Name every draft, and put Vertrag before Angehoerige
Two things the dashboard and the hire wizard were getting wrong.
The drafts card named the author only on other people's drafts. With
foreign and own rows side by side that reads as an inconsistency, not as
information: the eye has to work out that a missing name means "mine".
Now every row says it, "von mir" on the own ones -- the same wording the
notes in the bell already use.
In the wizard, Angehoerige stood before Vertrag. What a contract is made
of -- entry date, working days, a fixed term -- is on paper before the
conversation happens; relatives the person brings along, often on the
first day. The optional step came before the one the hire rests on.
Swapping them meant touching the part that would have broken silently:
the per-step validation was a positional list that had to line up with
STEP_LABELS by hand. Reordered labels alone would have left the checks
where they were -- "Weiter" on Vertrag would have validated the
relatives and waved an empty entry date through, until the database
refused it at the end. The checks are keyed by step name now, so they
travel with the step.
Drafts saved before this land on the step number they stored, which now
points at a different page. Nothing is lost -- the payload carries every
field -- but somebody resuming an older draft may open on Vertrag where
they left Angehoerige.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@
The picker in the bell now governs both lists, so note_subscriptions is
renamed to colleague_subscriptions -- a name that only mentions notes would
mislead the next reader.
Reading and writing a draft now reach differently far. hire_drafts_owner
(for all) is split into four policies: select lets in your own drafts and
those of the people you added, while insert/update/delete stay with the
owner. A draft is unfinished work with no lock and no history; two people
writing into the same row would overwrite each other silently.
That split forces a change in the actions: a policy does not reject a write,
it lets it hit no rows. saveHireDraft and deleteHireDraft now read the row
count instead of reporting success over a row that never changed.
The card shows a foreign draft with its author and without Fortsetzen or
Loeschen -- offering a button that reliably ends in a database error is a
promise without cover.
check-schema-types.mjs learns `alter table ... rename to`; without it the
drift check reports one rename as two errors.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
- components/hire/: 4-step Hire Wizard (Person/Position/Vertrag/
Zusammenfassung) matching sec4.4, with a HireWizardProvider context so it
can be opened both from the global "+ Neueinstellung" button and from a
"Fortsetzen" link on a saved draft.
- actions/hireDrafts.ts: save/delete hire_drafts (owner-scoped RLS already
in place from Phase 1). Dashboard now shows the "Entwuerfe" card the
Phase 1 plan deferred, since the wizard it depends on now exists.
- lib/positions.ts: shared open-positions loader (position number, org
breadcrumb, resolved manager name) used by both the wizard and (later)
the Positions page.
Two real bugs found via live testing and fixed in supabase/functions.sql:
1. hire_employee/rehire_employee: a two-branch CASE returning bare string
literals defaults to `text`, not the target enum, so `status = case
when ... then 'Geplant' else 'Aktiv' end` failed against the
employment_status column. Fixed with an explicit ::employment_status
cast on the whole CASE expression.
2. Postgres precedence gotcha: ->> and || sit at the *same* precedence
tier and left-associate, so `payload->>'first_name' || ' ' ||
payload->>'last_name'` does not group the way it reads - it tries to
apply ->> to an intermediate text value and fails with "operator does
not exist: text ->> unknown". Fixed by parenthesizing every ->>'...'
expression that participates in a || chain.
Also fixed: hire_employee referenced v_position.title outside the branch
that assigns v_position, raising "record not assigned" whenever a hire
wasn't tied to a position_id; extracted a v_job_title variable instead.
Verified live end-to-end: wizard search -> select position -> submit
creates the employee, closes the position, and writes matching
employee_history + audit_log rows atomically.
