An incident created on test failure already inherits its assignees from the
test case owners, but the resolution workflow always entered NewStage and
nothing ever advanced it. The incident stayed in New, and because the TCRS
mirror only carries assignee details on Assigned records, the Incident
Manager rendered it as unassigned even though the task had assignees.

Drive the workflow's own "assign" transition right after the incident task
is created when assignees resolved, so the incident lands on the Assigned
stage with the owner attached and the TCRS record carries the assignee.
Incidents without owners are untouched and stay in New.

Stage and transition ids of TestCaseResolutionTaskWorkflow move into
IncidentWorkflowStages so the TCRS sync handler and the new assignment path
cannot drift apart.

Log a warning when an incident has assignees but is not at the New stage,
so a broken assumption about synchronous workflow start surfaces instead of
silently leaving the incident unassigned.

Assert the typed assignee id on the Assigned TCRS record rather than
substring-matching the serialized details blob, which could pass on an
incidental match elsewhere in the payload.

The incident raised by the pipeline re-run in "Resolving incident & re-run
pipeline" is created after the table has an owner, so it now opens as Assigned
rather than New and the hardcoded expectation no longer holds.

Give acknowledgeTask an initialStatus parameter defaulting to New, and pass
Assigned at that one call site. The other three call sites act on incidents
raised before ownership is claimed and keep the default.

…-assigned

…-assigned

table3 is given an owner in beforeAll, so the incident raised by its
Consistency failure is auto-assigned on creation and never opens as New.

The Assigned stage offers only reassign and resolve, so an auto-assigned
incident cannot be acknowledged. The rerun step in "Resolving incident &
re-run pipeline" only needs to confirm a fresh incident exists before checking
the open/closed counts, so assert its status directly.

Adds verifyIncidentStatus for that and restores acknowledgeTask to its original
shape, since every remaining caller acts on an incident raised before the table
gained an owner.

…-assigned

Callers go on to openIncidentTaskTab, whose Incident tab only exists on the
test case details page. acknowledgeTask navigated there as a side effect;
verifyIncidentStatus replaced it at one call site and stopped short, leaving
the run on the table's Data Quality tab.