mirror of
https://github.com/yusufipk/dikte.git
synced 2026-09-11 10:56:10 +00:00
Keep the front where the dictation started on macOS
Recording goes through ffmpeg's avfoundation input, and opening a capture
session there brings the process that did it to the front. ffmpeg is a
child of Dikte with no bundle of its own, so macOS credits the move to
Dikte: the window the user was dictating into goes inactive, its caret
stops, its title bar greys out, and the Cmd+V at the end of the run lands
somewhere other than the document it was meant for. Measured with a
TextEdit document in front:
press the shortcut front = TextEdit
recorder.start returns front = TextEdit
89 ms later front = Dikte
Nothing about the capture session can be asked not to do this. Starting
ffmpeg in its own session, and clearing __CFBundleIdentifier from its
environment, were both tried and both measured to make no difference, so
it is undone instead: the application in front is noted before the
indicator goes up, and a short watch puts it back the moment Dikte takes
the front. Measured at 99 ms from the moment it is taken, against the
title bar staying grey for the whole dictation before. All three ways in
do it, dictation, agent and meeting, since all three open the same
capture.
The indicator had a share of the same problem and needed AppKit for it
too, so mac_window.py carries both. An NSPanel is hidden by the system
the moment its application stops being the active one, which for a
dictation indicator is immediately, and ordering one to the front brings
its application with it unless it carries the nonactivating style bit.
Neither is reachable through Qt.
The runtime is loaded in _appkit() rather than at import, the way paste.py
loads its frameworks in _macos_api(), so the module imports on a machine
with no AppKit and the tests below run there as well: 52 new tests, none
of them skipped anywhere.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
This commit is contained in:
Gökhan
@@ -440,6 +440,45 @@ class MacOS(ClipboardContract, DikteTest):
|
||||
self.assertFalse(paste.paste_ready())
|
||||
|
||||
|
||||
class MacPasteGoesWhereTheDictationStarted(MacOS):
|
||||
"""The keys land in the frontmost window, so the front is what decides
|
||||
where a transcript ends up."""
|
||||
|
||||
def setUp(self):
|
||||
super().setUp()
|
||||
from dikte import mac_window
|
||||
self.mac_window = mac_window
|
||||
self.activated = []
|
||||
self.patch_attr(mac_window, "activate", self.activated.append)
|
||||
|
||||
def frontmost(self, dikte_is):
|
||||
self.patch_attr(self.mac_window, "is_frontmost", lambda: dikte_is)
|
||||
|
||||
def test_a_dikte_that_took_the_front_hands_it_back_before_pressing(self):
|
||||
self.frontmost(True)
|
||||
paste.press("cmd+v", focus=4242)
|
||||
self.assertEqual(self.activated, [4242])
|
||||
self.assertEqual([event for _, event in self.api.posted], [1001, 1002])
|
||||
|
||||
def test_another_application_in_front_is_where_the_user_went_and_is_left(self):
|
||||
self.frontmost(False)
|
||||
paste.press("cmd+v", focus=4242)
|
||||
self.assertEqual(self.activated, [])
|
||||
|
||||
def test_a_run_that_remembered_nobody_asks_nothing(self):
|
||||
self.frontmost(True)
|
||||
paste.press("cmd+v")
|
||||
self.assertEqual(self.activated, [])
|
||||
|
||||
def test_the_front_is_handed_back_only_once_macos_trusts_dikte(self):
|
||||
"""Pulling the user out of their window and then failing to type would
|
||||
be the worst of both."""
|
||||
self.frontmost(True)
|
||||
self.api.trusted = False
|
||||
with self.assertRaises(paste.PasteError):
|
||||
paste.press("cmd+v", focus=4242)
|
||||
self.assertEqual(self.activated, [])
|
||||
|
||||
class MacClipboardSnapshot(DikteTest):
|
||||
def test_every_native_type_is_restored_and_the_files_are_removed(self):
|
||||
directory = tempfile.mkdtemp(prefix="dikte-test-clipboard-")
|
||||
|
||||
@@ -488,6 +488,288 @@ class KdeSettings(Settings):
|
||||
self.assertFalse(window.evdev_enabled.isHidden())
|
||||
|
||||
|
||||
|
||||
|
||||
class FakeAppKit:
|
||||
"""The Objective-C runtime, answering rather than being asked.
|
||||
|
||||
Stands in for what mac_window._appkit() loads, so what the indicator's
|
||||
window would have been sent can be read off `told` on any machine. The
|
||||
selectors come back as the names they were registered under, which is what
|
||||
lets the tests below name the message they mean.
|
||||
"""
|
||||
|
||||
def __init__(self, window=4242, panel=True, mask=0xe):
|
||||
self.objc = self
|
||||
self.window, self.panel, self.mask = window, panel, mask
|
||||
self.told = {}
|
||||
|
||||
def objc_getClass(self, name):
|
||||
return 1 if name == b"NSPanel" else 0
|
||||
|
||||
def selector(self, name):
|
||||
return name.decode()
|
||||
|
||||
def shared(self, _class_name, _selector):
|
||||
return 1
|
||||
|
||||
def ask(self, _to, selector):
|
||||
return self.window if selector == "window" else 0
|
||||
|
||||
def ask_unsigned(self, _to, _selector):
|
||||
return self.mask
|
||||
|
||||
def ask_of_class(self, _to, _selector, _klass):
|
||||
return self.panel
|
||||
|
||||
def tell_bool(self, _to, selector, value):
|
||||
self.told[selector] = value
|
||||
|
||||
tell_unsigned = tell_bool
|
||||
|
||||
|
||||
class MacIndicatorWindow(DikteTest):
|
||||
"""Which of the three AppKit settings the indicator's window is sent.
|
||||
|
||||
The nonactivating bit is the one worth a test of its own: it is legal on an
|
||||
NSPanel and nowhere else, and sending it to a plain NSWindow raises an
|
||||
Objective-C exception, which through ctypes is not something Python can
|
||||
catch. It takes the whole process down. `_is_panel` is the only thing
|
||||
standing between the two, so what it answers has to decide what is sent.
|
||||
"""
|
||||
|
||||
def setUp(self):
|
||||
super().setUp()
|
||||
from dikte import mac_window
|
||||
self.mac_window = mac_window
|
||||
self.patch_attr(mac_window.QGuiApplication, "platformName",
|
||||
staticmethod(lambda: "cocoa"))
|
||||
|
||||
def told(self, **kwargs):
|
||||
"""What keep_on_screen sends an indicator, against this AppKit."""
|
||||
appkit = FakeAppKit(**kwargs)
|
||||
self.patch_attr(self.mac_window, "_appkit", lambda: appkit)
|
||||
widget = overlay_module.Overlay()
|
||||
self.addCleanup(widget.deleteLater)
|
||||
self.addCleanup(widget.close)
|
||||
self.answered = self.mac_window.keep_on_screen(widget)
|
||||
return appkit.told
|
||||
|
||||
def test_a_window_that_is_not_a_panel_is_never_sent_the_style_mask(self):
|
||||
"""The message that would kill the process. The other two still go."""
|
||||
told = self.told(panel=False)
|
||||
self.assertTrue(self.answered)
|
||||
self.assertNotIn("setStyleMask:", told)
|
||||
self.assertIs(told["setHidesOnDeactivate:"], False)
|
||||
self.assertEqual(told["setCollectionBehavior:"],
|
||||
self.mac_window.BEHAVIOUR)
|
||||
|
||||
def test_a_panel_without_the_bit_is_sent_the_mask_with_it_added(self):
|
||||
told = self.told(panel=True, mask=0xe)
|
||||
self.assertEqual(told["setStyleMask:"],
|
||||
0xe | self.mac_window.NONACTIVATING_PANEL)
|
||||
|
||||
def test_a_panel_that_already_has_the_bit_is_left_alone(self):
|
||||
"""Every dictation runs this again, and a mask Cocoa did not need is a
|
||||
window it rebuilds underneath the indicator."""
|
||||
told = self.told(panel=True,
|
||||
mask=0xe | self.mac_window.NONACTIVATING_PANEL)
|
||||
self.assertNotIn("setStyleMask:", told)
|
||||
|
||||
def test_a_window_the_view_does_not_have_yet_is_not_messaged(self):
|
||||
self.assertEqual(self.told(window=0), {})
|
||||
self.assertFalse(self.answered)
|
||||
|
||||
def test_nothing_is_sent_anywhere_but_cocoa(self):
|
||||
"""Every other platform hands out a winId that means something else
|
||||
entirely, and messaging it crashes the run."""
|
||||
self.patch_attr(self.mac_window.QGuiApplication, "platformName",
|
||||
staticmethod(lambda: "offscreen"))
|
||||
self.assertEqual(self.told(), {})
|
||||
self.assertFalse(self.answered)
|
||||
|
||||
|
||||
class GivingTheFrontBack(DikteTest):
|
||||
"""Opening the microphone brings Dikte to the front, and the window the
|
||||
user was dictating into goes inactive with the caret in it. Nothing can be
|
||||
asked of the capture session, so the front is put back afterwards.
|
||||
|
||||
The watch is driven by hand here: what matters is what it decides, not how
|
||||
long Qt takes to tick.
|
||||
"""
|
||||
|
||||
def setUp(self):
|
||||
super().setUp()
|
||||
from dikte import app as dikte_module
|
||||
from dikte import mac_window
|
||||
self.dikte = dikte_module
|
||||
self.activated = []
|
||||
self.patch_attr(mac_window, "activate", self.activated.append)
|
||||
self.ticks = []
|
||||
outer = self
|
||||
|
||||
class FakeTimer:
|
||||
"""Records what it was asked to do and hands over the tick."""
|
||||
|
||||
def __init__(self, _parent):
|
||||
self.interval = None
|
||||
self.running = False
|
||||
outer.ticks.append(self)
|
||||
|
||||
def setInterval(self, milliseconds):
|
||||
self.interval = milliseconds
|
||||
|
||||
def start(self):
|
||||
self.running = True
|
||||
|
||||
def stop(self):
|
||||
self.running = False
|
||||
|
||||
@property
|
||||
def timeout(self):
|
||||
return self
|
||||
|
||||
def connect(self, slot):
|
||||
self.tick = slot
|
||||
|
||||
self.patch_attr(dikte_module, "QTimer", FakeTimer)
|
||||
|
||||
class BareDikte:
|
||||
"""As much of the application as this one method touches."""
|
||||
|
||||
app = None
|
||||
_front_watch = None
|
||||
_the_front = dikte_module.Dikte._the_front
|
||||
_give_the_front_back = dikte_module.Dikte._give_the_front_back
|
||||
_stop_watching_the_front = dikte_module.Dikte._stop_watching_the_front
|
||||
|
||||
self.bare = BareDikte
|
||||
|
||||
def watching(self, was_in_front, dikte_in_front, on=None):
|
||||
from dikte import mac_window
|
||||
self.patch_attr(mac_window, "is_frontmost", lambda: dikte_in_front)
|
||||
dikte = on if on is not None else self.bare()
|
||||
dikte._give_the_front_back(was_in_front)
|
||||
return self.ticks[-1] if self.ticks else None
|
||||
|
||||
def test_the_front_goes_back_to_whoever_had_it(self):
|
||||
watch = self.watching(4242, dikte_in_front=True)
|
||||
watch.tick()
|
||||
self.assertEqual(self.activated, [4242])
|
||||
self.assertFalse(watch.running) # one is enough; it stops watching
|
||||
|
||||
def test_a_front_that_was_never_taken_is_left_where_it_is(self):
|
||||
"""The microphone does not always take it, and pulling an application
|
||||
forward that is already there is one flicker for nothing."""
|
||||
watch = self.watching(4242, dikte_in_front=False)
|
||||
watch.tick()
|
||||
self.assertEqual(self.activated, [])
|
||||
self.assertTrue(watch.running) # still waiting for the moment
|
||||
|
||||
def test_it_gives_up_rather_than_watching_for_ever(self):
|
||||
watch = self.watching(4242, dikte_in_front=False)
|
||||
with mock.patch.object(self.dikte.time, "monotonic",
|
||||
return_value=self.dikte.time.monotonic() + 60):
|
||||
watch.tick()
|
||||
self.assertFalse(watch.running)
|
||||
self.assertEqual(self.activated, [])
|
||||
|
||||
def test_a_dictation_started_in_dikte_itself_watches_nothing(self):
|
||||
"""Settings is a window of ours, and the front is already where it
|
||||
belongs."""
|
||||
self.assertIsNone(self.watching(os.getpid(), dikte_in_front=True))
|
||||
|
||||
def test_a_second_recording_calls_off_the_watch_the_first_one_left(self):
|
||||
"""The older watch remembers where the older recording started, and by
|
||||
now that is the wrong window to be pulling forward."""
|
||||
dikte = self.bare()
|
||||
first = self.watching(4242, dikte_in_front=False, on=dikte)
|
||||
second = self.watching(1111, dikte_in_front=False, on=dikte)
|
||||
self.assertFalse(first.running)
|
||||
self.assertTrue(second.running)
|
||||
second.tick() # and the survivor is the new one
|
||||
self.patch_attr(self.mac_window_module(), "is_frontmost", lambda: True)
|
||||
second.tick()
|
||||
self.assertEqual(self.activated, [1111])
|
||||
|
||||
def test_a_recording_nobody_needs_watching_for_still_calls_off_the_old_one(self):
|
||||
"""Starting the next one from Dikte's own window is not a reason to
|
||||
leave the last one's watch running."""
|
||||
dikte = self.bare()
|
||||
first = self.watching(4242, dikte_in_front=False, on=dikte)
|
||||
self.watching(os.getpid(), dikte_in_front=False, on=dikte)
|
||||
self.assertFalse(first.running)
|
||||
self.assertIsNone(dikte._front_watch)
|
||||
|
||||
def test_nobody_in_front_is_nobody_to_go_back_to(self):
|
||||
self.assertIsNone(self.watching(None, dikte_in_front=True))
|
||||
|
||||
def mac_window_module(self):
|
||||
from dikte import mac_window
|
||||
return mac_window
|
||||
|
||||
|
||||
class EveryRecordingProtectsTheFront(DikteTest):
|
||||
"""Three ways in, dictation, agent and meeting, and all three open the same
|
||||
avfoundation capture, so all three take the front the same way. What is
|
||||
checked here is the order: the front has to be noted before the microphone
|
||||
is opened, and the watch armed after, or there is nothing to go back to.
|
||||
"""
|
||||
|
||||
def setUp(self):
|
||||
super().setUp()
|
||||
from dikte import app as dikte_module
|
||||
self.dikte = dikte_module
|
||||
self.order = []
|
||||
|
||||
def app(self, **attributes):
|
||||
"""A Dikte that records the order it does things in, and nothing else.
|
||||
|
||||
The methods under test are called unbound against it, so the stand-ins
|
||||
go on the object rather than on the class.
|
||||
"""
|
||||
dikte = mock.Mock(**attributes)
|
||||
dikte._run_id = 0
|
||||
dikte.conf = {"mic_target": "", "max_seconds": 60,
|
||||
"meeting_mic_target": "", "meeting_system_target": "",
|
||||
"meeting_max_seconds": 60}
|
||||
dikte._the_front.side_effect = lambda: self.order.append("noted") or 4242
|
||||
dikte._give_the_front_back.side_effect = (
|
||||
lambda pid: self.order.append(f"watching {pid}"))
|
||||
dikte._begin_recording = (
|
||||
lambda owner: self.dikte.Dikte._begin_recording(dikte, owner))
|
||||
dikte.recorder.start.side_effect = (
|
||||
lambda *_a: self.order.append("microphone"))
|
||||
dikte.meeting_recorder.start.side_effect = (
|
||||
lambda *_a: self.order.append("microphone"))
|
||||
return dikte
|
||||
|
||||
def test_a_dictation_notes_the_front_before_the_indicator_is_even_shown(self):
|
||||
dikte = self.app(state=self.dikte.IDLE, recording=False)
|
||||
dikte.overlay.show_recording.side_effect = (
|
||||
lambda *_a: self.order.append("indicator"))
|
||||
self.dikte.Dikte.start(dikte)
|
||||
self.assertEqual(self.order,
|
||||
["noted", "indicator", "microphone", "watching 4242"])
|
||||
|
||||
def test_the_agent_does_the_same(self):
|
||||
dikte = self.app(ask_state=self.dikte.IDLE, recording=False)
|
||||
self.dikte.Dikte.start_ask(dikte)
|
||||
self.assertEqual(self.order, ["noted", "microphone", "watching 4242"])
|
||||
|
||||
def test_a_meeting_does_the_same(self):
|
||||
"""The one most worth protecting: the user is in a call."""
|
||||
dikte = self.app(meeting_state=self.dikte.M_IDLE)
|
||||
dikte.meeting_recorder.active = True
|
||||
self.dikte.Dikte.start_meeting(dikte)
|
||||
self.assertEqual(self.order, ["noted", "microphone", "watching 4242"])
|
||||
|
||||
def test_a_meeting_whose_microphone_never_opened_watches_nothing(self):
|
||||
dikte = self.app(meeting_state=self.dikte.M_IDLE)
|
||||
dikte.meeting_recorder.active = False
|
||||
self.dikte.Dikte.start_meeting(dikte)
|
||||
self.assertEqual(self.order, ["noted", "microphone"])
|
||||
|
||||
class Overlay(DikteTest):
|
||||
def overlay(self, **kwargs):
|
||||
widget = overlay_module.Overlay(**kwargs)
|
||||
|
||||
+13
-3
@@ -32,7 +32,8 @@ class Chain(DikteTest):
|
||||
transcript="uh, book it for Thursday",
|
||||
cleaned="Book it for Thursday.",
|
||||
cleanup_error=None, answer=("Booked.", ""), rms=None,
|
||||
clipboard=b"what was there before", paste_error=None):
|
||||
clipboard=b"what was there before", paste_error=None,
|
||||
focus=None):
|
||||
pipeline = worker.Pipeline(self.conf)
|
||||
done, failures, stages, cancels = [], [], [], []
|
||||
pipeline.finished.connect(lambda *args: done.append(args))
|
||||
@@ -60,7 +61,8 @@ class Chain(DikteTest):
|
||||
"copy": copy, "copy_bytes": copy_bytes, "press": press,
|
||||
"read_clipboard": read_clipboard}
|
||||
pipeline._work(self.wav, duration,
|
||||
self.rms if rms is None else rms, ask, paste_override)
|
||||
self.rms if rms is None else rms, ask, paste_override,
|
||||
focus)
|
||||
return {"done": done, "failures": failures, "stages": stages,
|
||||
"cancelled": cancels, **calls}
|
||||
|
||||
@@ -72,7 +74,15 @@ class Chain(DikteTest):
|
||||
self.assertEqual(run["done"][0],
|
||||
("uh, book it for Thursday", "Book it for Thursday.", ""))
|
||||
run["copy"].assert_called_once_with("Book it for Thursday.")
|
||||
run["press"].assert_called_once_with(self.conf["paste_shortcut"])
|
||||
run["press"].assert_called_once_with(self.conf["paste_shortcut"],
|
||||
focus=None)
|
||||
|
||||
def test_the_paste_is_told_where_the_dictation_started(self):
|
||||
"""Whoever was in front when the recording began is where the keys are
|
||||
meant to go, and the press is the only part that can act on it."""
|
||||
run = self.run_chain(focus=4242)
|
||||
run["press"].assert_called_once_with(self.conf["paste_shortcut"],
|
||||
focus=4242)
|
||||
|
||||
def test_the_stages_are_named_as_they_happen(self):
|
||||
run = self.run_chain()
|
||||
|
||||
Reference in New Issue
Block a user