sway: 1.6.1 -> 1.7 #151902

primeos · 2021-12-23T19:22:37Z

Motivation for this change

TODOs:

The terminal emulator in the default config file has been changed to foot.
Fix the VM test
Testing
Squash the commits

Known issues:

Chromium/Electron: Ozone/Wayland is broken

When still running Sway 1.6.1 and swaynag is from 1.7 it will crash with:

[michael@quorra:~]$ swaynag -m test
wl_registry@2: error 0: invalid version for global wl_output (39): have 3, wanted 4
00:00:00.000 [swaynag/swaynag.c:438] Error during outputs init.

Dropping Alacritty from the NixOS module can require configuration changes

Things done

Synthetica9 · 2021-12-24T09:30:07Z

I'm this close 🤏🏻 to suggesting we just add a `sway_default_terminal` symlink and patch the config to use that instead of keeping all old terminals around for ~5 releases

…

On Thu, 23 Dec 2021, 20:38 ofborg[bot], ***@***.***> wrote: @ofborg <https://github.com/ofborg>[bot] requested your review on: #151902 <#151902> sway: 1.6.1 -> 1.7. — Reply to this email directly, view it on GitHub <#151902 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABV7PJ25OTSPSDZWFYOUUZTUSN3BXANCNFSM5KVLSJNQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because your review was requested.Message ID: ***@***.***>

Synthetica9 · 2021-12-28T09:09:29Z

Test also broke, see here:

Synthetica9 · 2021-12-28T15:25:18Z

vscod{e,ium} with ozone platform wayland also seem broken, but I assume that's a wlroots thing rather than a Sway thing

primeos · 2021-12-28T16:25:31Z

@Synthetica9 thanks for the testing and feedback! :) Unfortunately I'm currently quite behind with testing as I wanted for Hydra to build more of staging-next but it's progressing slowly / will take a while (I might switch to cherry-picking for now).

I'm this close 🤏🏻 to suggesting we just add a sway_default_terminal
symlink and patch the config to use that instead of keeping all old
terminals around for ~5 releases

That would also be an option. But then again I'm hoping that this'll be the last change of the default terminal emulator for a while. Last time we dropped rxvt-unicode quite late but already mentioned that foot will become the new default: eb8a694

Given that sway_default_terminal won't help with copied/adapted configurations I'd suggest we simply make the change right away this time. It isn't ideal but then again every Sway user should have no issue figuring the problem out (it'll likely be a bit annoying and especially unexpected that $mod+Enter will likely be broken but usually everyone should use some sort of dmenu program to quickly investigate without even having to leave the Sway session; it's quite convenient that most Sway users should be power-users :D).

Test also broke, see here:

I've added it to the TODOs, thanks. The full (debug) output should hopefully contain enough information to find a fix/workaround.

vscod{e,ium} with ozone platform wayland also seem broken, but I assume that's a wlroots thing rather than a Sway thing

Oh, that doesn't sound good... :o Hopefully it doesn't affect the current Chromium as well. Unfortunately Ozone/Wayland is still far from stable (I forgot the official status but IIRC Igalia still considers it alpha or beta).

Edit: VM test output

machine # [   13.256664] systemd[874]: Started D-Bus User Message Bus.
machine # [   15.563809] systemd[1]: Created slice Slice /system/systemd-coredump.
machine # [   15.572926] systemd[1]: Started Process Core Dump (PID 979/UID 0).
machine # [   16.034810] systemd-coredump[980]: Process 926 (sway) of user 1000 dumped core.
machine #
[...]
machine # Found module sway without build-id.
machine # Stack trace of thread 926:
machine # #0  0x00007f0f67b8bbaa raise (libc.so.6 + 0x3bbaa)
machine # #1  0x000000000045d1ba _sway_assert (sway + 0x5d1ba)
machine # #2  0x000000000045787d view_map (sway + 0x5787d)
machine # #3  0x000000000045b361 handle_map (sway + 0x5b361)
machine # #4  0x00007f0f67e255ec wlr_signal_emit_safe (libwlroots.so.10 + 0x8b5ec)
machine # #5  0x00007f0f67e2baa4 xwayland_surface_role_commit (libwlroots.so.10 + 0x91aa4)
machine # #6  0x00007f0f67e1b676 surface_commit_state (libwlroots.so.10 + 0x81676)
machine # #7  0x00007f0f6747286a ffi_call_unix64 (libffi.so.8 + 0x786a)
machine # #8  0x00007f0f674719c2 ffi_call_int (libffi.so.8 + 0x69c2)
machine # #9  0x00007f0f67e826a2 wl_closure_invoke (libwayland-server.so.0 + 0xd6a2)
machine # #10 0x00007f0f67e7db92 wl_client_connection_data (libwayland-server.so.0 + 0x8b92)
machine # #11 0x00007f0f67e80632 wl_event_loop_dispatch (libwayland-server.so.0 + 0xb632)
machine # #12 0x00007f0f67e7e2b5 wl_display_run (libwayland-server.so.0 + 0x92b5)
machine # #13 0x0000000000410cc5 main (sway + 0x10cc5)
machine # #14 0x00007f0f67b77780 __libc_start_main (libc.so.6 + 0x27780)
machine # #15 0x0000000000410e8a _start (sway + 0x10e8a)
machine #
machine # Stack trace of thread 928:
machine # #0  0x00007f0f67b42cf5 __futex_abstimed_wait_common64 (libpthread.so.0 + 0x14cf5)
machine # #1  0x00007f0f67b3cc22 pthread_cond_wait@@GLIBC_2.3.2 (libpthread.so.0 + 0xec22)
machine # #2  0x00007f0f651206bb util_queue_thread_func (kms_swrast_dri.so + 0x1cf6bb)
machine # #3  0x00007f0f65120317 impl_thrd_routine (kms_swrast_dri.so + 0x1cf317)
machine # #4  0x00007f0f67b36d40 start_thread (libpthread.so.0 + 0x8d40)
machine # #5  0x00007f0f67c4a43f __clone (libc.so.6 + 0xfa43f)
machine #
machine # [   16.233414] systemd[1]: [email protected]: Deactivated successfully.

Sway (normal output):

libEGL warning: DRI2: failed to create dri screen
00:00:00.079 [ERROR] [wlr] [EGL] command: eglInitialize, error: EGL_NOT_INITIALIZED (0x3001), message: "DRI2: failed to create screen"
libEGL warning: NEEDS EXTENSION: falling back to kms_swrast
00:00:00.901 [ERROR] [wlr] [render/allocator/gbm.c:114] gbm_bo_create failed
00:00:00.901 [ERROR] [wlr] [render/swapchain.c:109] Failed to allocate buffer
[...] // Repeats many times
00:00:00.904 [ERROR] [sway/config/output.c:501] Failed to commit output Virtual-1
00:00:00.006 [INFO] [swaybar/tray/icon.c:348] Warning: no icon themes loaded
libEGL warning: DRI2: failed to create dri screen
libEGL warning: NEEDS EXTENSION: falling back to kms_swrast
glamor: No eglstream capable devices found
Refusing to try glamor on llvmpipe
EGL setup failed, disabling glamor
Failed to initialize glamor, falling back to sw
00:00:05.633 [ERROR] [sway/tree/view.c:590] select_workspace:Expected to find a workspace
(EE) failed to read Wayland events: Broken pipe
X connection to :0 broken (explicit kill or server shutdown).

Synthetica9 · 2021-12-29T17:41:53Z

Chromium also broke, definitely a blocker then:

$ chromium --verbose --enable-features=UseOzonePlatform --ozone-platf
orm=wayland
interface 'wl_output' has no event 4
[1229/184110.716325:ERROR:process_memory_range.cc(75)] read out of range
fish: Job 1, 'chromium --verbose --enable-fea…' terminated by signal SIGTRAP (Trace or breakpoint trap)
$ rwhich chromium
/nix/store/ds5pnbnjhdb96azyisgmkikhw0g0c65q-chromium-96.0.4664.110/bin/chromium

Currently looking if I can get the master of another wlroots compositor to work to confirm this is indeed a wlroots bug.

tadeokondrak · 2021-12-29T17:53:38Z

Chromium also broke, definitely a blocker then:

This is a Chromium issue: https://bugs.chromium.org/p/chromium/issues/detail?id=1279574 (patch)

Synthetica9 · 2021-12-29T17:57:58Z

This is a Chromium issue: bugs.chromium.org/p/chromium/issues/detail?id=1279574 (patch)

Yikes, this might mean Electron apps may stay broken for a while (especially since we don't build most of those)...

primeos · 2022-01-02T12:49:43Z

I've "fixed" the VM test by switching to the new Pixman renderer. It's more of a workaround than a fix but we already had to use software rendering before and this should work more reliably for the VM test.

I currently lack time for a more thorough analysis but from a quick test wlroot's TinyWL doesn't seem affected.

Yikes, this might mean Electron apps may stay broken for a while (especially since we don't build most of those)...

That is indeed really unfortunate... :o
I haven't seen a backport of the fixes yet so they might only land in Chromium 99 (chromium/chromium@dd4c3dd and chromium/chromium@a84b79d). Chromium 99 is scheduled to be released on 2022-03-01 so we should backport the fixes. For Electron apps it's even worse. The next version (Electron 17) should use Chromium 98 so we'd have to wait for Electron 18 (which should be based on Chromium 100 - at least that Chromium release will be only a month later).

primeos · 2022-01-02T12:50:04Z

@ofborg test sway

2

Synthetica9 · 2022-01-02T21:00:27Z

Updated the test to use foot.

That is indeed really unfortunate... :o
[...]

Would also be possible to fix this from wlroots perhaps? Honestly, I wouldn't feel comfortable upgrading (the default version of) Sway to 1.7 if something I think quite a lot of users use would knowingly break without a good fix... I know I would be staying on Sway 1.6 (for now) if the full version of 1.7 were to release today

jackinloadup · 2022-01-04T06:53:55Z

Sounds perfect for unstable. I'd really like to use it personally due to some of the fixes. It doesn't sound like I'll be affected by any of the unfortunate chromium issues. Let me see if I can figure out how to apply patches 😉 Gotta get DRM leasing for VR headsets. 🤞

Synthetica9 · 2022-01-04T15:30:27Z

Sounds perfect for unstable.

While nixos unstable is a rolling-release distro, in my experience it is also one of the most stable rolling-release distros out there. I'd like to avoid knowingly breaking things if I can avoid it.

It doesn't sound like I'll be affected by any of the unfortunate chromium issues. [...] Gotta get DRM leasing for VR headsets.

Haha yeah I guess I'm biased towards the chromium issue because I use a lot of electron-based apps and have a mixed-scaling setup, which isn't great with XWayland so I use as many wayland-native apps as I can get away with :P

Let me see if I can figure out how to apply patches 😉

The problem is that we don't build a lot of the electron apps ourselves, but grab pre-built binaries for them, so it would be difficult to patch those.

primeos · 2022-01-04T19:07:51Z

@Synthetica9 thanks for 0860d0d but can you please drop 5ec0950? It just makes working on this PR unnecessarily annoying (I intend to squash all commits at the end btw - assuming it doesn't really help to keep some changes separate - therefore this PR should also only contain the relevant changes for the update) and doesn't belong in this PR ("destroys" the diff; feel free to reformat it after that).

I'd like to avoid knowingly breaking things if I can avoid it.

+1, the name nixos-unstable is misleading, unfortunately. However, in this case I'm not sure yet how to proceed. One possibility would be to add sway_1_6 and sway_1_7 and start with sway = sway_1_6 if necessary. However, the NixOS module intentionally doesn't offer a package option so this is still not that convenient (requiring e.g. an overlay).

In any case the goal should be to (eventually) update Sway to version 1.7 by default.

The problem is that we don't build a lot of the electron apps ourselves, but grab pre-built binaries for them, so it would be difficult to patch those.

Agreed, the current status sucks for Electron users... :o

Would also be possible to fix this from wlroots perhaps?

I don't think they could apply a good workaround (without downgrading the supported wl_outputs version for all applications). I haven't looked at the details though. But given that they're aware of it and the wlroots bug report got closed (https://gitlab.freedesktop.org/wlroots/wlroots/-/issues/3344#note_1189198) I don't think there's any interest in a workaround (understandably).

I think the better option would be to try to get the Chromium patches backported to older releases and Electron (ideally that would already be happening as this should affect other Wayland compositors soon as well, I guess).

Synthetica9 · 2022-01-04T19:54:14Z

@Synthetica9 thanks for 0860d0d but can you please drop 5ec0950?

✅

One possibility would be to add sway_1_6 and sway_1_7 and start with sway = sway_1_6 if necessary.

Might honestly be the best option in this case, re-evaluate once all or most electron packages have the relevant patch.

However, the NixOS module intentionally doesn't offer a package option so this is still not that convenient (requiring e.g. an overlay).

(emphasis mine)

What's the reasoning behind this?

I don't think there's any interest in a workaround (understandably).

Yeah, I fully understand the sentiment of the wlroots people to not want to do such a workaround, but but it leaves us in a difficult situation downstream.

primeos

What's the reasoning behind this? (package option)

#50476 (it wasn't necessary so far) but it seems like I forgot about the last update: #50476 (comment)

So I guess it could make sense to add such an option now (too choose between Sway 1.6 and 1.7 - apart from that there's probably little use for it as overlays can already be used as an alternative and AFAIK there are no relevant Sway alternatives/forks).

Yeah, I fully understand the sentiment of the wlroots people to not want to do such a workaround, but but it leaves us in a difficult situation downstream.

Yes, unfortunately. Regarding Chromium: I tested the fixes with M98 (#153633) and it worked but I got a few other errors that broke hardware acceleration (but I need to check if that's a regression or not as I haven't used Chromium on that system for quite a while...).

nixos/tests/sway.nix

LunNova · 2022-01-14T02:35:37Z

Might be worth trying with WLR_RENDERER=vulkan VK_LAYER_PATH = "${pkgs.vulkan-validation-layers}/share/vulkan/explicit_layer.d" instead of pixman?

Synthetica9 · 2022-01-14T14:43:57Z

Might be worth trying with WLR_RENDERER=vulkan VK_LAYER_PATH = "${pkgs.vulkan-validation-layers}/share/vulkan/explicit_layer.d" instead of pixman?

Crashes with a similar error:

<<< Welcome to NixOS 22.05pre-git (x86_64) - tty2 >>>


machine login: alice (automatic login)


[alice@machine:~]$ sway
00:00:00.126 [wlr] [render/vulkan/renderer.c:1487] Could not match drm and vulkan device
00:00:00.126 [sway/server.c:79] Failed to create renderer

[alice@machine:~]$

primeos · 2022-01-15T13:47:20Z

I think Pixman should be fine for now. We needed WLR_RENDERER_ALLOW_SOFTWARE before so that wasn't ideal either.

For other renderers we should look into getting proper hardware acceleration to work:

nixpkgs/nixos/tests/sway.nix

Lines 20 to 23 in 02a162f

    
           # TODO: Investigate if we can get hardware acceleration to work (via 
        
           # virtio-gpu and Virgil). We currently have to use the Pixman software 
        
           # renderer since the GLES2 renderer doesn't work inside the VM (even 
        
           # with WLR_RENDERER_ALLOW_SOFTWARE):

I wouldn't consider it a high-priority TODO but feel free to have a look at it if someone has time and motivation.

IIRC it might at least require changes to Mesa as well since we currently don't even build VirGL by default (I enabled it once but it still wasn't enough - maybe that's changed though). For Vulkan there's also Venus now (https://docs.mesa3d.org/drivers/venus.html) but it requires Vulkan support from the host driver (so not appropriate for our VM tests; plus there are a lot of other requirements that we cannot satisfy).
Another problem might've been that 3D acceleration with QEMU requires the -sdl (-display sdl,gl=on) flag while our VM tests run headless. Anyway, not sure how the current situation looks but I definitely had 3D hardware-acceleration working with QEMU/KVM on NixOS (but that wasn't a headless setup).

In the end hardware acceleration might not be possible on a headless setup yet.

FlorianFranzen · 2022-01-16T13:19:34Z

Updating to a newer wlroots' version will also mean we will need this patch in all chromium based applications. Otherwise, due to the new protocol version, all these apps will crash with interface 'wl_output' has no event 4.

The patch was included in v99.0.4766, so it already hit electron master (v18 nightly, with request for backport) and chromium dev (#154668).

primeos · 2022-01-16T14:34:11Z

Yes, we already noticed and discussed it here (s. above). Chromium in Nixpkgs is also patched since today (#155203).

+1, the name nixos-unstable is misleading, unfortunately. However, in this case I'm not sure yet how to proceed. One possibility would be to add sway_1_6 and sway_1_7 and start with sway = sway_1_6 if necessary.

In the meantime I thought a bit more about this. I'm now for defaulting Sway to version 1.7. This isn't a Sway/wlroots bug, Chromium is now patched, and Electron apps would have to be used via XWayland in the meantime (IMO this is fair since Ozone/Wayland is still in alpha (or beta?)). It's certainly not ideal but IMO by far not enough to hold the update back.

If deemed necessary we could add a programs.sway.package option. Any opinions on this? Or maybe someone has a better idea?

Synthetica9 · 2022-01-16T14:34:52Z

Updating to a newer wlroots' version will also mean we will need this patch in all chromium based applications. Otherwise, due to the new protocol version, all these apps will crash with interface 'wl_output' has no event 4.

This was also discussed above, starting at #151902 (comment)

The patch was included in v99.0.4766, so it already hit electron master (v18 nightly, with request for backport) and chromium dev (#154668).

Thank you for keeping an eye on this, I hope this will trickle down faster than we expect!

Synthetica9 · 2022-01-16T19:46:14Z

@ofborg test sway

LunNova · 2022-01-17T22:35:06Z

For apps which bundle their own outdated electron versions, it's usually possible to use an alternative electron version.

Could consider doing this for apps which stop working with this update.

Example arch build doing this for discord https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=discord_arch_electron

chvp · 2022-01-23T15:48:01Z

1.7 final was released yesterday

Release notes: https://github.com/swaywm/sway/releases/tag/1.7 Notable (backward incompatible) changes: - The default terminal changed from Alacritty to foot Known issues: - `swaynag` will crash when Sway 1.6.1 is still running while the Nix package (and thus `swaynag`) is already updated to version 1.7. - The experimental Ozone/Wayland support of Electron apps will be broken for a while. Electron version 17 should work but the Chromium fixes haven't yet been backported to Electron version 16. NixOS module: programs.sway.extraPackages: The "alacritty" package was replaced with "foot". VM test: We switched from the OpenGL ES 2.0 renderer to Pixman. The terminal was also changed to foot but Alacritty is still used for the XWayland test (since foot doesn't support X11). Co-authored-by: Patrick Hilhorst <[email protected]>

primeos · 2022-01-23T18:46:31Z

Ok, so from my point of view this is ready to be merged as is. I'm running RC3 on my main setup since it was released and didn't notice any new issues so far (apart from not being able to use the Wayland support of Electron apps (but luckily I don't need any Electron apps anymore) - that isn't a Sway bug though and there's still XWayland or nested Wayland compositors). The diff also looks good to me (a few additional changes but still a clean diff). So I suggest we merge this :) Or are there any objections left?

Synthetica9 · 2022-01-23T18:49:37Z

Do we want to keep 1.6 around for the time being to accommodate Electron users?

…

On Sun, 23 Jan 2022, 19:46 Michael Weiss, ***@***.***> wrote: Ok, so from my point of view this is ready to be merged as is. I'm running RC3 on my main setup since it was released and didn't notice any new issues so far (apart from not being able to use the Wayland support of Electron apps (but luckily I don't need any Electron apps anymore) - that isn't a Sway bug though and there's still XWayland or nested Wayland compositors). The diff also looks good to me (a few additional changes but still a clean diff). So I suggest we merge this :) Or are there any objections left? — Reply to this email directly, view it on GitHub <#151902 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABV7PJZYWS3J3N2XD5QJOY3UXREJHANCNFSM5KVLSJNQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you were mentioned.Message ID: ***@***.***>

primeos · 2022-01-23T19:22:50Z

Do we want to keep 1.6 around for the time being to accommodate Electron
users?

That's the question. I've written about it in #151902 (comment) but there were not replies. I don't feel that it's necessary (we already provide override mechanics) and if we'd add one we'd also need programs.sway.package to make it usable without an overlay (and if users could figure out that they need to downgrade Sway to fix Electron they could very likely also do so via an overlay). And the biggest reason against it is that apparently the Wayland support of Electron 16 is also broken since 16.0.6 (or 16.0.5). At least I noticed that with Signal-Desktop and other Linux distributions are affected as well: e06082e. That should theoretically affect more Electron apps soon.

Anyway, I'm not objecting if someone thinks that having sway_1_6 would help (edit: although I would appreciate a draft or suggestion for an implementation as there are multiple approaches - using overrides would produce the smallest diff - alternatively we could also revisit this topic later, if necessary).

dmayle · 2022-01-23T19:28:44Z

Personally, I'm waiting for this to land. Electron works under XWayland, and there are other override mechanisms, so the only argument in favor of keeping sway_1_6 is to enable users to choose a third workaround mechanism as opposed to one of two others. Given the added maintenance cost, I can't find the value.

Synthetica9 · 2022-01-23T19:31:49Z

Okay, feel free to merge then. I'm opening a tracking issue for the Chromium situation

Princemachiavelli · 2022-01-23T19:48:41Z

Personally, I'm waiting for this to land. Electron works under XWayland, and there are other override mechanisms, so the only argument in favor of keeping sway_1_6 is to enable users to choose a third workaround mechanism as opposed to one of two others. Given the added maintenance cost, I can't find the value.

Agreed, my impression is having multiple explicit versions of packages really should be limited to packages with explicit versions that are often pinned such as programming languages/runtimes and databases i.e it's a configuration supported by upstream. If someone really needs to hold back a package completely there are methods for using overlays in their Nix config. I would rather see those methods become better documented and used more frequently than for the nixpkgs repository to accumulate more and more workarounds.

primeos · 2022-01-23T20:46:10Z

Ok, thanks for the quick feedback. Let's merge this then :) 🚀

I also want to thank everyone who helped via reviews/feedback, etc. (i.e. all participants of this PR).

Synthetica9 · 2022-01-23T22:11:10Z

Weird glitch I encountered in testing, and I have no idea if it's something that could pop up in the future. Just wanted to share it here, but assuming it was a weird one-of, can't reproduce it now.

Princemachiavelli · 2022-01-23T22:27:15Z

@Synthetica9 What GPU/GPU driver are you using? I'm an Vega 8/amdgpu and haven't had any regressions. Although sway --version does show "sway version 1.8-dev" instead of 1.7 . probably an issue with the release commit.

I've been using chromium and several Electron apps like VScode, Electronmail, and Freetube and they all seem to work so they must all be defaulting to XWayland.

The only thing I have noticed is that the new render_bit_depth option does not result in the GPU/display actually switching to a 10 bit profile; drm_info still shows the current bpc as 8. Even with the verbose and debug flags I don't see any errors related to switching the render/display profile so I'm not sure what the issue is.

Synthetica9 · 2022-01-23T22:28:45Z

@Synthetica9 What GPU/GPU driver are you using? I'm an Vega 8/amdgpu and haven't had any regressions. Although sway --version does show "sway version 1.8-dev" instead of 1.7 . probably an issue with the release commit.

This is in the QEMU test driver, on an Intel GPU.

primeos mentioned this pull request Dec 23, 2021

Remove wlroots_0_14 (tracking issue) #151311

Closed

17 tasks

ofborg bot requested review from Ma27 and Synthetica9 December 23, 2021 19:37

ofborg bot added 11.by: package-maintainer This PR was created by the maintainer of the package it changes 10.rebuild-darwin: 1-10 10.rebuild-linux: 1-10 labels Dec 23, 2021

primeos changed the base branch from staging to staging-next December 28, 2021 15:59

github-actions bot added 6.topic: nixos Issues or PRs affecting NixOS modules, or package usability issues specific to NixOS 8.has: module (update) This PR changes an existing module in `nixos/` labels Dec 28, 2021

Synthetica9 force-pushed the sway branch from 0860d0d to db07318 Compare January 4, 2022 19:49

primeos commented Jan 6, 2022

View reviewed changes

nixos/tests/sway.nix Outdated Show resolved Hide resolved

nixos/tests/sway.nix Outdated Show resolved Hide resolved

nixos/tests/sway.nix Show resolved Hide resolved

primeos changed the base branch from staging-next to master January 9, 2022 18:12

primeos force-pushed the sway branch from 52c0a2a to 1c96e3e Compare January 23, 2022 18:20

primeos force-pushed the sway branch from 1c96e3e to 2e719d1 Compare January 23, 2022 18:28

primeos marked this pull request as ready for review January 23, 2022 18:46

Synthetica9 mentioned this pull request Jan 23, 2022

Tracking issue: Electron: Ozone/Wayland broken with wlroots 0.15 #156352

Closed

11 tasks

primeos merged commit a3d847c into NixOS:master Jan 23, 2022

primeos mentioned this pull request Jan 23, 2022

sway: Optimal NixOS integration #57602

Open

22 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

sway: 1.6.1 -> 1.7 #151902

sway: 1.6.1 -> 1.7 #151902

primeos commented Dec 23, 2021 •

edited

Loading

Synthetica9 commented Dec 24, 2021 via email

Synthetica9 commented Dec 28, 2021

Synthetica9 commented Dec 28, 2021 •

edited

Loading

primeos commented Dec 28, 2021 •

edited

Loading

Synthetica9 commented Dec 29, 2021 •

edited

Loading

tadeokondrak commented Dec 29, 2021

Synthetica9 commented Dec 29, 2021

primeos commented Jan 2, 2022 •

edited

Loading

primeos commented Jan 2, 2022 •

edited

Loading

Synthetica9 commented Jan 2, 2022

jackinloadup commented Jan 4, 2022

Synthetica9 commented Jan 4, 2022

primeos commented Jan 4, 2022

Synthetica9 commented Jan 4, 2022

primeos left a comment

LunNova commented Jan 14, 2022

Synthetica9 commented Jan 14, 2022

primeos commented Jan 15, 2022 •

edited

Loading

FlorianFranzen commented Jan 16, 2022 •

edited

Loading

primeos commented Jan 16, 2022

Synthetica9 commented Jan 16, 2022

Synthetica9 commented Jan 16, 2022

LunNova commented Jan 17, 2022

chvp commented Jan 23, 2022

primeos commented Jan 23, 2022

Synthetica9 commented Jan 23, 2022 via email

primeos commented Jan 23, 2022 •

edited

Loading

dmayle commented Jan 23, 2022

Synthetica9 commented Jan 23, 2022

Princemachiavelli commented Jan 23, 2022

primeos commented Jan 23, 2022

Synthetica9 commented Jan 23, 2022

Princemachiavelli commented Jan 23, 2022

Synthetica9 commented Jan 23, 2022

sway: 1.6.1 -> 1.7 #151902

sway: 1.6.1 -> 1.7 #151902

Conversation

primeos commented Dec 23, 2021 • edited Loading

Motivation for this change

Things done

Synthetica9 commented Dec 24, 2021 via email

Synthetica9 commented Dec 28, 2021

Synthetica9 commented Dec 28, 2021 • edited Loading

primeos commented Dec 28, 2021 • edited Loading

Edit: VM test output

Synthetica9 commented Dec 29, 2021 • edited Loading

tadeokondrak commented Dec 29, 2021

Synthetica9 commented Dec 29, 2021

primeos commented Jan 2, 2022 • edited Loading

primeos commented Jan 2, 2022 • edited Loading

Synthetica9 commented Jan 2, 2022

jackinloadup commented Jan 4, 2022

Synthetica9 commented Jan 4, 2022

primeos commented Jan 4, 2022

Synthetica9 commented Jan 4, 2022

primeos left a comment

Choose a reason for hiding this comment

LunNova commented Jan 14, 2022

Synthetica9 commented Jan 14, 2022

primeos commented Jan 15, 2022 • edited Loading

FlorianFranzen commented Jan 16, 2022 • edited Loading

primeos commented Jan 16, 2022

Synthetica9 commented Jan 16, 2022

Synthetica9 commented Jan 16, 2022

LunNova commented Jan 17, 2022

chvp commented Jan 23, 2022

primeos commented Jan 23, 2022

Synthetica9 commented Jan 23, 2022 via email

primeos commented Jan 23, 2022 • edited Loading

dmayle commented Jan 23, 2022

Synthetica9 commented Jan 23, 2022

Princemachiavelli commented Jan 23, 2022

primeos commented Jan 23, 2022

Synthetica9 commented Jan 23, 2022

Princemachiavelli commented Jan 23, 2022

Synthetica9 commented Jan 23, 2022

primeos commented Dec 23, 2021 •

edited

Loading

Synthetica9 commented Dec 28, 2021 •

edited

Loading

primeos commented Dec 28, 2021 •

edited

Loading

Synthetica9 commented Dec 29, 2021 •

edited

Loading

primeos commented Jan 2, 2022 •

edited

Loading

primeos commented Jan 2, 2022 •

edited

Loading

primeos commented Jan 15, 2022 •

edited

Loading

FlorianFranzen commented Jan 16, 2022 •

edited

Loading

primeos commented Jan 23, 2022 •

edited

Loading