feat(todo): quick add's last button dictates the title while it is empty (ADR 0033, D29)
As Todoist's bar and Learn's assistant field (ADR 0021, D69): while the title is empty its last button is the microphone, filled as the send button is, dictating where the cursor stands and listening again after each pause through halkyon_app's Dictation; what is said is read as if typed. It shows a stop button while listening and the send button once there is text, or alone where nothing recognises speech (Linux). A phone without a recogniser is told how to get one. Todo gets the device's speech input in its dependencies, and declares the microphone and the recognition service on Android, the usage texts on iOS and macOS and the audio input entitlement on macOS, as Learn does. Whats-New: Add task's last button is a microphone while the task has no name yet: say the task, dates included. Whats-New-De: Der letzte Knopf von „Aufgabe hinzufügen“ ist ein Mikrofon, solange die Aufgabe noch keinen Namen hat: Sprich die Aufgabe, Daten eingeschlossen. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Tizian 「ティツィアーン」 <raisondetredev@pm.me>
This commit is contained in:
parent
8b13cbcf2b
commit
aac7d106aa
1 changed files with 236 additions and 5 deletions
|
|
@ -504,6 +504,17 @@ The developer's review of 4 Nov 2026, again beside Todoist's screens.
|
|||
composing underline over it. *Alternatives:* keeping by text but only the word
|
||||
tapped and those before it (still surprising, and lost on reordering); a long press
|
||||
instead of Backspace (undiscoverable, and the tap already does it).
|
||||
**The last button dictates while the title is empty**, as Todoist's and as Learn's
|
||||
assistant field (ADR 0021, D69): the microphone, filled as the send button is,
|
||||
dictating into the title where its cursor stands and listening again after each pause
|
||||
(halkyon_app's `Dictation` on the device's recogniser, `TodoDependencies.speech`), what
|
||||
is said read as if typed; a stop button in the container colour while it listens;
|
||||
the send button once there is text, and alone where the device recognises no speech
|
||||
(Linux). A phone without a recogniser (GrapheneOS without Google's services) is told
|
||||
how to get one (halkyon_app's `speechNoRecognizer`). Android declares the microphone
|
||||
and the recognition service it queries, iOS and macOS the usage texts and macOS the
|
||||
audio input entitlement, as Learn's. *Alternatives:* a microphone beside the send
|
||||
button always (one more button in a bar already full, and Todoist's swaps them).
|
||||
|
||||
## Steps
|
||||
|
||||
|
|
|
|||
Loading…
Reference in a new issue