You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 486e1d6
Browse filesBrowse the repository at this point in the historyBrowse files
fix(daemon): settle the first capture after any iOS open that did not read the app (#3376)
* fix(daemon): settle the first capture after any iOS open that did not read the app (#3367)
CI evidence across 30 runs of 01-settings.ad: 61 iOS opens reported 54
probe-failed, 7 observable, 0 unobservable, so the unobservable-only mark
from #3356 never fired. Every stale-tap failure followed a probe-failed
open. Mark the next capture for any defined verdict other than observable,
mirroring the open-policy fixed-delay condition; an open with no verdict
(physical device) still does not mark.
Co-Authored-By: Claude Sonnet 5.5 <[email protected]>
* docs(open): scope the post-open settle to opens that report a verdict
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* docs(daemon): describe the post-open mark by what the verdict proves, not its cause
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Sonnet 5.5 <[email protected]>
'gap: the response reports the dispatch only; only opt-in --verify/--settle capture post-action evidence into it. The deferred marks set after dispatch (Android snapshot freshness after press/click, post-gesture stabilization when the request sets postGestureStabilization) are judged by the next capture, never in this response. An iOS Simulator open whose launch watch ended unobservable marks the next capture in that session to wait for a quiet surface (markPostOpenStabilization); that steadies where a target is resolved, not whether the tap landed. The iOS ambiguous-failure corroboration reconsiders only a thrown runner error.',
169
+
'gap: the response reports the dispatch only; only opt-in --verify/--settle capture post-action evidence into it. The deferred marks set after dispatch (Android snapshot freshness after press/click, post-gesture stabilization when the request sets postGestureStabilization) are judged by the next capture, never in this response. An iOS Simulator open that did not read the launched app (any postOpenObservation other than observable) marks the next capture in that session to wait for a quiet surface (markPostOpenStabilization); that steadies where a target is resolved, not whether the tap landed. The iOS ambiguous-failure corroboration reconsiders only a thrown runner error.',
0 commit comments