maziggy dc6217f4e6 fix(vp): scope saved AMS mapping to the printer it was resolved against ...

Round-2 review fixes for #2700.

    Blocking: the toggle didn't actually gate the archive write. archive.py's
    promotion fired for any print_data carrying ams_mapping, but bambu_mqtt's
    request-topic interception captures ams_mapping unconditionally for every
    print source (slicer-direct LAN prints included). Since main.py's
    real-printer auto-archive path forwards the full MQTT payload as
    print_data, every archive on any install — VP or not — grew
    extra_data.slicer_ams_mapping. Fixed by replacing the print_data-sniffing
    with an explicit `slicer_ams_mapping` param on archive_print() that only
    the VP-queue path (already gated on save_ams_mapping) ever passes.

    Blocking: a saved mapping could get reused on a printer it was never
    resolved against — tray IDs only mean something relative to one printer's
    AMS layout. extra_data.slicer_ams_mapping is now stored as
    {mapping, printer_id} instead of a bare array:
    - add_to_queue's fallback only fires when the reprint's target printer_id
      matches the mapping's origin printer.
    - The frontend's archiveAmsMapping only surfaces (and the Mapping button
      only appears) when the print modal's selected printer matches too.
    - A model-based VP (target_printer_id=None, no MQTT bridge to any real
      printer) never stamps a mapping in the first place — there's no live AMS
      layout for the slicer to have resolved tray IDs against.

    Also from review:
    - Multi-plate archives now get the Mapping button too (the per-plate
      FilamentMapping loop was missing archiveAmsMapping entirely).
    - Added coverage for the previously-untested late-MQTT archive patch path
      (_restamp_recent_queue_item), including the model-based-VP skip case.
    - usingArchiveMapping now also resets on printer change, not just
      plate/archive (it already worked via the printer-scoping above, but is
      now an explicit dependency too).
    - The Mapping button's revert (OFF) now undoes only the slots it itself
      set, not every manual pick in scope — matches the comment above it.
    - Added a comment on why negative-value slots (external spool) are
      skipped rather than cleared when applying a saved mapping.

2026-08-02 09:48:32 +02:00

..

app

fix(vp): scope saved AMS mapping to the printer it was resolved against

2026-08-02 09:48:32 +02:00

tests

y fix(camera): rotate every still exactly once, and cover the sources that let ffmpeg write the file

2026-08-02 09:47:29 +02:00

__init__.py

…