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
This commit was created on GitHub.com and signed with GitHub’s verified signature.
Fixed
get-note-content by exact id no longer reports a note that exists as not_found
when Notes.app is slow (#316). Root cause: the AppleScript runner retries a timed-out
call, and when the retry then failed differently (Notes.app, still busy with the
abandoned event, answering "Can't get note id") that second error replaced the timeout,
so a 30-second AppleEvent timeout came back as a definitive not_found. The runner now
keeps the first timeout's diagnostic and appends the retry's error ("… A retry then
failed: …"), so the result stays timeout_indeterminate for every tool, not just this
one.
get-note-content by id reads existence and the password flag from the NoteStore
database (the same exact-ID read get-note-structure makes) when it holds the note, and
skips the AppleScript metadata step that timed out in #316. That leaves one Notes.app
round trip (the body) instead of two. Without Full Disk Access, or for a note not yet saved
to the database, the AppleScript metadata read runs as before. Content and contentHash
still come from the AppleScript body, so hashes stay compatible with every write guard.
A body read that Notes.app answers with "not found" is checked against the database:
when the store still holds the live note, the result is notes_unavailable (Notes.app
busy, retry), not not_found; when it does not, the result is not_found (it was
previously misclassified as verification_failed).
A body-read timeout on a note with no attachments no longer blames a large image. The
error says Notes.app was most likely busy, suggests a retry or a longer timeoutSeconds,
and points to get-note-structure for a read that does not need Notes.app.