Repository navigation
omamobile-settings 0.1.0: add a native device information foundation - #526
Conversation
Introduce an optional Qt Quick application with searchable, read-only device and software information. Keep navigation, theme observation, desktop launching and the single-instance D-Bus entrypoint together so native settings pages can be contributed independently with their backend contracts. Exercise navigation, atomic theme updates, resource loading, process teardown and the D-Bus entrypoint using neutral data and a private offscreen session. Install QtSvg in CI for the application's declared module and SVG icons. Co-Authored-By: Codex <codex@openai.com>
…oot cell wording (derekross#488) Two optional points from the review of derekross#472, and the record of slot B on Lef's phone. The roadmap dates the marked state: B stayed marked (0x35 read 06) from the local unit's last write until the first flash after the unit was masked on 2026-11-07, when it read 04. Since then a boot spends one of B's tries (retry count 3 after set_active, 2 after one boot), so the flash-before-the-last-try routine applies. slot-successful:b was not read after those flashes; Lef will add it after the next one. The bit is not cleared by hand. The DEVICES Boot cell says Lef's phone, not this phone. Signed-off-by: Lef <me@lefteros.com>
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>
Build the zram driver as a module, with zstd (Omarchy's algorithm) as the default and lz4 available. Omarchy's settings configure swap on zram on every aarch64 machine (zram-generator, 90-omarchy.conf and the vm.* tuning in 99-omarchy-sysctl.conf), and stock Android runs one on the Pixel 9a (half the RAM), but our kernel has no driver, so a phone has no swap at all. On the 9a (7.4 GB) a cargo build next to Brave and the shell left 177 MB, stalled the session for minutes and ended at the OOM killer, which took the terminal's app scope with it. device/tegu/NOTES.md records the test: with half the RAM as zram, 5 GB more allocated ran without an OOM kill and with SSH answering; with Omarchy's RAM-sized zram, 6 GB more allocated hung the phone until a restart. Stock Android's own setting on the 9a, and how to set it up by hand until the image ships zram-generator, are there too. Nothing changes at boot until something configures a zram device: the modules load only on demand. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
) The follow-ups from the review of derekross#463. `unlock` reads the lock state again right before it types the PIN, so an unlock in the meantime (a fingerprint, the person holding the phone) no longer sends the PIN and a Return to the focused app; the new test fails without that re-check. `help` prints the whole header, the tests run against an empty /proc of their own instead of the host's, the install line in the docs names 0.1.1, and a typo is fixed. Package test: 57 ok.
|
Thank you, @marselnenaj. I'll leave the full review until this is out of draft, but a few early questions now, so they don't get harder to change as it grows:
The About page reads only the model, OS, kernel, memory, storage and package versions, with no serials or addresses, and the screenshots use the preview data. Thank you for that. When it's ready, please mark it ready for review and add a smoke report and a screenshot from a phone. |
…up, and derekross#456's notes (derekross#509) From the review of derekross#456: - device.conf (and the package's copy): NV_BACKUP_EXTRA adds trusty_persist and trusty_userdata next to devinfo, as husky's package does since derekross#449, so tools/nv-backup --check shows when Android has written Trusty's storage. - NOTES.md: trusty_userdata in the list of per-phone partitions, and the provenance row for the new value. - BRINGUP-M2.md: don't use Suspend in M2.7 until husky's U6; linux-omarchy-mobile#206 spelled out; tools/phone-smoke's report from the M2.6 root (test 18). - ufs-probe.dtso: "If the board file's bootargs change". Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The OnePlus 6 (enchilada) ABL boots an Android boot image header v0: a gzip kernel with the DTB appended, no dtbo and no vendor_boot. Until now nothing in the tree could make one, so every path refused BOOT_METHOD=android-v0 and the port's first-light image was built by hand. load_device now accepts android-v0 with its own required keys and the six header fields (PAGESIZE, BASE and the offsets), checked as numbers; BOOT_IMAGES must be boot alone and boot_b must be on UFS_RW. images packs boot.img with the pinned mkbootimg.py: gzip -9n of Image with the (overlaid) DTB appended, the shared initramfs as the ramdisk, CMDLINE and the descriptor's offsets. mkdtboimg is only required for android-v4. The port stays RAM-only: fastboot, flash and push refuse android-v0 and point at `fastboot boot`, and the installer still refuses the method, until the slot-B install and its CONTRIBUTING.md 3.1 entry are agreed. Also in device/enchilada/NOTES.md and docs/DEVICES.md, from review on derekross#445 and derekross#432: link the maintainer's decision on restoring stock; say the firewall failure booted from boot_b after --set-active=b; and drop the guess that each rescue boot spends one of slot A's tries (after one more rescue boot A was still at 5 and B at 7). Tested: tests/boot-method (enchilada images reach the kernel check; fastboot, flash, push and both installer modes refuse; bad header fields refused; no phone access) and tests/images-layout, which gains an android-v0 layout test (one boot.img, header v0, page size and load addresses, kernel = gzip Image + DTB, ramdisk = initramfs, deterministic) while the gs201/gs101 baseline still matches 1e7a998 byte for byte. Against the 7.1.0-rc1-sdm845 build used for first light on 2026-11-04, images gives a boot.img byte-identical to the one that was booted then with `fastboot boot` (sha256 78d7bd1a...0d98c). Nothing was sent to the phone for this change. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…rekross#519) The Arch root on the phone has no /dev/disk/by-partlabel, so the M0 commands that read the radio partitions through it fail there: M0.1's checksum loop (both plans), M0.5's blockdev --getro check and 1.6's name-to-device listing. derekross#490 fixed the same lookup in FINGERPRINT-PLAN, and the maintainer asked for these two plans to get the same treatment. Each command now finds its partition the way FB1 and F19.0 do, by PARTNAME in /sys/class/block/*/uevent, and 1.6 lists those PARTNAME lines in place of the by-partlabel symlinks. 1.6's "take a fresh backup" note now points at tools/nv-backup, which finds the partitions the same way. No step does anything else differently: the same partitions are read, nothing is mounted or written. MODEM-PLAN's M0 result note (2026-10-12) records what was run that day and is left as it was. Checked: the changed shell snippets parse (bash -n, sh -n), the lookup picks sda5 for efs over efs_backup on a fake sysfs, tools/pr-check origin/main and tests/all. Docs only; no phone was used. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…s copies (derekross#520) Follow-up to Derek's review of derekross#490. F3's rollback pointed at F19.3's, which removes all of /var/lib/omamobile-fp. Once storageproxyd has run, that directory also holds Trusty's TD, so the rollback could delete it. F3's quote already said to run it only where storageproxyd never ran; it now has that as a real Stop line (F19.1's), and its rollback removes only what F3 writes on the phone: mfg_data.img and persist-ss/. F19.3's rollback is unchanged (the maintainer's to fix). derekross#490 dropped F3's laptop copy of mfg_data, which tools/nv-backup doesn't take. F3 takes it again with a read-only dd from the partition found by PARTNAME (this root has no by-partlabel), refused unless the block device reads ro=1, and checks its hash against F19.3's phone copy, so a phone without the 10-13 copy keeps one. FB1 still compared against the 10-12 backup, and FB1 and F19.0 each carried their own checksum loop. Both now run `tools/nv-backup --check "$B"` against the newest backup (B, now in section 6's variables): it finds the partitions by PARTNAME, refuses a writable or mounted one and reports SAME or DIFFERENT for each. Docs only; nothing ran on a phone. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Agreed on review size: this PR will stay foundation plus About; functional pages will follow separately with their backends. Complete application integration remains separate. The source comes from my own Settings worktree and is my MIT-licensed code. I will clarify that in the README, remove the unused D-Bus palette hint, and pass the package version through the build. I agree that Material with substituted colors is insufficient. I am preparing semantic theme tokens for colors, typography, surfaces and control geometry, with stable navigation/layout, so themes can customize the app consistently. I will bring an updated foundation preview for agreement before extending that design across page PRs. This remains draft until those changes and qualification are complete. I will add an actual phone screenshot and smoke evidence, distinguishing development-phone adaptations from the public head. The smaller foundation will not replace the complete app already installed on the phone. |
Remove the unused surface palette hint and its window ownership plumbing until a public consumer needs it. Clarify the own-worktree MIT provenance, and share the package version with production and session-fixture builds. Check the public version entrypoint without constructing hardware controllers. Co-Authored-By: Codex <codex@openai.com>
Replace the fixed Material skin with shared native Qt controls and semantic tokens from Omarchy colors.toml and shell.toml. Follow the system monospace font, theme typography and control states live while keeping the existing read-only navigation and About scope. Observe atomic theme replacement, preserve known colors on malformed input and keep control height responsive to font metrics. The foundation retains the package-derived version and does not add pages, helpers or service actions. Co-Authored-By: Codex <codex@openai.com>
…s#491) The extractor wrote straight into OUTDIR. The plans run it into the default out/vendor-firmware and tar that tree onto the phone, so a rerun for another image or device kept the earlier run's firmware and sensor registry files, and the firmware went into vendor-firmware.sha256 as if it were this run's. On ext4, debugfs exits 0 when a copy fails: for a missing file it writes nothing, so an older copy of the same name passed the size check, and for a short write (a full disk, a file size limit) it leaves a truncated file that passed too. A run that failed half way left a mix of old and new files. Each run now extracts into an empty staging directory inside OUTDIR (the -c camera APEX into a new file beside it), runs the checks there, checks that the checksum list names exactly the mapped files, and only then moves the previous outputs aside and the new ones into place: renames within OUTDIR, then the -c file. A failed run leaves the previous outputs and the -c file as they were; if a rename fails while publishing, what was already replaced is put back. The script replaces only what it writes: usr/lib/firmware/updates, usr/share/omamobile/sensors, vendor-firmware.sha256 and the -c file; the rest of OUTDIR stays. An ext4 copy fails on any debugfs message but its banner and the exact notes that it can't give a file the image's owner (as a user); the messages are read in the shell, so a nearly full disk can't hide them. A -c that names a directory, or a path inside the outputs the run replaces, is refused before anything is extracted, and so is an -o that resolves to / (/, //, /usr/.., a link to /), before the scratch directory is made. tools/phone-provision already extracts into a new directory per run; it calls the tool the same way as before. tests/extract-vendor-firmware (run by tests/all) runs the tool on made-up files with stand-in debugfs and fsck.erofs that print real debugfs's messages, and, where the host has mkfs.ext4 and debugfs, on a real ext4 image of the same files with a root-owned registry: a reused directory keeps no earlier firmware or registry file, a husky run drops a cheetah registry, and a missing file or registry, a short write (also under a file size limit), a failed sanity check, a missing camera APEX or a failed rename leaves the previous outputs byte for byte. Against the tool before this change, only 2 of its checks pass. No factory image was read and nothing ran on a phone. Found in an audit of the tools. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… on the firmware's default calibration without speaker.cal (derekross#496) Turns on omamobile-audio's OMAMOBILE_SPK_DEFAULT_CAL=1 for the Pixel 7, with a drop-in for omamobile-speaker-protection.service: some units have no speaker.cal in persist, and stock's HAL then runs the protection firmware on its tuning's default calibration. A phone that has speaker.cal still uses it, c9 and verify stay refused on the default, the volume ceilings are unchanged (stage media, 737/10), and /etc/omamobile/speaker.conf is read after the drop-in, so a line there still turns it off. Evidence: derekross#427, the contributor's comment of 2026-11-07 15:33 UTC (a start and a playback on 0.5.16, installed with pacman -U, on a Pixel 7 without speaker.cal), and with this drop-in: the service gets the setting, start logs "using the protection firmware's default calibration", stage media at 737/10, a playback, and the watchdog stays active. omamobile-audio: the one debugfs call whose message is matched ("File not found by ext2_lookup") runs under LC_ALL=C, so a phone in another language can't trip it.
…ocus (derekross#494) The 9a's main camera focuses through an onsemi LC898129, which now has a lens driver (linux-omarchy-mobile#213). Build it as a module next to the 7 Pro's SEM1215SA (CONFIG_VIDEO_LC898129=m), and turn on contrast autofocus in the S5KGN8 tuning file (libcamera 0057), with the coarse and fine steps of the 7 Pro's main camera. device/tegu/NOTES.md records where the register map came from, which registers were written while working it out (only the focus target), and what was measured on the phone: the controller's registers, the 12-bit focus target and its Hall read-back, which end is near, and omacam focusing on a subject at about 25 cm, on libcamera 4.22 with the Af lines added by hand. DEVICES.md adds the autofocus to the 9a's camera cell, in review with the kernel PR, and says the far end isn't checked yet. The board's lens node is in the notes; it waits for the 9a's camera device tree (linux-omarchy-mobile#164) and isn't in a pull request yet, as the camera cell says. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…tead of rebooting (derekross#506) The OnePlus 13's reboot-to passed "--help" to the kernel as the reboot reason, as the Pixels' did before derekross#459: the phone rebooted. An argument that starts with "-" is never a reason, so print the usage and exit (0 for -h and --help, 2 for anything else), with the same text as the Pixels'. It has no libc, so the usage goes out through write(2) and the exit through exit_group(2), and a small loop compares the option. Both copies (the package's and the initramfs's) stay identical, as test/all checks. Not run on a OnePlus 13. Built as the package builds it (gcc -nostdlib on an aarch64 Linux, a Pixel 7) and with clang for aarch64, then run with options only: --help and -h printed the usage and returned 0, -z and --bootloader returned 2. The build from main, run under qemu-aarch64 as a normal user, shows the old path: --help reached reboot(... RESTART2, "--help") (refused there, EPERM). test/all passes. I used AI help (Claude) to work on this. The version (0.1.1) was set at merge.
… handshake (derekross#504) Some phones take longer than ssh's default to finish the first SSH handshake while their USB link is still coming up. tools/phone-ssh and tools/phone-ssh-proxy now take the connect timeout from OMAMOBILE_SSH_CONNECT_TIMEOUT, defaulting to 6 s as before, so a slow phone can ask for more without making a dead connection slower to report everywhere else. tests/phone-ssh covers the default and the override, and that the proxy receives it.
…#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.
…loses (derekross#524) Flicking a card up to close it, or any window closing while recent apps are open, dropped the switcher back to the first (newest) card. The ListView's model is root.windows, a new JavaScript array whenever the toplevel list changes, and a new model puts currentIndex back to 0; the re-centring after the count change then centred that 0. The view now remembers the card the user is on, by its window and by its index, whenever a swipe ends or it centres a card, and comes back to it after the model changes: the same window if it is still open, else the card that took the closed one's place (clamped to the last). Opening recents still starts on the newest card. test/switcher-touch: test_closing_a_window_keeps_the_place swipes to the second card, closes another window (the same window stays current), then the current one (the next card, not the newest). It fails without the change ("the same window after another one closed") and passes with it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…kross#525) The Hotspot tile runs `nmcli device wifi hotspot`, whose connection uses ipv4.method shared. NetworkManager's shared mode runs dnsmasq for DHCP and DNS on the hotspot, and nothing in the root image pulls it in (only omamobile-waydroid depends on it, and images don't ship that). The access point comes up, then the activation fails: ip:shared4: could not start dnsmasq: Could not find "dnsmasq" binary state change: ip-config -> failed (reason 'ip-config-unavailable') and the phone goes back to its Wi-Fi network. Seen on a Pixel 9a; the missing binary is independent of the phone. With dnsmasq installed the same connection activates, gets 10.42.0.1/24 and dnsmasq serves 10.42.0.10-254. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Phone evidence for the proposed theme revision: I ran the exact AArch64 candidate binary ( Read-only
Optional observations: display, GPU, touch, buttons, Wi-Fi, battery, sound cards, sensors and thermal checks passed. Bluetooth had no controller; camera enumeration was skipped; a development service was failed. Those existing device findings are not resolved or attributed to this read-only app change. This phone has local Omarchy shell adaptations, so this is application-preview evidence on a development installation, not the clean installation test requested separately on #471. No package replacement, shell restart or reboot was performed. |
|
@derekross @marselnenaj a heads-up and a question, so we don't build the same thing twice. On my 8 Pro I have been trying a few Settings pages as a plugin of the shell (local only, nothing opened as a PR): keyboard haptics levels with an Off that silences feedbackd for the whole phone, a Change PIN page (the lock PIN is the account password, so it runs I won't open a Settings plugin next to this app. @derekross, where should pages like these go? If the native app is the place, I can offer the two backends as small scripts that a native page would call ( |
|
Thanks for coordinating this and documenting the fbcli stdin finding in #538. Small, tested haptics and PIN backends would be useful for native pages once the foundation and integration point are agreed with @derekross. This PR still contains only the foundation and About page; it does not include a parallel PIN-change implementation. My local development app already has feedbackd profile/strength controls and a separate Back-gesture feedback preference. That preference suppresses only the Back event and leaves the global profile unchanged. I will keep that distinction explicit rather than label a global silent profile as keyboard-only Off. Calls and alarms need their own qualification before making broader promises. Update: I reproduced your stdin finding using the genuine feedbackd 0.8.9 CLI and GLib with an inert feedback implementation, then the actual Quickshell execDetached caller. Detached /dev/null input ended the synthetic event immediately; an empty closed pipe allowed natural completion. I have installed the minimal caller correction on my development phone and verified the shell restarted with its existing authentication, feedback preference and global profile preserved. I have not measured the physical motor effect, so this is not a vibration-strength or timing claim. This remains separate from the foundation PR. Credit for identifying the issue belongs to @Somnius and #538; I would also be happy to build on your small backend contributions once the maintainer settles the placement. |
|
To make the overlap more concrete, here is what is already implemented in my separate native development app. These pages are not all included in #526; this PR deliberately remains the foundation and About, with functional pages to follow as small contributions.
Some secondary entries still launch existing tools; replacing those with complete native pages is unfinished. Device brightness/refresh support, Bluetooth availability and recording performance also remain separate hardware/runtime qualifications. I do not want to present the existence of a settings row as completed device support. What #538 would add or improve:
I would prefer to reuse and credit small, tested backends rather than introduce duplicate haptics or PIN implementations. I can use a shared backend contract and use native pages if that is the maintainer's chosen integration point. Your current 8 Pro measurements should remain identified as device-specific until reproduced elsewhere. |
|
Updated to The complete GitHub test and policy jobs pass on this head: https://github.com/derekross/omarchy-mobile/actions/runs/42530331977. Local package validation also passed, including the 12 Qt entries and private-bus/version checks. The actual phone preview and read-only smoke evidence are in my earlier comment; that development-phone evidence is not a claim that #471's full migration sequence has been qualified. |
…ekross#559) Fourteen tools printed --help as a hard-coded line range of their header comment (sed -n '3,NNp'), the pattern that cut kernel-pr-check's header after derekross#287, derekross#288 and derekross#310 each edited it. On derekross#287 the maintainer suggested working the range out from the file; derekross#529 did that for kernel-pr-check alone. This moves all of them onto one helper, tools/lib/usage.sh: usage_from_header prints the comment lines from line 3 up to the first line that isn't a comment (derekross#529's awk line), with "# " taken off or, with --raw, kept. Tools load it only for --help, so the copies the tests make of a tool, with only the lib files it needs, still run. kernel-pr-check moves onto it too. Ten tools print byte for byte what they printed before (cmp of stdout, stderr and exit status, and of phone-demo's and phone-mirror's error usage). Five ranges were already wrong, and print the header now: fetch-toolchain and build-flame-probe printed two lines of code after it; install-phone cut --reinstall, --unsigned and its last paragraph, build-rootfs cut --channel, and release cut its check-bundle stage. build-packages, phone-demo and release print only the start of a longer header, so the helper takes a paragraph count (a bare "#" line ends one) and their output stays what the range meant. install-phone keeps the awk line inline, as nv-backup and phone-provision already do: make-bundle ships these three alone, without tools/lib (review lesson L-497). tests/tools-help fails if any tool still hard-codes the range, and runs --help on every tool that prints its header (the helper's users and the three inline ones): it must print the header, or whole paragraphs from its start, and no code, and a copy whose header has grown by a line must print that line too. It replaces derekross#529's single-tool check in tests/kernel-pr-check. On the base it fails on the range check, and the per-tool check fails for 13 of the 15 tools. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Derek Ross <derekross@gmail.com>
…he 7 Pro's modem starts at boot by default (derekross#574) * AGENTS.md: stop the modem before a push or a reboot-to reboot-to is a raw reboot syscall and bin/omamobile push ends in it, so neither stops omamobile-modem.service. Its clean stop never runs, the modem-up marker stays, and on the 7 Pro, where the modem starts at boot, the next boot skips the modem once as if the last one had hung; a second unclean end in a row holds it off until --retry. That happened on 2026-11-08 after a reboot-to bootloader taken to charge a flat battery. Add a bullet to the push notes in section 2: stop the unit first, under the human's go for the reboot, or use systemctl reboot when no reason is needed; what the stop costs today (about 40 s, logged as failed, marker still removed); why this is not the restart the modem gotchas forbid; and that the open derekross#403 would make reboot-to count as a clean stop. The reboot-to gotcha and the NV overlay note in section 4 point to it; the NV note keeps nv-overlay capture as the narrower alternative. Part of derekross#532 (its docs request only; the watcher and SIGTERM fixes are package changes). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * AGENTS.md, MODEM-PLAN: modem at boot by default, derekross#403's reach, B4 Addresses the review on derekross#574. - AGENTS.md, the modem restart gotcha: it still called the 7 Pro's stack at boot opt-in (`systemctl enable omamobile-modem-boot.target`), which stopped being true with derekross#398. It now says the stack starts by default since device-cheetah 0.1.43 (MODEM-PLAN M4.6b), and that `systemctl mask` keeps it off from the next boot, as the target, the package's install message and the device README say: `disable` doesn't remove device-cheetah's vendor wants link. - AGENTS.md section 2: derekross#403 doesn't make `reboot-to` a clean stop. Its note lets the next boot's start clear a `modem-up` marker from the boot `reboot-to` ended, but the CP's ordered power-off and the NV capture still run only in the unit's stop, so the advice to stop the modem first stands with derekross#403 merged too. - MODEM-PLAN M4.6 B4 rebooted with `reboot-to omamobile` and expected the clean stop's "modem stopped (...): modem-up removed", which a raw reboot syscall can't produce. The step is there to show that a reboot's stop clears the marker, so it now reboots through systemd, as N5 does, and its pass line takes today's "timeout" result (derekross#532). The recorded results are unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * MODEM-PLAN: M4.6b B8 and B9b reboot through systemd, as B4 now does Part of the review on derekross#574. B8 plants another boot's modem-up marker so that "the reboot's clean stop keeps a marker that isn't this boot's", and its pass line reads the previous boot's "is another boot's ... kept"; B9b's planted reboot rests on the same stop hook, and its last "clean reboot" checks that `--stopped` removed the refused copies. All three rebooted with `reboot-to omamobile`, which stops no unit, so the stop hook they test never ran: the same gap as B4. They now reboot with `systemctl reboot`. The recorded results are unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…derekross#516) Calculate notification capacity from the sheet's resting position instead of its animated position. The shade still follows sheet movement, while expansion and screen geometry continue to update its available room. Exercise the production geometry expressions under real Qt bindings. The baseline changes room and height during reveal; the candidate keeps both stable and preserves final portrait, landscape and short-screen geometry. A control restoring the old calculation fails the assertions. Keep the change independent of host interfaces, safe-area extensions, backend scheduling and hardware support. The package increment is proposed from the current public main and can be adjusted by the maintainer at merge. No phone frame-rate or runtime qualification is claimed. Co-authored-by: Codex <codex@openai.com> Co-authored-by: Derek Ross <derekross@gmail.com>
…ekross#518) * omamobile-shell 0.2.102: defer open-time backend refresh until settled Opening Quick Settings starts eleven backend reads together while the sheet animates. Queue those existing read entry points after reveal and tracked sheet interactions settle, with one start per 48 ms. Coalesce unfinished open requests and cancel pending work when the sheet closes. Keep direct actions and their confirmation reads immediate. Exercise the production queue declarations under real Qt bindings for settling, staggering, cancellation, coalescing and scroll/edit gates. A synchronous burst control fails the timing and coalescing assertions. The queue tracks reveal animation, background drag, the four Flickables and edit mode. It does not track slider presses or serialize process completion. Existing polling and in-flight completion remain unchanged. No physical frame-rate improvement is claimed. The proposed package version must be allocated again at merge. Co-Authored-By: Codex <codex@openai.com> * omamobile-shell: prioritize visible Quick Settings reads Dispatch notifications first, then DND and brightness, so the shade does not wait behind less visible status reads on every open. Editing itself no longer blocks the queue until Done; actual reveal, drag and scrolling still pause it. Keep action readbacks immediate and preserve close cancellation/coalescing. Actual Qt tests cover first-screen order and idle edit-mode opens as well as spacing and gesture gates. Restoring either old behavior fails its test. Co-Authored-By: Codex <codex@openai.com> * omamobile-shell: name refresh probes in the scheduler fixture Record refresh entry points by name instead of their former numeric slots. This makes the notification-first order and immediate DND assertions readable without changing the production scheduler, its timing or gesture gates. The real offscreen Qt fixture passes all nine entries. The synchronous burst, old ordering and restored edit gate still fail their targeted assertions. Co-Authored-By: Codex <codex@openai.com> --------- Co-authored-by: Codex <codex@openai.com> Co-authored-by: Derek Ross <derekross@gmail.com>
…oss#521) * omamobile-shell: tick once when the Back gesture arms Use the same optional feedbackd button-pressed event as the launcher hold. Latch feedback per drag activation so crossing the threshold repeatedly does not request repeated ticks. Recognition, Escape dispatch, retreat semantics, fullscreen exclusion and compositor lock exclusion remain unchanged. Exercise exact source signal handlers in offscreen Qt without invoking feedbackd. Include both edges, inactive changes, re-crossing and retreat. Co-Authored-By: Codex <codex@openai.com> * omamobile-shell: make the Back haptic fixture readable Use normal Python indentation, separate statements and explicit imports. Remove unused json without changing the extracted QML handlers or tested gesture behavior. Six Qt entries pass and removing the latch still fails the once-per-gesture case. Co-Authored-By: Codex <codex@openai.com> * omamobile-shell 0.2.111: the version this lands as --------- Co-authored-by: Codex <codex@openai.com> Co-authored-by: Derek Ross <derekross@gmail.com>
…ekross#508) * omamobile-shell 0.2.102: serialize power key actions A second short press can arrive while the first is still locking and powering the panel off. Serialize short actions so the later press sees the completed transition. Use the captured release time for debounce, and close the lock descriptor in the action subprocess so background helpers cannot keep it held. The focused host regression fails against main with the wrong screen state and passes with this fix, covering wake, debounce, brightness preservation and lock descriptor inheritance. Co-Authored-By: Codex <noreply@openai.com> * omamobile-shell 0.2.102: bound the power-key lock wait Drop a short press with a diagnostic after five seconds waiting for the previous action. Keep display transitions serialized without queueing later presses indefinitely behind a stuck helper. The regression fails against the previous unbounded wait and passes with the timeout, preserving debounce/display state and recovering after the holder exits. Restrict its PATH to explicit utilities and desktop stubs. Keep the existing draft version; the maintainer will assign it at merge. Co-Authored-By: Codex <noreply@openai.com> --------- Co-authored-by: Gr-linkk <280405187+Gr-linkk@users.noreply.github.com> Co-authored-by: Codex <noreply@openai.com> Co-authored-by: Derek Ross <derekross@gmail.com>
…e screen and over the drawer (derekross#557) With the keyboard up on the home screen (the dock's search pill) or over the drawer (its search field), the keyboard covers the dock, and with it the swipe up. The strip wasn't mapped there, so the only way out was the keyboard's Hide key first. While the keyboard is up the strip now shows on the home screen and over the open drawer too. Its swipe up hides the keyboard and opens the drawer from the home screen (as the dock's swipe does), or closes the drawer over it. Over the drawer a long swipe or a hold opens no recent apps: in the first version it did, and the drawer stayed half open over the app picked from them. The dock still reserves nothing on the home screen, so the strip is the only reserving bottom window there. test/pill-touch: the keyboard on home brings the strip (alone), a swipe in it on home opens the drawer and hides the keyboard, and over the drawer it closes the drawer with no recent apps. test/all follows the new condition. Claude-Session: https://claude.ai/code/session_01176vSdGN6JWgiwDLYXe1ES Co-authored-by: den-kloot <1522975+den-kloot@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Derek Ross <derekross@gmail.com>
…s#462) * omamobile-gnss 0.2.8: keep the hang marker through shutdown Clear the 9a marker only after systemd has stopped the GNSS service cleanly. A failed stop or unexpected exit retains it so a later location request cannot repeat a hang. * omamobile-gnss: on the 9a, --retry by hand clears the marker and says what starts GNSS Run by hand, omamobile-gnss-start --retry cleared the hang marker and then refused with "no systemd invocation ID", and the skip message read as if --retry started GNSS. On the 9a it now says that the marker is cleared and that the next location request starts GNSS, in its unit; the skip message says the same. The 7 Pro's messages are unchanged. The test covers the by-hand retry: marker gone, nothing started. * omamobile-gnss: the skip messages give the retry's full path omamobile-gnss-start lives in /usr/lib/omamobile-vendor, not on PATH, so the two 9a messages that name its --retry now give the path.
…ck (derekross#568) A device package can give omamobile-modem its own ordered list of link modules (link.conf). The file is read as data, never sourced: it is refused, before the sleep inhibitor and the bring-up latch, when it is a symlink, holds a NUL or other binary byte, or names anything but module names and their parameters. Without the file the starter loads what it loaded before; the Pixel 7 Pro's path is unchanged (pcie-gs201-cp allow_hw=1).
…'s numbers (derekross#541) binderfs's major is assigned when the kernel boots (507 on the cheetah kernel), so lxc-omamobile.conf carried `c 507:*` and omamobile-waydroid-check compared it with /proc/devices before every start. LXC 7 can take the nodes by path instead (`device_cgroup_rule_parse_devpath()` in cgfsng.c): it stat()s the path when the container starts and uses that major and minor. The conf now has `/dev/binderfs/binder`, `/dev/binderfs/vndbinder` and `/dev/binderfs/hwbinder` and no hard-coded major. This is also narrower: the old line allowed every minor of the binder major, including binder-control. If a node is missing when the container starts, lxc-start fails with "Failed to parse device rule" instead of starting an Android that cannot boot. binderfs is mounted by Waydroid only when the container starts, so the check cannot stat the nodes; it wants the three lines in the conf instead. omamobile-waydroid-check also does for the render node what it did for binder: it reads the Panthor node's major:minor from sysfs and wants the matching `c MAJ:MIN rwm` line (`c 226:128` today). While doing that it stopped printing a wrong path (/dev/renderD128; the node is /dev/dri/renderD128). test/all covers the three path lines (each one deleted in turn, and the old `c 507:*` form not satisfying it), no hard-coded `c N:*` besides tty and pts, and the render node matching and not matching. With the check limited to binder, the render comparison never failing, or a binder major put back in the conf, the suite fails. Tested on a Pixel 7 Pro (GE2AE) running `7.3.0-rc6-omamobile-push1-g4cdfc1aa4b6b`, one session start each (the package's files copied to the phone for the run, then the originals put back, with matching sha256): - With the files of the previous head (`c 507:*`): container and session running, sys.boot_completed=1, servicemanager, hwservicemanager and vndservicemanager running. - With this change: the same. surfaceflinger running, ro.hardware.egl is mesa, the check passes with the render node 226:128 and the three binder lines. - Inside the container with this change, `toybox mknod` (mknod only, nodes never read or opened): refused for `b 7:0`, `c 1:1`, `c 507:4` (binder-control) and `c 507:8`; allowed for `c 507:5`, `c 507:6` and `c 507:7`, which are binder, hwbinder and vndbinder on this boot. Not covered: only one boot of one kernel; a start after a kernel update that moves the binder major; a start where binderfs is not yet mounted (the expected failure, not provoked); suspend and wake, and apps, with the session running. Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
…bmeta source (derekross#511) The Xiaomi 14's notes record how the phone was installed (OrangeFox, slot B only: the installer's own console shows it, so slot A was never written), the partition mapping, where vbmeta came from, that the installer dropped vendor_boot's AVB footer on slot B, and the plan for the next backup from the guarded initramfs.
…ory (derekross#586) Run by a relative path from somewhere other than the repository root (cd ..; omarchy-mobile/tools/build-packages --help, or ./release --help from tools/), the eight tools that cd to the repository before they read their options failed --help: they passed "$0", which is relative to the directory they had just left, and awk couldn't open it (exit 2). Derek pointed this out on derekross#559 and suggested usage_from_header as the place to fix it. usage_from_header now takes only FILE's name and reads that tool from the tools directory above tools/lib/usage.sh. The helper's own path is still valid when it runs, because every caller has just sourced it from its current directory, so this fixes all eight callers in one place with no call site changed. The tools that resolve their own path first (fetch-toolchain, kernel-pr-check, publish-kernel) or don't cd before --help (build-flame-probe, phone-smoke, phone-demo), and the bundle's inline-awk tools (install-phone, nv-backup, phone-provision), already worked; their files are unchanged. phone-smoke and phone-demo still load the helper only on the help path. tests/tools-help now runs every tool's --help from the repository's parent and from tools/ by a relative path and compares the output and exit code with a run from the root. On main it fails at tools/build-packages. --help from the root is byte-identical before and after for all 17 tools (16 exit 0, phone-mirror exit 2). Tested: tests/tools-help, tools/pr-check origin/main, tests/kernel-pr-check, tests/install-phone, tests/make-bundle, and tests/all, which passes except tests/release-upload. That suite fails the same way on main because the worktree path contains "/.". Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…#589) CPU_MAX advertises SVE but still traps ADDVL before the HAL builder can run. Enable CPACR_EL1 ZEN and general FP/SIMD access plus CPTR_EL3 EZ, preserving unrelated bits and leaving vector length and EL2 state unchanged. Exercise the actual Emu constructor with an entirely synthetic ELF containing ADDVL, RDVL and RET. The unchanged extractor raises UC_ERR_EXCEPTION; the patched extractor returns the vector-length-dependent result. Keep optional dependency skips explicit and distinguish integer SVE from FP access in separate gate fixtures. Co-authored-by: Codex <codex@openai.com>
The 2026-11-08 entry gave only the release, 7.1.0-rc1-sdm845, and the header of these notes describes two builds of it. The boot used the first-light kernel: the initramfs was packed against the build from 2026-11-04, not the netfilter build from 2026-11-05. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…#583) Document why a failed or empty SMSC reply is incomplete with a present SIM, while the no-SIM error remains an expected result. Follow up on the review of derekross#569 without changing probe behavior or its package version. Co-authored-by: OpenAI Codex <noreply@openai.com>
…oss#499) * CONTRIBUTING: add Dipper single-slot recovery entry Name the sole boot partition the installer may write and the physical fastboot button path, as required before any single-slot port is flashed. * CONTRIBUTING: group single-slot recovery entries * CONTRIBUTING: record Dipper forced restart * CONTRIBUTING: identify Dipper restart test context * CONTRIBUTING: record Dipper power-hold outcome * CONTRIBUTING: confirm Dipper restart under RAM kernel
…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>
…oot hook does (derekross#576) A fresh install's initramfs rescue SSH had no owner key until the first run of omamobile-boot-update's hook, which then rewrote vendor_kernel_boot (derekross#564). tools/install-phone now appends the owner's authorized_keys to the initramfs in its local copy of slot B's image, with the hook's own steps, so the installed image is byte for byte what the hook builds from the same package: checked on the 2026.11.08 cheetah image under umask 077, 022 and 002. Each step stops the install on failure. The write path is unchanged: only slot B's images and userdata. The bundle's README lists lz4 and cpio, and CI installs android-tools so tests/install-phone-initramfs runs there. Fixes derekross#564.
…rekross#587) * bin/omamobile, tools, rootfs: no pipe into grep -q under pipefail The follow-up to derekross#575, which fixed the test scripts: the same pattern in the laptop-side scripts that set pipefail. grep -q exits at the first match, a writer still writing gets SIGPIPE, and pipefail fails the pipeline (AGENTS.md section 4, Userspace). Each site now uses grep -c ... >/dev/null, which reads all its input, with the same pattern, flags and truth value. The one that loses in practice is tools/fingerprint-capture's wait loop: sed prints the whole log from the step's start, and with the end line early in a verbose kmsg stream the old check never saw it and the step ran to its timeout. The others have short writers (adb get-state, android-info.txt, 64 bytes of the TA, the keyring, the image's passwd) that fit the pipe buffer before grep exits; they change so that no script keeps the pattern. Two of them failed open: build-inside's user-account check and signing.sh's revoked-key check would have passed on a SIGPIPE. tests/grep-q-pipes now scans bin/, tools/ and rootfs/ as well, plus tools/lib (sourced by pipefail scripts), and skips comment lines. It leaves out five stage tools whose remaining hits are command strings that tools/phone-ssh runs in the phone's shell, which has no pipefail. Nothing under packages/ or initramfs/ changes. Tested: the race harness (old 0/200, new 200/200 per form; the fingerprint-capture loop old 0/100, new 100/100), tests/grep-q-pipes with an injected pipe (fails, names the line) and a commented one (passes), tools/pr-check, tests/all. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * tests/grep-q-pipes: scan the package tests too, and the two pipes they held The scan's 'packages/*/test' pathspec matched no file, so no package test was ever scanned (91 scripts). With 'packages/*/test/*' it reads 164, and found two pipes into grep -q that landed since: device-husky's speaker_protection check (derekross#453) and omamobile-waydroid's render-node check (derekross#541). Both now read the whole pipe. Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Derek Ross <derekross@gmail.com>
# Conflicts: # .github/workflows/checks.yml
|
Merged as omamobile-settings 0.1.0. On my Pixel 7 Pro (Omarchy 4.0.6): About shows the phone's own model, kernel and versions, a second launch from the launcher brings back the same window, and the look sits well beside Omarchy's menu. Thank you, @marselnenaj. The page PRs can build on this now, with pages living in this app as decided above. |
…erekross#526) A new optional package: a Qt Quick Settings app with a searchable home, a read-only About page (model, OS, kernel, architecture, memory, storage use and the Omarchy, shell and own versions), a look taken from the current Omarchy theme's colours, fonts and rounding, and a single-instance session D-Bus service, org.omamobile.Settings, with Open(page). Settings pages will live in this app, and the shell opens them. Run on the maintainer's Pixel 7 Pro on 2026-11-09 (Omarchy 4.0.6): About shows the phone's own details, a second launch brings back the same window, and the look sits well beside Omarchy's menu. Co-Authored-By: Codex <codex@openai.com>
|
@derekross @marselnenaj closing the loop on my question here: I'm going with omamobile-settings for my pages instead of my own local settings package and shell-plugin pages. I've dropped the local package from my phone, and my pages are PRs on this app now: haptics #629, change PIN #630, lock timeout #866 and keyboard layouts #892, plus the tests guard in #930. Quick toggles stay in the shade, as decided above. |








Add an optional native Settings foundation with mobile navigation, search and a read-only About page. Omarchy's existing palette, font and control settings drive the appearance through shared native Qt controls. Functional settings pages will follow as separate contributions with their backends.
The About page reports model, operating system, kernel, architecture, memory, storage and package versions. Unknown values stay unavailable. Repeat launches route to the existing window through session D-Bus;
--previewsupplies neutral device information and--checkvalidates the catalog. The application version is supplied by the package build.Appearance
The revision uses Qt's Basic templates with native controls styled from existing
colors.tomlandshell.tomlscalar settings. Fonts and control heights update with the theme while navigation stays consistent. Atomic theme changes are watched, and malformed primary colors retain the last valid palette. The adapter does not evaluate arbitrary TOML expressions or gradients.Dark/light preview gallery.
Validation
Try the preview
The source comes from Marsel Nenaj's own MIT-licensed Settings worktree, with Codex assistance. Original Omarchy API/theme attribution is retained in the README and license.