v0.23.1

Upgrading

Deploy and run migrations as usual. This release adds no schema of its own.

An installation that already has the Enterprise module working is unaffected — its tables exist. This matters for installations that have not installed it yet, and for every new one.

---

Changed

• Admin → Modules → Install uses the shared drag-and-drop file picker — the ZIP box only opened a
  file chooser on click; it now takes a dropped file too (FileDropZone, as in the importer), posts
  through Inertia so a rejected file shows its error under the drop zone, and shows upload progress.
  The upload is validated by a Form Request (InstallModuleRequest) instead of inline.
• Operator Workstation view follows the selected workstation — switching from Queue to
  Workstation with a workstation picked (e.g. ?workstation=10) now shows only the orders whose
  current step runs there — plus not-yet-started orders whose first step is there, so they can be
  started from that station — and only that workstation's machine state, with an "All workstations"
  link back to the whole line. per_line tracking keeps the whole-line view, and a workstation
  account's assignment alone doesn't filter it — only an actual selection does. The Queue view shows
  those not-yet-started orders as their own "To start at …" cards below "Ready at …", and its
  polling count includes them.

Fixed

• Installing a module from a ZIP in Admin → Modules always failed with "Could not open ZIP file": the upload was stored on the local disk (storage/app/private) but the installer was handed a storage/app/… path, so the file was never found and each attempt leaked its archive.
• Enabling a module did not create the tables it ships, so every screen it contributes answered
  500 with "relation … does not exist" the moment it was opened. Enabling ran migrate, but a
  module's migrations are registered by its service provider, providers are registered at boot, and
  the process doing the enabling booted with the module switched off — so migrate only ever saw
  the application's own paths. It is now pointed at the module's own directory explicitly. Together
  with the ZIP fix above this makes Admin → Modules → Install usable end to end; before, an install
  reported success and then failed on first use.
• Sidebar: a group stayed unhighlighted (and collapsed) on pages a module added to it, and a module's own group never highlighted at all — only core pages lit their group up. Module links now extend the group's match list, so the breadcrumb trail finds them too.
• Uploads over PHP's 32 MB post_max_size (e.g. a backup archive for restore, which the app accepts up to 500 MB) crashed with a bare PostTooLargeException page. The Docker image now allows 512 MB and an oversized body is reported as a flash error (413 for JSON clients) instead.
• Operator Queue: the "All" workstation chip now actually clears the selection (it fell back to
  the workstation remembered in the session).
• Live lists stopped updating after a create or delete until a browser refresh — returning to
  the same list (e.g. the redirect after adding a work order in /admin/work-orders) mounted a new
  synced collection while the old one was still shutting down, and the old one's echo.leave()
  unsubscribed the channel both shared. Collection channels are now reference-counted
  (lib/sharedChannels.js): a channel is left only when its last user releases it, which also stops
  an unmounting list from cutting off the app-wide work_orders_active / issues_open feeds.
• Phantom rows in live lists after a rolled-back write — collection deltas were broadcast from
  model events mid-transaction, so a write that failed later in the same transaction still pushed
  its row into every open list, where it stayed until a refresh. Deltas are now sent after the
  transaction commits and dropped on rollback.
• Operator Workstation view: the machine-state picker uses the shared Dropdown instead of a
  native <select>, and structured extra-data values render as readable text instead of
  [object Object].