Handle dict-shaped reporter events when deriving watcher host failures.
- parse reporter JSON events with type/message/severity fields
- preserve existing support for list-shaped event records
- record the false-PASS failure mode in ATVM automation run learnings
Prefer failing and terminal host results when merging supplemental
subrun artifacts so a later partial artifact cannot overwrite a more
authoritative outcome.
Also point ATVM agent guidance at the ATVM-specific git guide and add
that guide to document draft-only commit handling and the default SSH
command pattern.
Handle Currents "Cloud Run Finished" table rows where the trailing
"s" in long duration values wraps onto its own continuation line.
Instead of appending that standalone continuation to the end of the
host row, drop the wrapped row and rely on the existing duration
parser to accept values without the trailing "s". This preserves all
host rows in parent summary parsing for completed non-categorized runs.
Also record the failure mode and recovery guidance in the ATVM run
learnings doc.
Teach generated-spec TEST FLOW parsing to honor test-install-only runtime gates, including single-quoted Cypress.env checks, and suppress plugin branches when install-only mode disables them.
Document the 2026-04-16 install-only flow mismatch so future watcher updates keep Mattermost TEST FLOW aligned with the actual generated spec path.
Tighten `Cloud Run Finished` parsing so categorized watcher recovery only starts on real summary headers, stops at the grouped run's `Recorded Run` URL, and parses only actual summary-table host rows.
Also merge grouped per-host reporter artifacts into categorized recovery so completed grouped batches keep the correct host membership and Mattermost posts remain stable for both categorized and non-categorized runs.
Adjust parent-run summary parsing so the watcher does not stop at the "Recorded Run" detection log line and miss later host rows in Mattermost status output.
Document the 2026-04-15 failure mode in ATVM run learnings so future watcher/reporting work preserves the broader Cloud Run Finished block parsing rule.
- skip report artifacts that disappear during watcher post-processing
- use safe mtime handling for XML and mochawesome discovery
- prevent watcher crashes from transient FileNotFoundError races
- update the non-categorized watcher path to fall back to the parent Cloud Run Finished summary when only the final check-xml-files reporter XML is available
- prevent single-host runs from being marked TERMINATED without host results when the controller log already contains the real host outcome
- allow the watcher to populate host results and post Mattermost successfully for completed single-host runs with reused build names
- update the ATVM status template to replace the FUNCTIONALLY section with a TEST FLOW section that shows the generic numbered run steps once for the whole test
- update the ATVM automation guide to describe TEST FLOW as the standard status-report section instead of FUNCTIONALLY
- update the watcher-generated status output so Mattermost and local status responses use the same TEST FLOW section
- add a 2026-03-27 run learning recording the move from FUNCTIONALLY to TEST FLOW for future ATVM reporting
- update the categorized watcher to fall back to earlier per-spec Run Finished entries when a grouped Cloud Run Finished section is present but does not contain host rows
- broaden duration parsing so colon-form timings like 13:59 and 15:32 from per-spec Cypress summaries are converted into seconds correctly
- make grouped Mattermost posting resilient across categorized batches whose cloud-summary output is incomplete or inconsistent, not just a single distro family
- update the categorized watcher to recognize grouped Cloud Run Finished summaries that are followed by the controller log message as well as the older literal form
- allow grouped host summaries like the multi-host ubuntu batch to be parsed and posted even when the alternate controller log format is used
- prevent completed grouped runs from being left in RUNNING state solely because the watcher missed the alternate grouped summary terminator
- update the categorized watcher to parse all ATVM host rows from a grouped Cloud Run Finished block instead of only the first host row
- allow a later terminal grouped summary to replace an earlier RUNNING state for the same logical group so grouped posts are emitted with the full host set
- fix multi-host same-group reporting so batches like ubuntu22.04 plus ubuntu24.04 are posted as one grouped result with both hosts included
- update the categorized watcher to mark a grouped subrun complete once its final check-xml-files.ts artifact exists and the next active host belongs to a different inferred group
- prevent a finished group such as ubuntu from being kept in RUNNING state after execution has already advanced into the next category
- allow grouped Mattermost posts to be emitted at category boundaries instead of being blocked by stale active-group carryover
- update the watcher to merge consecutive categorized child build ids into one logical grouped subrun when they resolve to the same real host-based group
- prevent one real grouped run from being fragmented across multiple watcher subruns just because the raw ci-build-id label changes between hosts
- keep grouped Mattermost posting aligned with the inferred group identity instead of the unstable raw child build id
- update the categorized watcher to use the latest matching host reporter artifact written before grouped finalization so grouped results can be posted as soon as the host run is actually done
- reduce dependence on the later parent Cloud Run Finished summary block for grouped-run completion reporting
- keep grouped runs from waiting until near parent-run completion when real host-level completion evidence is already available
- keep categorized grouped sub-runs in RUNNING state when only the grouped check-xml-files.ts artifact has appeared and the parent log has not yet exposed the real host summary
- prevent blank Mattermost grouped-run posts with empty host tables, zero counts, and zero/invalid timing caused by posting too early
- require the categorized watcher to wait for host-level grouped summary data before treating the grouped run as complete
- update the watcher to stop trusting misleading categorized child build labels when they do not match the host/spec actually being executed
- infer the reported categorized group name from the actual host being run, so mismatched labels like ubuntu-batch for a Red Hat host are corrected in status reporting
- document the categorized watcher workaround in the ATVM guide, watcher design, and watcher README without changing the underlying ATVM runner scripts
