Repository navigation
omamobile-shell 0.2.125: the shade and the bar read Omarchy 4.5's notification store - #482
Conversation
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.
main already carries 0.2.101 (derekross#464), and derekross#482 and derekross#420 hold 0.2.102 and 0.2.103.
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.
|
My development phone already has a local 4.5 reader, so I compared its extra deletion changes with this PR instead of submitting another reader. The incremental public consumer is now #533, targeting your shell-notifications-45 branch. It waits for successful deletion replies, uses one daemon-owned snapshot for keyed Clear all, preserves new arrivals and retains rows on failure or unsupported IPC. Both legacy command/model tests still pass; the complete shell suite passes 297 checks with no failures or skips. Independent review passed. Its small provider prerequisite is marselnenaj@0247da2, based on #471. Both actual-package provider suites pass. I will submit that commit as a separate upstream PR after #471 merges. The capability dependency and typed reply enforce the prerequisite; release versions remain for allocation at merge. |
6be78a3 to
36458fa
Compare
36458fa to
9783a78
Compare
…ification store The review of #471 found that the bar's unread count and the quick-settings shade read Omarchy 4.0's notification files (one <timestamp>-<id>.json per notification under ~/.local/state/omarchy/notifications/ and history/), which Omarchy 4.5 no longer writes: its Store.js keeps live.json (toasts on screen) and history.json (closed rows, newest first, 100 over 24 hours), one JSON array each, rows named by a key with `ts` in seconds, and its migration 1792648436 deletes the 4.0 files. On 4.5 the shade would show nothing, the badge would never count, and Clear all would call a `clear` that 4.5 lacks. Both layouts are now read, chosen by what exists, not by version: where history.json exists the shade reads live.json and history.json, and Clear all is dismissAll followed by `forget <key>` for every notification read (the 4.5 version of patch 0019 in #471 takes the key); the bar counts history.json rows newer than the shade's mark, watching the file and re-reading it every 30 s, since the service replaces it with rename(2). Without history.json, every command and count is the one 0.2.101 runs. Tested: test/all (the offscreen touch tests included); new rows in test/notifications-model.js and test/status-model.js in 4.5.0-3's format, made by its own Store.js; test/notifications-store.sh runs the shade's read and Clear all commands from QuickSettings.qml on both layouts with a fake qs; test/quicksettings-render now draws the light theme over 4.5 files and checks that all three rows show (run in a sandbox like the OWL container, with Omarchy 4.0.6's shell: 3 rows on both layouts; main's plugin shows 0 on 4.5). Mutations of each new branch fail a test. Not run on a phone. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
9783a78 to
e6a2020
Compare
…was checked against (#670) [omamobile] held omarchy and omarchy-settings at 4.0.6 while the patch series didn't apply to 4.5 (#433, #434). The 4.5 family, the update's stand-ins and the held migrations are in (#661, #662, #665, #482), a rehearsal image updated from 4.0.6 in one and in two steps, and the 7 Pro took the update through its Updates sheet on 2026-11-11: exit 0, 28 patches verified, 98 migrations with 24 held and none failed, phone-smoke passing after the restart. The pin moves rather than goes: a phone takes a newer Omarchy only after the series was checked against it. Phones on 4.0.6 move to 4.5.0-3 at their next omarchy update; legacy-4.0.6/ stays for them until then. Both files were checked against upstream's signatures. Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
…he other node tests do (#671) test/notifications-store.sh (#482) runs NotificationsModel.js under node with no check for it, so test/all failed on ms01, which has no node, and the development publish stopped at omamobile-shell 0.2.125 on 2026-11-11. Every other node test in test/all skips there; this one now does too. CI has node and still runs it. Run on ms01 with the change: the test reports SKIP and the rest of test/all goes on. Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
On Omarchy 4.5 the notification service keeps the toasts on screen in live.json and the closed ones in history.json, rows with a key, where 4.0 kept one file per notification. phone-demo still looked for 4.0's files, so on a 4.5 phone (the Pixel 7 Pro since 2026-11-11): - notify and clear found no demo notifications to take away, so the demo's toasts stayed in the shade and on the lock glance after clear, against the tool's promise to remove exactly what it added; - the Do Not Disturb warning never fired (it read notifications.json; 4.5 keeps the state in quiet.json); - restore-notes copied a backup's 4.0 history/ and images/ directories into a store that never reads them. The tool now tells the layouts apart the way the shade does (history.json is there from the 4.5 service's first start). On 4.5 it finds the demo's rows by key and takes them away through Omarchy's own forget call; only when the shell doesn't answer does it rewrite the two files itself. It counts rows that are still there instead of reporting them gone, reads Do Not Disturb from quiet.json, and copies nothing from a backup into a 4.5 store. 4.0 behaves as before. tests/phone-demo-notes runs the tool's phone side against both layouts in a made-up home. Not run on a phone: seed and clear change what the phone shows, and wait for the maintainer. Found by the round 52 review of #482. Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
On Omarchy 4.5 the notification service keeps the toasts on screen in live.json and the closed ones in history.json, rows with a key, where 4.0 kept one file per notification. phone-demo still looked for 4.0's files, so on a 4.5 phone (the Pixel 7 Pro since 2026-11-11): - notify and clear found no demo notifications to take away, so the demo's toasts stayed in the shade and on the lock glance after clear, against the tool's promise to remove exactly what it added; - the Do Not Disturb warning never fired (it read notifications.json; 4.5 keeps the state in quiet.json); - restore-notes copied a backup's 4.0 history/ and images/ directories into a store that never reads them. The tool now tells the layouts apart the way the shade does (history.json is there from the 4.5 service's first start). On 4.5 it finds the demo's rows by key and takes them away through Omarchy's own forget call; only when the shell doesn't answer does it rewrite the two files itself. It counts rows that are still there instead of reporting them gone, reads Do Not Disturb from quiet.json, and copies nothing from a backup into a 4.5 store. 4.0 behaves as before. tests/phone-demo-notes runs the tool's phone side against both layouts in a made-up home. Not run on a phone: seed and clear change what the phone shows, and wait for the maintainer. Found by the round 52 review of #482. Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
…#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.
…ification store (derekross#482) The review of derekross#471 found that the bar's unread count and the quick-settings shade read Omarchy 4.0's notification files (one <timestamp>-<id>.json per notification under ~/.local/state/omarchy/notifications/ and history/), which Omarchy 4.5 no longer writes: its Store.js keeps live.json (toasts on screen) and history.json (closed rows, newest first, 100 over 24 hours), one JSON array each, rows named by a key with `ts` in seconds, and its migration 1792648436 deletes the 4.0 files. On 4.5 the shade would show nothing, the badge would never count, and Clear all would call a `clear` that 4.5 lacks. Both layouts are now read, chosen by what exists, not by version: where history.json exists the shade reads live.json and history.json, and Clear all is dismissAll followed by `forget <key>` for every notification read (the 4.5 version of patch 0019 in derekross#471 takes the key); the bar counts history.json rows newer than the shade's mark, watching the file and re-reading it every 30 s, since the service replaces it with rename(2). Without history.json, every command and count is the one 0.2.101 runs. Tested: test/all (the offscreen touch tests included); new rows in test/notifications-model.js and test/status-model.js in 4.5.0-3's format, made by its own Store.js; test/notifications-store.sh runs the shade's read and Clear all commands from QuickSettings.qml on both layouts with a fake qs; test/quicksettings-render now draws the light theme over 4.5 files and checks that all three rows show (run in a sandbox like the OWL container, with Omarchy 4.0.6's shell: 3 rows on both layouts; main's plugin shows 0 on 4.5). Mutations of each new branch fail a test. Not run on a phone. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…was checked against (derekross#670) [omamobile] held omarchy and omarchy-settings at 4.0.6 while the patch series didn't apply to 4.5 (derekross#433, derekross#434). The 4.5 family, the update's stand-ins and the held migrations are in (derekross#661, derekross#662, derekross#665, derekross#482), a rehearsal image updated from 4.0.6 in one and in two steps, and the 7 Pro took the update through its Updates sheet on 2026-11-11: exit 0, 28 patches verified, 98 migrations with 24 held and none failed, phone-smoke passing after the restart. The pin moves rather than goes: a phone takes a newer Omarchy only after the series was checked against it. Phones on 4.0.6 move to 4.5.0-3 at their next omarchy update; legacy-4.0.6/ stays for them until then. Both files were checked against upstream's signatures. Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
…he other node tests do (derekross#671) test/notifications-store.sh (derekross#482) runs NotificationsModel.js under node with no check for it, so test/all failed on ms01, which has no node, and the development publish stopped at omamobile-shell 0.2.125 on 2026-11-11. Every other node test in test/all skips there; this one now does too. CI has node and still runs it. Run on ms01 with the change: the test reports SKIP and the rest of test/all goes on. Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
…oss#686) On Omarchy 4.5 the notification service keeps the toasts on screen in live.json and the closed ones in history.json, rows with a key, where 4.0 kept one file per notification. phone-demo still looked for 4.0's files, so on a 4.5 phone (the Pixel 7 Pro since 2026-11-11): - notify and clear found no demo notifications to take away, so the demo's toasts stayed in the shade and on the lock glance after clear, against the tool's promise to remove exactly what it added; - the Do Not Disturb warning never fired (it read notifications.json; 4.5 keeps the state in quiet.json); - restore-notes copied a backup's 4.0 history/ and images/ directories into a store that never reads them. The tool now tells the layouts apart the way the shade does (history.json is there from the 4.5 service's first start). On 4.5 it finds the demo's rows by key and takes them away through Omarchy's own forget call; only when the shell doesn't answer does it rewrite the two files itself. It counts rows that are still there instead of reporting them gone, reads Do Not Disturb from quiet.json, and copies nothing from a backup into a 4.5 store. 4.0 behaves as before. tests/phone-demo-notes runs the tool's phone side against both layouts in a made-up home. Not run on a phone: seed and clear change what the phone shows, and wait for the maintainer. Found by the round 52 review of derekross#482. Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
Preparation for the move to Omarchy 4.5 (#434). Not run on a phone. Every phone runs Omarchy 4.0.6 today (pinned in
rootfs/packages.pinned), and on 4.0.6 this changes nothing in effect.Why
The #471 review found that the bar's unread count and the quick-settings shade read Omarchy 4.0's files, which 4.5 no longer writes. From the real packages:
~/.local/state/omarchy/notifications/<ts>-<id>.json(on screen),…/history/<ts>-<id>.json(closed, newest 10)~/.local/state/omarchy/notifications/live.json(on screen),history.json(closed, newest first, 100 rows / 24 h),quiet.jsontimestampin ms (NotificationLogic.jsserializePopup)key(n…),tsin seconds, history rows alsoclosed_ts/closed_reason(Store.jspersistable/closedEntry/serialise)Service.qmlfile queue (persistPopupFile,archivePopupFileFor,writeHistoryFile)Service.qmlholds the rows in memory and rewrites each file whole, atomically (FileViewatomicWrites)dismissAll,clear,showHistory,dismiss <summary>;forget <ts>-<id>.jsonfrom our patch 0019dismissAll,showHistory,dismiss <summary>, noclear;forget <key>from #471's 4.5 version of 00191792648436.shdeletes the 4.0 per-notification files,history/andimages/On 4.5 the shade would show nothing (each line is an array, which the model skipped), the badge would never count, and Clear all would call a
clearthat doesn't exist.What
The layout is picked by what exists, not by the version string:
history.jsonexists → 4.5.NotificationsModel.js:parse()reads a line holding an array as 4.5 rows (storeEntry: the key is both the row key and the forget name,ts×1000 rounded) and a line holding an object as one 4.0 file, as before.shown()also returns every forget name it read.QuickSettings.qml: the read takeslive.jsonandhistory.jsonwhenhistory.jsonexists, and otherwise runs the old loop unchanged. Clear all on 4.5 isdismissAll, thenforget <key>for every notification read (dismissed toasts keep their keys in history). Otherwise it'sdismissAll; clearas before.Bar.qml/StatusModel.unreadStored: whilehistory.jsonloads, the badge counts its rows newer than the shade's mark. The file is watched (FileViewwatchChanges) and also re-read every 30 s, because the service replaces it with rename(2). Without the file, the count comes from thehistory/folder model as before.Tests
test/allpasses (run in a bubblewrap sandbox with no network, D-Bus, Wayland or/run), offscreen touch tests and qmlformat included.test/notifications-model.jsandtest/status-model.js. The rows have 4.5.0-3's shape, made with its ownStore.js, redacted verification-code row included.test/notifications-store.shtakes the shade's read and Clear all commands out ofQuickSettings.qmland runs them on both layouts with a fakeqs. On 4.5 it includes 4.0 leftovers, which must not show, and the empty files the service first creates.test/quicksettings-render: the light theme now draws over 4.5 files and the dark one over 4.0 files, with a new check that all three rows show. Run in the same sandbox in place of the OWL container, against Omarchy 4.0.6's shell: 3 rows on both layouts. Main's plugin shows 0 rows on the 4.5 fixtures.Store.jsparse into the expected rows, tap actions and unread counts.history.jsonchecks, forget in Clear all, the names passed to Clear all).Open points
forget <key>). Without it, a swiped row comes back after the 15 s grace and Clear all only dismisses the toasts.qs ipcper notification read, and history can hold up to 100, so it can take a few seconds.FileViewwatch notices the atomic replace ofhistory.jsonon the phone isn't verified. The 30 s re-read is the fallback.🤖 Generated with Claude Code