DeleteUserWithCascadeAsync never removed UserPermissionOverrides rows, and
the FK on UserId is Restrict. Once an override was saved for a member, the
admin hard-delete (DeleteUserAsync -> DeleteUserWithCascadeAsync) failed the
foreign key inside the transaction and the controller surfaced it as a 400,
so the user was never deleted. Add an explicit ExecuteDelete on
UserPermissionOverrides before the user is removed, matching how UserRoles is
already cleared in the same method (no schema change).
SetOverrideAsync was check-then-insert on the composite key with no
DbUpdateException handling, so two concurrent PUTs for the same
(UserId, PermissionKey) let the loser violate PK_UserPermissionOverrides and
return 500. Catch the conflict, detach the pending insert, and converge by
updating the persisted row to the caller's requested state.
* refactor(api): enforce service and data-service boundaries
* refactor(api): complete feature service boundaries
* refactor(identity): enforce service and data boundaries
* refactor(vendors): enforce service and data boundaries
* refactor(workorders): enforce service and data boundaries
* refactor(backend): enforce architecture and optimize dispatch
* style(backend): format changed architecture files
* fix(architecture): address backend review follow-ups
* fix(backend): sanitize exception disclosure in changed API endpoints
Replace raw exception-message disclosure (ex.Message) returned to API
callers with a stable sanitized public message plus correlated structured
internal logging, across the endpoints changed in this PR.
- Add SanitizedErrors helper: logs the original exception at Error with a
generated correlation id and returns a stable public message referencing
it so support can trace without exposing internals.
- Inject ILogger<T> into the 14 changed controllers and route every
ex.Message/dbex.Message disclosure through the helper, preserving status
codes, response shapes, and business data (e.g. OpenWorkOrders).
- Leave FluentValidation (vex.Errors) and existing fixed-message catches
untouched; out-of-scope controllers (Account/Contact/Employee/Asset/
PMSchedule) are unchanged.
- Add focused tests proving internal exception text is not returned and
that Error logging carrying the original exception is invoked.
* fix(architecture): abstract job run state access
* style: format board update service
* test: use collection assertion idiom
