Commit graph

6 commits

Author SHA1 Message Date
3daf2fcdc5 Verify (don't order) phone number; fix DECT location lookup endpoint
Some checks failed
CI / verify (push) Has been cancelled
Two bugs surfaced by the same failed finalize run for store 7311.

1. addPhoneNumbersToLocation was POSTing to /locations/{id}/numbers,
   which Webex treats as "provision a NEW number" — that triggered
   duplicate PSTN orders against numbers AE had already manually
   ordered through Control Hub, and the provider rejected them with
   ORDER_CREATION_FAILED. AE's workflow is manual ordering, so this
   step now VERIFIES the number's current state in the org via
   GET /telephony/config/numbers?phoneNumber=..., classifies the
   outcome (already here / pending / wrong location / not in org),
   and either succeeds silently or throws with a specific
   Control-Hub fix-it instruction. The bot never triggers a PSTN
   order now.

2. findDectNetworkInLocation was hitting
   GET /telephony/config/locations/{id}/dectNetworks, which Webex
   does not implement (returns HTTP 404 "No static resource ..."
   unconditionally). That silently broke both the finalize
   idempotency pre-check and the /provisionDect fallback, so
   /provisionDect kept showing the "create network" card for stores
   that already had one, then 409'd on the create attempt. Switched
   to the real endpoint GET /telephony/config/dectNetworks with a
   locationId filter. Also added translateCreateDectError to
   rewrite the deeply-nested 27453 "access code in use" 409 blob
   into an actionable "already exists, re-run /provisionDect"
   message.

Extracted both error/state translators as pure exported functions
(classifyNumberAssignment, translateCreateDectError) with unit-test
coverage locking down the exact Webex payload shapes.

README: updated /finalizeStore to reflect the manual-order
expectation and specific fix-it instructions on mismatch.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-09 13:53:38 -04:00
261d11f29d Harden /finalizeStore against pending-order and DECT-404 noise
Some checks are pending
CI / verify (push) Waiting to run
A finalize run against a store whose phone number still had a pending
carrier order surfaced three unrelated-looking 400s (add-number,
caller ID = LOCATION_NUMBER, auto attendant "number already used")
that all traced back to the number not being attached, plus a
spurious "DECT network pre-check failed" warn when the location had
never had a DECT network before.

Fixes:
- addPhoneNumbersToLocation now translates the raw
  NUMBER_HAS_PENDING_ORDERS payload into an actionable operator
  message ("wait for the carrier order, then re-run /finalizeStore").
  Original error preserved as .cause. Extracted as pure
  translateAddNumberError so it can be unit-tested.
- finalizeStore marks the phone-number-add step { critical: true } so
  finalize aborts immediately on that failure instead of cascading
  three downstream errors that bury the root cause.
- findDectNetworkInLocation treats HTTP 404 as "no networks yet"
  (Webex returns 404 for that state, not an empty list), so the
  finalize pre-check stops warning on the normal first-time path.
  The remaining warn now includes status code for anything that does
  reach it.

Tests: 4 new cases in test/locations.test.js locking down the
translator behavior against the exact Webex payload shape.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-09 12:35:00 -04:00
203850fd69 Fix requestJson dropping Content-Type on POSTs (HTTP 415)
Some checks are pending
CI / verify (push) Waiting to run
requestJson merged Content-Type: application/json into a headers object,
then spread `...extra` after `headers` in the fetch options — which let
extra.headers (always passed by webexJson for Authorization) overwrite
the merged object, stripping Content-Type entirely. Undici then
defaulted the string body's Content-Type to text/plain;charset=UTF-8,
and strict Webex endpoints like POST /locations rejected the request
with HTTP 415. Latent since the node-fetch -> native fetch migration.

Reorder to `{ method, ...extra, headers }` so the explicit merged
headers win, and add a test/http.test.js suite (7 tests) that pins the
merge behaviour down so this can't regress silently again.

Also: update the npm test script to `test/**/*.test.js` glob. Node
22.23 no longer accepts a bare `test/` directory as a positional arg
to `--test` (fails with ERR_UNSUPPORTED_DIR_IMPORT).

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-09 12:07:15 -04:00
6b585bf33f Add /provisionPhone for wired desk phones (7841/7821)
Some checks failed
CI / verify (push) Has been cancelled
New /provisionPhone <storeNumber> command manages the store user's
wired desk phones via the same card pattern as /provisionDect:
multi-select checkbox list that doubles as the display, MACs render
as AA:BB:CC:DD:EE:FF, model dropdown for add (defaults to 7841),
side-by-side Add / Remove-checked actions, and every removal goes
through an explicit confirm card. Adds are idempotent (MACs already
registered are skipped) and bulk MAC input is supported.

DECT handsets are filtered out of the display so /provisionPhone and
/provisionDect coexist cleanly on the same store user without
overlapping responsibilities.

Refactors:
- Extract MAC helpers (normalize/format/display) to src/utils/mac.js
  so both DECT and wired-phone flows share one implementation. dect.js
  re-exports for backward compat with existing consumers.
- Add parseEmailArg helper; migrate /userInfo to use it.

Tests: pure-logic coverage for WIRED_PHONE_MODELS,
isSupportedWiredModel, filterWiredPhones (DECT/model filter), and
parseEmailArg.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-07 12:39:06 -04:00
5dce057ce7 Add /provisionDect + create DECT network in /finalizeStore
Some checks are pending
CI / verify (push) Waiting to run
New /provisionDect slash command manages DECT basestations and handsets
for a store via a single card: multi-select checkbox list doubles as
the display, MACs render as AA:BB:CC:DD:EE:FF, Add/Remove sit
side-by-side per section, and every removal goes through an explicit
confirm card. Handsets always auto-pair (no bind-to-basestation input)
so they roam.

/finalizeStore now idempotently creates the "Store XXXX" DECT network
(DBS-210) with the per-store default access code, so new stores are
DECT-ready the moment finalize completes. Location-scoped lookup
(findDectNetworkInLocation) handles both the finalize idempotency
check and the /provisionDect fallback for freshly-created empty
networks. Non-critical: a store can still go live if the DECT step
fails, and /provisionDect keeps a recovery "create network" prompt
for legacy stores.

Pure-logic helpers (generateDectAccessCode, MAC normalize/format/
display, dectNetworkName) are unit-tested via node:test.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-07 07:29:21 -04:00
6de5392e90 Add node:test suite and Gitea Actions CI
Some checks are pending
CI / verify (push) Waiting to run
- 30-test suite (node --test, no framework) covering the exact functions
  that have regressed in previous rounds:
    * parseStoreArg / parseStoreNumber / storeEmail (helpers)
    * greetingForBrand (brand fallback)
    * formatE911Address / formatSuite (google response reducer)
    * parseNextLink (RFC-5988 pagination — now exported)
    * runStep success / non-critical / critical / no-bot paths
- test/setup.js stubs required env vars so any src/ module can be
  imported cleanly; loaded via --import once per test process
- npm run test wired up; directory form works on Node 20 + 22
- .gitea/workflows/ci.yml runs lint + format:check + test on push/PR
  to main, on Node 22 to match the runtime image
- Exclude test/ and .gitea/ from the Docker build context
- Exclude docker/ from bot Prettier scope (remote-agent is its own
  sub-project with its own tooling)

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-07-06 16:07:02 -04:00