Initial implementation of cosmic-workspace-unstable-v1 #2030

Consolatis · 2024-07-26T01:22:25Z

I finally have decided to not wait another 3 years until the official workspace protocol is accepted. This goes very much against my own believes that labwc should only use wlroots or wayland protocols but version 1 of the cosmic-workspaces protocol is basically the same as what is proposed in the upstream protocol. Using the cosmic variant also allows using the protocol without hiding it behind feature flags as the upstream protocol may change in incompatible ways before being accepted.

This PR is an initial implementation of the protocol, mostly written in a compositor agnostic way so there is the potential to also propose this implementation for wlroots with minor changes at some point (e.g. mostly znew() vs calloc() and wl_list_append() vs wl_list_insert(x.prev, y)).

The integration part of the protocol into labwc is purposefully minimal, it only tells the client about the workspaces available, their state name and allows the client to activate them. The protocol theoretically also allows for clients to configure the workspaces and the protocol implementation allows for that but the labwc integration does not.

Draft PR for sfwbar and Waybar to have something for testing:

Test config for Waybar

~/.config/waybar/config

    "cosmic/workspaces": {
        "format": "{name}",
        "on-click": "activate",
        "all-outputs": true,
    },

~/.config/waybar/style.css (modified the existing button:hover)

button:hover,
#workspaces button.active {
    background: inherit;
    box-shadow: inset 0 -3px #ffffff;
}

tokyo4j · 2024-07-26T16:17:10Z

I don't have any strong preference here since I usually use only one workspace and the integration part is minimal as you said, but what will we do when the ext-workspace-unstable is finally merged and wlroots adds implementation for it? Will we temporarily support both of cosmic-workspace-unstable-v1 and ext-workspace-unstable for compatibility? (just for question)

Consolatis · 2024-07-26T17:26:00Z

what will we do when the ext-workspace-unstable is finally merged and wlroots adds implementation for it? Will we temporarily support both of cosmic-workspace-unstable-v1 and ext-workspace-unstable for compatibility? (just for question)

Yes, depending on the availability of clients that would then support ext-workspaces the idea would be to support both for labwc.

johanmalm · 2024-07-26T20:17:55Z

Fine with me - in terms of scope. Not looked at code.
I share your verdict and had come to the same conclusion.

Consolatis · 2024-07-27T19:47:48Z

Waybar PR: Alexays/Waybar#3481

01micko · 2024-07-29T11:00:10Z

Test run, seems to work fine - screencap, is this what I'm supposed to be looking at? (methinks yes)

Back on my Slack-current install with updated latest stuff so it builds nice :)

Consolatis · 2024-07-29T12:13:59Z

is this what I'm supposed to be looking at? (methinks yes)

Yes, at least with the stock sfwbar.config or a derivation of it. The 1-4 on the left look like the pinned workspaces configured in the sfwbar config by default. They can be removed by commenting out the pins line in the pager section.

01micko · 2024-07-29T12:21:43Z

is this what I'm supposed to be looking at? (methinks yes)

Yes, at least with the stock sfwbar.config or a derivation of it. The 1-4 on the left look like the pinned workspaces configured in the sfwbar config by default. They can be removed by commenting out the pins line in the pager section.

Yeah I figured I had to use stock first up, so just renamed my user config. I'll play with it some more later.

jlindgren90 · 2024-08-01T09:57:07Z

There is also https://wayland.app/protocols/kde-plasma-virtual-desktop - not sure of the pros/cons of following Cosmic vs. KDE here.

droc12345 · 2024-08-01T10:11:41Z

The only comment I have about adding different implementations like these is
to try and keep it so that it can be easily removed, if need be.

stefonarch · 2024-08-01T11:58:50Z

There is also https://wayland.app/protocols/kde-plasma-virtual-desktop - not sure of the pros/cons of following Cosmic vs. KDE here.

The lxqt-panel will use this one if the compositor is kwin_wayland. Desktop pager and "Move to Desktop x" in the windows button right click menu of the taskbar are working fine with it.

Consolatis · 2024-08-01T18:29:54Z

There is also https://wayland.app/protocols/kde-plasma-virtual-desktop - not sure of the pros/cons of following Cosmic vs. KDE here.

From a quick look:

The cosmic variant is an older version of the protocol submitted at wayland-protocols and should be easy to adjust to the official one once merged
The KDE one looks completely independent and I am not aware of any upstream efforts.
The KDE one is also way less flexible. E.g. it doesn't seem to notify about any output <--> workspace relation, doesn't allow the compositor to send independent workspaces per output nor allow a client to change the workspaces on a defined output. It also doesn't provide any form of capability systems so a panel can't disable options not supported by the compositor.
The upstream / cosmic variant allows for output independent workspaces via the group concept and has a capability system.
Additionally the upstream / cosmic one also supports slightly different concepts for workspaces like tags (globally or per output) as multiple workspaces can be active at the same time.

jlindgren90 · 2024-08-02T02:59:48Z

If the cosmic protocol is closer than the upstream MR, then that's reason enough, you've convinced me :)

My only other thought: there was a wlr protocol proposed 5(!) years ago, https://gitlab.freedesktop.org/wlroots/wlr-protocols/-/merge_requests/35. It was closed in favor of the upstream wayland-protocols MR, and there's been no discussion on it recently. Would it be worth checking with the wlroots folks if it could be revived, given that the upstream MR seems to have stalled?

I do realize wlroots development also happens slowly... if there is no interest there, then I'm fine with adopting the cosmic namespace variant. But it might be worth at least asking.

Consolatis · 2024-08-02T07:16:58Z

My only other thought: there was a wlr protocol proposed 5(!) years ago, https://gitlab.freedesktop.org/wlroots/wlr-protocols/-/merge_requests/35. It was closed in favor of the upstream wayland-protocols MR, and there's been no discussion on it recently. Would it be worth checking with the wlroots folks if it could be revived, given that the upstream MR seems to have stalled?

That is an even older variant of the current upstream MR (and thus also an older version of the cosmic vendored one). I don't think we want yet another version of the same protocol floating around. My hope is that either the upstream one is merged within the next year or if that is not the case that wlroots will accept an implementation of the cosmic variant instead (for example the protocol implemenation in this PR that could be slightly adapted for wlroots). In the meantime, compositors and panels can add support for the cosmic vendored one which can be very easily adapted for the upstream MR once merged (as of now I think it only replaced the wl_array capabilities with a proper single uint32_t bitset).

I do realize wlroots development also happens slowly... if there is no interest there, then I'm fine with adopting the cosmic namespace variant. But it might be worth at least asking.

As both cosmic and wlroots are members of the wayland-protocols governance body and protocols in the ext- namespace only need 2 ACKS from members I do not think that wlroots would accept a new wlr vendored protocol. At that point it would make way more sense to get the upstream one merged. See also https://gitlab.freedesktop.org/wlroots/wlr-protocols#submitting-new-protocols.

jlindgren90 · 2024-08-02T09:01:16Z

My hope is that either the upstream one is merged within the next year or if that is not the case that wlroots will accept an implementation of the cosmic variant instead (e.g. for example this one that could be slightly adopted for wlroots).

Okay, this makes sense... a wlroots implementation would continue to use the "cosmic-" protocol namespace then (until the "ext-" version is merged). I am good with this approach.

johanmalm

Have left a bunch of review comments - but other than those it LGTM. Not tested.

A couple of general points:

Should we prefix with wlr_cosmic_? Couldn't we just do cosmic. For example wlr_cosmic_workspace_group_output_enter()
We ought to consider ways to not lose created workspace on Reconfigure. Can do that on separate PR though if we can think of something.

johanmalm · 2024-08-02T19:02:45Z

src/protocols/cosmic_workspaces/cosmic-workspaces.c

+
+/*
+ * .--------------------.
+ * | TODO |


Should this TODO still be here?

Will work on the remaining items tomorrow and then remove the TODO but none of them should cause any issues.

src/protocols/cosmic_workspaces/cosmic-workspaces.c

src/workspaces.c

src/protocols/cosmic_workspaces/transactions.c

Consolatis · 2024-08-02T21:04:22Z

Have left a bunch of review comments - but other than those it LGTM. Not tested.

Thanks.

Should we prefix with wlr_cosmic_? Couldn't we just do cosmic. For example wlr_cosmic_workspace_group_output_enter()

I've changed everything to a lab_cosmic_ prefix, the wlr_ one was intended to upstream the protocol implementation at some point but that is really easy to fix with some sed 's/lab_cosmic/wlr_cosmic/g'.

We ought to consider ways to not lose created workspace on Reconfigure. Can do that on separate PR though if we can think of something.

I've removed the feature for clients to create workspaces completely as its unclear who is in control of the final workspace configuration. We could consider adding that back at some point (it just requires reacting to two signals).

include/common/array.h

johanmalm · 2024-08-03T14:20:59Z

include/common/array.h

+ * @_val: the item to add to the array
+ * Note: asserts() when the memory allocation for the new item failed
+ */
+#define xwl_array_add(_arr, _val) do { \


This is indeed a nice wrapper. I now get it 😄

Suggest that we avoid assert() and just exit() with an error message.

Suggest that we call it array_add() becase:

xwl makes me think of XWayland

It actually does more than wl_array_add() die-on-null

Also suggest adding the following to the comment in the header to make it easier to wrap one's head around it:

Let us illustrate the function of this macro by an example: uint32_t value = 5; array_add(array, value); ...is the equivalent of the code below which is how you would otherwise use the wl_array API: uint32_t *elm = wl_array_add(array, sizeof(uint32_t)); if (!elm) { perror("failed to allocate memory"); exit(EXIT_FAILURE); } *elm = value;

Renamed to array_add(), replaced the assert with exit() and added your comment.

johanmalm · 2024-08-03T14:23:46Z

Have tested with waybar. Works well.

LGTM other than the suggestion I just left on the array-macro.

g7fya · 2024-08-08T01:55:03Z

hello, installed labwc and sfwbar from Consolatis repositorie, work great. However, is there a way where sfwbar taskbar to show only windows from the current workspace?

Consolatis · 2024-08-08T02:00:10Z

is there a way where sfwbar taskbar to show only windows from the current workspace?

No, that requires a further protocol to match windows with workspaces. This one only supports controlling the workspaces themselves.

LBCrion · 2024-08-08T10:03:43Z

No, that requires a further protocol to match windows with workspaces. This one only supports controlling the workspaces themselves.

We could implement cosmic-toplevel-info-unstable and cosmic-toplevel-management-unstable with a fallback to wlr-foreign-top-level in theory. Although I suspect that's out of scope for now ;)

Consolatis force-pushed the wip/cosmic_workspaces branch from 1972260 to e6aec2c Compare July 26, 2024 01:34

Consolatis mentioned this pull request Jul 26, 2024

Initial implementation of cosmic-workspace-unstable-v1 LBCrion/sfwbar#212

Draft

stefonarch mentioned this pull request Jul 26, 2024

Wlroots taskbar lxqt/lxqt-panel#2046

Draft

Consolatis force-pushed the wip/cosmic_workspaces branch from e6aec2c to 8c7e451 Compare July 26, 2024 22:52

Consolatis mentioned this pull request Jul 27, 2024

Add cosmic/workspaces module Alexays/Waybar#3481

Draft

workspaces: slight struct reordering

d9866aa

Consolatis force-pushed the wip/cosmic_workspaces branch from 94b744d to 2f71733 Compare August 2, 2024 17:37

johanmalm reviewed Aug 2, 2024

View reviewed changes

Consolatis force-pushed the wip/cosmic_workspaces branch from 2f71733 to e4f63a0 Compare August 2, 2024 20:55

johanmalm reviewed Aug 2, 2024

View reviewed changes

include/common/array.h Outdated Show resolved Hide resolved

Consolatis force-pushed the wip/cosmic_workspaces branch 4 times, most recently from 2bfc868 to 8567dd3 Compare August 3, 2024 03:04

johanmalm reviewed Aug 3, 2024

View reviewed changes

Consolatis force-pushed the wip/cosmic_workspaces branch from 8567dd3 to 4df9c26 Compare August 3, 2024 16:22

Consolatis added 3 commits August 3, 2024 18:25

common/array.h: add array_add() helper

a9daaa0

cosmic-workspaces: protocol implementation

31f4336

cosmic-workspaces: labwc integration

904e0d2

Consolatis force-pushed the wip/cosmic_workspaces branch from 4df9c26 to 904e0d2 Compare August 3, 2024 16:25

01micko mentioned this pull request Aug 4, 2024

Plan for release 0.8.0 #2046

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Initial implementation of cosmic-workspace-unstable-v1 #2030

Initial implementation of cosmic-workspace-unstable-v1 #2030

Consolatis commented Jul 26, 2024 •

edited

Loading

tokyo4j commented Jul 26, 2024

Consolatis commented Jul 26, 2024

johanmalm commented Jul 26, 2024

Consolatis commented Jul 27, 2024

01micko commented Jul 29, 2024

Consolatis commented Jul 29, 2024 •

edited

Loading

01micko commented Jul 29, 2024

jlindgren90 commented Aug 1, 2024

droc12345 commented Aug 1, 2024

stefonarch commented Aug 1, 2024

Consolatis commented Aug 1, 2024 •

edited

Loading

jlindgren90 commented Aug 2, 2024

Consolatis commented Aug 2, 2024 •

edited

Loading

jlindgren90 commented Aug 2, 2024

johanmalm left a comment

johanmalm Aug 2, 2024

Consolatis Aug 2, 2024

Consolatis commented Aug 2, 2024

johanmalm Aug 3, 2024 •

edited

Loading

Consolatis Aug 3, 2024

johanmalm commented Aug 3, 2024

g7fya commented Aug 8, 2024

Consolatis commented Aug 8, 2024

LBCrion commented Aug 8, 2024

Initial implementation of cosmic-workspace-unstable-v1 #2030

Are you sure you want to change the base?

Initial implementation of cosmic-workspace-unstable-v1 #2030

Conversation

Consolatis commented Jul 26, 2024 • edited Loading

tokyo4j commented Jul 26, 2024

Consolatis commented Jul 26, 2024

johanmalm commented Jul 26, 2024

Consolatis commented Jul 27, 2024

01micko commented Jul 29, 2024

Consolatis commented Jul 29, 2024 • edited Loading

01micko commented Jul 29, 2024

jlindgren90 commented Aug 1, 2024

droc12345 commented Aug 1, 2024

stefonarch commented Aug 1, 2024

Consolatis commented Aug 1, 2024 • edited Loading

jlindgren90 commented Aug 2, 2024

Consolatis commented Aug 2, 2024 • edited Loading

jlindgren90 commented Aug 2, 2024

johanmalm left a comment

Choose a reason for hiding this comment

johanmalm Aug 2, 2024

Choose a reason for hiding this comment

Consolatis Aug 2, 2024

Choose a reason for hiding this comment

Consolatis commented Aug 2, 2024

johanmalm Aug 3, 2024 • edited Loading

Choose a reason for hiding this comment

Consolatis Aug 3, 2024

Choose a reason for hiding this comment

johanmalm commented Aug 3, 2024

g7fya commented Aug 8, 2024

Consolatis commented Aug 8, 2024

LBCrion commented Aug 8, 2024

Consolatis commented Jul 26, 2024 •

edited

Loading

Consolatis commented Jul 29, 2024 •

edited

Loading

Consolatis commented Aug 1, 2024 •

edited

Loading

Consolatis commented Aug 2, 2024 •

edited

Loading

johanmalm Aug 3, 2024 •

edited

Loading