Repository navigation
omamobile-omarchy 0.1.17, omamobile-shell 0.2.41: the lock glance after hands-on testing - #181
Merged
Merged
Conversation
… screen goes dark
On the Pixel 7 Pro the glance came back after 15 s on the PIN view, but a
power-button press or a tap that woke the screen went straight to the PIN
view, and "sometimes the view is reset". The lock blanks the display
itself 10 s after the last input, before the glance's 15 s return, and
after that the phone may sleep (omamobile-sleepd), which freezes QML
timers, so the return never ran. Nothing that turned the screen off told
the glance, and the touch that woke a dark screen was delivered to the
lock and counted as the glance's tap to show the PIN view.
Patch 0023 (revised in place, as it was merged today and nothing has
built on it) now brings the glance back, with an empty field, whenever
the locked display goes dark:
- the lock's own blank bumps a blankSerial in the Service;
- a new "omarchy-shell lock dark" call does the same for what blanks it
from outside, without touching the display itself.
The Service keeps displayDark until the display is lit through the lock
again (runWake). While it is dark, a touch or swipe on the glance only
lights the screen through the lock, and so does a touch that starts
within 400 ms of a wake. The 15 s return is now counted on the wall clock
(a 1 s check against when the PIN view last had input), as the lock's
blank timer is, so a sleep in between can't freeze it. The computer lock
and the fingerprint PAM flow are unchanged.
From the maintainer's hands-on test, a tap no longer opens the PIN view.
He touched a notification to read it and got the PIN pad. Now only a
swipe up opens it, from anywhere, a card included: a clear drag of 8 %
of the screen's height, or a quick flick. A typed key and fingerprint
unlock still work as before. A tap on a card opens it to more of its
text (summary up to 2 lines, body up to 4) and runs nothing; a second
tap closes it. A tap elsewhere does nothing, and open cards close when
the glance comes back. A sideways swipe on a card dismisses it, as on a
phone and in the shade. The first 12 px of a drag decide its way: mostly
sideways is a card swipe, mostly vertical is the swipe up, from a card
too. The card slides as a toast does (patch 0017's finishSwipe: past
35 % or a flick it flies off, else it springs back). That one
notification is then forgotten through the notifications service's
forget call (patch 0019), the call the shade makes, so it leaves the
history and the shade too. Nothing of it runs, the next one moves up
from the "more" line, and the last one leaves "No new notifications".
He also asked for the glance to show the notifications the phone's
shade shows. It left out Omarchy's own messages (omarchy-action), bare
notify-send pings and anything over 24 h old, and grouped one card per
app, so three Omarchy messages showed as one. The shade
(omamobile.quicksettings NotificationsModel) lists the newest ten of the
toasts on screen and the history, which Omarchy's service trims to ten
(historyLimit). It has no age limit, leaves no app out and skips only
rows with neither summary nor body. The glance now:
- uses the same rows;
- shows each notification as its own card, newest first, as many as fit
above the hint (up to three);
- counts every one not shown in "N more from A and B" (or "A, B and 2
others");
- gives each card its age ("2m") as its detail;
- names omarchy-action "Omarchy" (with its glyph, as NotificationCard
draws it) and notify-send "Notifications".
Do-not-disturb now follows the toasts rather than hiding every card:
only what still toasts shows (Omarchy's own messages, and a critical
notify-send, NotificationLogic.shouldBypassDnd).
omamobile-shell 0.2.41 makes the calls:
- omamobile-keys calls "lock dark" before the power key turns a locked
screen off, and before it powers off the panel behind a new lock;
- omamobile-sleepd calls it, bounded by a 3 s timeout, just before each
suspend, for a screen that went dark some other way.
An older lock without the call ignores it.
It also fills the glance's weather line, which had nothing to read.
The maintainer's weather is a community home screen widget that fetches
on its own and keeps nothing to borrow, so the shell now owns the
weather, whatever widgets are placed. A new omamobile-weather user
service runs at login and every 30 minutes (omamobile-weather.timer,
enabled for every graphical session). It fetches Open-Meteo's current
weather, with the same API and words as the shell's weather widget (a
test checks the words agree with HomeModel.weatherText for every code).
It writes ~/.local/state/omamobile/weather.json atomically:
{"at", "temperature" (rounded), "unit", "text", "source"}. Where, in
order:
- the phone's position: GeoClue at city accuracy, at most every 3 h,
asked only while the Location switch runs GeoClue's agent, with a hard
timeout. The client is where-am-i, which omamobile-gnss lets through,
so a timer never raises a prompt. A failed fix falls back to the last
one, up to a day old.
- the setting ~/.config/omamobile/weather.json ({"city"} or coordinates,
"unit", "location": "auto" or "fixed"; fixed puts it before the
position).
- the home screen today widget's city (its owl state), or the weather
widget's named place.
A city is looked up once with Open-Meteo's geocoding and kept by name,
and "Town, ST" picks the US state. The unit is the setting's, else the
locale's (°F in the US). There is one try per run, and a failure leaves
the old reading. There is no IP geolocation, and nothing is fetched
without a place. The lock still hides a reading over 3 h old.
Tests:
- test/lock-glance:
- a blank on the PIN view returns to the glance and clears the PIN;
- a wake tap and a wake swipe on a dark display don't show the PIN
view, and neither does a swipe that starts right after a wake;
- a frozen timer (the wall clock 20 s on) returns to the glance;
- taps on a card or on empty space don't show the PIN view;
- a card opens and closes, and only that card;
- a swipe up from a card shows the PIN view;
- a short, slow drag doesn't;
- three notifications from one app are three cards, and five are three
cards plus "2 more from Mail and GitHub";
- a sideways swipe on a card dismisses it (the forget names its file)
without showing the PIN view, and the next one moves up;
- a short one springs back;
- a mostly vertical swipe from a card shows the PIN view;
- the last one dismissed, with hidden text, leaves "No new
notifications";
- the shade's ten, Omarchy's own and 30 h old notifications show, and
do-not-disturb shows only what still toasts.
- test/patches.py checks the Service's blank, dark and wake paths.
- omamobile-shell: keys.sh and sleepd.sh check the call order.
weather.py checks:
- the words against HomeModel.weatherText;
- the settings and units;
- the order of the sources, with GeoClue stubbed (unusable, so the
widget's city or the setting; fixed, so the setting; a fix reused for
3 h and up to a day after a failure);
- the JSON shape and source, the city cache;
- one try that keeps the old reading;
- where-am-i's output.
Both packages' test/all and tools/pr-check pass. The series applies
with --fuzz=0 to the real aarch64 Omarchy 4.0.6-1 (20 applied).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
derekross
added a commit
that referenced
this pull request
Oct 30, 2026
… setting (#185) On the Pixel 7 Pro the lock glance showed a 24-hour time while the phone bar and the home clock widget showed 12-hour. The bar's clockFormat in ~/.config/omarchy/shell.json is "ddd MMM d h:mm AP". Two faults: - The big glance time used the Qt format "h:mm". Qt counts "h" from 0 to 23 unless the same format has an AM/PM marker, so even with the bar's format read, 2 pm showed as "14:02". It is now built from the hour (h % 12 || 12, with minutes) and still leaves AM/PM out. The small clock over the PIN view keeps "h:mm AP". - shell.json was only read while the glance was active (path gated on the keypad toggle) and asynchronously. A failed read, such as Omarchy replacing the file, cleared the format, and the fallback was the session's Qt locale, which is C/POSIX for the shell and so 24-hour. It is now read at once (blockLoading, a small file), whatever the lock shows, again on every change and whenever the glance comes back. A read that fails or catches a half-written file keeps the last format. The 12/24-hour rule stays the one the bar and the home clock use (omamobile-shell HomeModel.systemTimeFormat): quoted literals ignored, an AM/PM marker or "h" without "H" means 12-hour. Without a clockFormat the lock now follows the bar's own default, HH:mm, so all three still agree; it no longer falls back to the Qt locale. Patch 0023 is revised in place, as #181 did after #176. Tests (test/lock-glance, now under LC_ALL=C as the shell runs): - the maintainer's format is 12-hour before the glance is active, with a 1-12 big time without AM/PM and a small clock with it; - a half-written or missing shell.json keeps the last format; - HH:mm and no clockFormat are 24-hour; - the rule (twelveHour, quoted literals), glanceTime at 14:02, 00:05, 12:30 and 9:41, and barClockFormat's parse. The FileView stand-in now reads at once with blockLoading, as Quickshell does. test/all and tools/pr-check pass, and the series applies with --fuzz=0 to the real aarch64 Omarchy 4.0.6-1 (20 applied). Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
This was referenced Nov 1, 2026
nesmeck
added a commit
to nesmeck/omarchy-mobile
that referenced
this pull request
Nov 4, 2026
angelrenefm's unit passed every D3 step on linux-omarchy-mobile#181's head (issue derekross#342), as derekross asked on derekross#181: stage 0 found ABL's state as on the first unit, 60 Hz with no storm, 871 flips with no timeouts, and screen_off=2 six times out of six. The plan records it with his kernel release. The D3a paragraph now says its kernel lacked the padded-row commit, and D3c's that it had all five, for the series' new order. The package ships as 0.1.7: 0.1.6 is omarchy-mobile#396's AVB footers, which merges first. The results keep that the build tested was numbered 0.1.5. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
nesmeck
added a commit
to nesmeck/omarchy-mobile
that referenced
this pull request
Nov 4, 2026
angelrenefm's unit passed every D3 step on linux-omarchy-mobile#181's head (issue derekross#342), as derekross asked on derekross#181: stage 0 found ABL's state as on the first unit, 60 Hz with no storm, 871 flips with no timeouts, and screen_off=2 six times out of six. The plan records it with his kernel release. The D3a paragraph now says its kernel lacked the padded-row commit, and D3c's that it had all five, for the series' new order. The package ships as 0.1.7: 0.1.6 is omarchy-mobile#396's AVB footers, which merges first. The results keep that the build tested was numbered 0.1.5. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
robertzas
pushed a commit
to robertzas/omarchy-mobile
that referenced
this pull request
Nov 14, 2026
… screen goes dark (derekross#181) On the Pixel 7 Pro the glance came back after 15 s on the PIN view, but a power-button press or a tap that woke the screen went straight to the PIN view, and "sometimes the view is reset". The lock blanks the display itself 10 s after the last input, before the glance's 15 s return, and after that the phone may sleep (omamobile-sleepd), which freezes QML timers, so the return never ran. Nothing that turned the screen off told the glance, and the touch that woke a dark screen was delivered to the lock and counted as the glance's tap to show the PIN view. Patch 0023 (revised in place, as it was merged today and nothing has built on it) now brings the glance back, with an empty field, whenever the locked display goes dark: - the lock's own blank bumps a blankSerial in the Service; - a new "omarchy-shell lock dark" call does the same for what blanks it from outside, without touching the display itself. The Service keeps displayDark until the display is lit through the lock again (runWake). While it is dark, a touch or swipe on the glance only lights the screen through the lock, and so does a touch that starts within 400 ms of a wake. The 15 s return is now counted on the wall clock (a 1 s check against when the PIN view last had input), as the lock's blank timer is, so a sleep in between can't freeze it. The computer lock and the fingerprint PAM flow are unchanged. From the maintainer's hands-on test, a tap no longer opens the PIN view. He touched a notification to read it and got the PIN pad. Now only a swipe up opens it, from anywhere, a card included: a clear drag of 8 % of the screen's height, or a quick flick. A typed key and fingerprint unlock still work as before. A tap on a card opens it to more of its text (summary up to 2 lines, body up to 4) and runs nothing; a second tap closes it. A tap elsewhere does nothing, and open cards close when the glance comes back. A sideways swipe on a card dismisses it, as on a phone and in the shade. The first 12 px of a drag decide its way: mostly sideways is a card swipe, mostly vertical is the swipe up, from a card too. The card slides as a toast does (patch 0017's finishSwipe: past 35 % or a flick it flies off, else it springs back). That one notification is then forgotten through the notifications service's forget call (patch 0019), the call the shade makes, so it leaves the history and the shade too. Nothing of it runs, the next one moves up from the "more" line, and the last one leaves "No new notifications". He also asked for the glance to show the notifications the phone's shade shows. It left out Omarchy's own messages (omarchy-action), bare notify-send pings and anything over 24 h old, and grouped one card per app, so three Omarchy messages showed as one. The shade (omamobile.quicksettings NotificationsModel) lists the newest ten of the toasts on screen and the history, which Omarchy's service trims to ten (historyLimit). It has no age limit, leaves no app out and skips only rows with neither summary nor body. The glance now: - uses the same rows; - shows each notification as its own card, newest first, as many as fit above the hint (up to three); - counts every one not shown in "N more from A and B" (or "A, B and 2 others"); - gives each card its age ("2m") as its detail; - names omarchy-action "Omarchy" (with its glyph, as NotificationCard draws it) and notify-send "Notifications". Do-not-disturb now follows the toasts rather than hiding every card: only what still toasts shows (Omarchy's own messages, and a critical notify-send, NotificationLogic.shouldBypassDnd). omamobile-shell 0.2.41 makes the calls: - omamobile-keys calls "lock dark" before the power key turns a locked screen off, and before it powers off the panel behind a new lock; - omamobile-sleepd calls it, bounded by a 3 s timeout, just before each suspend, for a screen that went dark some other way. An older lock without the call ignores it. It also fills the glance's weather line, which had nothing to read. The maintainer's weather is a community home screen widget that fetches on its own and keeps nothing to borrow, so the shell now owns the weather, whatever widgets are placed. A new omamobile-weather user service runs at login and every 30 minutes (omamobile-weather.timer, enabled for every graphical session). It fetches Open-Meteo's current weather, with the same API and words as the shell's weather widget (a test checks the words agree with HomeModel.weatherText for every code). It writes ~/.local/state/omamobile/weather.json atomically: {"at", "temperature" (rounded), "unit", "text", "source"}. Where, in order: - the phone's position: GeoClue at city accuracy, at most every 3 h, asked only while the Location switch runs GeoClue's agent, with a hard timeout. The client is where-am-i, which omamobile-gnss lets through, so a timer never raises a prompt. A failed fix falls back to the last one, up to a day old. - the setting ~/.config/omamobile/weather.json ({"city"} or coordinates, "unit", "location": "auto" or "fixed"; fixed puts it before the position). - the home screen today widget's city (its owl state), or the weather widget's named place. A city is looked up once with Open-Meteo's geocoding and kept by name, and "Town, ST" picks the US state. The unit is the setting's, else the locale's (°F in the US). There is one try per run, and a failure leaves the old reading. There is no IP geolocation, and nothing is fetched without a place. The lock still hides a reading over 3 h old. Tests: - test/lock-glance: - a blank on the PIN view returns to the glance and clears the PIN; - a wake tap and a wake swipe on a dark display don't show the PIN view, and neither does a swipe that starts right after a wake; - a frozen timer (the wall clock 20 s on) returns to the glance; - taps on a card or on empty space don't show the PIN view; - a card opens and closes, and only that card; - a swipe up from a card shows the PIN view; - a short, slow drag doesn't; - three notifications from one app are three cards, and five are three cards plus "2 more from Mail and GitHub"; - a sideways swipe on a card dismisses it (the forget names its file) without showing the PIN view, and the next one moves up; - a short one springs back; - a mostly vertical swipe from a card shows the PIN view; - the last one dismissed, with hidden text, leaves "No new notifications"; - the shade's ten, Omarchy's own and 30 h old notifications show, and do-not-disturb shows only what still toasts. - test/patches.py checks the Service's blank, dark and wake paths. - omamobile-shell: keys.sh and sleepd.sh check the call order. weather.py checks: - the words against HomeModel.weatherText; - the settings and units; - the order of the sources, with GeoClue stubbed (unusable, so the widget's city or the setting; fixed, so the setting; a fix reused for 3 h and up to a day after a failure); - the JSON shape and source, the city cache; - one try that keeps the old reading; - where-am-i's output. Both packages' test/all and tools/pr-check pass. The series applies with --fuzz=0 to the real aarch64 Omarchy 4.0.6-1 (20 applied). Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
robertzas
pushed a commit
to robertzas/omarchy-mobile
that referenced
this pull request
Nov 14, 2026
… setting (derekross#185) On the Pixel 7 Pro the lock glance showed a 24-hour time while the phone bar and the home clock widget showed 12-hour. The bar's clockFormat in ~/.config/omarchy/shell.json is "ddd MMM d h:mm AP". Two faults: - The big glance time used the Qt format "h:mm". Qt counts "h" from 0 to 23 unless the same format has an AM/PM marker, so even with the bar's format read, 2 pm showed as "14:02". It is now built from the hour (h % 12 || 12, with minutes) and still leaves AM/PM out. The small clock over the PIN view keeps "h:mm AP". - shell.json was only read while the glance was active (path gated on the keypad toggle) and asynchronously. A failed read, such as Omarchy replacing the file, cleared the format, and the fallback was the session's Qt locale, which is C/POSIX for the shell and so 24-hour. It is now read at once (blockLoading, a small file), whatever the lock shows, again on every change and whenever the glance comes back. A read that fails or catches a half-written file keeps the last format. The 12/24-hour rule stays the one the bar and the home clock use (omamobile-shell HomeModel.systemTimeFormat): quoted literals ignored, an AM/PM marker or "h" without "H" means 12-hour. Without a clockFormat the lock now follows the bar's own default, HH:mm, so all three still agree; it no longer falls back to the Qt locale. Patch 0023 is revised in place, as derekross#181 did after derekross#176. Tests (test/lock-glance, now under LC_ALL=C as the shell runs): - the maintainer's format is 12-hour before the glance is active, with a 1-12 big time without AM/PM and a small clock with it; - a half-written or missing shell.json keeps the last format; - HH:mm and no clockFormat are 24-hour; - the rule (twelveHour, quoted literals), glanceTime at 14:02, 00:05, 12:30 and 9:41, and barClockFormat's parse. The FileView stand-in now reads at once with blockLoading, as Quickshell does. test/all and tools/pr-check pass, and the series applies with --fuzz=0 to the real aarch64 Omarchy 4.0.6-1 (20 applied). Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
cx3301J
added a commit
to cx3301J/omarchy-mobile
that referenced
this pull request
Nov 19, 2026
The 7a's bootloader leaves 0x1cf8 and 0x2a82306d in DSIM0 at 60 Hz; derekross#181's Pixel 7 row is the same pair. Assisted-by: DeepSeek Signed-off-by: cx3301J <cx3301J@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up to #176, from the maintainer testing the glance lock on the Pixel 7 Pro.
omarchy-shell lock dark(called by omamobile-keys before the power key turns the screen off, and by omamobile-sleepd before a suspend) and the wake path all return to the glance.omamobile-weatherservice and timer (every 30 min, at login and on resume) writes~/.local/state/omamobile/weather.json, which the glance reads. The lock itself makes no network calls.omamobile-weather set CITY [F|C]andomamobile-weather show.Patch 0023 is revised in place, since #176 merged today and nothing builds on it.
Tests:
test/allpasses: the series applies--fuzz=0to the real aarch64 Omarchy 4.0.6-1, and lock-glance runs 34/34.test/allpasses, including the new weather tests and the keys/sleepd ordering.tools/pr-checkpasses.Tested on the Pixel 7 Pro (2026-10-30, kernel
7.3.0-rc6-omamobile-push1-gaae3b058ef98):source": "widget", °F from the locale).Open follow-up: a suspend logind starts without sleepd (upower's critical-battery action) doesn't call
lock dark. #134 already patchesomarchy-system-sleep-lock, so it should add the call there when it's rebased.🤖 Generated with Claude Code