Repository navigation
omamobile-shell 0.2.135: an optional device.lua after the shared phone rules - #204
Conversation
|
Thanks, Nate. The hook is small and leaves every current phone alone. Please rebase onto main (the shell is at 0.2.45 and the next few numbers are taken by open PRs, so take the next free version when it's ready), wrap the |
|
A friendly check-in: this has been quiet for over a week and now conflicts with main. If you'd like to continue, a rebase onto main will get it reviewed again. If I don't hear back within a week I'll close it to keep the queue tidy; it can be reopened any time. |
|
Still active. I'm rebasing this onto main now. As you asked, it lands together with the comet board DTS, which I'm rebasing onto the zumapro stack that came in with linux-omarchy-mobile#187. |
Keep the initial Comet descriptor, recovery record and boot policy separate from the daily integration package. Follow current main for shared USB, mirrorlist, repository hook and reboot-to files. No hardware service starts on installation and no devinfo writer or boot-success unit is shipped. Install, status and README consistently keep boot-success disabled under the maintainer's issue derekross#143 policy. Leave SOC_COMPATIBLE commented out, as tegu does, so install-phone and flash refuse comet until the public board DTB lands: comet has no stock slot to fall back to. The package depends on omamobile-shell without a version; the descriptor scale it reads comes with derekross#204. Public board DTS and module sources, an exact candidate build and an owner smoke report remain prerequisites. The notes label older private-kernel evidence. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Thanks, Nate, and good to hear it's still active. Nothing new has arrived on this branch yet, so my points from 10-31 stand. Since the base is now two weeks old, here is what the rebase will meet:
It still lands together with #203. |
…e rules A phone with more than one panel or its own input devices (the Pixel 9 Pro Fold's inner and outer screens) needs session settings no other phone has. hypr/phone.lua now loads /usr/share/omamobile/hypr/device.lua after the shared rules when a device package installs one, so that package can carry them without a device branch in phone.lua. The load runs under pcall: a syntax or runtime error in the device file is printed (Lua's print, on Hyprland's output) and the shared rules stay in force. With no device file nothing changes, so every phone on main is unaffected. test/phone-rules.lua loads the real phone.lua with that one path redirected to a temporary file, and test/all runs it with a valid file, a syntax error and a runtime error; the default run has none. Each case checks the device file ran only when valid and the shared rules still hold. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
096f8f2 to
a5d5718
Compare
|
Rebased onto main (c462357) and cut down as you suggested. Point by point:
It still lands together with #203. Comet's own |
|
Thanks, Nate. This is exactly the cut I asked for, and it is ready as far as the code goes. The scale seed is gone and Two things are left, both on my side:
It no longer depends on #203 for anything, so I may merge it ahead of it. Comet's own |
A Hyprland screen shader, pow(rgb, 1.55) then x0.88, from blind rounds by eye on the 8 Pro's panel. The cause of the flat mid-tones is unknown, so this is a stopgap. It is set from /usr/share/omamobile/hypr/device.lua, which the phone.lua hook from derekross#204 loads; this depends on derekross#204. It is off by default: ~/.config/omamobile/screen-look-on turns it on, and without it device.lua doesn't call hl.config. glslang and lua become checkdepends for the test.
derekross
left a comment
There was a problem hiding this comment.
CI ran on this head and both jobs pass, your device.lua case among them.
One thing changed since my comment this morning: another pull request took omamobile-shell 0.2.134 on main a little later, so yours now needs 0.2.135. As I said, that one is mine: I'll merge main into your branch, set the version, retitle the pull request and merge once CI is green on that push. Nothing is needed from you.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
What and why
A phone with more than one panel, or input devices of its own, needs session settings that no other phone has. The Pixel 9 Pro Fold (comet, #203) has an inner and an outer screen.
hypr/phone.luanow loads/usr/share/omamobile/hypr/device.luaafter the shared rules when a device package installs one, so that package can carry those settings without a device branch inphone.lua.pcall. A syntax or runtime error in the device file is printed asomamobile: /usr/share/omamobile/hypr/device.lua: <error>(Luaprint, so on Hyprland's output), and the shared rules stay in force.phone.lua(setup-user adds it tolooknfeel.luaonce) gets the hook, for new and existing users alike.omamobile-setup-useris untouched.This lands together with #203. Comet's own
device.lua(its inner-panel touch mapping) follows as a small device-comet change once both are in.Device and build
Checks (no phone needed)
tools/pr-check upstream/mainpassestests/all: not run for this change. It touches only omamobile-shell, andtests/packagesfor that package is below.packages/omamobile-shell/test/allin the Arch Linux ARM build container, withluainstalled: passes, excepttest/power-overlap.sh(the last test in the file), which fails the same way on unchanged main in this container, as a timing failure under arm64 emulation. In one earlier runtest/sleepd.shtimed out too; it passes when run alone, on main and on this branchtest/phone-rules.luawithDEVICE_HOOK_CASEunset,valid,syntaxandruntime: all pass. The device file runs only when valid, and every shared-rule check still passes.Co-Authored-By:trailerOn the phone
device.lua. The test covers this under the Lua stub of Hyprland's API, not in a live session.userdata, weakens the storage guard, or reads registers from userspace.🤖 Generated with Claude Code