phosh issueshttps://source.puri.sm/Librem5/phosh/-/issues2021-08-09T08:20:39Zhttps://source.puri.sm/Librem5/phosh/-/issues/2Built-in hardware accelerated screencaster with ponies2021-08-09T08:20:39ZGuido GuntherBuilt-in hardware accelerated screencaster with poniesAs far as I know, in the Wayland world, any traditional screencasting (and screenshotting) tool will not work, by design. It needs to be done by the compositor (or display server, or whatever it is in this case).
For documentation, tuto...As far as I know, in the Wayland world, any traditional screencasting (and screenshotting) tool will not work, by design. It needs to be done by the compositor (or display server, or whatever it is in this case).
For documentation, tutorials and public demonstration purposes, it would be really useful to have a built-in screencasting tool.
Ideally:
* Configurable recording framerate
* Configurable codec (lossless, VP9, VP8), multithreaded by default
Ponies on rainbows:
* Hardware-accelerated encoding built-in, otherwise with VP9 you'll grind to a halt because it's so slow to encode (GNOME Shell can't keep up with VP9 and will fill your RAM, for example). Also you'll certainly eat the battery.
* Sound mixing from the microphone or from the pulseaudio "playback monitor" virtual device (whatever the phone is playing) or from microphone + playback monitor.
* Ability to record only an area of the screen instead of the full screenhttps://source.puri.sm/Librem5/phosh/-/issues/8Time-based (and optionally location-based) do-not-disturb mode2021-08-09T08:20:39ZGuido GuntherTime-based (and optionally location-based) do-not-disturb modeI want the phone to be quiet (vibration mode, or maybe even completely silent) at certain user-defined hours (ex: 22h to 8h) to facilitate sleep and also prevent tech addiction. Other platforms have that, often with the ability to allow ...I want the phone to be quiet (vibration mode, or maybe even completely silent) at certain user-defined hours (ex: 22h to 8h) to facilitate sleep and also prevent tech addiction. Other platforms have that, often with the ability to allow exceptions for some "priority" contacts or applications that the user may decide to allow. As I assume notifications would be handled by phosh (?), I'm filing it here...
Additionally, I have the usecase where I want the ability to configure location-based connectivity+sound+notifications behavior (but that's a broader feature request, not sure where to request it) so that for example when I arrive in the viscinity of my home's wifi, or the office's wifi, my phone goes silent (and/or, in the case of the broader feature request: auto-disables some connectivity features, for example turning off the baseband, bluetooth, nfc, etc., to conserve energy and privacy)https://source.puri.sm/Librem5/phosh/-/issues/17Background image code should use GnomeBG2021-08-09T08:20:39ZGuido GuntherBackground image code should use GnomeBGfrom libgnome-desktop (https://developer.gnome.org/gnome-desktop3/3.26/gnome-desktop3-gnome-bg.html)from libgnome-desktop (https://developer.gnome.org/gnome-desktop3/3.26/gnome-desktop3-gnome-bg.html)https://source.puri.sm/Librem5/phosh/-/issues/47Accessibility support2021-08-09T08:20:39ZPeter VágnerAccessibility supportAccessibility is not yet a thing on mobile.
Explore and explain how much of current GNOME accessibility features can be reused for mobile experience.
Might it be doable to start at-spi2-core and expect at least some preliminary accessibi...Accessibility is not yet a thing on mobile.
Explore and explain how much of current GNOME accessibility features can be reused for mobile experience.
Might it be doable to start at-spi2-core and expect at least some preliminary accessibility out of ATK aware apps such as those GTK+ based?
How at-spi2-core can be started while testing?
I know GNOME accessibility is not yet touch support ready however disabled users are also dreaming for open device and great platform so I'd at least like to see some hopes cleared up at this point.
Another key point is that it is great to think accessibility features from the begining not add it as an after thought.https://source.puri.sm/Librem5/phosh/-/issues/49Implement org.freedesktop.impl.portal.Access for flatpak2021-08-09T08:20:39ZGuido GuntherImplement org.freedesktop.impl.portal.Access for flatpakSimilar to GcrSystemprompter we want to implement the above DBus API to present a proper access dialog for flatpak. See *src/access.c* in xdg-desktop-portal-gtk and *js/ui/accessDialog.js* in gnome-shell.Similar to GcrSystemprompter we want to implement the above DBus API to present a proper access dialog for flatpak. See *src/access.c* in xdg-desktop-portal-gtk and *js/ui/accessDialog.js* in gnome-shell.https://source.puri.sm/Librem5/phosh/-/issues/59Phosh crashing Gnome3 desktop when running - allow to skip `G_BUS_NAME_OWNER_...2021-08-09T08:20:39ZMartin ChangPhosh crashing Gnome3 desktop when running - allow to skip `G_BUS_NAME_OWNER_FLAGS_REPLACE`# What problem did you encounter
Launching Phosh with `../wlroots/_build/rootston/rootston -E '_build/run -U' -C ./data/rootston.ini` crashes Gnome3 desktop (Gnome will force a logout) on both Wayland and X11.
## What is the current be...# What problem did you encounter
Launching Phosh with `../wlroots/_build/rootston/rootston -E '_build/run -U' -C ./data/rootston.ini` crashes Gnome3 desktop (Gnome will force a logout) on both Wayland and X11.
## What is the current behaviour?
Phosh crashes Gnome whenever phosh is launched. Even when Phosh is running in a different VT than Gnome is. For example, Running Gnome at VT1 and Phosh under Weston in VT2. Gnome still crashes.
## What is the expected behaviour?
Phosh should not crash Gnome.
## How to reproduce
1. Launch Gnome3 as desktop
2. Build Phosh.
3. Run Phosh with the command `../wlroots/_build/rootston/rootston -E '_build/run -U' -C ./data/rootston.ini`
4. Gnome crash.
# Which version did you encounter the bug in?
- [x] I Compiled it myself. If you compiled phosh from source please provide the
git revision via e.g. by running ``git log -1 --pretty=oneline`` and pasting
the output below.
- [ ] I used the precompiled Debian package (e.g. by running a prebuilt
image). Please paste the output oof ``dpkg -s phosh`` below.
```
Phosh Version: caa2536a3730855f0f7d15020d4d1407914bd86c (HEAD -> master, origin/master, origin/HEAD)
```
# What hardware are you running phosh on?
- [ ] amd64 qemu image
- [ ] Librem5 devkit
- [x] nested compositor. If so please give the full command you start phosh
with.
- [x] other (please elaborte)
Running on a fresh Arch Linux installation on an Acer Spin5 51GN (I5 8250U, HD 520 (GTX 1050, not enabled)) with touch screen. Running Gnome 3.30.2
```
Command: ../wlroots/_build/rootston/rootston -E '_build/run -U' -C ./data/rootston.ini
```
# Releveant logfiles
Compositor log: [log.txt](/uploads/ee10762bd2ecf46781ab2a738b784b9b/log.txt)https://source.puri.sm/Librem5/phosh/-/issues/61Group Dialogs with the appliation main window2021-08-09T08:20:40ZGuido GuntherGroup Dialogs with the appliation main windowApplication modal dialogs should not show up as separate windows in the window list (this needs support from the compositor).
See https://source.puri.sm/Librem5/chatty/issues/89Application modal dialogs should not show up as separate windows in the window list (this needs support from the compositor).
See https://source.puri.sm/Librem5/chatty/issues/89https://source.puri.sm/Librem5/phosh/-/issues/62Switch gamma control to wlr-gamma-control-unstable-v12021-08-09T08:20:40ZemersionSwitch gamma control to wlr-gamma-control-unstable-v1The old protocol has a few issues. The new one is here:
https://github.com/swaywm/wlr-protocols/blob/master/unstable/wlr-gamma-control-unstable-v1.xmlThe old protocol has a few issues. The new one is here:
https://github.com/swaywm/wlr-protocols/blob/master/unstable/wlr-gamma-control-unstable-v1.xmlhttps://source.puri.sm/Librem5/phosh/-/issues/65Investigate GTK4 switch2021-08-09T08:20:40ZGuido GuntherInvestigate GTK4 switchJust a tracker bug with issues that we notice when building against GTK+4
Currently blocked by https://gitlab.gnome.org/GNOME/gtk/merge_requests/713Just a tracker bug with issues that we notice when building against GTK+4
Currently blocked by https://gitlab.gnome.org/GNOME/gtk/merge_requests/713https://source.puri.sm/Librem5/phosh/-/issues/68Switch the app from the bottom bar2021-08-09T08:20:40ZAdrien PlazasSwitch the app from the bottom barEpic: https://source.puri.sm/Librem5/use-cases/issues/26
We should be able to switch to the previous/next app wit the appropriate gesture in the bottom bar (App Switching).Epic: https://source.puri.sm/Librem5/use-cases/issues/26
We should be able to switch to the previous/next app wit the appropriate gesture in the bottom bar (App Switching).https://source.puri.sm/Librem5/phosh/-/issues/77Allow to take screenshots2021-08-09T08:20:40ZGuido GuntherAllow to take screenshotsRegular Screenshots would be a first step towards !181 . I find myself using `wlroots/_build/examples/screenshot` rather often.
We could do that outside phosh in a separate process.
Relevant protocol https://github.com/swaywm/wlr-proto...Regular Screenshots would be a first step towards !181 . I find myself using `wlroots/_build/examples/screenshot` rather often.
We could do that outside phosh in a separate process.
Relevant protocol https://github.com/swaywm/wlr-protocols/blob/master/unstable/wlr-screencopy-unstable-v1.xml
TODOs after !782:
- [ ] Capture all outputs
- [ ] capture single window
- [ ] allow to select screen areahttps://source.puri.sm/Librem5/phosh/-/issues/92Add compositor as subproject2021-08-09T08:20:40ZGuido GuntherAdd compositor as subprojectOptionally fetching and building the compositor would make things simpler for people running nested.Optionally fetching and building the compositor would make things simpler for people running nested.https://source.puri.sm/Librem5/phosh/-/issues/93Add matrix bot2021-08-09T08:20:40ZGuido GuntherAdd matrix botI'd be nice to have matrix bot with MRs and bug notifications forwarded to the phosh channelI'd be nice to have matrix bot with MRs and bug notifications forwarded to the phosh channelhttps://source.puri.sm/Librem5/phosh/-/issues/97Allow passwords with arbitrary characters2021-08-09T08:20:40ZBart RibbersAllow passwords with arbitrary charactersCurrently Phosh only allows passwords with purely numbers in them. It would be nice however to allow any arbitrary character, like letters, punctuation, symbols, etc. This would probably require integration with `virtboard`.Currently Phosh only allows passwords with purely numbers in them. It would be nice however to allow any arbitrary character, like letters, punctuation, symbols, etc. This would probably require integration with `virtboard`.https://source.puri.sm/Librem5/phosh/-/issues/103Improve AM/PM time layout on lockscreen2021-08-09T08:20:40ZTobias BernardImprove AM/PM time layout on lockscreenDisplaying AM/PM in the same font size next to the time looks a bit odd, because the string is very long.
![image](/uploads/4733d2ef9fcaedd2b13df44addf8aa9c/image.png)
Potential solutions include
- styling AM/PM differently (smaller si...Displaying AM/PM in the same font size next to the time looks a bit odd, because the string is very long.
![image](/uploads/4733d2ef9fcaedd2b13df44addf8aa9c/image.png)
Potential solutions include
- styling AM/PM differently (smaller size, different font weight) to de-emphasize it compared to the time itself and make the whole thing less wide
- removing the AM/PM label completely, since presumably people know which one they set and don't change it very often
Personally I'd lean towards the second solution, since it's simpler both visually and implementation-wise, and seems to work fine on other mobile platforms (e.g. the iPhone does it).
![image](/uploads/29062e97ee9fcab6a1d331d1e20f4e98/image.png)https://source.puri.sm/Librem5/phosh/-/issues/112hide/unhide lock screen2021-08-09T08:20:41ZGuido Guntherhide/unhide lock screencurrently we do a `gtk_widget_{new,destroy}()` to build / hide the lockscreen. We should just toggle visibility to reduce the setup cost and flickering.currently we do a `gtk_widget_{new,destroy}()` to build / hide the lockscreen. We should just toggle visibility to reduce the setup cost and flickering.https://source.puri.sm/Librem5/phosh/-/issues/117Restart gnome-session if it gave up2021-08-09T08:20:41ZGuido GuntherRestart gnome-session if it gave upCurrently if gnome-session gives up due to a failed component (see https://source.puri.sm/Librem5/squeekboard/issues/26) we end up with a black screen. Make systemd restart the session in that case.Currently if gnome-session gives up due to a failed component (see https://source.puri.sm/Librem5/squeekboard/issues/26) we end up with a black screen. Make systemd restart the session in that case.https://source.puri.sm/Librem5/phosh/-/issues/126split out layer surface code2021-08-09T08:20:41ZGuido Gunthersplit out layer surface codesqueekboard wants to use it as well: https://source.puri.sm/Librem5/squeekboard/issues/39#note_44466squeekboard wants to use it as well: https://source.puri.sm/Librem5/squeekboard/issues/39#note_44466https://source.puri.sm/Librem5/phosh/-/issues/139support output mirroring2021-08-09T08:20:41ZGuido Gunthersupport output mirroringespecially useful when doing presentations with the phone and an externally hooked up output.especially useful when doing presentations with the phone and an externally hooked up output.https://source.puri.sm/Librem5/phosh/-/issues/143show/hide top menu by swiping (one-finger) down/up2021-08-09T08:20:41ZMartin Kepplingershow/hide top menu by swiping (one-finger) down/upThe only intuitive behaviour IMO to show the menu from the top bar, is to slide / swipe down using 1 finger (that part is semi-ok right now already, only the "tip" is recognized though) and to hide it again, swip up, starting basically a...The only intuitive behaviour IMO to show the menu from the top bar, is to slide / swipe down using 1 finger (that part is semi-ok right now already, only the "tip" is recognized though) and to hide it again, swip up, starting basically anywhere, while it is open.
**Nobody I showed it to would have guessed that the top-menu closes by tapping the top bar again.**https://source.puri.sm/Librem5/phosh/-/issues/145lockscreen: what should go there2021-08-09T08:20:41ZGuido Guntherlockscreen: what should go therethere's
- clock/date (alread there)
- battery/wifi/wwan icons
- [ ] https://source.puri.sm/Librem5/phosh/issues/94
- [ ] https://source.puri.sm/Librem5/phosh/issues/12
but there's likely more to come. This issue is there to get input ...there's
- clock/date (alread there)
- battery/wifi/wwan icons
- [ ] https://source.puri.sm/Librem5/phosh/issues/94
- [ ] https://source.puri.sm/Librem5/phosh/issues/12
but there's likely more to come. This issue is there to get input for future directions so we can design proper interfaces.
This is not a poll! Be sure you're familiar with other open phosh bugs before commenting here.
/cc @sebastian.krzyszkowiak @bob.hamhttps://source.puri.sm/Librem5/phosh/-/issues/162Overview shows focus outline on widgets when interacting with it2021-08-09T08:20:41ZSebastian KrzyszkowiakOverview shows focus outline on widgets when interacting with it![Screenshot_20190901_183039](/uploads/519fcc8b33679cdf6221a21d029f1079/Screenshot_20190901_183039.png)
![Screenshot_20190901_183109](/uploads/38dc449652f1e5cac5e2836333e0c5ea/Screenshot_20190901_183109.png)
This doesn't always happen ...![Screenshot_20190901_183039](/uploads/519fcc8b33679cdf6221a21d029f1079/Screenshot_20190901_183039.png)
![Screenshot_20190901_183109](/uploads/38dc449652f1e5cac5e2836333e0c5ea/Screenshot_20190901_183109.png)
This doesn't always happen right away, but starts to be shown after some playing with the UI. Happens on devkit as well.https://source.puri.sm/Librem5/phosh/-/issues/166notifications: allow to show on lock screen2021-08-09T08:20:41ZGuido Gunthernotifications: allow to show on lock screenthis needs privacy handling as well so there's more to it than just the widget. See e.g. (`org.gnome.desktop.notifications` `show-in-lock-screen`)this needs privacy handling as well so there's more to it than just the widget. See e.g. (`org.gnome.desktop.notifications` `show-in-lock-screen`)https://source.puri.sm/Librem5/phosh/-/issues/168notifications: Handle application specific notification settings2023-05-18T18:54:28ZGuido Gunthernotifications: Handle application specific notification settings`org.gnome.desktop.notifications.application.<name>`
- [x] `show-banners` (see !675)
- [ ] `enable` (this one can just go on top of !675 and make sense to have). See !863
----
- [ ] `show-in-lockscreen` needs support first (for persis...`org.gnome.desktop.notifications.application.<name>`
- [x] `show-banners` (see !675)
- [ ] `enable` (this one can just go on top of !675 and make sense to have). See !863
----
- [ ] `show-in-lockscreen` needs support first (for persistent lockscreen notifications we could have banners easily)
- [ ] `force-expand` (we always expand atm)
- [ ] `enable-sound-alert` (we don't handle sound alert at all atm)
- [ ] `details-in-lockscreen` (we don't show notifications on the lock screen atm)https://source.puri.sm/Librem5/phosh/-/issues/169notifications: handle x-gnome-privacy-scope2021-08-09T08:20:41ZGuido Gunthernotifications: handle x-gnome-privacy-scopehttps://source.puri.sm/Librem5/phosh/-/issues/183Instant feedback when opening an app2021-08-09T08:20:42ZTobias BernardInstant feedback when opening an appStarting an app can take a while, so after clicking the icon in the app grid you currently see the previously active app first, and then when the app starts it overlays that app. This can lead to thinking that opening the app didn't work...Starting an app can take a while, so after clicking the icon in the app grid you currently see the previously active app first, and then when the app starts it overlays that app. This can lead to thinking that opening the app didn't work, so I've found myself going back to the overview and try to open it again.
Ideally the app window should open immediately (like on iOS for example), even if it contains just a spinner for a few seconds. I'd be interested in a general GNOME-wide pattern for this, but there might also be short-term solutions we could do on the phone only, since we only have a single fullscreen window at a time. I think Endless has something like this, where they have a separate fake window containing the app icon and a spinner, which closes when the real window opens.Guido GuntherGuido Guntherhttps://source.puri.sm/Librem5/phosh/-/issues/188Keyboard covers shell dialog2021-08-09T08:20:42ZTobias BernardKeyboard covers shell dialogShell dialogs are centered, so when the keyboard opens it can partially obscure them:
![Screenshot_from_2019-09-23_10-50-10](/uploads/33ab55789f5e1aac3afd828134cc6e3c/Screenshot_from_2019-09-23_10-50-10.png)
Ideally we'd center the dia...Shell dialogs are centered, so when the keyboard opens it can partially obscure them:
![Screenshot_from_2019-09-23_10-50-10](/uploads/33ab55789f5e1aac3afd828134cc6e3c/Screenshot_from_2019-09-23_10-50-10.png)
Ideally we'd center the dialog vertically on the available space that isn't covered by the keyboard.https://source.puri.sm/Librem5/phosh/-/issues/196Follow-up from "Scroll to current activity in overview"2021-08-09T08:20:42ZSebastian KrzyszkowiakFollow-up from "Scroll to current activity in overview"The following discussion from !317 should be addressed:
- [ ] @guido.gunther started a [discussion](https://source.puri.sm/Librem5/phosh/merge_requests/317#note_55372): (+2 comments)
> since state in the wayland protocol is a bit ...The following discussion from !317 should be addressed:
- [ ] @guido.gunther started a [discussion](https://source.puri.sm/Librem5/phosh/merge_requests/317#note_55372): (+2 comments)
> since state in the wayland protocol is a bit mask i wonder if this should be an enum and we want to grab the whole state on state change which wouldn't be much more effort? we likely at least want to know about full screen but maximized might also be useful.https://source.puri.sm/Librem5/phosh/-/issues/206Keyboard language indicator papercuts2021-08-09T08:20:42ZTobias BernardKeyboard language indicator papercutsThe keyboard language indicator in the system area should only be visible when both of these conditions are true:
- More than one language is configured in the Region & Language settings
- A physical keyboard is plugged in
Also, there s...The keyboard language indicator in the system area should only be visible when both of these conditions are true:
- More than one language is configured in the Region & Language settings
- A physical keyboard is plugged in
Also, there should be a right margin on the label (e.g. 6px), as it's currently too close to the battery icon:
![image](/uploads/517558c977472fdfe97ceb8b6c26b370/image.png)https://source.puri.sm/Librem5/phosh/-/issues/208App drawer: app grid layout2021-08-09T08:20:42ZTobias BernardApp drawer: app grid layoutIf we're going to support larger screens better we need a way for the app drawer to work well across these.
In my view, this is the best path forward at the moment:
- We keep a fixed number of apps per page, regardless of form factor (o...If we're going to support larger screens better we need a way for the app drawer to work well across these.
In my view, this is the best path forward at the moment:
- We keep a fixed number of apps per page, regardless of form factor (otherwise we lose the benefits of pagination for finding apps)
- The apps on a page flow from the top left to the bottom right, and there can be empty spots at the end
- Apps can be freely re-arranged and moved between pages via drag & drop
- 24 seems like a good number of apps per page, because it's in the ballpark of what other systems have (usually 20-30 per page, regardless of screen size) and it's divisible by 3, 4, 6, and 8 (which allows for flexible re-layouting)
- No fixed favorites row, as that would not work if the number of icons per row change. Instead, you can freely move icons around on the pages, and the first row of apps on the first page is a sort of "implicit" set of favorites
So on phones we'd have 4x6 apps per page:
![image](/uploads/a7af0a74164a3112b55c7b38d1c621c2/image.png)
On larger, landscape screens we'd have the same number of apps, but laid out differently, e.g. 6x4 or 8x3:
![image](/uploads/0102d2540fa9e370f25a9a80b399806d/image.png)https://source.puri.sm/Librem5/phosh/-/issues/221Notifications without body don't show up2021-08-09T08:20:42ZSebastian KrzyszkowiakNotifications without body don't show upNotifications contain summary and optional body. However, in Phosh notifications without body fail to show up properly.
To reproduce:
`notify-send "lorem ipsum"`
You'll see that something (single pixel line?) shows up on the screen, b...Notifications contain summary and optional body. However, in Phosh notifications without body fail to show up properly.
To reproduce:
`notify-send "lorem ipsum"`
You'll see that something (single pixel line?) shows up on the screen, but hides after a split second.
Notifications with body (like `notify-send "lorem ipsum" "dolor sit amet"`) work fine.https://source.puri.sm/Librem5/phosh/-/issues/223Only show favorite apps in overview2021-08-09T08:20:42ZTobias BernardOnly show favorite apps in overviewThe app cards in the overview are currently smaller than they should be, and we show a scrolling version of the app grid below that, which is a bit weird.
We could move the overview a bit closer to the designs relatively easily, if we m...The app cards in the overview are currently smaller than they should be, and we show a scrolling version of the app grid below that, which is a bit weird.
We could move the overview a bit closer to the designs relatively easily, if we make the app cards larger, and only show the first row of apps below that. Swiping down or clicking the search entry would move down to the full app grid.
This intermediate state would look something like this:
![intermediate](/uploads/e18ac767f02dd6261cd07e4392be12d9/intermediate.png)
Longer-term design direction for reference:
![overview](/uploads/7981d44951da5ee088c9bd803889e97e/overview.png)https://source.puri.sm/Librem5/phosh/-/issues/230Prevent selection of passcode GtkEntry on the lock screen2021-08-09T08:20:42ZSebastian KrzyszkowiakPrevent selection of passcode GtkEntry on the lock screenIt's unnecessary, doesn't interact well with the paginator and only leads to glitches:
![lock2](/uploads/a77a7bf2de6612b7958d7978c23b71d0/lock2.png)
![lock3](/uploads/edd93cf469c414568864d25c02cf981c/lock3.png)
![lock4](/uploads/312d325...It's unnecessary, doesn't interact well with the paginator and only leads to glitches:
![lock2](/uploads/a77a7bf2de6612b7958d7978c23b71d0/lock2.png)
![lock3](/uploads/edd93cf469c414568864d25c02cf981c/lock3.png)
![lock4](/uploads/312d325fa0df792a3680374bcc841fb2/lock4.png)
![lock](/uploads/de7b8f62317b133beaf9a7897a49bba2/lock.png)
https://source.puri.sm/Librem5/phosh/-/issues/239Lockscreen's animated arrow causes the whole surface to be damaged2021-08-09T08:20:42ZSebastian KrzyszkowiakLockscreen's animated arrow causes the whole surface to be damagedNot a high priority, but there sure must be some way to make gtk behave ;)
```
[794442.359] wl_callback@125.done(9728034)
[794458.750] -> wl_surface@35.attach(wl_buffer@43, 0, 0)
[794459.022] -> wl_surface@35.set_buffer_scale(2)
[7944...Not a high priority, but there sure must be some way to make gtk behave ;)
```
[794442.359] wl_callback@125.done(9728034)
[794458.750] -> wl_surface@35.attach(wl_buffer@43, 0, 0)
[794459.022] -> wl_surface@35.set_buffer_scale(2)
[794459.068] -> wl_surface@35.damage(0, 0, 360, 720)
[794459.268] -> wl_surface@35.frame(new id wl_callback@125)
[794459.328] -> wl_surface@35.commit()
[794469.792] wl_buffer@43.release()
[794474.019] wl_display@1.delete_id(125)
```https://source.puri.sm/Librem5/phosh/-/issues/246Notifications Stack On Top Of Each Other2021-08-09T08:20:42ZKyle RankinNotifications Stack On Top Of Each Other# Problem Statement
Currently when more than one notification occurs at a time, the most recent notification covers others up so that one can only view a single notification.
# Expected Behavior
Like with Gnome Shell notifications, new ...# Problem Statement
Currently when more than one notification occurs at a time, the most recent notification covers others up so that one can only view a single notification.
# Expected Behavior
Like with Gnome Shell notifications, new notifications appear below existing notifications and when a notification goes away, any notifications below it move up to fill its space.
# Steps to Reproduce
* `notify-send test1 test1; notify-send test2 test2; notify-send test3 test3`
* Note that they all quickly appear but only test3 is visible
* Tapping a notification makes it disappear and reveals notifications below it that haven't timed out.https://source.puri.sm/Librem5/phosh/-/issues/256lockup (no touch events in shell) but terminal keyboard still working2021-08-09T08:20:42ZTodd Weaverlockup (no touch events in shell) but terminal keyboard still working# What problem did you encounter
Phosh seemed to lockup (not taking any touch events), but the terminal keyboard still was working
## How to reproduce
![librem-5-birch-keyboard-touch-works-phosh-does-not-2020-01-28](/uploads/3db071805...# What problem did you encounter
Phosh seemed to lockup (not taking any touch events), but the terminal keyboard still was working
## How to reproduce
![librem-5-birch-keyboard-touch-works-phosh-does-not-2020-01-28](/uploads/3db071805e098db7bb71773af54b6da0/librem-5-birch-keyboard-touch-works-phosh-does-not-2020-01-28.webm)
# Which version did you encounter the bug in?
```
purism@l5b:~$ uname -a
Linux l5b 5.3.0-librem5-h1 #1 SMP PREEMPT Sat Jan 25 09:55:23 CET 2020 aarch64 GNU/Linux
purism@l5b:~$ dpkg -s phosh | grep Version
Version: 0.1.7
purism@l5b:~$ dpkg -s phoc | grep Version
Version: 0.1.5pureos1
purism@l5b:~$ dpkg -s squeekboard | grep Version
Version: 1.8.0
purism@l5b:~$ uptime
07:54:26 up 42 min, 2 users, load average: 0.26, 0.16, 0.13
purism@l5b:~$ date
Tue 28 Jan 2020 07:54:27 AM PST
purism@l5b:~$
```
# What hardware are you running phosh on?
- Librem 5 Birchhttps://source.puri.sm/Librem5/phosh/-/issues/259Have icon(s) for non-Wi-Fi network interfaces2021-08-09T08:20:43ZBob HamHave icon(s) for non-Wi-Fi network interfacesIt would be useful to have an icon indicating that a non-Wi-fi network interface (USB, Ethernet, whatever) is up, similar to the Wi-Fi icon.It would be useful to have an icon indicating that a non-Wi-fi network interface (USB, Ethernet, whatever) is up, similar to the Wi-Fi icon.https://source.puri.sm/Librem5/phosh/-/issues/266Backlight slider does not reliably set the brightness2021-08-09T08:20:43ZSebastian KrzyszkowiakBacklight slider does not reliably set the brightnessSince !422, quick slides over the slider can result in the slider following the finger, but the actual screen brightness staying at the old position.
To easily see what happens there I recommend opening the power page in g-c-c and playi...Since !422, quick slides over the slider can result in the slider following the finger, but the actual screen brightness staying at the old position.
To easily see what happens there I recommend opening the power page in g-c-c and playing with phosh's brightness slider a bit. The two sliders (one in phosh and one in g-c-c) should be always in sync with each other, but it's actually easy to have them showing completely different values (with the g-c-c one telling he truth).
![brightness](/uploads/1b4298a92f6eddbe419c7923f309b658/brightness.png)https://source.puri.sm/Librem5/phosh/-/issues/267GSM indicator does not work if phosh is started with no ModemManager running2021-08-09T08:20:43ZSebastian KrzyszkowiakGSM indicator does not work if phosh is started with no ModemManager runningTo reproduce:
```
sudo systemctl disable --now ModemManager
# the GSM indicator disappears
sudo systemctl restart phosh
# there's no GSM indicator, as expected
sudo systemctl enable --now ModemManager
# no GSM indicator appears, even th...To reproduce:
```
sudo systemctl disable --now ModemManager
# the GSM indicator disappears
sudo systemctl restart phosh
# there's no GSM indicator, as expected
sudo systemctl enable --now ModemManager
# no GSM indicator appears, even though it should
```
Just restarting ModemManager with GSM indicator already working does not trigger this bug.https://source.puri.sm/Librem5/phosh/-/issues/270Wifimanager: Use specifc device instead of device with active connection2021-08-09T08:20:43ZJulian SparberWifimanager: Use specifc device instead of device with active connectionCurrently we are using the wifi device with an active connection for the state shown in the top bar.
This approach has at least this two issues:
* When more then one wifi device is available and connected: the user has no way to know ...Currently we are using the wifi device with an active connection for the state shown in the top bar.
This approach has at least this two issues:
* When more then one wifi device is available and connected: the user has no way to know which device state is shown.
(The device which connected first is used.)
* We can't show an disabled state since we don't consider disconnected devices and therefore we have no device. (will be needed for the quicksettings.
I would suggest to use the first available wifi device which solves the second issue. And solve the first issue at a later point since not even upstream GNOME handles this case.https://source.puri.sm/Librem5/phosh/-/issues/271StatusIcon: Wifi status flashes offline when disabling wifi2021-08-09T08:20:43ZJulian SparberStatusIcon: Wifi status flashes offline when disabling wifiThe Wifi icon in the top bar flashes for a really short amount of time the offline icon before disappearing. I would expect it to go directly to hidden without the extra step over the offline icon.
![doc_2020-02-19_13-29-35](/uploads/2d...The Wifi icon in the top bar flashes for a really short amount of time the offline icon before disappearing. I would expect it to go directly to hidden without the extra step over the offline icon.
![doc_2020-02-19_13-29-35](/uploads/2d02caa4b55002d32b4193c82597abda/doc_2020-02-19_13-29-35.mp4)https://source.puri.sm/Librem5/phosh/-/issues/2723G/4G/etc status bar indicator styling2021-08-09T08:20:43ZTobias Bernard3G/4G/etc status bar indicator stylingThe 3G/4G/etc indicator label in the status bar is currently the same text size as other labels in the status bar, but using the regular font weight.
I'd restyle it a bit:
- bold font weight
- slightly smaller font size to make it feel ...The 3G/4G/etc indicator label in the status bar is currently the same text size as other labels in the status bar, but using the regular font weight.
I'd restyle it a bit:
- bold font weight
- slightly smaller font size to make it feel like part of the icon more
- given the previous point, make sure it's still vertically centeredhttps://source.puri.sm/Librem5/phosh/-/issues/277Quicksettings: Some times they remain highlighted2021-08-09T08:20:43ZJulian SparberQuicksettings: Some times they remain highlightedNot really sure why this happens. Some times when clicking a quick settings the button remains highlighted. Quickly pressing the button a few times causes the issue most of the time. This doesn't happen on the desktop with a mouse.
![ima...Not really sure why this happens. Some times when clicking a quick settings the button remains highlighted. Quickly pressing the button a few times causes the issue most of the time. This doesn't happen on the desktop with a mouse.
![image](/uploads/b8302ffe086fcf048bac51367607a456/image.png)https://source.puri.sm/Librem5/phosh/-/issues/279Settings causes keyboard to hide permanently2021-08-09T08:20:43ZDorota CzaplejewiczSettings causes keyboard to hide permanently# What problem did you encounter
Opening settings changes keyboard behaviour.
This seems to stem from the settings bar calling the DBus method for hiding the keyboard: https://source.puri.sm/Librem5/phosh/blob/master/src/shell.c#L204
...# What problem did you encounter
Opening settings changes keyboard behaviour.
This seems to stem from the settings bar calling the DBus method for hiding the keyboard: https://source.puri.sm/Librem5/phosh/blob/master/src/shell.c#L204
I don't think this is the right approach to getting the keyboard out of the way. Before observing this, I assumed that the settings layer takes focus, and that's what makes the keyboard go away, in absence of any editable fields. It would also make sense because the settings are effectively modal, and contain widgets that can be operable with a keyboard in principle. If we worry about editable fields popping the keyboard back in, this should be tackled separately.
The DBus API for hiding/revealing the keyboard is IMO the wrong tool to use it, because its resource (visibility) management is very manual causing bugs like this one. I consider the API only a crutch to give the user control because we will never reliably pop up for odd apps.
## What is the current behaviour?
Settings hides the keyboard forcefully, and then never brings it back.
## How to reproduce
1. Open the terminal (or anything with a text fiels)
2. Open the text field: the keyboard pops up
3. Click the settings bar: settings pops up, keyboard becomes hidden
4. Click the text field on the application
Expected behaviour: settings hides, keyboard returns.
Actual behaviour: settings hides.
# Which version did you encounter the bug in?
Version: 0.1.8 from amber
# What hardware are you running phosh on?
- [ ] Librem5
# Releveant logfiles
Squeekboard's WAYLAND_DEBUG=1 log starts before clicking the settings bar, ends after clicing the application:
```
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091791,261] -> wl_surface@32.destroy()
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091791,460] -> wl_buffer@34.destroy()
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091791,489] -> wl_shm_pool@36.destroy()
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091792,178] -> zwlr_layer_surface_v1@37.destroy()
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091868,748] wl_display@1.delete_id(32)
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091868,835] wl_display@1.delete_id(34)
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091868,866] wl_display@1.delete_id(36)
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091868,899] wl_display@1.delete_id(37)
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091868,928] zwp_input_method_v2@24.surrounding_text("", 0, 0)
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091869,057] zwp_input_method_v2@24.text_change_cause(0)
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091869,098] zwp_input_method_v2@24.content_type(0, 0)
Mar 01 14:54:57 pureos sm.puri.OSK0.desktop[1510]: [1091869,160] zwp_input_method_v2@24.done()
```https://source.puri.sm/Librem5/phosh/-/issues/280Can't shut down when there is a notification2021-08-09T08:20:43ZAdrien PlazasCan't shut down when there is a notificationI had notifications but wanted to shut the Librem 5 down, so I opened the top tray, touched the power off button, but it did nothing. I had to close the notifications first for the power off button to work. I didn't try other buttons, I ...I had notifications but wanted to shut the Librem 5 down, so I opened the top tray, touched the power off button, but it did nothing. I had to close the notifications first for the power off button to work. I didn't try other buttons, I suspect it would be the same.https://source.puri.sm/Librem5/phosh/-/issues/282show application names for the pinned / top applications2021-08-09T08:20:43ZMartin Kepplingershow application names for the pinned / top applicationsI'm always wondering why the "top" pinned applicaions don't show the name while all others do. IMO this should be consistent for all app icons.
We could also hide all and make the name for an app visible on "touch==1" so that it's visib...I'm always wondering why the "top" pinned applicaions don't show the name while all others do. IMO this should be consistent for all app icons.
We could also hide all and make the name for an app visible on "touch==1" so that it's visible while holding the finger down. Then one can choose to swipe away (move the whole overview drawer in the future I guess) or remove the finger in order to complete the "tap" and start the application.https://source.puri.sm/Librem5/phosh/-/issues/283turn off backlight when proximity is on2021-08-09T08:20:44ZGuido Guntherturn off backlight when proximity is onwith proximity we dimm but turn off the backlight - with wlr-power-management being integrated we can do this now.with proximity we dimm but turn off the backlight - with wlr-power-management being integrated we can do this now.https://source.puri.sm/Librem5/phosh/-/issues/284Show feedback status on lock screen / top bar2021-08-09T08:20:44ZGuido GuntherShow feedback status on lock screen / top barWhen pulling the phone out of the pocket i often wonder if it would ring when a call came in, showing the feedback icon next to the other icons would mitigate this. Same goes for the top bar but to a lesser extend.
/cc @tobias.bernard s...When pulling the phone out of the pocket i often wonder if it would ring when a call came in, showing the feedback icon next to the other icons would mitigate this. Same goes for the top bar but to a lesser extend.
/cc @tobias.bernard since i did not see it in any of the mockups yet.https://source.puri.sm/Librem5/phosh/-/issues/286Notification bubble: minor visual issues2021-08-09T12:03:34ZTobias BernardNotification bubble: minor visual issuesThe notification bubbles look quite good already, but there are a few small papercuts, specifically:
- The app symbolic icon should be dimmed
- There should be a little bit less horizontal space between the symbolic icon and app name (ma...The notification bubbles look quite good already, but there are a few small papercuts, specifically:
- The app symbolic icon should be dimmed
- There should be a little bit less horizontal space between the symbolic icon and app name (maybe 2 or 3px less)
- There's a bit too much vertical space below the app name/icon. I'd reduce that by a few pixels.
![image](/uploads/95ac27ace02f4f5efad0f44317cfad36/image.png)https://source.puri.sm/Librem5/phosh/-/issues/290Connect to upower async2021-08-09T08:20:44ZGuido GuntherConnect to upower asyncWe use `up_client_new_full` which calls `up_exported_daemon_proxy_new_for_bus_sync` and there's currently no async version in libupower-glib.We use `up_client_new_full` which calls `up_exported_daemon_proxy_new_for_bus_sync` and there's currently no async version in libupower-glib.https://source.puri.sm/Librem5/phosh/-/issues/293turn off screen by double tap on lockscreen2021-08-09T08:20:44ZMartin Kepplingerturn off screen by double tap on lockscreenIn order to prevent wear-out of the hardware power button, people like to use other means for turning off the screen. I'm used to do it by double-tapping on the lockscreen (on my Android phone). That enables to look at the clock and turn...In order to prevent wear-out of the hardware power button, people like to use other means for turning off the screen. I'm used to do it by double-tapping on the lockscreen (on my Android phone). That enables to look at the clock and turn off the screen by only using the power button once (to turn on).https://source.puri.sm/Librem5/phosh/-/issues/298move settings to settings/2021-08-09T08:20:44ZGuido Gunthermove settings to settings/The settings menu grows rather complex hence we have a settings/ folder. For unknown reasons `QuickSettings` and its derived widgets don't live there so we should move them. Let's do that past !452 to not break that MR.The settings menu grows rather complex hence we have a settings/ folder. For unknown reasons `QuickSettings` and its derived widgets don't live there so we should move them. Let's do that past !452 to not break that MR.https://source.puri.sm/Librem5/phosh/-/issues/306Also blank screen directly after boot when idle for x seconds2021-08-09T08:20:44ZGuido GuntherAlso blank screen directly after boot when idle for x secondsRight after booting the phone when the user does nothing, the screen does not fully blank (it only dims). As soon as the user unlocks once everything works as expected (first dim, then lock). We also blank correctly on the lock screen af...Right after booting the phone when the user does nothing, the screen does not fully blank (it only dims). As soon as the user unlocks once everything works as expected (first dim, then lock). We also blank correctly on the lock screen afterwards so only initial boot is broken. We can fix this by notifying g-s-d after a timeout once the shell is fully up.https://source.puri.sm/Librem5/phosh/-/issues/307settings: Make speaker icon larger and touch sensitive2021-08-09T08:20:44ZGuido Gunthersettings: Make speaker icon larger and touch sensitiveit could then open up sound settings right away.it could then open up sound settings right away.https://source.puri.sm/Librem5/phosh/-/issues/314media-player: add progress bar2021-08-09T08:20:44ZGuido Gunthermedia-player: add progress barin the end i forgot to add a progress bar that shows the elapsed time of the currently played song.in the end i forgot to add a progress bar that shows the elapsed time of the currently played song.https://source.puri.sm/Librem5/phosh/-/issues/322Use dbusmock in tests2021-08-09T08:20:44ZGuido GuntherUse dbusmock in testsgiven the amount of dbus services we care about using `dbusmock` would make sense.given the amount of dbus services we care about using `dbusmock` would make sense.https://source.puri.sm/Librem5/phosh/-/issues/324Window thumbnails all rendered next to each other2021-08-09T08:20:44ZGuido GuntherWindow thumbnails all rendered next to each otherCurrently we e.g. don't stack dialogs on top of their parents - this did not change with thumbnails (!481) it just becomes more visible since thumbnails have contents now (yay!) and so dialogs have the same height as their parents in the...Currently we e.g. don't stack dialogs on top of their parents - this did not change with thumbnails (!481) it just becomes more visible since thumbnails have contents now (yay!) and so dialogs have the same height as their parents in the overview which looks somewhat irritating.
(well known issue and needs compositor work, just filing this since i think we did not track it anywhere yet and since this affects the relations of `PhoshToplevel` and `PhoshThumbnail` classes)https://source.puri.sm/Librem5/phosh/-/issues/327Rework PhoshActivity and PhoshOverview2021-08-09T08:20:45ZSebastian KrzyszkowiakRework PhoshActivity and PhoshOverviewCurrently `PhoshOverview` manually manages the relationship between `PhoshToplevel` (a representation of xdg-foreign-toplevel) and `PhoshActivity` (a display-only widget with an activity card), by tracking them with `g_object_{set/get}_d...Currently `PhoshOverview` manually manages the relationship between `PhoshToplevel` (a representation of xdg-foreign-toplevel) and `PhoshActivity` (a display-only widget with an activity card), by tracking them with `g_object_{set/get}_data}`, requesting and passing window thumbnails etc.
This is far from ideal - `PhoshOverview` is already quite a messy thing having multiple responsibilities. However, the most obvious way of fixing that - letting `PhoshActivity` consume and handle `PhoshToplevel` by itself - isn't good either, since 1:1 relationship between `PhoshActivity` and `PhoshToplevel` is very temporary (see #324; and potentially other things in the future like splash screens or suspended activites).
We need a new entity responsible for representing and managing a single activity. I'd propose renaming current `PhoshActivity` into `PhoshActivityCard`, and creating a new `PhoshActivity` class that would then handle any actual logic by itself, allowing `PhoshOverview` to not care about activity implementation details.
This way, `PhoshActivityCard` could just represent the state of any `PhoshActivity` given to it, regardless of its internal implementation, freeing `PhoshOverview` from unwanted responsibilities. This also opens a path for compositing multiple thumbnails onto a single activity, as needed for window stacks (#324), without cluttering classes that shouldn't need to care about it, since this whole logic could be then properly contained in implementation of `PhoshActivity`.
This may also touch `PhoshThumbnail`, since currently it sits in a rather awkward place between Overview, Toplevel and Activity.https://source.puri.sm/Librem5/phosh/-/issues/330Multitasking cards: Drop labels2021-08-09T08:20:45ZTobias BernardMultitasking cards: Drop labelsNow that we have proper window thumbnails, we no longer need the label stopgap at the bottom to have windows be distinguishable. Instead, we can move the app icon to the center and have it overlap the bottom edge, like in the mockups:
!...Now that we have proper window thumbnails, we no longer need the label stopgap at the bottom to have windows be distinguishable. Instead, we can move the app icon to the center and have it overlap the bottom edge, like in the mockups:
![image](/uploads/189d4114e59857e22ec48dabb56d6d9c/image.png)
Since the thumbnails in the overview are currently still quite a bit smaller than in the mockups, I'd go with 32px icons (rather than 64) for now.https://source.puri.sm/Librem5/phosh/-/issues/331Settings: Power Off menu papercuts2021-08-09T08:20:45ZTobias BernardSettings: Power Off menu papercutsThe current power menu is functionally fine, but needs some styling fixes:
- [ ] Add 18px margin on top and right
- [ ] Increase padding inside the button so it's 40x40 px
- [ ] Make the button circular
- [x] Inside the popover, add 6px...The current power menu is functionally fine, but needs some styling fixes:
- [ ] Add 18px margin on top and right
- [ ] Increase padding inside the button so it's 40x40 px
- [ ] Make the button circular
- [x] Inside the popover, add 6px of margin all around
- [x] Add 6px of vertical padding on each menu item in the popover
Mockup:
![image](/uploads/bbe49ef67fd2aa412d48790afa299d62/image.png)https://source.puri.sm/Librem5/phosh/-/issues/333Feature request: Quick unlock on lock screen2021-08-09T08:20:45ZMarcoFeature request: Quick unlock on lock screenMy android phone has a feature called "Quick unlock" which sets the lock screen to immediately unlock when the correct PIN is entered. There is no need to press "Unlock" or similar.
May this be possible to be integrated into phosh lock ...My android phone has a feature called "Quick unlock" which sets the lock screen to immediately unlock when the correct PIN is entered. There is no need to press "Unlock" or similar.
May this be possible to be integrated into phosh lock screen? This would reduce the time unlocking takes every time by at least one button press :)https://source.puri.sm/Librem5/phosh/-/issues/334Add a popover to window thumbnails to enable automatic scaling?2021-08-09T08:20:45ZGuido GuntherAdd a popover to window thumbnails to enable automatic scaling?I'm on the fence on this one and it's certainly not high priority but doing a long press on a window thumbnail in phosh that shows a popover (which can get other usefull stuff as well) to enable automatic scaling of an app could help use...I'm on the fence on this one and it's certainly not high priority but doing a long press on a window thumbnail in phosh that shows a popover (which can get other usefull stuff as well) to enable automatic scaling of an app could help users forced to use non scaling apps (~~would need an extension to the private wayland protocol~~)https://source.puri.sm/Librem5/phosh/-/issues/336Asynchronous app search2021-08-09T08:20:45ZSebastian KrzyszkowiakAsynchronous app searchWith almost 100 launcher icons search in phosh becomes very slow on the phone, taking about a second to filter out the view on each key press.
The UI should react immediately regardless of the dataset size, possibly showing some spinner...With almost 100 launcher icons search in phosh becomes very slow on the phone, taking about a second to filter out the view on each key press.
The UI should react immediately regardless of the dataset size, possibly showing some spinner while it gets updated.https://source.puri.sm/Librem5/phosh/-/issues/338Wallpaper does not fill screen when fractional scaling is used2021-08-09T08:20:45ZFederico Amedeo IzzoWallpaper does not fill screen when fractional scaling is used# What problem did you encounter
I'm using Phosh on a Samsung Galaxy A3 2015 (samsung-a3ulte) with PostmarketOS
I've set the scaling factor to 1.5 in `/etc/phosh/rootston.ini`
And I've noticed that the wallpaper did not fill the screen a...# What problem did you encounter
I'm using Phosh on a Samsung Galaxy A3 2015 (samsung-a3ulte) with PostmarketOS
I've set the scaling factor to 1.5 in `/etc/phosh/rootston.ini`
And I've noticed that the wallpaper did not fill the screen after setting the scale factor, while it did before.
`/etc/phosh/rootston.ini`
```
[cursor:seat0]
map-to-output:DSI-1
[output:DSI-1]
scale = 1.5
```
## What is the current behaviour?
The wallpaper does not fill the screen.
## What is the expected behaviour?
The wallpaper fills the screen.
## How to reproduce
* Install PostmarketOS with Phosh on any device or \
Install PostmarketOS with Phosh on [QEMU target](https://wiki.postmarketos.org/wiki/QEMU_amd64_(qemu-amd64)) (no device required)
* Set a fractional scaling factor in `/etc/phosh/rootston.ini` or `/usr/share/phosh/phoc.ini`, for example `scale=1.5`
* The wallpaper does not fill the screen
![IMG_20200711_090806](/uploads/25718d4c618a7c043514c8882f7e2703/IMG_20200711_090806.jpg)
# Which version did you encounter the bug in?
- [ ] I Compiled it myself. If you compiled phosh from source please provide the
git revision via e.g. by running ``git log -1 --pretty=oneline`` and pasting
the output below.
- [ ] I used the precompiled Debian package (e.g. by running a prebuilt
image). Please paste the output oof ``dpkg -s phosh`` below.
- [x] I used the precompiled PostmarketOS package
```
samsung-a3ulte:~$ apk info phosh
phosh-0.4.0-r0 description:
Shell PoC for the Librem5
phosh-0.4.0-r0 webpage:
https://source.puri.sm/Librem5/phosh
phosh-0.4.0-r0 installed size:
880640
```
# What hardware are you running phosh on?
- [ ] amd64 qemu image
- [ ] Librem5 devkit
- [ ] nested compositor. If so please give the full command you start phosh
with.
- [x] other
I'm using Phosh on a Samsung Galaxy A3 2015 (samsung-a3ulte) with PostmarketOS. \
The phone has a 540x960 screen and this is [its wiki page](https://wiki.postmarketos.org/wiki/Samsung_Galaxy_A3_2015_(samsung-a3ulte)).
# Releveant logfiles
Please provide relevant logs. You can e.g. the logs since last boot read
with ``journalctl -b 0``.
* Kernel log: [dmesg](/uploads/9fab193b0a2813bb0402224a9f300554/dmesg)
* System log: [logread](/uploads/a843015c8daca7ecb283614a368718f3/logread)https://source.puri.sm/Librem5/phosh/-/issues/349Hook on focus change2021-08-09T08:20:45ZFederico CerattoHook on focus changeHello and thanks for developing Phosh.
It would be useful to have Phosh call a hook on change of focused application, ideally with the PIDs of the current and previous application (and, optionally, their binary paths).
Alternatively, it...Hello and thanks for developing Phosh.
It would be useful to have Phosh call a hook on change of focused application, ideally with the PIDs of the current and previous application (and, optionally, their binary paths).
Alternatively, it could use a signal, a named fifo, a d-bus message...
This could allow implementing battery optimizations including dynamically changing scheduling priorities, SIGSTOP/SIGCONT, changing swappines & so on.https://source.puri.sm/Librem5/phosh/-/issues/351Notification with timeout aren't being removed from the message tray2021-08-09T08:20:45ZGuido GuntherNotification with timeout aren't being removed from the message traycan be checked with `notify-send -t 3000 asdf sadf` - the notification sticks in the tray even after 3 seconds.can be checked with `notify-send -t 3000 asdf sadf` - the notification sticks in the tray even after 3 seconds.https://source.puri.sm/Librem5/phosh/-/issues/352'Really' Persistent notifications2021-08-09T08:20:45ZGuido Gunther'Really' Persistent notificationsNotifications aren't e.g. stored across reboots / poweroffs. We should keep them around.Notifications aren't e.g. stored across reboots / poweroffs. We should keep them around.https://source.puri.sm/Librem5/phosh/-/issues/354Topbar widgets/cusotmization2021-08-09T08:20:46ZKonrad DybcioTopbar widgets/cusotmizationThere's some free space on the top bar (mainly on the right side), which could be used for displaying things like temperature, cpu frequency, unread notification count or really anything else.There's some free space on the top bar (mainly on the right side), which could be used for displaying things like temperature, cpu frequency, unread notification count or really anything else.https://source.puri.sm/Librem5/phosh/-/issues/355Application icon not removed from favorites when app is removed2021-08-09T08:20:46ZJoao AzevedoApplication icon not removed from favorites when app is removed# What problem did you encounter
When deleting an application whose desktop Icon was in the favorites bar, while the app icon was removed, an artifact from the desktop shortcut remained.
To be completely removed either:
- Remove from ...# What problem did you encounter
When deleting an application whose desktop Icon was in the favorites bar, while the app icon was removed, an artifact from the desktop shortcut remained.
To be completely removed either:
- Remove from favorites
- Restart `phoc`
## What is the current behaviour?
When deleting an application remove the desktop Icon from the favorites Bar without need of restarting.
## What is the expected behaviour?
## How to reproduce
Please provide steps to reproduce the issue:
- Add an app to favorites
- Delete the app from the OS
This was with lollypop (flatpak) after deleting the app:
![2020-08-03-183831](/uploads/cbf7262e265ab49f9df46760b867b4c2/2020-08-03-183831.png)
![2020-08-03-183853](/uploads/a1481804e0b703139a354dce1b4b277b/2020-08-03-183853.png)
This was with GNOME-MPV from amber-main repos:
![2020-08-28-011819](/uploads/07afad50d70e90edc00ffcd00f2fa312/2020-08-28-011819.png)
# Which version did you encounter the bug in?
- [ ] I Compiled it myself. If you compiled phosh from source please provide the
git revision via e.g. by running ``git log -1 --pretty=oneline`` and pasting
the output below.
- [x] I used the precompiled Debian package (e.g. by running a prebuilt
image). Please paste the output oof ``dpkg -s phosh`` below.
```
Phosh Version: 0.4.3
```
# What hardware are you running phosh on?
- [ ] amd64 qemu image
- [x] Librem5
- [ ] nested compositor. If so please give the full command you start phosh
with.
- [ ] other (please elaborate)
# Releveant logfiles
Please provide relevant logs. You can e.g. the logs since last boot read
with ``journalctl -b 0``.https://source.puri.sm/Librem5/phosh/-/issues/368decouple lockscreen from shell2021-08-09T08:20:46ZMis012decouple lockscreen from shellOn desktop, it is typical that lockscreen is separate from shell (e.g sway - swaylock, gnome - gnome-screensaver/gdm) - which allows for changing lockscreen without changing the shell.
Funnily enough, even android has support for custom...On desktop, it is typical that lockscreen is separate from shell (e.g sway - swaylock, gnome - gnome-screensaver/gdm) - which allows for changing lockscreen without changing the shell.
Funnily enough, even android has support for custom lockscreens.
Would it be possible to decouple the lockscreen from the rest of the shell to allow for using a different lockscreen?https://source.puri.sm/Librem5/phosh/-/issues/370Lock screen: MPRIS control visuals2021-08-09T08:20:46ZTobias BernardLock screen: MPRIS control visualsThe current lock screen MPRIS controls look a bit off, I'd
- move them below the time/date/status block (and move that block up a bit)
- left-align the text
- move the two text lines down a little bit and closer together, so as to appea...The current lock screen MPRIS controls look a bit off, I'd
- move them below the time/date/status block (and move that block up a bit)
- left-align the text
- move the two text lines down a little bit and closer together, so as to appear centered on the cover vertically
- round the corners on the widget (I'd try 9 or 12px)
- Move the skip buttons a bit towards the center, they feel to spread out
- Use the same amount of padding vertically and horizontally on the container
![image](/uploads/9d911e95831b90d324ff7d962fbd4911/image.png)Julian SparberJulian Sparberhttps://source.puri.sm/Librem5/phosh/-/issues/371Settings: MPRIS controls isn't correctly rounded off2021-08-09T08:20:46ZJulian SparberSettings: MPRIS controls isn't correctly rounded offThe widget isn't fully round and the button should be rounded off as well (iirc)
![image](/uploads/3bedba501a6093f5ec5a4d88788f6fdf/image.png)
CC: @tobias.bernardThe widget isn't fully round and the button should be rounded off as well (iirc)
![image](/uploads/3bedba501a6093f5ec5a4d88788f6fdf/image.png)
CC: @tobias.bernardJulian SparberJulian Sparberhttps://source.puri.sm/Librem5/phosh/-/issues/375media-player: strange state when used with Shortwave2021-08-09T08:20:46ZJulian Sparbermedia-player: strange state when used with Shortwave# What problem did you encounter
When using Shortwave (from flathub) the media-player widget does show without any radio station playing (only on startup).
The play button is insensitive, but the next and prev buttons are sensitive but d...# What problem did you encounter
When using Shortwave (from flathub) the media-player widget does show without any radio station playing (only on startup).
The play button is insensitive, but the next and prev buttons are sensitive but don't do anything. Also the widget shows "Unknown Title" and "Unknown Artist".
![open_shortwave](/uploads/4c557e72c3ec8afe6e049f6f9cfbe86d/open_shortwave.png)
After pressing play for a radio station (in-app) the correct name and subtitle are shown. When pressing pause (in-app and media player widget) the media-player widget disappears but the notification drawer doesn't resize. (after closing/opening the drawer it uses only the needed space).
![shortwave-hidden-media-player-after-pause](/uploads/a4f665b9a6d0a3a73f23ce816e90cd7c/shortwave-hidden-media-player-after-pause.png)
When pressing the play button again from inside the app the widget shows up again in the notification drawer, but the title and subtitle in the media player widget remains empty (but it shows the correct cover). After some time the widget is filled in and shows correctly.
![shortwave-empty-text](/uploads/a6f8868b0f51ff9b1814d1279f159208/shortwave-empty-text.png)
## What is the expected behaviour?
- The media player widget shouldn't show Unknown Artist/Title nor shouldn't it be empty.
- When the media player widget is hidden the notification drawer should resize.
- In general I would expect the widget to behave the same way as in GNOME.
## How to reproduce
- install shortwave from flathub via flatpak (couldn't reproduce the issue in Rhythmbox from flatpak)
- add a radio station to shortwave
- do the things described above.
# Which version did you encounter the bug in?
- [ ] I Compiled it myself. If you compiled phosh from source please provide the
git revision via e.g. by running ``git log -1 --pretty=oneline`` and pasting
the output below.
- [X] I used the precompiled Debian package (e.g. by running a prebuilt
image). Please paste the output oof ``dpkg -s phosh`` below.
```
Package: phosh
Status: install ok installed
Priority: optional
Section: x11
Installed-Size: 1162
Maintainer: Guido Günther <agx@sigxcpu.org>
Architecture: arm64
Version: 0.4.4
Provides: notification-daemon, polkit-1-auth-agent
Depends: dconf-gsettings-backend | gsettings-backend, libc6 (>= 2.17), libcairo2 (>= 1.2.4), libfeedback-0.0-0 (>= 0.0.0+git20200707), libgcr-base-3-1 (>= 3.8.0), libgcr-ui-3-1 (>= 3.8.0), libgdk-pixbuf2.0-0 (>= 2.23.0), libglib2.0-0 (>= 2.53.2), libgnome-desktop-3-17 (>= 3.17.92), libgtk-3-0 (>= 3.21.5), libhandy-1-0 (>= 1.0.0), libnm0 (>= 1.0.0), libpam0g (>= 0.99.7.1), libpango-1.0-0 (>= 1.37.2), libpolkit-agent-1-0 (>= 0.99), libpolkit-gobject-1-0 (>= 0.94), libpulse-mainloop-glib0 (>= 0.99.1), libpulse0 (>= 0.99.1), libsecret-1-0 (>= 0.7), libupower-glib3 (>= 0.99.4-3~), libwayland-client0 (>= 1.9.91), fonts-lato, gsettings-desktop-schemas, phoc (>= 0.4.0)
Recommends: feedbackd, iio-sensor-proxy, gnome-session
Description: Pure Wayland shell for mobile devices
Phosh is a simple shell for Wayland compositors speaking the layer-surface
protocol. It currently supports
.
* a lockscreen
* brightness control and nighlight
* the gcr system-prompter interface
* acting as a polkit auth agent
* enough of org.gnome.Mutter.DisplayConfig to make gnome-settings-daemon happy
* a homebutton that toggles a simple favorites menu
* status icons for battery, wwan and wifi
.
If you're not working on a Wayland compositor then this package is likely not
very useful for you.
Homepage: https://source.puri.sm/Librem5/phosh
```
# What hardware are you running phosh on?
- [ ] amd64 qemu image
- [x] Librem5 devkit
- [ ] nested compositor. If so please give the full command you start phosh
with.
- [ ] other (please elaborte)https://source.puri.sm/Librem5/phosh/-/issues/376Settings panel should have same animation as the app grid for opening/closing2021-08-09T08:20:46ZJulian SparberSettings panel should have same animation as the app grid for opening/closingThe notification drawer doesn't have an animation when opening. I think we could use the same code as for the app grid. since https://source.puri.sm/Librem5/phosh/-/issues/155 is done.
Booth animation should also stick to the finger but...The notification drawer doesn't have an animation when opening. I think we could use the same code as for the app grid. since https://source.puri.sm/Librem5/phosh/-/issues/155 is done.
Booth animation should also stick to the finger but that depends on https://source.puri.sm/Librem5/phoc/-/issues/82
CC: @guido.guntherhttps://source.puri.sm/Librem5/phosh/-/issues/377Notifications don't disappear with an animation2021-08-09T08:20:46ZJulian SparberNotifications don't disappear with an animationThe notifications slide in from the top but then just disappear. I think they should slide back up.
cc: @tobias.bernardThe notifications slide in from the top but then just disappear. I think they should slide back up.
cc: @tobias.bernardhttps://source.puri.sm/Librem5/phosh/-/issues/378PhoshPanel(Top panel) and PhoshHome should use a common base class2021-08-09T08:20:46ZJulian SparberPhoshPanel(Top panel) and PhoshHome should use a common base classThe top panel and the home panel are similar but have quite some code differences.
I think they should use the same base class and use the same animation till we get gestures.
Also it would be better to have a more descriptive name for ...The top panel and the home panel are similar but have quite some code differences.
I think they should use the same base class and use the same animation till we get gestures.
Also it would be better to have a more descriptive name for the top panel, which would free up the `PhoshPanel` name for the base class.
CC: @zbrownhttps://source.puri.sm/Librem5/phosh/-/issues/380UI design: In horizontal (landscape) mode the top and bottom bar takes 20% of...2021-08-09T08:20:46ZdohacyUI design: In horizontal (landscape) mode the top and bottom bar takes 20% of the screen height# What problem did you encounter
There is not enough room for activities in the current landscape layout
## What is the current behaviour?
Together the top and bottom bar takes 20% of the available screen height.
My screen has a heigh...# What problem did you encounter
There is not enough room for activities in the current landscape layout
## What is the current behaviour?
Together the top and bottom bar takes 20% of the available screen height.
My screen has a height of 720 pixels, the bottom bar takes 80px, the top takes 65px.
![lackofroom](/uploads/fc22c861d3886f410639155f1bb57505/lackofroom.png)
## What is the expected behaviour?
More room for activities
## How to reproduce
Activate landscape mode.
# Discussion
Moving the bottom bar to the left or right as seen on android would allow to free 11% of screen spaces.
The information displayed on the bottom bar do not necessitate to be readable, in the opposite of the top bar.
Mockup:
![mockup](/uploads/f048de442db80b5ebda68b504dc3891e/mockup.png)
## Possible future issue
It's possible the animation bringing up the bottom drawer would seem less natural. Hard to say without testing first.
# Which version did you encounter the bug in?
I used the precompiled Debian package.
```
Package: phosh
Status: install ok installed
Priority: optional
Section: x11
Installed-Size: 1163
Maintainer: DebianOnMobile Maintainers <debian-on-mobile-maintainers@alioth-lists.debian.net>
Architecture: arm64
Version: 0.4.4-1mobian1
Provides: notification-daemon, polkit-1-auth-agent
Depends: dconf-gsettings-backend | gsettings-backend, libc6 (>= 2.17), libcairo2 (>= 1.2.4), libfeedback-0.0-0 (>= 0.0.0+git20200305), libgcr-base-3-1 (>= 3.8.0), libgcr-ui-3-1 (>= 3.8.0), libgdk-pixbuf2.0-0 (>= 2.23.0), libglib2.0-0 (>= 2.53.2), libgnome-desktop-3-19 (>= 3.17.92), libgtk-3-0 (>= 3.21.5), libhandy-1-0 (>= 1.0.0), libnm0 (>= 1.0.0), libpam0g (>= 0.99.7.1), libpango-1.0-0 (>= 1.37.2), libpolkit-agent-1-0 (>= 0.99), libpolkit-gobject-1-0 (>= 0.94), libpulse-mainloop-glib0 (>= 0.99.1), libpulse0 (>= 0.99.1), libsecret-1-0 (>= 0.7), libupower-glib3 (>= 0.99.4-3~), libwayland-client0 (>= 1.9.91), fonts-lato, gsettings-desktop-schemas, phoc (>= 0.4.0)
```
# What hardware are you running phosh on?
- Pinephone (mobian)https://source.puri.sm/Librem5/phosh/-/issues/384Crash when closing windows2021-08-09T08:20:47ZSebastian KrzyszkowiakCrash when closing windowsMost likely related to the switch to libhandy-1, possibly even a libhandy bug:
```
(gdb) bt
#0 0x0000ffffb925e624 in hdy_carousel_box_get_page_at_position (self=0xaaaae855bc20, position=0) at ../src/hdy-carousel-box.c:1745
#1 0x0000ff...Most likely related to the switch to libhandy-1, possibly even a libhandy bug:
```
(gdb) bt
#0 0x0000ffffb925e624 in hdy_carousel_box_get_page_at_position (self=0xaaaae855bc20, position=0) at ../src/hdy-carousel-box.c:1745
#1 0x0000ffffb925e6d0 in hdy_carousel_box_get_current_page_index (self=0xaaaae855bc20) at ../src/hdy-carousel-box.c:1765
#2 0x0000ffffb925a208 in animation_stopped_cb (self=0xaaaae855bab0, box=<optimized out>) at ../src/hdy-carousel.c:205
#3 0x0000ffffb94e6c5c in () at /lib/aarch64-linux-gnu/libgobject-2.0.so.0
#4 0x0000ffffb9503690 in g_signal_emit_valist () at /lib/aarch64-linux-gnu/libgobject-2.0.so.0
#5 0x0000ffffb9503b98 in g_signal_emit () at /lib/aarch64-linux-gnu/libgobject-2.0.so.0
#6 0x0000ffffb925e54c in animate_position (frame_clock=0xaaaae80795d0, self=0xaaaae855bc20) at ../src/hdy-carousel-box.c:327
#7 0x0000ffffb925e54c in animation_cb (widget=<optimized out>, frame_clock=0xaaaae80795d0, user_data=<optimized out>) at ../src/hdy-carousel-box.c:436
#8 0x0000ffffb9e280e0 in () at /lib/aarch64-linux-gnu/libgtk-3.so.0
#9 0x0000ffffb94e69f8 in g_closure_invoke () at /lib/aarch64-linux-gnu/libgobject-2.0.so.0
#10 0x0000ffffb94fb2b8 in () at /lib/aarch64-linux-gnu/libgobject-2.0.so.0
#11 0x0000ffffb950362c in g_signal_emit_valist () at /lib/aarch64-linux-gnu/libgobject-2.0.so.0
#12 0x0000ffffb9503b98 in g_signal_emit () at /lib/aarch64-linux-gnu/libgobject-2.0.so.0
#13 0x0000ffffb99be8a8 in () at /lib/aarch64-linux-gnu/libgdk-3.so.0
#14 0x0000ffffb99a79b0 in () at /lib/aarch64-linux-gnu/libgdk-3.so.0
#15 0x0000ffffb93f30e4 in () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#16 0x0000ffffb93f251c in g_main_context_dispatch () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#17 0x0000ffffb93f28e8 in () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#18 0x0000ffffb93f2c80 in g_main_loop_run () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#19 0x0000ffffb9cdaa60 in gtk_main () at /lib/aarch64-linux-gnu/libgtk-3.so.0
#20 0x0000aaaabb3d2a1c in main (argc=<optimized out>, argv=<optimized out>) at ../src/main.c:96
```https://source.puri.sm/Librem5/phosh/-/issues/388Make the lockscreen adaptive/work in landscape orientation2021-08-09T08:20:47ZArnaud FerrarisMake the lockscreen adaptive/work in landscape orientationCurrently the lockscreen is portrait-only, meaning if the user selects the landscape orientation, phosh changes the display orientation when locking and unlocking the screen.
This is counter-intuitive and introduces unnecessary complexi...Currently the lockscreen is portrait-only, meaning if the user selects the landscape orientation, phosh changes the display orientation when locking and unlocking the screen.
This is counter-intuitive and introduces unnecessary complexity in the lockscreen code (need to save current transform, switch to portrait, then re-apply transform when unlocking).
Having the lockscreen work in landscape orientation would require a landscape mode design so it can fit in 360px height (cc @tobias.bernard ), probably by putting blocks side-by-side. A few ideas:
- default view: have the clock/icons block on the left and media-player/notification area on the right
- unlock view: label, input field and "Unlock" button on the left side of the screen, keypad on the right
I'd be willing to work on that if/when we settle on a design.https://source.puri.sm/Librem5/phosh/-/issues/393Detect physical keyboard on tablets2021-08-09T08:20:47ZanonimnoDetect physical keyboard on tabletsHello,
I am running phosh version 0.4.5-1mobian1 on current nightly image of mobian on Pinetab.
My Pinetab came with the additional keyboard.
Two issues occur:
* Althogh the additional keyboard is attached, recognized and working, th...Hello,
I am running phosh version 0.4.5-1mobian1 on current nightly image of mobian on Pinetab.
My Pinetab came with the additional keyboard.
Two issues occur:
* Althogh the additional keyboard is attached, recognized and working, the onscreen keyboard is shown everytime that somewhere text can be entered.
Expected behaviour is, that when mechanical keyboard is attached that the onscreen keyboard is not shown/trigged. Attached you will find the output of journalctl with multiple attachment and disattachment of the keyboard.
* Minor additional issue is, that per default the backlight of the mechanical keyboard is on. It can be disabled with a key stroke, works as expected, but after the standby the backlight is on again.
If I can help debugging, I am happy to do it.
Regards
Anonimno
[keyboard_attach.log](/uploads/5d0fbbea1ba905e38913f8aa50a0790f/keyboard_attach.log)https://source.puri.sm/Librem5/phosh/-/issues/395fr_IT: wrong keyboard indicator used2021-08-09T08:20:47ZFabio Tomatfr_IT: wrong keyboard indicator usedswitching from Friulian keyboard to Italian keyboard the notification on the top left corner won't change, changing to other languages works (Friuli is part of Italy so is it related to a "fur_IT and it_IT" issue?)switching from Friulian keyboard to Italian keyboard the notification on the top left corner won't change, changing to other languages works (Friuli is part of Italy so is it related to a "fur_IT and it_IT" issue?)https://source.puri.sm/Librem5/phosh/-/issues/396In the app drawer, the 2nd screen is frozen2021-08-09T08:20:47ZDavid HamnerIn the app drawer, the 2nd screen is frozenOS: Byzantium
Desktop: Phosh
(Using 2 screens) if the main screen is in the app drawer, the 2nd screen is unresponsive.
Perhaps clicking on the 2nd screen should exit the app drawer without doing anything.OS: Byzantium
Desktop: Phosh
(Using 2 screens) if the main screen is in the app drawer, the 2nd screen is unresponsive.
Perhaps clicking on the 2nd screen should exit the app drawer without doing anything.https://source.puri.sm/Librem5/phosh/-/issues/397Initial lockscreen should unlock gnome keyring2021-08-09T08:20:47ZOliver SmithInitial lockscreen should unlock gnome keyringPhosh will display an initial lockscreen, if it was not [started from a display manager](https://source.puri.sm/Librem5/phosh/-/blob/d0be67133f36b09324de86637abaaf0d69189e4c/src/shell.c#L1054). In that case, it should also set up the PAM...Phosh will display an initial lockscreen, if it was not [started from a display manager](https://source.puri.sm/Librem5/phosh/-/blob/d0be67133f36b09324de86637abaaf0d69189e4c/src/shell.c#L1054). In that case, it should also set up the PAM session as @guido.gunther explained in #phosh. With PAM, the password typed in by the user would be used to unlock gnome-keyring.
> If you want phosh to show the initial lock screen you also expect it to handle e.g. unlocking gnome-keyring, ... (which we currenlty don't do but should) [...]
> phosh will never open gnome-keyring itself, it would just resort to pam [...]
Related: brainstorming about how PAM could be handled with/without full disk encryption and with displaying as few password prompts as possible: [pmaports#823 (comment)](https://gitlab.com/postmarketOS/pmaports/-/issues/823#note_433581347)https://source.puri.sm/Librem5/phosh/-/issues/400Hide logout button based on gnome-session's CanSwitchUser2021-08-09T08:20:47ZOliver SmithHide logout button based on gnome-session's CanSwitchUserAs requested by @guido.gunther, creating this issue, and a separate one for turning session into a GObject: #399
> ollieparanoid: would a patch to hide the logout button with a gconf option get accepted?
>
> ollieparanoid: context: in p...As requested by @guido.gunther, creating this issue, and a separate one for turning session into a GObject: #399
> ollieparanoid: would a patch to hide the logout button with a gconf option get accepted?
>
> ollieparanoid: context: in postmarketOS, the button is not useful in the default, single UI install. If users want to use pmOS with multiple UIs, we have a guide that explains how to adjust the install and they could enable the button then.
>
> Guido: there's no logout button - there's only a menu entry in a popover and that gets disabled automatically when the session does not support logout.
>
> Guido: a patch that hides that entry (instead of disabling) it when the session does not support logout is fine.
>
> craftyguy: hmm, so we could signal to phosh somehow that logout isn't supported?
>
> Guido: it currently allows logout when started from a DM https://source.puri.sm/Librem5/phosh/-/blob/master/src/panel.c#L343 - currently gnome-session does not have a CanLogout flag (only `CanShutdown) - it be cool to have that. Otherwise we need to come up with something else that let's the shell decide when to show logout.
>
> Guido: (gsetting worst case)
>
> Guido: but then we're surely looking towards user switching in phosh (although not short term)
>
> [...]
>
> ollieparanoid: gnome-session apparently has CanSwitchUser, this looks suitable
>
> ollieparanoid: so in theory... in phosh, add src/session.c:phosh_session_can_switch_user(), which does the dbus call to gnome session, and has a callback parameter, and call that from src/panel.c:phosh_panel_constructed() - the callback would then hide the logout button if the session can't swich the user?
>
> Guido: basially yes, it would be better to turn session into a GObject (it's from the early times) and then just listen to notify::can-switch-user) - similar to what we do other things (e.g. network connectivity, torch, ...), this is also needed for logout dialogs, ...https://source.puri.sm/Librem5/phosh/-/issues/405Sync up with systemds 'docked' property2021-08-09T08:20:47ZGuido GuntherSync up with systemds 'docked' propertylogind has a docked property, maybe we can move docking detection there in the long run.logind has a docked property, maybe we can move docking detection there in the long run.https://source.puri.sm/Librem5/phosh/-/issues/410should support search2021-08-09T08:20:47ZGuido Gunthershould support searchwe should support searching for documents, etc. via search providers (see !303, https://developer.gnome.org/SearchProvider/)we should support searching for documents, etc. via search providers (see !303, https://developer.gnome.org/SearchProvider/)https://source.puri.sm/Librem5/phosh/-/issues/412Errantly tap the home bar & opening the app grid while typing.2021-08-09T08:20:47ZC WoodErrantly tap the home bar & opening the app grid while typing.## What is the current behavior?
Errantly tapping the home bar and opening the app grid when using the on-screen keyboard.
Problem occurs when trying to use any key in the bottom row of the on-screen key
## What is the expected behav...## What is the current behavior?
Errantly tapping the home bar and opening the app grid when using the on-screen keyboard.
Problem occurs when trying to use any key in the bottom row of the on-screen key
## What is the expected behavior?
I expect either that the hit area for the home bar would be reduced or the key hit area on the on-screen keyboard to be increased when the keyboard is deployed.
## How to reproduce
Deploy the on-screen keyboard and use any key on the bottom row. It isn't persistent, so it's probably a miss guided tap, but reducing the hit area on the home bar may reduce the number of errant taps.
# Which version did you encounter the bug in?
Manjaro 5.9.1-6-MANJARO-ARM Beta 1
Phosh?
# What hardware are you running phosh on?
PinePhone v1.2https://source.puri.sm/Librem5/phosh/-/issues/414Need access to power menu from lock screen2021-08-09T08:20:48ZChance RosenthalNeed access to power menu from lock screen(Referred from Manjaro-ARM/PinePhone repo, crashes experienced on new PinePhone CE.)
I've had a couple of undiagnosed crashes that resulted in the phone becoming stuck at the lock screen, and failing to respond over ssh. You know, a pro...(Referred from Manjaro-ARM/PinePhone repo, crashes experienced on new PinePhone CE.)
I've had a couple of undiagnosed crashes that resulted in the phone becoming stuck at the lock screen, and failing to respond over ssh. You know, a proper crash =P
I'd like to be able to at least *try* to shut down gracefully, but it seems the power menu is only available from the tray, which isn't available when the phone is locked. Generally, I expect a longish press of the power button to bring up the power menu on a smartphone. That could be a solution, or it could just respond to a tap on the icons.
Hard-killing a system that's still sorta responding (in that it almost-kinda unlocks) is nerve-wracking.https://source.puri.sm/Librem5/phosh/-/issues/417Support output mirroring2021-08-09T08:20:48ZGuido GuntherSupport output mirroringg-c-c has a 'mirror mode' (mirroring displays).
Related #138 but let's have a separte issue.g-c-c has a 'mirror mode' (mirroring displays).
Related #138 but let's have a separte issue.https://source.puri.sm/Librem5/phosh/-/issues/418Implement org.freedesktop.impl.Access API2021-08-09T08:20:48ZGuido GuntherImplement org.freedesktop.impl.Access APIThis is used by flatpak to ask for permissions, see https://bugzilla.gnome.org/show_bug.cgi?id=768669
- Test script: https://bug768669.bugzilla-attachments.gnome.org/attachment.cgi?id=331605
- implementation in gtk desktop portal: https...This is used by flatpak to ask for permissions, see https://bugzilla.gnome.org/show_bug.cgi?id=768669
- Test script: https://bug768669.bugzilla-attachments.gnome.org/attachment.cgi?id=331605
- implementation in gtk desktop portal: https://github.com/flatpak/xdg-desktop-portal-gtk/blob/master/src/access.c can be tested via
```
TITLE="Give Firefox Microphone Access?"
SUBTITLE="Firefox wants to use your microphone."
BODY="Access to your microphone can be changed at any time from the privacy settings."
OPTIONS="{'icon': <'audio-input-microphone-symbolic'>,
'deny_label': <'Cancel'>,
'grant_label': <'Turn On'>}"
gdbus call --session --dest org.freedesktop.impl.portal.desktop.gtk \
--object-path /org/freedesktop/portal/desktop \
--method org.freedesktop.impl.portal.Access.AccessDialog \
/org/gnome/Shell/Access org.gnome.Shell '' \
"$TITLE" "$SUBTITLE" "$BODY" "$OPTIONS"
```Guido GuntherGuido Guntherhttps://source.puri.sm/Librem5/phosh/-/issues/421Phosh enables docked mode on its own2021-08-09T08:20:48ZSebastian KrzyszkowiakPhosh enables docked mode on its ownIt happened to me twice now that phosh has just out of the sudden decided to go into docked mode, even though I haven't plugged anything in, without giving me ability to override that choice.
![docked](/uploads/a15186961127f97a8fd625f5b...It happened to me twice now that phosh has just out of the sudden decided to go into docked mode, even though I haven't plugged anything in, without giving me ability to override that choice.
![docked](/uploads/a15186961127f97a8fd625f5ba9f50bb/docked.png)
Logs from the moment it happened don't seem very useful:
```
lis 10 05:59:09 dogwood org.gnome.Geary[773]: ![crt] 05:59:09.0569 Handy[NET|SER|REP|CNV|PER|SQL|NRM|DES]:update_decoration_layouts: assertion 'HDY_IS_HEADER_GROUP (self)' failed
lis 10 05:59:09 dogwood org.gnome.Geary[773]: ![crt] 05:59:09.0574 Handy[NET|SER|REP|CNV|PER|SQL|NRM|DES]:update_decoration_layouts: assertion 'HDY_IS_HEADER_GROUP (self)' failed
lis 10 05:59:09 dogwood phosh[825]: invalid unclassed pointer in cast to 'GtkWidget'
lis 10 05:59:09 dogwood phosh[825]: gtk_widget_grab_focus: assertion 'GTK_IS_WIDGET (widget)' failed
lis 10 05:59:09 dogwood feedbackd[854]: Failed to erase vibra effect with id 14: Invalid argument
```https://source.puri.sm/Librem5/phosh/-/issues/422phosh crashes with wlroots 0.12.0: error on 0 dimension without anchors2021-08-09T08:20:48ZOliver Smithphosh crashes with wlroots 0.12.0: error on 0 dimension without anchorsAfter upgrading from wlroots 0.11.0 to 0.12.0, phosh 0.5.1 will crash ~1 second after starting.
phoc version: 0.4.4
Log with `G_MESSAGES_DEBUG=phoc-wlroots`:
<details>
```
dbus-daemon[11709]: [session uid=10000 pid=11709] Successfully...After upgrading from wlroots 0.11.0 to 0.12.0, phosh 0.5.1 will crash ~1 second after starting.
phoc version: 0.4.4
Log with `G_MESSAGES_DEBUG=phoc-wlroots`:
<details>
```
dbus-daemon[11709]: [session uid=10000 pid=11709] Successfully activated service 'ca.desrt.dconf'
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.111: [types/wlr_surface.c:666] New wlr_surface 0xffff82ed3280 (res 0xffff830e7400)
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.112: [types/wlr_surface.c:666] New wlr_surface 0xffff82ed3660 (res 0xffff830e7f40)
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.112: [types/wlr_layer_shell_v1.c:458] new layer_surface 0xffff82f4fe30 (res 0xffff8390ee20)
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.112: [../src/layer_shell.c:747] new layer surface: namespace phosh layer 2 anchor 13 size 0x32 margin 0,0,0,0
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.256: [types/wlr_surface.c:666] New wlr_surface 0xffff82e5e400 (res 0xffff83107640) (phoc:11711): phoc-wlroots-DEBUG: 19:49:22.256: [types/wlr_surface.c:666] New wlr_surface 0xffff82e5e810 (res 0xffff83107eb0)
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.256: [types/wlr_surface.c:666] New wlr_surface 0xffff82e5ec00 (res 0xffff82f47400)
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.256: [types/wlr_surface.c:666] New wlr_surface 0xffff82e5efd0 (res 0xffff82f47880)
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.256: [types/wlr_layer_shell_v1.c:458] new layer_surface 0xffff82f4ffc0 (res 0xffff82f47d00)
phoc-wlroots-INFO: 19:49:22.257: [wayland] error in client communication (pid 11773)
(phoc:11711): phoc-wlroots-DEBUG: 19:49:22.257: [../src/cursor.c:696] entered surface 0, lx: 100.000000, ly: 100.000000, sx: 0.000000, sy: 0.000000
Gdk-Message: 19:49:22.282: Error 71 (Protocol error) dispatching to Wayland display.
```
</details>
I ran bisect and found this to be the breaking commit:
https://github.com/swaywm/wlroots/commit/8dec751a6d84335fb04288b8efab6dd5c90288d3Sebastian KrzyszkowiakSebastian Krzyszkowiakhttps://source.puri.sm/Librem5/phosh/-/issues/424Crash when switching all screens off with wlr-randr / GNOME Settings2021-08-09T08:20:48ZMarcoCrash when switching all screens off with wlr-randr / GNOME Settings# What problem did you encounter
Phosh crashs when switching screen off and on with wlr-randr.
## What is the current behaviour?
Phosh crashs.
## What is the expected behaviour?
Phosh keeps running as before.
## How to reproduce
1...# What problem did you encounter
Phosh crashs when switching screen off and on with wlr-randr.
## What is the current behaviour?
Phosh crashs.
## What is the expected behaviour?
Phosh keeps running as before.
## How to reproduce
1. `wlr-randr --output DSI-1 --off`
2. `wlr-randr --output DSI-1 --on`
when i run this on my PineTab i can see phosh crash
# Which version did you encounter the bug in?
- [ ] I Compiled it myself. If you compiled phosh from source please provide the
git revision via e.g. by running ``git log -1 --pretty=oneline`` and pasting
the output below.
- [x] I used the precompiled Debian package (e.g. by running a prebuilt
image). Please paste the output oof ``dpkg -s phosh`` below.
```
Phosh Version: 0.5.1-1
```
# What hardware are you running phosh on?
- [ ] amd64 qemu image
- [ ] Librem5 devkit
- [ ] nested compositor. If so please give the full command you start phosh
with.
- [x] other: Pine64 PineTab
# Releveant logfiles
```
Nov 13 22:26:58 pinetab sm.puri.Phosh.desktop[3242]: **
Nov 13 22:26:58 pinetab sm.puri.Phosh.desktop[3242]: phosh-shell:ERROR:../phosh-v0.5.1/src/shell.c:571:on_monitor_removed: assertion failed: (priv->primary_monitor && priv->primary_monitor != monitor)
Nov 13 22:26:58 pinetab sm.puri.Phosh.desktop[3242]: Bail out! phosh-shell:ERROR:../phosh-v0.5.1/src/shell.c:571:on_monitor_removed: assertion failed: (priv->primary_monitor && priv->primary_monitor != monitor)
Nov 13 22:26:58 pinetab phoc[3101]: [backend/drm/atomic.c:56] DSI-1: Atomic commit failed (pageflip): Invalid argument
Nov 13 22:26:58 pinetab systemd[1]: Created slice system-systemd\x2dcoredump.slice.
Nov 13 22:26:58 pinetab systemd[1]: Started Process Core Dump (PID 3779/UID 0).
Nov 13 22:26:58 pinetab systemd[1]: Started Process Core Dump (PID 3781/UID 0).
Nov 13 22:27:02 pinetab polkitd[3165]: Unregistered Authentication Agent for unix-session:c1 (system bus name :1.37, object path /org/freedesktop/PolicyKit1/AuthenticationAgent, locale de_DE.UTF-8) (disconnected from bus)
Nov 13 22:27:02 pinetab gnome-session[3219]: gnome-session-binary[3219]: WARNING: Could not get session path for session. Check that logind is properly installed and pam_systemd is getting used at login.
Nov 13 22:27:02 pinetab gnome-session-binary[3219]: WARNING: Could not get session path for session. Check that logind is properly installed and pam_systemd is getting used at login.
Nov 13 22:27:02 pinetab gnome-keyring-daemon[3233]: couldn't prompt for password: The operation was cancelled
Nov 13 22:27:02 pinetab gnome-keyring-d[3233]: couldn't prompt for password: The operation was cancelled
Nov 13 22:27:02 pinetab phoc[3101]: [backend/drm/atomic.c:56] DSI-1: Atomic commit failed (pageflip): Invalid argument
Nov 13 22:27:02 pinetab gnome-session[3219]: gnome-session-binary[3219]: WARNING: Application 'sm.puri.Phosh.desktop' killed by signal 6
Nov 13 22:27:02 pinetab gnome-session-binary[3219]: WARNING: Application 'sm.puri.Phosh.desktop' killed by signal 6
Nov 13 22:27:04 pinetab systemd-coredump[3782]: Process 3503 (calls) of user 1000 dumped core.
Stack trace of thread 3503:
#0 0x0000ffff9d0c66c8 raise (libc.so.6 + 0x376c8)
#1 0x0000ffff9d0b31cc abort (libc.so.6 + 0x241cc)
#2 0x0000ffff9d23eae4 n/a (libwayland-client.so.0 + 0xbae4)
#3 0x0000ffff9d23c934 n/a (libwayland-client.so.0 + 0x9934)
#4 0x0000ffff9d238b98 n/a (libwayland-client.so.0 + 0x5b98)
#5 0x0000ffff9d23a400 wl_display_dispatch_queue_pending (libwayland-client.so.0 + 0x7400)
#6 0x0000ffff9d86174c n/a (libgdk-3.so.0 + 0xb074c)
#7 0x0000ffff9d7ea4e8 gdk_display_get_event (libgdk-3.so.0 + 0x394e8)
#8 0x0000ffff9d86131c n/a (libgdk-3.so.0 + 0xb031c)
#9 0x0000ffff9e1efad0 g_main_context_dispatch (libglib-2.0.so.0 + 0x61ad0)
#10 0x0000ffff9e1f00b0 n/a (libglib-2.0.so.0 + 0x620b0)
#11 0x0000ffff9e1f01b4 g_main_context_iteration (libglib-2.0.so.0 + 0x621b4)
#12 0x0000ffff9d64e910 g_application_run (libgio-2.0.so.0 + 0x110910)
#13 0x0000aaaab8491de8 main (calls + 0xbde8)
#14 0x0000ffff9d0b3538 __libc_start_main (libc.so.6 + 0x24538)
#15 0x0000aaaab8491e48 _start (calls + 0xbe48)
#16 0x0000aaaab8491e48 _start (calls + 0xbe48)
Nov 13 22:27:04 pinetab systemd[1]: systemd-coredump@0-3779-0.service: Succeeded.
Nov 13 22:27:05 pinetab gnome-session-binary[3219]: WARNING: Application 'sm.puri.Phosh.desktop' killed by signal 5
Nov 13 22:27:05 pinetab gnome-session[3219]: gnome-session-binary[3219]: WARNING: Application 'sm.puri.Phosh.desktop' killed by signal 5
Nov 13 22:27:06 pinetab phoc[3101]: [backend/drm/drm.c:971] Skipping pageflip on output 'DSI-1'
Nov 13 22:27:06 pinetab systemd-coredump[3843]: Process 3835 (phosh) of user 1000 dumped core.
Stack trace of thread 3835:
#0 0x0000ffff866d39d8 raise (libpthread.so.0 + 0x139d8)
#1 0x0000aaaad4a343c8 n/a (phosh + 0x8a3c8)
#2 0x0000ffff86c66b80 g_log_structured_array (libglib-2.0.so.0 + 0x69b80)
#3 0x0000ffff86c67788 g_log_structured_standard (libglib-2.0.so.0 + 0x6a788)
#4 0x0000aaaad4a3dbd4 n/a (phosh + 0x93bd4)
#5 0x0000ffff86d7e26c n/a (libgobject-2.0.so.0 + 0x1d26c)
#6 0x0000ffff86d7fc64 g_object_new_with_properties (libgobject-2.0.so.0 + 0x1ec64)
#7 0x0000ffff86d806a0 g_object_new (libgobject-2.0.so.0 + 0x1f6a0)
#8 0x0000aaaad4a3d070 n/a (phosh + 0x93070)
#9 0x0000aaaad49c75e0 main (phosh + 0x1d5e0)
#10 0x0000ffff86570538 __libc_start_main (libc.so.6 + 0x24538)
#11 0x0000aaaad49c776c n/a (phosh + 0x1d76c)
#12 0x0000aaaad49c776c n/a (phosh + 0x1d76c)
Nov 13 22:27:06 pinetab systemd[1]: systemd-coredump@2-3842-0.service: Succeeded.
Nov 13 22:27:08 pinetab systemd-coredump[3785]: Process 3242 (phosh) of user 1000 dumped core.
Stack trace of thread 3242:
#0 0x0000ffff9755e6c8 raise (libc.so.6 + 0x376c8)
#1 0x0000ffff9754b1cc abort (libc.so.6 + 0x241cc)
#2 0x0000ffff97c684fc g_assertion_message (libglib-2.0.so.0 + 0x904fc)
#3 0x0000ffff97c68570 g_assertion_message_expr (libglib-2.0.so.0 + 0x90570)
#4 0x0000aaaaacacede4 n/a (phosh + 0x6cde4)
#5 0x0000ffff97d52760 g_closure_invoke (libgobject-2.0.so.0 + 0x16760)
#6 0x0000ffff97d67cb4 n/a (libgobject-2.0.so.0 + 0x2bcb4)
#7 0x0000ffff97d6e6dc g_signal_emit_valist (libgobject-2.0.so.0 + 0x326dc)
#8 0x0000ffff97d6ed9c g_signal_emit (libgobject-2.0.so.0 + 0x32d9c)
#9 0x0000aaaaacac2214 n/a (phosh + 0x60214)
#10 0x0000ffff97d52760 g_closure_invoke (libgobject-2.0.so.0 + 0x16760)
#11 0x0000ffff97d67cb4 n/a (libgobject-2.0.so.0 + 0x2bcb4)
#12 0x0000ffff97d6e6dc g_signal_emit_valist (libgobject-2.0.so.0 + 0x326dc)
#13 0x0000ffff97d6ed9c g_signal_emit (libgobject-2.0.so.0 + 0x32d9c)
#14 0x0000ffff97d57ec0 n/a (libgobject-2.0.so.0 + 0x1bec0)
#15 0x0000ffff97d5a8c4 g_object_notify_by_pspec (libgobject-2.0.so.0 + 0x1e8c4)
#16 0x0000ffff96918178 n/a (libffi.so.7 + 0x6178)
#17 0x0000ffff96917868 n/a (libffi.so.7 + 0x5868)
#18 0x0000ffff977b9880 n/a (libwayland-client.so.0 + 0x9880)
#19 0x0000ffff977b5b98 n/a (libwayland-client.so.0 + 0x5b98)
#20 0x0000ffff977b7400 wl_display_dispatch_queue_pending (libwayland-client.so.0 + 0x7400)
#21 0x0000ffff9826474c n/a (libgdk-3.so.0 + 0xb074c)
#22 0x0000ffff981ed4e8 gdk_display_get_event (libgdk-3.so.0 + 0x394e8)
#23 0x0000ffff9826431c n/a (libgdk-3.so.0 + 0xb031c)
#24 0x0000ffff97c39ad0 g_main_context_dispatch (libglib-2.0.so.0 + 0x61ad0)
#25 0x0000ffff97c3a0b0 n/a (libglib-2.0.so.0 + 0x620b0)
#26 0x0000ffff97c3a4c4 g_main_loop_run (libglib-2.0.so.0 + 0x624c4)
#27 0x0000ffff985a964c gtk_main (libgtk-3.so.0 + 0x2c864c)
#28 0x0000aaaaaca7f610 main (phosh + 0x1d610)
#29 0x0000ffff9754b538 __libc_start_main (libc.so.6 + 0x24538)
#30 0x0000aaaaaca7f76c n/a (phosh + 0x1d76c)
#31 0x0000aaaaaca7f76c n/a (phosh + 0x1d76c)
Nov 13 22:27:08 pinetab systemd[1]: systemd-coredump@1-3781-0.service: Succeeded.
```https://source.puri.sm/Librem5/phosh/-/issues/431Scramble lockscreen layout2021-08-09T08:20:48ZVladScramble lockscreen layoutIt's a feature implemented in LineageOS ([1](https://review.lineageos.org/c/LineageOS/android_packages_apps_Settings/+/225678),[2](https://review.lineageos.org/c/LineageOS/android_frameworks_base/+/225679/27))
The main idea is to increas...It's a feature implemented in LineageOS ([1](https://review.lineageos.org/c/LineageOS/android_packages_apps_Settings/+/225678),[2](https://review.lineageos.org/c/LineageOS/android_frameworks_base/+/225679/27))
The main idea is to increase security by shuffling lockscreen keyboard layout.
(go from 1-9 0 to random every time).
It works because it decreases chance of a replay attack.
Has some effect against both brute forcing fingerprints left on the screen and looking over the sholder.https://source.puri.sm/Librem5/phosh/-/issues/434Phosh flick app switching doesn't always switch (threshold too hard)2021-08-09T08:20:48ZTodd WeaverPhosh flick app switching doesn't always switch (threshold too hard)# What problem did you encounter
Phosh app switching doesn't always get to the new app
## What is the current behaviour?
harder flick required
## What is the expected behaviour?
flick motion scrolls through the app drawers easier
#...# What problem did you encounter
Phosh app switching doesn't always get to the new app
## What is the current behaviour?
harder flick required
## What is the expected behaviour?
flick motion scrolls through the app drawers easier
## How to reproduce
![purism-librem-5-evergreen-phosh-app-drawer-2020-11-23](/uploads/afb25fa2be83194eb7896837a7f70749/purism-librem-5-evergreen-phosh-app-drawer-2020-11-23.webm)
# Which version did you encounter the bug in?
```
purism@pureos:~$ uname -a
Linux pureos 5.9.0-1-librem5 #1 SMP PREEMPT Fri Nov 13 03:06:12 PST 2020 aarch6x
purism@pureos:~$ dpkg -s phosh | grep Version
Version: 0.6.0
purism@pureos:~$
```
# What hardware are you running phosh on?
- Librem 5 Evergreenhttps://source.puri.sm/Librem5/phosh/-/issues/441App switcher removes focus ability from second screen2021-08-09T08:20:48ZTodd WeaverApp switcher removes focus ability from second screen# What problem did you encounter
Focus ability is lost when app switches is opened on the phone
## What is the expected behaviour?
Not sure, probably that apps can still gain focus on the second screen.
## How to reproduce
1. Attach...# What problem did you encounter
Focus ability is lost when app switches is opened on the phone
## What is the expected behaviour?
Not sure, probably that apps can still gain focus on the second screen.
## How to reproduce
1. Attach second screen
2. Launch app into second screen (drag it there)
3. Open App Switcher on Phone
4. Try to move second screen app or have it focus.
![purism-librem-5-pureos-phosh-second-screen-focus-2020-11-28](/uploads/7ac3e5cdd379a61bdfcf89a35ec8139d/purism-librem-5-pureos-phosh-second-screen-focus-2020-11-28.webm)
# Which version did you encounter the bug in?
```
purism@pureos:~$ uname -a
Linux pureos 5.9.0-1-librem5 #1 SMP PREEMPT Fri Nov 27 19:15:53 UTC 2020 aarch64 GNU/Linux
purism@pureos:~$ dpkg -s phosh | grep Version
Version: 0.6.0
purism@pureos:~$ dpkg -s phoc | grep Version
Version: 0.5.1
purism@pureos:~$
```
# What hardware are you running phosh on?
- Librem 5 Evergreenhttps://source.puri.sm/Librem5/phosh/-/issues/442Unable to clear terminal notification (unable to clear any notification unles...2021-08-09T08:20:49ZTodd WeaverUnable to clear terminal notification (unable to clear any notification unless you 'activate' it)# What problem did you encounter
Cannot clear a notification
## What is the current behaviour?
Unable to clear a notification unless you activate the notification button
## What is the expected behaviour?
Swipe clear notifications p...# What problem did you encounter
Cannot clear a notification
## What is the current behaviour?
Unable to clear a notification unless you activate the notification button
## What is the expected behaviour?
Swipe clear notifications probably.
## How to reproduce
1. Run a terminal command
2. Let it complete
3. Get the notification
4. Cannot clear it (video below is docked)
![purism-librem-5-pureos-phosh-notification-not-clearing-2020-11-28](/uploads/e9bea72d2932b433b9a9b7045c4e9754/purism-librem-5-pureos-phosh-notification-not-clearing-2020-11-28.webm)
# Which version did you encounter the bug in?
```
purism@pureos:~$ uname -a
Linux pureos 5.9.0-1-librem5 #1 SMP PREEMPT Fri Nov 27 19:15:53 UTC 2020 aarch64 GNU/Linux
purism@pureos:~$ dpkg -s phosh | grep Version
Version: 0.6.0
purism@pureos:~$ dpkg -s phoc | grep Version
Version: 0.5.1
purism@pureos:~$```
# What hardware are you running phosh on?
- Librem 5 Evergreenhttps://source.puri.sm/Librem5/phosh/-/issues/445Puts lockscreen on both outputs on NextDock2021-08-09T08:20:49ZGuido GuntherPuts lockscreen on both outputs on NextDockWhen moving the primary display to the external screen on the NextDock and then hitting the phones power button we get a lockscreen on both outputs while it should be only on the external one. Unlocking will then just unlock the internal...When moving the primary display to the external screen on the NextDock and then hitting the phones power button we get a lockscreen on both outputs while it should be only on the external one. Unlocking will then just unlock the internal screen - funny enough it does not happen with the external screen i'm using here. I'll have a look.https://source.puri.sm/Librem5/phosh/-/issues/447background: Load background async2021-08-09T08:20:49ZGuido Guntherbackground: Load background asyncWhile pixbuf creation is async the actual `g_file_read()` in `load_background` load isn't. This should speed up startup.
Noticed while looking at !673 While pixbuf creation is async the actual `g_file_read()` in `load_background` load isn't. This should speed up startup.
Noticed while looking at !673 https://source.puri.sm/Librem5/phosh/-/issues/448Mentions docs/ from README2021-08-09T08:20:49ZGuido GuntherMentions docs/ from READMEwe don't mention there's additional gtk-doc in docs/ incuding the markdown files which are readable as is.
Also mention where the docs can be browsed online.we don't mention there's additional gtk-doc in docs/ incuding the markdown files which are readable as is.
Also mention where the docs can be browsed online.