pro.pdf, qig/udw-pro/, qig/udm-pro/).
  • docs/DEVICE_REGISTRY.md — generated from that JSON, so it inherits the same conflict in the issue-130 row.

Both narratives are true and neither replaces the other; concatenating two 3 KB missing strings into one cell is an editorial call, not a rebase fix, and dropping either side loses sourced work (research/ubiquiti-udw-pro-doc-slug-controls.md, added by this PR, is cited by the branch text and would be orphaned if main's version wins outright).

The rest of the PR is clean and I can land it as soon as the text question is answered: devices/ubiquiti/UDM-Pro.json (both 10G cages), devices/ubiquiti/unifi-dream-machine-pro.json → deleted, README.md, docs/CONNECTORS.md, docs/wiki/Home.md, scripts/build-wiki.js, scripts/merge-research.js.

Decision needed: keep main's #2051 census text for issue 130 and add the doc-slug findings as a separate checked/sources increment, or publish both narratives side by side in the same field? After that, docs/DEVICE_REGISTRY.md gets regenerated, not hand-merged.

Conflict is not mechanical — **needs Jay's decision** before this can land. Rebasing onto current main gives 2 conflicted files, both over **issue #130 (UDW-PRO)**: - `research/open-device-requests.json` — the `missing` string and the `sources` array for entry `130`. main now carries #2051's grant-census text (`SWX` grantee enumeration, the EU store tree, "no Pro hides inside the parent grant", sources `fccid.io/SWX` pages 1–5 + the Wayback `uacc-psu-udw` capture). This branch carries a different investigation of the same field (`dl.ui.com` naming conventions: `udm-pro.pdf` 200/353,553 B proves `_ds` is not a convention, the five bare-slug probings that all 302 to techspecs, `qig/udw-pro/` and `qig/udwpro/` absent, sources `dl.ui.com/ds/udm-pro.pdf`, `udw-pro.pdf`, `qig/udw-pro/`, `qig/udm-pro/`). - `docs/DEVICE_REGISTRY.md` — generated from that JSON, so it inherits the same conflict in the issue-130 row. Both narratives are true and neither replaces the other; concatenating two 3 KB `missing` strings into one cell is an editorial call, not a rebase fix, and dropping either side loses sourced work (`research/ubiquiti-udw-pro-doc-slug-controls.md`, added by this PR, is cited by the branch text and would be orphaned if main's version wins outright). The rest of the PR is clean and I can land it as soon as the text question is answered: `devices/ubiquiti/UDM-Pro.json` (both 10G cages), `devices/ubiquiti/unifi-dream-machine-pro.json` → deleted, `README.md`, `docs/CONNECTORS.md`, `docs/wiki/Home.md`, `scripts/build-wiki.js`, `scripts/merge-research.js`. Decision needed: keep main's #2051 census text for issue 130 and add the doc-slug findings as a separate `checked`/`sources` increment, or publish both narratives side by side in the same field? After that, `docs/DEVICE_REGISTRY.md` gets regenerated, not hand-merged.
Author
Owner

surgery run 2026-11-19 03:0xZ — queued for rebase, and it overlaps #2076.

Local git merge-tree against be18b65d: conflicts in docs/DEVICE_REGISTRY.md (generated — regenerated, not merged) and research/open-device-requests.json (issue 130, the UDW-PRO entry).

That entry was just rewritten by #2076 (cc8e6074), which landed research/ubiquiti-udw-pro-absent-from-vendor-catalogue.md and the Chile//compliance/ evidence — this PR carries research/ubiquiti-udw-pro-doc-slug-controls.md and its own 130 text. The two research notes are different documents; the JSON entry is one field, so the same "which paragraph" question as #2057 applies. I will rebase the code/registry parts next run and bring the 130 text here for a decision rather than pick a winner.

Note: head d5e15d71 has 0 check statuses (no pull_request run created — same zero-status class as #2069/#2073 before their re-push), 4h after it was opened.

**surgery run 2026-11-19 03:0xZ — queued for rebase, and it overlaps #2076.** Local `git merge-tree` against `be18b65d`: conflicts in `docs/DEVICE_REGISTRY.md` (generated — regenerated, not merged) and `research/open-device-requests.json` (issue 130, the UDW-PRO entry). That entry was just rewritten by **#2076** (`cc8e6074`), which landed `research/ubiquiti-udw-pro-absent-from-vendor-catalogue.md` and the Chile/`/compliance/` evidence — this PR carries `research/ubiquiti-udw-pro-doc-slug-controls.md` and its own 130 text. The two research notes are different documents; the JSON entry is one field, so the same "which paragraph" question as #2057 applies. I will rebase the code/registry parts next run and bring the 130 text here for a decision rather than pick a winner. Note: head `d5e15d71` has **0 check statuses** (no pull_request run created — same zero-status class as #2069/#2073 before their re-push), 4h after it was opened.
Author
Owner

Status note, 2026-11-19 03:55Z.

The no checks half of this PR is not about this PR. It is issue #2082: a head pushed by the rebase bot (d5e15d71, committer packetforge-bot) gets no workflow run at all, which is why the last runs here are for eeeccb21 (2026-11-18 23:11Z) and nothing since. The cure I verified twice tonight is in #2082: rebase onto current main (0 behind) and push with a non-bot committer; the full 8-check run appears in ~100 s.

The merge conflict is the real blocker and it is not mechanical:

  • research/open-device-requests.json - a whole-file JSON rewrite on both sides, not an adjacent-insertion conflict. The two sides disagree about the text of request 130 and request 803. main now carries the #2076/#2079 wording for 130 (nine storefronts from the page's own alternativeStores, quick-start key resolving to artwork) plus #2072's 803 paragraph. This PR carries its own long rewrite of both. Picking one side drops research; keeping both duplicates the entry. That is Jay's call, per the same rule that applies to #2057.
  • docs/DEVICE_REGISTRY.md - generated, so regeneration would be mechanical (npm run regen:registry-doc) after the JSON is decided.

No change made to the branch this run.

Status note, 2026-11-19 03:55Z. The `no checks` half of this PR is not about this PR. It is issue #2082: a head pushed by the rebase bot (`d5e15d71`, committer `packetforge-bot`) gets no workflow run at all, which is why the last runs here are for `eeeccb21` (2026-11-18 23:11Z) and nothing since. The cure I verified twice tonight is in #2082: rebase onto *current* `main` (0 behind) and push with a non-bot committer; the full 8-check run appears in ~100 s. The merge conflict is the real blocker and it is not mechanical: - `research/open-device-requests.json` - a whole-file JSON rewrite on both sides, not an adjacent-insertion conflict. The two sides disagree about the text of request **130** and request **803**. main now carries the #2076/#2079 wording for 130 (nine storefronts from the page's own `alternativeStores`, quick-start key resolving to artwork) plus #2072's 803 paragraph. This PR carries its own long rewrite of both. Picking one side drops research; keeping both duplicates the entry. That is **Jay's call**, per the same rule that applies to #2057. - `docs/DEVICE_REGISTRY.md` - generated, so regeneration would be mechanical (`npm run regen:registry-doc`) *after* the JSON is decided. No change made to the branch this run.
Author
Owner

Superseded by #2093, which carries this PR's substance rebased onto current main. This branch has been mergeable: false since it was opened - the regenerated-count hunks in README.md, docs/CONNECTORS.md, docs/wiki/Home.md and scripts/build-wiki.js have since moved under it. No force-push from this account, so the fix lands in #2093 instead: devices/ubiquiti/UDM-Pro.json with 9x 1G RJ45 + 2x 10G SFP+ and both spellings as aliases, unifi-dream-machine-pro.json removed, scripts/merge-research.js corrected, plus research/ubiquiti-udw-pro-doc-slug-controls.md re-measured today (every row reproduced byte for byte). Closing this one is safe once #2093 is merged.

Superseded by #2093, which carries this PR's substance rebased onto current main. This branch has been `mergeable: false` since it was opened - the regenerated-count hunks in `README.md`, `docs/CONNECTORS.md`, `docs/wiki/Home.md` and `scripts/build-wiki.js` have since moved under it. No force-push from this account, so the fix lands in #2093 instead: `devices/ubiquiti/UDM-Pro.json` with 9x 1G RJ45 + 2x 10G SFP+ and both spellings as aliases, `unifi-dream-machine-pro.json` removed, `scripts/merge-research.js` corrected, plus `research/ubiquiti-udw-pro-doc-slug-controls.md` re-measured today (every row reproduced byte for byte). Closing this one is safe once #2093 is merged.
Author
Owner

Closed as superseded. #2093 landed this change on main as 73ea47f4 (devices(ubiquiti): one UDM-Pro row with two 10G cages, not a 10G copper port (#130)): same file set, UDM-Pro.json with 9x 1G RJ45 + 2x 10G SFP+ and both spellings as aliases, unifi-dream-machine-pro.json removed, the regenerated-count hunks re-derived on current main. Diffed this head against main for every file it touches - the only residue is the earlier "read 2026-11-18" wording and the 7-line re-verification note, both superseded by #2093's 2026-11-19 re-measurement. Nothing to rebase.

Closed as superseded. #2093 landed this change on main as 73ea47f4 (`devices(ubiquiti): one UDM-Pro row with two 10G cages, not a 10G copper port (#130)`): same file set, `UDM-Pro.json` with 9x 1G RJ45 + 2x 10G SFP+ and both spellings as aliases, `unifi-dream-machine-pro.json` removed, the regenerated-count hunks re-derived on current main. Diffed this head against main for every file it touches - the only residue is the earlier "read 2026-11-18" wording and the 7-line re-verification note, both superseded by #2093's 2026-11-19 re-measurement. Nothing to rebase.
jaylouisw closed this pull request 2026-11-19 05:09:57 +00:00

Pull request closed

Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
2 participants
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!2049
No description provided.