Failures from the defaults branch PR: https://github.com/BrowserSync/bslive/actions/runs/22034914640/job/63666068435
All 9 failures are caused by two issues introduced during experiments on the defaults branch.
The server now emits a BsLiveRules event ("bslive rules violated") on startup for watch-based configs. The test utility code doesn't recognise this event kind and logs "cannot continue - probably an unsupported type", causing watch-graceful to timeout waiting for expected messages.
Failing test:
graceful_exit.spec.ts:15— "graceful exit on timeout" (Error: 1000 timeout expired)
Files:
- tests/utils.ts — line ~117-121, the
loose.safeParsefallback logs but doesn't handleBsLiveRules - crates/bsnext_input/src/lib.rs —
BsLiveRulesErrordefinition (~line 294) - crates/bsnext_system/src/start/start_kind/start_from_inputs.rs — where the event is emitted
The OutputLine payload shape changed. Previously each output line had "prefix": "[run]" and no task_id. Now:
prefixis a random short ID like"[M3CpiT]"instead of the fixed"[run]"- A new
task_idfield is present in the payload
This breaks all assertions that check the exact shape of OutputLine objects.
Failing tests:
watcher.before.spec.ts:15— "running before" (expected"[run]", got random prefix + extratask_idfield)watcher.cancellation.spec.ts:15— "cancellation of sibling tasks" (can't find"starting succeeding task"line)watcher.fail.spec.ts:15— "ignoring failures in a task sequence" (extratask_idfield)watcher.patterns.spec.ts:15— "patterns in globs" (extratask_idfield)watcher.spec.ts:95— "custom output for index.html" (random prefix +task_id)watcher.spec.ts:115— "custom output for 01.txt" (extratask_id)watcher.spec.ts:135— "custom output for 02.txt" (extratask_id)watcher.spec.ts:155— "without prefix" (extratask_id)
Files:
- tests/watcher.before.spec.ts
- tests/watcher.cancellation.spec.ts
- tests/watcher.fail.spec.ts
- tests/watcher.patterns.spec.ts
- tests/watcher.spec.ts
- tests/graceful_exit.spec.ts
Both issues stem from changes to task execution on the defaults branch. To fix: either revert the output format changes (prefix/task_id) or update the test expectations and utils to handle the new shape.