Skip to content

Commit 7466adb

Browse files
committed
docs(changelog): the Apple paths a preserved recording depends on to be retryable
1 parent 0148439 commit 7466adb

1 file changed

Lines changed: 10 additions & 3 deletions

File tree

‎CHANGELOG.md‎

Lines changed: 10 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -52,9 +52,16 @@
5252
`dispose-on-failed-finish` themselves, because their retries re-read a log file and a status file that
5353
cleanup never touches. A finish also states why it was asked for: `capture` may preserve, while
5454
`disposal` — session teardown, the one caller that owns forced cleanup — disposes whatever a failed
55-
finish left. Expect one trade: while a preserved recording is open, `record start` on that device is
56-
refused until its session runs `record stop` or closes. `perf stop` no longer memoizes a refused
57-
finish, so its second attempt re-pulls the trace the first one preserved.
55+
finish left. Two Apple paths had to become genuinely retryable for that to be true: the runner path
56+
memoized its refused stop and handed the same rejection to every later `record stop` without asking
57+
the runner again, and the simulator path rejected a recording purely because `simctl recordVideo` had
58+
exited non-zero — discarding a video the recorder had already written and leaving an exit code no
59+
retry could change. A stopped recorder is now an observation rather than a failure: the exit is
60+
disclosed on the completion and the file is collected, so those recordings export instead of erroring.
61+
Expect one trade: while a preserved recording is open, `record start` on that device is refused until
62+
its session runs `record stop` or closes, and a file no retry can rescue — a recorder killed mid-write
63+
— now says so with `recording-output-unplayable` and names session close as the way out. `perf stop`
64+
no longer memoizes a refused finish, so its second attempt re-pulls the trace the first one preserved.
5865

5966
- Changed: `record stop` no longer carries a start-trim step no recorder could arm. The trim cut the
6067
interval between recorder start and target-app readiness, but the runner's `recordStart` answer has

0 commit comments

Comments
 (0)