Skip to content

omamobile-omarchy 0.1.37, omamobile-shell 0.2.103: double tap to wake, a single touch leaves the screen dark - #420

Merged
derekross merged 7 commits into
derekross:mainfrom
angelrenefm:panther-doubletap
Nov 9, 2026
Merged

derekross merged 7 commits into
derekross:mainfrom
angelrenefm:panther-doubletap

Conversation

@angelrenefm

Copy link
Copy Markdown
Contributor

A dark phone lit up on any touch: Hyprland's misc:mouse_move_enables_dpms (on in Omarchy) and the glance's wake-only touch (0023) both woke on the first contact, so it lit up in a pocket or under a palm. Android wakes on a double tap. This does the same, in two commits:

  1. omamobile-omarchy, patch 0029 (lock): on a touch-only lock (the lock-keypad toggle) a lone touch on the dark screen does nothing, and a second one within 400 ms lights it through the lock as before; neither touch does anything else. A computer's lock (keypad off) wakes on any click or mouse move as before.
  2. omamobile-shell:
    • hypr/phone.lua: misc.mouse_move_enables_dpms = false, next to key_press_enables_dpms.
    • bin/omamobile-doubletap, a user unit for every graphical session (wants link, as omamobile-sleepd): reads the touchscreen without a grab and on a double tap (0.4 s, 14 % of the width) calls omamobile-keys double-tap, which wakes a dark screen as a short power press does, locked or not, and leaves a lit one alone. This covers the unlocked phone with the screen off, where no lock is up to see the taps.
    • The touchscreen is found by its capabilities (INPUT_PROP_DIRECT, BTN_TOUCH, ABS_MT_POSITION_X), not by driver name, and the distance comes from its own X range (EVIOCGABS). Nothing in it is specific to one phone.

Version numbers are left to you.

This changes every phone

The Pixel 7 Pro, the 9a and every other phone with omamobile-shell stop waking on a single touch. Power still wakes as before. If you'd rather not change everyone's phone at once, it could be an Omarchy toggle (say omamobile-doubletap, on by default, as Android has it). The toggle would have to switch all three parts: phone.lua, the watcher, and the lock's darkTapWakes(). I left it out to keep this small. Tell me if you want it.

Tested

On a Pixel 7 (panther), Omarchy 4.0.6, by hand before this branch: the same lock change, mouse_move_enables_dpms = false in the user's input.lua, and an earlier version of the watcher run as a system service. With the phone locked (PIN) and its screen turned off with the power key, before it suspended: a single touch left it dark and a double tap woke it. The owner confirmed both by hand. The unlocked dark screen (the watcher's case) was not tried, nor a phone already in s2idle (the touchscreen doesn't wake it).

The packaged versions in this PR are not run on a phone yet. They differ from what ran there in three ways:

  • the watcher finds the touchscreen by its capabilities, not by the name FT3658;
  • it asks omamobile-keys double-tap (the power key's screen check: bl_power or Hyprland's dpmsStatus), not gs-decon's stats;
  • it is a user unit, not a system service.

Checks, on the laptop (Fedora, Qt 6 qmltestrunner):

  • tools/pr-check origin/main: all ok.
  • packages/omamobile-omarchy/test/lock-glance against Omarchy 4.0.6-1 (aarch64) with the series and 0029: Totals: 56 passed, 0 failed. Without 0029 the three new tests fail (53 passed, 3 failed).
  • packages/omamobile-shell: test/doubletap.py 13/13, test/keys.sh, test/phone-rules.lua, and test/all: exit 0, 301 ok.
  • packages/omamobile-omarchy/test/patches.py against the real Omarchy 4.0.6-1 package: 98 ok, 0 failed. The series applies with --fuzz=0 (28 applied, 0 already in Omarchy), the new 0029 check passes, and the glance runs under qmltestrunner with 56 passed.
  • packages/omamobile-omarchy/test/all: everything for 0029 passes. test/update.sh fails on this laptop with timeout ("start: the PIN is taken"), and it fails the same way on a clean origin/main checkout, so it isn't from this PR.

Not checked

  • Any phone other than the Pixel 7, including the Pixel 7 Pro (Synaptics touch) and the 9a. On the 9a, omamobile-keys already says the lock's own wake doesn't power the panel. The watcher's omamobile-keys call should cover that, but it hasn't been tried.
  • Whether Hyprland passes the first tap on a dark, unlocked screen to the app under it. The Pixel 7 test didn't look for this.
  • A phone in s2idle: the touchscreen is not a wake source, so a double tap does nothing until Power wakes it.
  • Battery cost of the watcher: it sleeps in read() and wakes only for touch events. Not measured.
  • A phone whose touchscreen driver doesn't report BTN_TOUCH (pointer emulation). The watcher wouldn't find it and would log "no touchscreen yet". On such a phone a single touch would also no longer wake an unlocked dark screen.

I used AI help (Claude) to work on this.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @angelrenefm. This is a careful piece of work. The watcher finds the touchscreen by its capabilities rather than a driver name, takes the distance from the device's own range, and logs once instead of every five seconds. The unlocked wake is now shared with the power key instead of copied, and you list exactly what ran, what differs from it and what you didn't check. That made this easy to review. I checked the mechanism on our kernels: the Pixel 7 Pro's Synaptics driver, the 9a's Goodix and your FT3658 all report INPUT_PROP_DIRECT, BTN_TOUCH and ABS_MT_POSITION_X, so discovery works on all three. I also checked Omarchy's idle service, which ends an idle cycle on any touch and wakes the display: on our phone profile it locks at once, so it doesn't bring the single-touch wake back.

Because this changes how every phone wakes, the Pixel 7 Pro included, I need three things before I merge:

  1. A run of this branch as packaged. Please install both packages from this head on your Pixel 7 and check that systemctl --user status omamobile-doubletap says it's watching your touchscreen. Then try the locked and the unlocked dark screen and paste a tools/phone-smoke report. After that I'll run it on my 7 Pro before merging.
  2. The unlocked dark screen. I looked at Hyprland 0.57's touch path. onTouchDown sends the touch to the focused window whatever the DPMS state, and mouse_move_enables_dpms only decides whether the display also turns on. So with this change, touches on a dark unlocked phone go to the app under the finger with nothing lit, and a double tap lands there as a blind double tap. The lock is fine, since its surface takes the touches and 0029 swallows them. Could you check this on your phone? I see three ways to handle it, and I'm open to others: lock first when the screen is dark and unlocked (the watcher could run the power key's lock path), keep Hyprland's touch wake while unlocked, or test it and say plainly that it's accepted.
  3. Versions. Please bump omamobile-omarchy to 0.1.33 and omamobile-shell to the next free version (0.2.102 if omamobile-shell: show the battery’s reported state #421 lands first), and put both in the title.

Some smaller ideas, none blocking:

  • Android only wakes on a double tap when the proximity sensor says "far". The sensor proxy already offers ProximityNear, and checking it would cover the pocket case better than any timing.
  • On a phone with two touchscreens (the Pixel Fold) the watcher only watches the first one. Watching every matching device would cover both.
  • The lock's window runs press to press and the watcher's runs lift to press, so on a locked phone the watcher can wake where the lock said no. It's harmless, but a sentence in the README would help.
  • A double tap can't wake the phone from sleep, so with the sleep toggle on it works only in the first 30 s after the screen goes off. Could the README say "while awake"?

Thanks again! I'm looking forward to this one.

A touch on the dark lock screen lit it: the glance's wake-only touch
(0023) and Hyprland's mouse_move_enables_dpms both woke on the first
contact, so a phone in a pocket or brushed by a palm lit up. Android
wakes on a double tap.

Patch 0029: on a touch-only lock (the lock-keypad toggle) a lone touch
on the dark screen does nothing, and a second one within 400 ms of it
lights the screen through the lock as before; neither touch does
anything else. LockView.darkTapWakes() keeps the time of the first tap
and is used by the glance's touch area and by the lock's own MouseArea.
A computer's lock (keypad off) wakes on any click or mouse move as
before.

Tests: test/lock-glance.qml now expects a double tap to wake, lone taps
400 ms apart and a swipe to leave it dark, a double tap on the media
controls to press nothing, and one click to wake a computer's lock;
test/patches.py checks the patched files. Without 0029 three of the new
tests fail (control run).

This changes the lock on every phone. The compositor half is in
omamobile-shell (next commit).

I used AI help (Claude) to work on this.
… dark

Hyprland's misc:mouse_move_enables_dpms (on in Omarchy) lit a dark
phone on any touch, so it lit up in a pocket. Android wakes on a double
tap; this does the same on every phone with omamobile-shell.

- hypr/phone.lua turns mouse_move_enables_dpms off, next to
  key_press_enables_dpms.
- bin/omamobile-doubletap, a user unit for every graphical session
  (graphical-session.target wants link, as omamobile-sleepd): reads the
  touchscreen's evdev node without a grab (the user is in the input
  group) and on two taps within 0.4 s and 14 % of the screen's width
  (150 px of 1080) calls omamobile-keys double-tap. The touchscreen is
  found by its capabilities (INPUT_PROP_DIRECT, BTN_TOUCH,
  ABS_MT_POSITION_X), not by a driver name, and the distance comes from
  its own X range (EVIOCGABS), so nothing in it is a Pixel's. With no
  touchscreen it waits for one; an unreadable node is logged once.
- omamobile-keys double-tap: a dark screen wakes as a short power press
  wakes it (the lock's wake and the panel when locked, the panel when
  unlocked); a lit one is left alone. The unlocked wake is now one
  function shared with the power key. It doesn't write the power-wake
  marker, so the next power press still counts.

The lock wakes on a double tap itself (patches/omarchy/0029); this
watcher should also light the panel where the lock's wake alone doesn't
(tegu, as omamobile-keys notes; not tried there). Without 0029 the lock
still wakes on one touch, as before.

Tests: test/doubletap.py (the detector on synthetic events: quick
pairs, slow pairs, far pairs, long presses, triples; the touchscreen
picked from a fake /sys over a touchpad and a key device),
test/keys.sh (double-tap by lock and screen state), test/phone-rules.lua
(both dpms settings), and test/all checks the unit and its wants link.
Mutation controls: a 4 s gap, and a double-tap that skips the screen
check, each fail their test.

I used AI help (Claude) to work on this.
…letap (0029)

The lock recognised the double tap itself, so it woke in a pocket too:
it has no proximity reading. omamobile-shell's omamobile-doubletap
already wakes the lock through its IPC on the same double tap. Now no
touch wakes a dark touch-only lock; the daemon does, and can skip it
while the proximity sensor reads near. A computer's lock wakes on any
input as before.

lock-glance.qml: two taps on the dark lock wake nothing and press no
media button; the IPC wake shows the glance. Against the real Omarchy
4.0.6 package, test/patches.py's qmltestrunner run gives 56 passed, 0
failed; with the old 0029 the two rewritten tests fail.

Not run on the phone.

I used AI help (Claude) to work on this.
When the phone has a proximity sensor (iio-sensor-proxy's
net.hadess.SensorProxy with HasProximity; omamobile-sensors on the AoC
phones), omamobile-doubletap claims it on the first touch on a dark
screen and skips a double tap while it reads near. The claim ends 2 s
after the last touch, so the sensor runs only around taps. Without the
service or the sensor nothing changes.

test/doubletap.py: 22 ok, with a fake sensor proxy (no bus).

Not run on the phone: the Pixel 7's sensors are not up yet, and no 7 Pro
was at hand.

I used AI help (Claude) to work on this.
The case derekross#420's review left open: Hyprland sends a touch to the window
under it whatever the screen's state, so on a dark unlocked phone a tap,
or a double tap, landed in an app no one could see. omamobile-doubletap
now looks at the screen four times a second and, when it goes dark while
unlocked (an idle blank, the panel powered off by hand), runs
omamobile-keys dark-unlocked, which locks behind the dark panel; the
lock's surface then takes the touches. Only where the screen's state can
be told (a DSI connector). The proximity claim's release moves into the
same one-second wait.

README: the dark-unlocked lock, that it works while the phone is awake
(a double tap can't wake it from sleep), and that it times a double tap
from lift to touch while the lock times touch to touch.

omamobile-omarchy 0.1.37, omamobile-shell 0.2.103 (0.1.36 and 0.2.102
are claimed by derekross#471 and derekross#482).

I used AI help (Claude) to work on this.
@angelrenefm angelrenefm changed the title omamobile-omarchy, omamobile-shell: double tap to wake, a single touch leaves the screen dark omamobile-omarchy 0.1.37, omamobile-shell 0.2.103: double tap to wake, a single touch leaves the screen dark Nov 7, 2026
@angelrenefm

Copy link
Copy Markdown
Contributor Author

@derekross, thanks for the careful review, and sorry this took a few days. The branch is rebased onto today's main, and here is where each point stands.

1. A run on the Pixel 7. My Pixel 7 runs Omarchy 4.5 with our 4.5 patch series, while main holds Omarchy at 4.0.6, so this branch's omamobile-omarchy (0029 for the 4.0.6 lock) can't be installed on it as packaged; its lock already carries the same change for 4.5. The shell side ran as packaged: this head's omamobile-doubletap, omamobile-keys and the user unit, installed over the phone's own, then put back.

  • systemctl --user status omamobile-doubletap: active (running) from /usr/lib/systemd/user/omamobile-doubletap.service; it logs watching /dev/input/event1, the FocalTech FT3658.
  • Locked and dark: powered off with the power key, then a double tap with a real finger woke it to the lock screen. No power key in the log for that wake.
  • Unlocked and dark: omarchy-brightness-display off on an unlocked phone. It locked by itself; the lock was up 1.1 s after the screen went off (details under point 2).
  • tools/phone-smoke --here, on that boot:

phone-smoke: unknown

Kernel 7.3.0-rc6-omamobile-push1-g44048a9aa795, tools at unknown, 2026-11-07T21:16Z.

Check Result Detail
boot (required) ok kernel 7.3.0-rc6-omamobile-push1-g44048a9aa795, slot b, up 21928s
storage guard (required) ok 55 partitions; only userdata and slot B's boot images writable
root (required) ok ext4 on userdata, 95G free
kernel health (required) ok no oops or BUG; 0 WARN, 6 error lines in dmesg (6 known, 0 new)
systemd units ok none failed
USB gadget (required) ok bound to 11210000.usb
display ok card2-DSI-1 connected, 1080x2400
GPU ok render node on panthor
session (required) ok Hyprland and the Omarchy shell running
lock ok PAM file present, the shell can lock (passwordPam true, locked true)
touch ok FocalTech FT3658
buttons ok gpio-keys
Wi-Fi ok wlp1s0 on brcmfmac, link up
Bluetooth ok hci0, powered true
battery ok 91% Discharging, 32 °C
sound cards ok googleaocsndcar
sensors ok Accelerometer AmbientLight Proximity
cameras skip no libcamera tools
thermal ok hottest zone 49 °C

Packages: hyprland 0.57.0-3.1, libcamera 0.7.2-4.22, mesa 1:26.2.4-1.4, omacam 0.3.3-1, omamessages 0.2.2-1, omamobile-apps 0.2.2-1, omamobile-audio 0.5.16-1, omamobile-device-panther 0.1.9.2-1, omamobile-firstboot 0.1.3-1, omamobile-gnss 0.2.7-1, omamobile-keyring 20261026-1, omamobile-mirror 0.1.1-1, omamobile-omarchy 0.1.32-32, omamobile-owl-widgets 0.1.2-1, omamobile-sensors 0.2.1-1, omamobile-shell 0.2.100-88, omaphone 0.2.3-1, omarchy 4.5.0-3, omatel-agent 0.2.4-2, quickshell 0.3.2.r7.g2b4501a-1

The packages it lists are my local builds (Omarchy 4.5 and the shell I run), not this head's.

2. The unlocked dark screen. I took your first option: a dark phone is a locked one, as on Android. omamobile-doubletap looks at the screen's DPMS four times a second. When it goes dark while unlocked (an idle blank, the panel powered off by hand), it runs omamobile-keys dark-unlocked, which locks behind the dark panel, so the lock's surface takes the touches. It only does this where it can tell the screen's state (a DSI connector). Measured on the phone: locked 1.42 s after the screen went off when it looked once a second, 1.10 s at four times a second. Most of that is the lock itself coming up ("screen-stabilizing" to session-locked), and a touch in that window still reaches the app. Tests: test/keys.sh covers dark and unlocked (locks, then the lock hears it is dark), lit, and already locked; test/doubletap.py covers the DSI check.

3. Versions. omamobile-omarchy 0.1.37 and omamobile-shell 0.2.103 (0.1.36 and 0.2.102 are claimed by #471 and #482). The title says so.

Your smaller ideas:

  • Proximity: in. The double tap is skipped while the proximity sensor reads near. The claim runs only from the first dark touch until 2 s after the last, and without the sensor it wakes as before.
  • README: it now says it works while the phone is awake, and that the watcher times lift to touch while the lock times touch to touch.
  • Two touchscreens: not done; it still watches the first one.

Not checked: this head's omamobile-omarchy on Omarchy 4.0.6, a Fold, and the proximity skip with the phone in a pocket.

I used AI help (Claude) to work on this.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @angelrenefm, and no need to apologise for the wait. This came back better than I asked. Locking a screen that goes dark while unlocked is the Android behaviour I hoped for. Moving the double-tap decision into the daemon so the proximity check can apply to the lock as well is a good call. And your report says exactly what ran where, including what couldn't. The tests are thorough too: I applied the series with 0029 to Omarchy 4.0.6 myself (29 applied at --fuzz=0), and 0029 changes exactly the two places your commit describes.

Since your phone runs 4.5, the packaged 4.0.6 run has to be mine, and I'll do it on the 7 Pro before merging, as planned. Before that, a few things:

  1. Text left over from before the third commit. Since "leave the dark lock's double tap to omamobile-doubletap", the lock no longer recognises a double tap itself, but four places still say it does: the daemon's docstring (line 7), the double-tap comment in omamobile-keys (line 326), the comment in phone.lua (line 131), and in the README row, the sentence about the watcher timing lift to touch while the lock times touch to touch. That sentence can go now. Could you also refresh the PR description? It still describes 0029 as the 400 ms lock wake and the watcher as sleeping in read(). I'll write the squash message myself, since the first commit's body describes the earlier design.
  2. omamobile-omarchy 0.1.39: select 4.0.6 and 4.5 patch families #471 and 0029. omamobile-omarchy 0.1.39: select 4.0.6 and 4.5 patch families #471 (Omarchy 4.0.6 and 4.5 patch families) moves the 4.0.6 series to legacy-4.0.6/ and makes patches/ the 4.5 series. Git would merge your new patches/0029-… cleanly into what is by then the 4.5 series, so 4.0.6 phones would lose it and 4.5 would get a 4.0.6 patch. Whichever of the two is ready first lands first, and the other one moves 0029 into legacy-4.0.6/: if omamobile-omarchy 0.1.39: select 4.0.6 and 4.5 patch families #471 lands first, please put 0029 there in this PR. If you're willing to add your 4.5 version of the lock change as well, that would complete it. I'll mention it on omamobile-omarchy 0.1.39: select 4.0.6 and 4.5 patch families #471 too.
  3. The four-times-a-second check. With the sleep toggle off, which is the default, a phone with its screen off stays awake, and the poll adds four wakeups a second for as long as it does. I'll measure it in my run. If you'd like to make it cheaper, it could look less often while the screen is dark, since all it needs to catch then is the screen coming back on.

Some smaller points, none blocking:

  • Proximity with the panel off. The sensor proxy says the light and proximity sensor under the panel only converts with the display's state, and it sends "off" exactly when your daemon claims it. It also keeps the last reading across a release. So the guard may never see "near", or may keep an old "near" for a moment. Your smoke report lists a proximity sensor on that boot, so if you can, try covering the top of the phone and double tapping: it should be ignored, then wake once uncovered.
  • Version skew. With 0029, a lit lock that still thinks it's dark ignores every touch. That can happen if omamobile-omarchy is updated without the new shell. Hyprland would then light the panel on a touch, and Power would turn it off again. A conflicts=('omamobile-shell<0.2.103') line in omamobile-omarchy would rule it out.
  • The window before the lock. Please say in the README that a touch in the second or so before the lock is up still reaches the app.
  • Several panels. screen_dark() reads "dark" if any DSI connector is off, and the keys' check does the same for monitors. On a phone with a second panel that would claim proximity on every touch, and a double tap in an app could switch the other panel on. A comment for whoever ports a Fold would be enough for now.
  • Versions. I'll set them at merge from what's free then. If this lands before omamobile-shell 0.2.125: the shade and the bar read Omarchy 4.5's notification store #482 and omamobile-omarchy 0.1.39: select 4.0.6 and 4.5 patch families #471, that will be shell 0.2.102 and omarchy 0.1.36.

Thanks again. I'm looking forward to running it!

derekross added a commit that referenced this pull request Nov 8, 2026
Thirty-one lessons from round 44's reviews, confirmed by the maintainer, plus a second instance of L-454 from #493. #420's and #471's reviews found the same patch-series lesson from either side; it is one entry (L-550).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
derekross added a commit to Gr-linkk/omarchy-mobile that referenced this pull request Nov 8, 2026
main already carries 0.2.101 (derekross#464), and derekross#482 and derekross#420 hold 0.2.102 and
0.2.103.
derekross pushed a commit that referenced this pull request Nov 8, 2026
The bar, the home widget and quick settings follow UPower's own
battery.state instead of "charger online": charging and time estimates
show only when the battery reports that state, and pending and unknown
states are shown as such, so external power no longer implies charging.

The bar's Full glyph loses its bolt too (on the 7 Pro as well as in
quick settings): the icon helper chose the charging-full glyph for Full
even when charging was false, and a Pixel 6 reporting Full at 91%
showed the bolt. The battery-state test covers it. Tested on a Pixel 6
in all three states on all three surfaces.

0.2.104: main already carries 0.2.101 (#464), and #482 and #420 hold
0.2.102 and 0.2.103.
# Conflicts:
#	packages/omamobile-shell/version
… (review of derekross#420)

Since this branch's third commit the lock never times a double tap: 0029
swallows every touch while the locked display is dark, and only
omamobile-doubletap wakes it. Four comments and the README still said the
lock recognises a double tap itself.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
derekross added a commit that referenced this pull request Nov 9, 2026
test/weather.py (since 0.2.90, #366) runs plugins/omamobile.home/HomeModel.js
under node, but the PKGBUILD never said so. CI's runner and the build image
have node, so nothing noticed until tools/phone-install-package built the
shell on the Pixel 7 Pro, which has no node: check() failed at
"omamobile-weather" and the package wasn't installed (2026-11-08, while
installing #420 for its test). phone-install-package installs a package's
checkdepends for the build and removes them afterwards, so with this line
the phone built and installed it and nodejs was gone again.

No version change: the package's files are the same.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@derekross
derekross merged commit 71b47a6 into derekross:main Nov 9, 2026
2 checks passed
derekross added a commit to jkjoplin/omarchy-mobile that referenced this pull request Nov 9, 2026
derekross added a commit that referenced this pull request Nov 9, 2026
* omamobile-shell 0.2.107: Grok Bot fits the phone's screen

Grok Bot (grok-bot from [omarchy], x.ai's agent, an Electron app) sets
a minimum window size of 512x520 (SAND_MIN_WINDOW_SIZE in its main
process). The Pixel 7 Pro is 480 px wide in logical pixels, so the app
laid itself out 512 px wide inside a 480 px window and ran off the right
edge: the ends of messages, the header buttons and the composer's
microphone were cut off. Chromium had the same problem and setup-user
gives it --force-device-scale-factor=0.83.

setup-user now does the same for Grok Bot, whose launcher reads
~/.config/grok-bot-flags.conf. The factor comes from the panel's mode
(the DSI connector in /sys/class/drm) and the monitor scale setup-user
seeds, so the layout gets about the 578 px Chromium gets on the 7 Pro:
0.83 on the 7 Pro (480 px), 0.70 on the 9a (405 px), 0.62 for a
1080 px panel at scale 3 (360 px), where 0.83 would leave 434 px, still
under 512. With no readable mode it falls back to 0.83. Once, and a
factor the user's file already has stays theirs.

Tested on a Pixel 7 Pro (Omarchy 4.5.0-3, omamobile-shell 0.2.100):
the factor by hand made the window fit (screenshots before and after);
the new step, run alone on the phone against a scratch home, read
1440x3120 and wrote 0.83. test/setup-user.sh covers the 7 Pro, the 9a,
a 1080x2400 panel, no mode, a user's own factor and two logins; its
first new check fails against the old script.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TTVicoVA5YdXzobkXGujF9

* omamobile-shell 0.2.108: renumber above main's 0.2.107 (#420)

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Derek Ross <derekross@gmail.com>
derekross added a commit that referenced this pull request Nov 9, 2026
Round 45's confirmed-candidate lessons from the omarchy-mobile and
linux-omarchy-mobile reviews, plus five from the maintainer session of
2026-11-08 (the four build steps a test kernel needs for push, reboot-to
and the modem's clean-stop marker, pacman steps on a non-Arch build
host, a new dependency's example config and the identifier scan, and
checkdepends). Six candidates that were second instances of existing
lessons became "Again in round 45" sentences on L-333, L-427, L-430,
L-454, L-532 and L-547. C-32 and C-33 correct round 44 on #420's 0029
(applies to omarchy 4.5.0-3 at fuzz 0) and on kernel #219's test phone
(the cover letter names it).

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
derekross added a commit that referenced this pull request Nov 9, 2026
Round 46's confirmed candidates from the reviews of the open
omarchy-mobile and linux-omarchy-mobile pull requests, the round's own
process slips (shared refs and scratch files deleted by glob, a
maintainer fixup that skipped the package's test/all, an amend on a
contributor branch) and the 2026.11.08 release PRs (#547, #548, #549,
#550, #560). Repeats of existing lessons extend them instead: L-45
(kernel #222), L-183 (#409), L-426 (the round's version collisions,
now set by hand at merge) and L-550 (#471 with #420). C-34 and C-35
correct our #290 shell suggestion and round 45's claim that
tests/dipper-probe guards anything.

Stacked on #551 (round 45's lessons).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
derekross added a commit that referenced this pull request Nov 9, 2026
Round 46's confirmed candidates from the reviews of the open
omarchy-mobile and linux-omarchy-mobile pull requests, the round's own
process slips (shared refs and scratch files deleted by glob, a
maintainer fixup that skipped the package's test/all, an amend on a
contributor branch) and the 2026.11.08 release PRs (#547, #548, #549,
#550, #560). Repeats of existing lessons extend them instead: L-45
(kernel #222), L-183 (#409), L-426 (the round's version collisions,
now set by hand at merge) and L-550 (#471 with #420). C-34 and C-35
correct our #290 shell suggestion and round 45's claim that
tests/dipper-probe guards anything.

Stacked on #551 (round 45's lessons).

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
derekross pushed a commit that referenced this pull request Nov 10, 2026
Omarchy 4.5 changed the files most of our patches touch, so one series
no longer fits both. omamobile-omarchy now carries two families and picks
by the installed Omarchy version: patches/ is the 26-patch port to
Omarchy 4.5.0, and legacy-4.0.6/ is the 4.0.6 series as it was, 29
patches with #420's double tap to wake (0029). Any other Omarchy version
is refused, so a new Omarchy never gets a half-applied series.

Checked against the signed aarch64 packages: 29 apply on omarchy 4.0.6-1
and 26 on 4.5.0-3, each with --fuzz=0, and the package's tests pass on
both. The 4.5 family ran on a contributor's Pixel 7 Pro on Omarchy
4.5.0-3. On the maintainer's Pixel 7 Pro (Omarchy 4.0.6, 0.1.37
installed), the upgrade to 0.1.39 on 2026-11-09 rewrote no Omarchy file:
status and verify report 29 patches from legacy-4.0.6, the log says
"series already applied", and the shell didn't restart.

Co-Authored-By: Codex <codex@openai.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
Thirty-one lessons from round 44's reviews, confirmed by the maintainer, plus a second instance of L-454 from derekross#493. derekross#420's and derekross#471's reviews found the same patch-series lesson from either side; it is one entry (L-550).

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
…#421)

The bar, the home widget and quick settings follow UPower's own
battery.state instead of "charger online": charging and time estimates
show only when the battery reports that state, and pending and unknown
states are shown as such, so external power no longer implies charging.

The bar's Full glyph loses its bolt too (on the 7 Pro as well as in
quick settings): the icon helper chose the charging-full glyph for Full
even when charging was false, and a Pixel 6 reporting Full at 91%
showed the bolt. The battery-state test covers it. Tested on a Pixel 6
in all three states on all three surfaces.

0.2.104: main already carries 0.2.101 (derekross#464), and derekross#482 and derekross#420 hold
0.2.102 and 0.2.103.
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
test/weather.py (since 0.2.90, derekross#366) runs plugins/omamobile.home/HomeModel.js
under node, but the PKGBUILD never said so. CI's runner and the build image
have node, so nothing noticed until tools/phone-install-package built the
shell on the Pixel 7 Pro, which has no node: check() failed at
"omamobile-weather" and the package wasn't installed (2026-11-08, while
installing derekross#420 for its test). phone-install-package installs a package's
checkdepends for the build and removes them afterwards, so with this line
the phone built and installed it and nodejs was gone again.

No version change: the package's files are the same.

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
…, a single touch leaves the screen dark (derekross#420)

A phone in a pocket woke on any touch: the locked screen lit on the
first contact, and a brush of the palm did the same. Now a double tap
wakes it, as on Android, and a single touch leaves it dark.

omamobile-shell's new omamobile-doubletap (a user unit for every
graphical session) reads the touchscreen without a grab: the first
input device with INPUT_PROP_DIRECT, BTN_TOUCH and ABS_MT_POSITION_X,
with the user in the input group. It times two taps from one lift to the
next touch. While the screen is dark it wakes the lock through its IPC
(0009), as the power key does, unless the proximity sensor reads near.
Lit, it does nothing, so apps keep their own double taps. Hyprland's
misc:mouse_move_enables_dpms is off on the phone (hypr/phone.lua), so the
first touch no longer lights the panel before the lock sees it.

omamobile-omarchy's 0029 makes the locked, dark lock ignore every touch
(the glance's wake-only touch from 0023): only omamobile-doubletap or the
power key wakes it. On a computer (keypad off) a click or a mouse move
wakes the lock as before. A dark, unlocked screen locks within about
1.5 s, so a tap after that reaches nothing.

Tested on the maintainer's Pixel 7 Pro on 2026-11-08 (Omarchy 4.0.6,
both packages built from this branch merged onto main, 29 patches
applied): a single touch while dark stayed dark; a double tap lit the
lock; covered, a double tap didn't wake, and uncovered it did; a tap on
a dark unlocked screen reached nothing and Power woke to the lock; a
short press locked once; double taps inside an app behaved as before.
The watcher used 16.9 s of CPU in 49 minutes of that session.

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
)

* omamobile-shell 0.2.107: Grok Bot fits the phone's screen

Grok Bot (grok-bot from [omarchy], x.ai's agent, an Electron app) sets
a minimum window size of 512x520 (SAND_MIN_WINDOW_SIZE in its main
process). The Pixel 7 Pro is 480 px wide in logical pixels, so the app
laid itself out 512 px wide inside a 480 px window and ran off the right
edge: the ends of messages, the header buttons and the composer's
microphone were cut off. Chromium had the same problem and setup-user
gives it --force-device-scale-factor=0.83.

setup-user now does the same for Grok Bot, whose launcher reads
~/.config/grok-bot-flags.conf. The factor comes from the panel's mode
(the DSI connector in /sys/class/drm) and the monitor scale setup-user
seeds, so the layout gets about the 578 px Chromium gets on the 7 Pro:
0.83 on the 7 Pro (480 px), 0.70 on the 9a (405 px), 0.62 for a
1080 px panel at scale 3 (360 px), where 0.83 would leave 434 px, still
under 512. With no readable mode it falls back to 0.83. Once, and a
factor the user's file already has stays theirs.

Tested on a Pixel 7 Pro (Omarchy 4.5.0-3, omamobile-shell 0.2.100):
the factor by hand made the window fit (screenshots before and after);
the new step, run alone on the phone against a scratch home, read
1440x3120 and wrote 0.83. test/setup-user.sh covers the 7 Pro, the 9a,
a 1080x2400 panel, no mode, a user's own factor and two logins; its
first new check fails against the old script.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TTVicoVA5YdXzobkXGujF9

* omamobile-shell 0.2.108: renumber above main's 0.2.107 (derekross#420)

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Derek Ross <derekross@gmail.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
Round 45's confirmed-candidate lessons from the omarchy-mobile and
linux-omarchy-mobile reviews, plus five from the maintainer session of
2026-11-08 (the four build steps a test kernel needs for push, reboot-to
and the modem's clean-stop marker, pacman steps on a non-Arch build
host, a new dependency's example config and the identifier scan, and
checkdepends). Six candidates that were second instances of existing
lessons became "Again in round 45" sentences on L-333, L-427, L-430,
L-454, L-532 and L-547. C-32 and C-33 correct round 44 on derekross#420's 0029
(applies to omarchy 4.5.0-3 at fuzz 0) and on kernel derekross#219's test phone
(the cover letter names it).

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
Round 46's confirmed candidates from the reviews of the open
omarchy-mobile and linux-omarchy-mobile pull requests, the round's own
process slips (shared refs and scratch files deleted by glob, a
maintainer fixup that skipped the package's test/all, an amend on a
contributor branch) and the 2026.11.08 release PRs (derekross#547, derekross#548, derekross#549,
derekross#550, derekross#560). Repeats of existing lessons extend them instead: L-45
(kernel derekross#222), L-183 (derekross#409), L-426 (the round's version collisions,
now set by hand at merge) and L-550 (derekross#471 with derekross#420). C-34 and C-35
correct our derekross#290 shell suggestion and round 45's claim that
tests/dipper-probe guards anything.

Stacked on derekross#551 (round 45's lessons).

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
…oss#471)

Omarchy 4.5 changed the files most of our patches touch, so one series
no longer fits both. omamobile-omarchy now carries two families and picks
by the installed Omarchy version: patches/ is the 26-patch port to
Omarchy 4.5.0, and legacy-4.0.6/ is the 4.0.6 series as it was, 29
patches with derekross#420's double tap to wake (0029). Any other Omarchy version
is refused, so a new Omarchy never gets a half-applied series.

Checked against the signed aarch64 packages: 29 apply on omarchy 4.0.6-1
and 26 on 4.5.0-3, each with --fuzz=0, and the package's tests pass on
both. The 4.5 family ran on a contributor's Pixel 7 Pro on Omarchy
4.5.0-3. On the maintainer's Pixel 7 Pro (Omarchy 4.0.6, 0.1.37
installed), the upgrade to 0.1.39 on 2026-11-09 rewrote no Omarchy file:
status and verify report 29 patches from legacy-4.0.6, the log says
"series already applied", and the shell didn't restart.

Co-Authored-By: Codex <codex@openai.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants