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 dd6844a
Browse filesBrowse the repository at this point in the historyBrowse files
feat(catalog): a confirmation that carries its own wait, and a text message
ASK_CONFIRM moves from ValueKind.TEXT to a new ValueKind.CONFIRM: the
question and the seconds it waits for an answer are one decision, and the
editor draws one control per kind, so the wait had to become part of the
kind rather than a special case for one action. Bounds are 5-60 s;
Action.number carries the wait and 0 — every rule saved before the field
existed — means ASK_CONFIRM_DEFAULT_SECONDS (10), not the floor.
SEND_SMS is new, with ValueKind.SMS: the number in text, the contact it was
picked from in displayName, the message in payload. Where it goes was read
off the head unit's own APKs rather than guessed. IBtCall carries twenty-six
transactions and not one of them is a message, so the vendor hands-free
service cannot send one; the system settings service, on the other hand,
manages a "MAP Client" profile (18, UUIDs MAP/MNS/MAS), which is the
Bluetooth Message Access Profile in the role a car takes. So the message
leaves through the paired phone.
No @SupportedOn on SEND_SMS: MAP is AOSP, not SAIC routing, so there is
nothing to declare per firmware generation. Whether it works is a bind read
at runtime, which is the same rule the vendor services already follow.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
0 commit comments