aboutsummaryrefslogtreecommitdiffstats
path: root/desktop
AgeCommit message (Collapse)AuthorFilesLines
7 daysfix(user): the avatar is ~/.user-icon.jpg, as hyprlock and SDDM useDanilo M.3-7/+32
The page showed AccountsService's icon, an older picture than the one the rest of the desktop uses: hyprlock, the udt SDDM theme (pointed at it by udt-accent) and the rofi power menu all read ~/.user-icon.jpg. get now prefers that file, and set converts the pick to JPEG there, through a temp file and a rename, before handing it to AccountsService so both agree. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
7 daysfix(user): the page retries partial saves and survives leaving mid-saveDanilo M.3-7/+15
A key whose write half-failed (AccountsService took it, git did not) reloads as unchanged and Save went dead. A key with an error now counts as changed, and editing a field clears its error. Processes were parented to the page, so leaving it mid-save killed userctl between stores. They hang off the module now, passed in as owner (not the module's id, which would bind to itself). busy now clears when the reload lands, so Save cannot be clicked again before saved is fresh, and the fields and avatar picker are disabled while busy. Avatar paths with # or % broke the preview under a bare file:// prefix; each segment is now URI-encoded for the Image source (page and row), while draft.avatar stays a plain path for userctl. Dropped a redundant saved clear in load. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
7 daysdocs(user): the module's README and the busctl escaping noteDanilo M.2-0/+31
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
7 daysfeat(user): a drawer page for the user's detailsDanilo M.3-0/+287
Name, email, avatar, language, location and the git signing key, each saved through userctl to every store that holds it. Only changed fields are sent; a field whose write failed keeps its edit and its error, so the next Save retries it. A failed userctl get reports a single error= line. The page shows it in red under the avatar, leaves saved and draft untouched, and disables Save until a later load succeeds. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
7 daysstyle(desktop): the notification comment back on its listDanilo M.1-3/+3
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
7 daysfeat(desktop): a greeting row above the notificationsDanilo M.3-2/+115
Avatar, Hello <login> and the real name, re-read whenever the grid view shows again so an edit is visible on the way back from the page. Module gains tiled, so a module can own a page without a tile. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
7 daysfeat(user): helpers for the user pageDanilo M.2-0/+93
Parsing userctl's output and locale -a, and the git name preview, kept out of QML so node can test them. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
7 daysfix(user): userctl survives leading dashes and reports failed readsDanilo M.2-15/+37
busctl and kwriteconfig6 parse a value starting with '-' as an option, so a location like '-x' failed; '--' before the value ends option parsing (checked against both binaries). get ran without pipefail, so a failing busctl gave empty fields and exit 0, which the page would show as a blank form. It now reads all fields first and, on failure, prints one error= line and exits 1. keys listed revoked, expired or invalid primaries, keys that cannot sign, and could label a key with a revoked uid. It now skips those. Unsetting the signing key uses --unset-all, since git exits 5 on a multi-valued key with --unset. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
7 daysfeat(user): userctl reads and writes the user's detailsDanilo M.2-0/+258
One writer for the stores other programs read: AccountsService (and with it the passwd GECOS LibreOffice prefills from), KDE's emaildefaults and the global git config, where the surname goes in as an initial. busctl is asked for JSON because its plain output escapes non-ASCII as octal. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
7 daysfeat(perf): the performance page moves to the drawer, with ~/.cacheDanilo M.10-155/+174
The island's performance icon and panel are gone: the drawer page covers them. With only one host left, the panel drops its three-column grid for one card per row at a larger size, since the page scrolls; temperatures get a card of their own as tiles. The files move from island/perf to modules/perf. A cache card ports conky's: ~/.cache and its five largest children, measured with du each time the page opens (28ms over 2 GiB here) rather than read from conky's 15-minute sampler file. The largest is yellow, red past half the total, as conky had it. The warning dot now shows only on the drawer tile. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
7 daysdocs(spec): user details in the drawerDanilo M.4-0/+0
A drawer row greets the login user and opens a page that edits name, email, avatar, language and location, written where other programs already read them: AccountsService (and through it the passwd GECOS LibreOffice reads), KDE's emaildefaults and the global git config, with the surname shortened to an initial for git. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysfeat(sound): per-app volume on the drawer page, output to 150%Danilo M.3-84/+157
The island spec put per-app volume on the drawer page, and it was never built. Each application playing audio now gets a row with its name, what it is playing (two browser tabs are otherwise identical rows), a click-to-mute level and its own slider. Streams are matched on PwNodeType.AudioOutStream. Quickshell reports isSink true for a playback stream, so the obvious isStream && !isSink filter found nothing while Firefox played. Their properties are empty until bound, so the page carries a PwObjectTracker for them, which stops with the drawer. speech-dispatcher-dummy is always present and never audible, and is hidden. The output slider runs to 150%, the ceiling the volume keybind already uses (wpctl -l 1.5), with a tick at 100%. Input and apps stop at 100%: a boosted microphone only clips. The meter became an inline component shared by both, and wheel-up never lowers a level already past its maximum. Now playing is dropped from the page: the OSD already carries it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysfeat(desktop): the mouse back button steps out of the drawerDanilo M.1-5/+20
The side button (BTN_SIDE, Qt.BackButton) does what Escape does: closes the history, then leaves a page for the tiles, then closes the drawer. Both go through one back(). The first attempt let the press fall through to the full-surface click catcher under the panel, since nothing above it accepts that button; it never arrived while a page was open. A MouseArea on top with z: 1, taking only the back button, works, and left clicks still pass through it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysfeat(updates): a page that reads at a glanceDanilo M.1-30/+106
The first cut was small type in one long list: 471 Slackware upgrades pushed everything below them out of sight, and the repo tag, the versions and the name ran together across the row. Type is 2px larger and rows taller. A section over ten rows starts folded behind a clickable heading with its count. The version being moved to is bright, the one being left dim, both in Inconsolata so digits line up. Every other row is shaded to carry the eye across, and the repo is a pill in one colour per repo, never red, green, yellow or the accent, which mean something elsewhere. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysdocs(updates): the module's READMEDanilo M.1-0/+32
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysfeat(updates): the drawer tile and pageDanilo M.3-0/+230
The tile counts what can be installed now (Slackware and the binary repo) and what only needs a build, and goes accent when there is something to install. It runs updatesctl while the drawer is open and not otherwise; a result over 30 minutes old is followed by a refresh that runs nvchecker, never after a refresh, so a failing nvchecker cannot loop. The page has one section per next action: Slackware upgrades, new and removed first, then upgrade here, awaiting build, upstream newer, app updates, the order the work is done in. Each row names the package, the repo it comes from and the one version jump its action covers. A source that could not be read says so under the section it empties instead of reading as nothing pending. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysfix(updates): review findings in updatesctlDanilo M.2-12/+27
A git hash starting with a letter (3.1-a59215d) ordered behind its base, because any trailing alphabetic chunk counted as a pre-release; only alpha, beta, rc, pre and dev do now. Trailing zeros (1.14 against 1.14.0) compare equal instead of reading as an upstream bump. A bad line in the slackpkg blacklist is skipped rather than crashing the script and leaving the drawer with nothing. An AppImage keeps its key over a flatpak of the same name, as the spec orders them, and flatpak list has a timeout. A Slackware upgrade compares version and build, so an arch change alone is not a row. The spec's prose now matches the code: the upstream comparison includes the installed version. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysfix(updates): drop a redundant clause, test the mirror blacklistDanilo M.2-2/+3
The awaiting-build branch also required the build to be older than upstream, which cannot fail there: a build newer than installed was already taken as an upgrade, so the build is at most the installed version, which is behind upstream whenever a row is emitted. Removing it left every test passing, which is how it was found. The blacklisted testing/ entry sat after the main one in the fixture, so setdefault kept the main one and the mirror-side blacklist was never exercised; it now comes first, and dropping the check fails the test. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daystest(updates): a hung NFS read reports unreachableDanilo M.1-0/+12
A fifo with no writer blocks cat at open() the way a hard NFS mount blocks during a NAS outage. The repo section reports unreachable within the timeout and the mirror sections still fill. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
8 daysfeat(updates): Slackware upgrades, new and removed packagesDanilo M.2-1/+92
Compared mirror against installed, as upgrade-all does, not against slackpkg's local ChangeLog: `slackpkg update` replaces that without installing anything, so a diff against it reads clean after an update that was never applied. New and removed come from the mirror's ChangeLog, the newest Added or Removed line deciding, so a package added and later upgraded still counts as new. Blacklisted build tags and directories are skipped, as slackpkg skips them. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
8 daysfeat(updates): sort tracked packages into the next actionDanilo M.2-1/+351
Each nvchecker stanza lands in one section: upgrade here when the binary repo's PACKAGES.TXT holds a newer build than installed, awaiting build when the recipe (the user's or SBo's) has caught up but the build has not, upstream newer when the recipe lags, app updates for AppImages and flatpaks. Built comes from PACKAGES.TXT rather than the .info, which is why gamescope and kitty no longer read as installable: the repo still held the old build. The repo is on a hard NFS mount, so it is read by a child the timeout can kill. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
8 daysfeat(updates): version order for updatesctlDanilo M.2-0/+117
nvtable compares versions for equality, which cannot tell an installed 2.1.278 from an SBo recipe of 2.1.267 that is older than it. vercmp orders numeric chunks as numbers and treats a trailing alphabetic chunk as a pre-release, so a kernel suffix or git hash orders as newer than its base and never reads as behind. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
8 daysfeat(desktop): performance and clipboard pages, a scrolling tile gridDanilo M.13-31/+492
The tile grid sits in a Flickable capped at three rows, so the fourth row the new modules bring scrolls inside that box instead of eating the notification space; a thin accent bar in the margin shows only when a row is hidden. Tiles are reordered with the everyday ones first. Performance shows the island's PerfPanel in a Page, with a tile line of CPU and RAM and the warning dot. Perf.detailed is now a count of live PerfPanel instances instead of a binding in the status pill: the island and the drawer both host the panel, and a bool with two writers would let one clear it under the other. Clipboard reads cliphist when its page opens: search, copy on click or Enter, delete per row, clear all behind a 3 second confirm, image entries as thumbnails decoded once onto tmpfs. Entries reach sh as $1, never inside a command string. A copy closes the drawer through a new Page.dismiss signal the drawer connects like back. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysfeat(status): quiet mode mutes the Claude and opencode soundsDanilo M.6-9/+156
A new registry mode, status.quiet, with a row on the status page. Both tools will play their sounds through notify-sound, which asks statusctl and plays nothing while quiet is on, so their popups stay and only the sound goes. Presentation sets quiet and restores it afterwards, like dnd and nolock: a sound during a call or a game is what quiet is for. notify-sound ignores any argument it has no sound for, because the opencode notifier runs its command on every event, session_started and user_message included. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
8 daysrefactor(island): the controls panel goesDanilo M.9-619/+15
Each of its seven tiles opened a drawer page that already existed, and it went unused. SUPER alone will open the drawer instead. Removed with it: the status pill's controls icon, Island.toggleControls, the island IPC target, Controls.js and its test, and the controls list in Prefs. Prefs now holds only the pinned IP; a controls key left in desktop.json is still kept on write, because writes merge into the parsed document. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysdocs: the performance panelDanilo M.1-1/+38
What the panel shows, the dot's thresholds and their lack of hysteresis, why the public IP is pinned, and what is sampled always versus only while open. AGENTS.md gains two gotchas found building it: FileView reads /proc on reload(), and a hot reload does not resize an existing layer surface, which left the panel cut off after the height was raised. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfix(perf): review findings in the service and the panelDanilo M.5-16/+26
- The panel loads only while open. Its models are fresh arrays every sample, so a hidden instance rebuilt its rows every 2s for the whole session. - The public IP fetch never kills one already running, and waits 30 minutes from the last attempt, not the last success: a rate-limited or offline fetch was retried every minute. - curl -4, so an IPv6 answer cannot silently stall the VPN check. - CPU load sums only the first eight /proc/stat fields. guest and guest_nice are already inside user and nice, and counting them again read high while VMs ran. - df gets timeout -k; the VRAM scan is not restarted over itself. - The GPU name joins the host constants in Perf.qml. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfix(island): the window is tall enough for the performance panelDanilo M.1-1/+3
The performance grid runs past the old 360px surface and its disks card was cut off. The surface is click-through outside the pills, so the extra height costs nothing. A hot reload does not resize an existing layer surface; it takes a restart of the desktop component. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfeat(island): the status pill opens the performance panelDanilo M.2-7/+57
A system monitor icon before the controls icon opens the panel in the status pill's capsule at 600px. A dot on its corner shows the worst of temperatures, RAM, disks and the public IP check, yellow or red, so a problem shows without opening anything. The detailed sampling runs only while the panel is open. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfeat(perf): the performance panel's card gridDanilo M.1-0/+330
CPU with per-core bars, RAM and temperatures down the left; GPU and network beside it; disks under those two. Colours follow conky's bands so the numbers read as they did there. The pin beside the public IP stores the address shown, and the address turns red when a fresh one differs from the pin. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
10 daysfeat(perf): the performance serviceDanilo M.1-0/+295
CPU, RAM, temperatures and disks are sampled always, so the island's icon can warn with the panel closed; per-core load, the network rate, VRAM and the GPU fan only while the panel is open. hwmon chips are found by name, since their numbers move across boots. df runs under timeout so a dead NFS server keeps the last rows rather than blocking. The public IP keeps conky's cache file and 30 minute pace. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
10 daysfeat(desktop): Prefs holds a pinned public IPDanilo M.1-2/+17
perf.pinnedIp in desktop.json, set from the performance panel's pin button and written only on change. The address lives in the user's state file, never in the repo. setControls and setPinnedIp share one write so both keep keys this version does not know. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
10 daysfeat(perf): pure parsing and arithmetic for the performance panelDanilo M.2-0/+336
/proc/stat deltas, meminfo, df rows, the fdinfo VRAM sum, byte rates, hwmon lookup by chip name, the pubip cache in conky's format, and the bands that colour the panel and the icon. A reading that cannot be made is null, and null never raises the worst band, so a missing sensor cannot raise an alarm. Only a fresh public IP can mismatch the pinned one; a failed fetch goes stale instead. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
10 daysdocs(island): calendar edits touch only what changedDanilo M.1-2/+4
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfix(calendar): review findings in calctl and the editorDanilo M.7-59/+170
An edit now rewrites only the fields that changed. A save with nothing changed used to rebuild the alarm, turning ACTION:AUDIO into DISPLAY and dropping ACKNOWLEDGED, re-anchor the start to the local zone, and strip SUMMARY's parameters; the test edits a foreign event unchanged and checks every one survives. get, edit and delete name the calendar the event is in. The same invite synced into two calendars has one UID, and the panel acted on whichever file came first, so editing one copy could overwrite or delete the other. A custom rule's date cannot be moved from the panel: BYDAY=MO stayed on Monday while the start moved, and the series lost the day shown. UNTIL is read back in local time, so a 22:30 series keeps its last occurrence. Files are written 0644 like vdirsyncer's. Enter in the title no longer bypasses the busy guard, which could create an event twice, and a throwing callback no longer leaks its Process. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysdocs: the calendar panel and calctlDanilo M.1-1/+18
The island README gains the panel. AGENTS.md records why calctl reads the vdir itself, the local-time recurrence rule, the system Python it needs, and two traps the build hit: a QML method named escape fails to load, and a pill expanding for any panel grew beside the calendar. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfeat(calendar): a bell on agenda rows whose event has an alarmDanilo M.1-1/+12
Checking a day's reminders meant opening each event. The bell shows when the event carries its own VALARM, preset or custom. cal-notif's per-calendar defaults are not reflected: they are its configuration, not a property of the event. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfix(island): the status pill stays at rest while the calendar is openDanilo M.1-1/+1
It expanded for any panel, which was right while it hosted every panel there was. The calendar panel lives in the clock pill, so the status pill grew into an empty box beside it. It now expands for its own two. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfeat(island): the clock pill opens the calendar panelDanilo M.3-7/+45
A click on the pill opens the panel in its capsule, 620px wide, growing left from the gap so the status pill does not move. Focus goes into the panel so its text fields and Escape work; Escape closes the editor before the panel. The Escape hook is back(), not escape(): QML rejects a method named escape with "Illegal method name", since it shadows the global JS function, and the reload failed until it was renamed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfeat(calendar): month grid beside the day's agendaDanilo M.1-0/+323
Six weeks from Monday, a dot per calendar with something that day, today and the selection marked. The agenda lists the selected day, all-day first; a click opens the event in the editor, and while the editor is open the grid picks its dates instead of the selection. Escape closes the editor before the panel. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
10 daysfeat(calendar): the event formDanilo M.1-0/+356
Dates come from the month grid beside it; times are typed or scrolled in 15-minute steps. Repeat and alarm offer the presets and show custom, unselectable, for an event carrying something richer, which the save then leaves alone. A failed save keeps the form open with calctl's line. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
10 daysfeat(calendar): the calendar serviceDanilo M.1-0/+102
Runs calctl, one Process per call so a reload and a save never collide, and drops a list that a newer one overtook. Watches the calendar directories the way the mail count watches xapian, so an edit from khal or a cron sync repaints the panel too. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
10 daysfeat(calendar): pure date logic for the panelDanilo M.2-0/+279
Month cells from Monday, which days an event covers, the day's agenda order, the grid's dots, time stepping, and the editor's draft in both directions. Days are strings and arithmetic is in UTC, so the DST change at the end of October never moves a day; the test steps across it. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
10 daysfeat(calendar): calctl creates, edits and deletes eventsDanilo M.2-1/+237
Edits change only what the form owns: a custom rule, extra alarms and properties this tool does not know survive, which the test checks on an event with BYDAY, two alarms and an X- property. Writes go through a temporary file and a rename so khal, cal-notif and vdirsyncer never read half a file, and a vdirsyncer sync is started detached afterwards. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
10 daysfeat(calendar): calctl reads the vdir as JSONDanilo M.2-0/+444
The calendar list comes from khal's config, so there is no second list. Recurrence is expanded with dateutil in naive local time, which keeps a weekly 09:00 at 09:00 across the DST change; the test pins that on 26 October. A rule richer than a plain FREQ, or alarms other than one preset, read as custom, which is what lets an edit leave them alone. A file that fails to parse is skipped rather than blanking the month. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
10 daysfix(island): review findings in the controls panelDanilo M.4-7/+22
Edit mode let a right-click and the wheel through to the tile, so a right-click opened the drawer and scroll changed volume, against the spec. The drag area now takes both buttons and the wheel. A scroll up on a level already above the cap, such as a mic set to 120% elsewhere, snapped it down to the cap. The cap now only stops a step up. Prefs logs a failed write, which printErrors off had silenced: the new order would otherwise vanish at the next restart with no trace. Its parsed object is renamed from data to doc, so it cannot shadow a base property of that name. The spec now says what was built: JSON.parse rather than a JsonAdapter, and the helper list as it is. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysdocs: the controls panel, its preferences file and bindDanilo M.1-1/+24
The island README gains the panel. AGENTS.md records that a bare SUPER release bind is cancelled by drag and scroll on Hyprland 0.56.2, which was the open risk in the spec, and that FileView creates parent directories while Quickshell.stateDir is per-shell. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfeat(island): controls panel in the status pill, island IPCDanilo M.3-7/+56
A settings icon in the status pill opens the controls panel in the same capsule the sound panel uses. Right-click on a tile closes the panel and opens its drawer page. The island IPC target reveals the island and toggles the panel, for the SUPER release bind. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
10 daysfeat(island): controls grid with an edit modeDanilo M.1-0/+175
Four columns, the visible list from Prefs. Edit mode puts a MouseArea over each tile that drags it and drops it into the slot under its centre, a × that hides it, and a strip of hidden tiles with a + each. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
10 daysfeat(island): a control tile bound to its live stateDanilo M.1-0/+143
Wifi and bluetooth through the Quickshell services, DND, presentation and auto-lock through the Status registry, output and mic through the sound module's tracked nodes. A tile whose backing object has not arrived yet is dimmed and ignores input, which covers the two seconds before Networking and Bluetooth populate. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>