3 Commits

Author SHA1 Message Date
b9d5f8ef92 T8.4 - CR-012: the delivery location is the shared vocabulary plus fifty feet
Staging is not the pain; the last fifty feet are - the correct floor lay-down,
shark cage or conduit tree instead of material picked at will by whoever is
closest. The Kitting & MIMO section gains:

- Delivery Building / Floor / Sector: the SAME dependent pickers CR-004 built,
  through the same fillLocSelect (which learned an optional field-map instead
  of being copied), reading the same project location lists, storing PATHS.
  A parallel free-text location vocabulary is exactly what CR-004 removed;
  none was added.
- A free-text detail field for the specifics ("Shark cage 7, conduit tree C"),
  persisted as delivDetail.
- deliveryLoc, the composed display string (labels off the shared lists, then
  the detail after a dash) - which is what the CR-011 email already reads
  (kitting_body preferred deliveryLoc from day one, with mimoLoc as the
  pre-CR-012 fallback) and what the package printout now carries as its own
  Delivery Location row.

Verification (each probe run alone): kitting_check.py extended to 26/26 (the
delivery selects are asserted to offer the SAME option list as the CR-004
trio, values persist as paths, the printout carries the composed value);
kitting_notify_check 17/17 now asserting the mail carries CR-012's composed
value, not the fallback. Regression: locations_check 58/58.

Items: CR-012

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-19 12:11:05 -07:00
96387105f4 T8.2 - CR-010: the warehouse owner is an account, on the package
A named person owns fulfillment of the kit - today that is Paul Coonrod,
informally, and everything bottlenecks through him. The package now records
it explicitly:

- The Warehouse owner control is a dropdown of project members (the same
  roster the Owner picker reads). Picking someone stores BOTH the display
  name (kitOwner - exports and old renderers keep working) and the account id
  (kitOwnerId - the routing CR-011's notifications will read at T8.3).
- Confirmed Aug 18: the field lives ON the work package, not the project - a
  package retargeted to a different warehouse notifies the right person
  without touching the project. The wizard gets no field.
- A stored name with no matching account - typed before the field was
  account-backed, or someone since removed from the project - is KEPT as a
  selected "(no account)" option and round-trips unchanged. Removing someone
  from the project breaks nothing.
- The dashboard filters by warehouse owner, options drawn from the owners
  actually present in the data - a filter offering people with nothing to
  fulfill is noise.

Verification (each probe run alone): kitting_check.py extended to 21/21
(T8.2 section: picker, id+name persistence, orphan survival, board filter).
Regression: generalinfo_check 49/49.

Items: CR-010

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-19 12:00:14 -07:00
79c4a36c70 T8.1 - CR-009: kitting statuses are a set, and Micron EUV is not kitting
The statuses: the proposed five, adopted as proposed - Not Started, Picking,
Staged, In Transit, Delivered - as one named constant (KIT_STATUSES) building
the select. They describe fulfillment; the old four ('Open', 'In Progress',
'Kitted', 'Delivered') mixed fulfillment with workflow. A value stored before
the set existed is kept, selected, and shown as "(legacy)" - CR-016's rule
that renamed vocabularies must not orphan recorded data - and round-trips
through collect unchanged.

Micron EUV: the sample SOP is the Micron configuration on record (CR-016 /
T5.7), and it now names kitting:false beside assets:false - off by CR-006
TOGGLE, in both the creator sample and the wizard sample. The section leaves
the form, the rail and the export; its data and model stay exactly where they
are (the probe loads the example package under the Micron sample and finds
its kitting values intact through collect). Any other SOP turns the section
on and it works fully - driven against sopA with everything enabled.

sections_check re-pointed, not relaxed: its sample-map pin said "naming only
assets"; it now says "naming exactly assets (CR-016) and kitting (CR-009)" -
still refusing any section that goes off without a recorded item behind it.

Verification (each probe run alone): NEW tests/kitting_check.py 14/14.
Regression: sections_check 95/95.

Items: CR-009

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-19 11:55:39 -07:00