Wayland-Protocols 1.10 Adds XDG-Output

[shmerl]

DEs might as well just agree on common underlying screen recording compositor library and define how layout is passed, aka. layout format and how to pass surfaces with zero copy. at that point all that compositors would only need the visual part where you define which windows are recorded in which layout.

That's the whole point. "Might agree" isn't really a practical approach, because as you can see no one agreed on it so far. It should be part of the standard protocols. And it better be, especially if security is concerned.

[justmy2cents]

By that point, then Wayland isn't necessary too.

These things need to be standardized and become quite clear to the rest, I think that's the point of Wayland: To clear the mess in X.

.Don't expect DEs/compositors agree on a "common underlying screen recording compositor library", that sound like a total disaster to happen. Currently DEs/WMs agree rarely and have NIH syndrome everywhere, instead sharing infrastructures in a more agnostic way (data indexing is only one of the few examples).

That would work in an ideal world, but not in the current ecosystem. Wayland needs to be strict, robust, feature complete and widely adopted.

fun fact. compositors are not part of wayland, now,... guess what screen recording is. i mean screen recording that does more than just record full output, aka. full featured OBS like screen recording. it is compositor

now you need to add access to sound as well and some other things.

it just doesn't fit in wayland.

and i think you miss the point about that library how i intended to sound it. it is not about recording solution. just common interfaces exposed by DE compositors for surface enumeration and passing those with 0 copy that software like OBS could effectively use

[timofonic]

that really isn't necessary to be in Wayland. i mean, it could be, but it is not even remotely needed

DEs might as well just agree on common underlying screen recording compositor library and define how layout is passed, aka. layout format and how to pass surfaces with zero copy. at that point all that compositors would only need the visual part where you define which windows are recorded in which layout.

By that point, then Wayland isn't necessary too.

These things need to be standardized and become quite clear to the rest, I think that's the point of Wayland: To clear the mess in X.

.Don't expect DEs/compositors agree on a "common underlying screen recording compositor library", that sound like a total disaster to happen. Currently DEs/WMs agree rarely and have NIH syndrome everywhere, instead sharing infrastructures in a more agnostic way (data indexing is only one of the few examples).

That would work in an ideal world, but not in the current ecosystem. Wayland needs to be strict, robust, feature complete and widely adopted.

[Gusar]

So why then is Gnome's the only worthy implementation, if not due to their protocol work?

[Gusar]

That would be a better name. At least it would tell who develop the protocol and the worthy compositor implementations.

That's some very, umm, interesting logic you got there: Gnome's is the only "worthy" implementation. Because they do all the work on the protocol, apparently. But when others do some work, that stuff is "not needed", so won't become part of the protocol. Thus assuring that only Gnome's work remains and therefore only their implementation will ever be "worthy".

Yep, that makes sense...

[kon14]

Rejected because it is not needed. Wayland was designed with CSD in mind. Scope creep is something you fight with napalm.

Design is always readjusted to fit users' needs, this is why software development continues after bug-fixing. Wayland as a protocol is still heavily under development. Wayland was designed with CSD in mind, not against SSD. Wayland was also designed for GBM, but now Gnome is merging EGL-streams support upstream iirc (?), the absurd short-term solution to Nvidia's monopoly abuses.

Either way, people are going to implement this whether it's accepted as a standard or not. Gnome's VETO will only result in duplication of effort and unnecessary incompatibilities.

How do you feel about cursor theme CSD? Do you also like how one has to install and select multiple themes for Gtk/Qt/Compositor(in case it defaults to X11 one)/etc for desktop/window applications? Does this look like the way forward to you?

[Gusar]

Rejected because it is not needed.

And here we have it, *exactly* the kind of "living in the bubble" mentality I mentioned. If it does not fit the one true way™ which Gnome is trying to shove down people's throats, it's "not needed". Well, tell that to all the people complaining about mpv having no borders. Or just rename Wayland to GnomeWindowingSystem and tell everyone else to take a hike and not bother.

[Gusar]

Made by RH/GNOME.

Server-side decorations protocol by KDE. Accepted by sway. Rejected by Gnome. Prefect proof of kon14's point about Gnome devs living in their own bubble and rejecting everything that comes from outside.

[kon14]

You are more than welcome to prove your point. Show which significant parts of Wayland have been designed by other DEs than GNOME.

Aside from my points not being all about the Wayland protocol or even Wayland in general, was that not precisely my exact question to you in the very post you quoted just now? Except you never answered that or my other questions.

To the best of my knowledge, Kristian Høgsberg has been the main individual contributor, in terms of commits, and founder of Wayland, with RedHat hiring Wayland devs for Fedora and therefore also supporting the Gnome desktop.

Can you link me to specific evidence that shows Gnome is in any way contributing significantly more than others? Being one of the first to implement Wayland means nothing by itself. Since I'm not the one to devalue Gnome regarding Wayland support specifically, you seem to reject others' attempts and I asked you this very question before I believe it's pretty sane to expect that you answer this question.

[kon14]

Umm why do think it is any good to have compositor diversity if the developers are unwilling to do the hard work at API and protocol level?

First things first implementing a protocol doesn't necessarily mean you've designed it yourself or aided at all in its design. Where do you base all your assumptions regarding Gnome being the only contributor to Wayland? Just cause KDE took a while to get the ball rolling on Plasma doesn't mean nobody else is working on Wayland support or that Gnome is in any way the only contributor to Wayland as a protocol. You should also realize how a lot of "Gnome" stuff actually comes out of Fedora and RH, much like how Wayland came to be in the first place.

Regarding compositor diversity and implementation diversity in general, it's something that's always good to have to some extent since it allows you to debug your own approach and realize whether a problem resides in drivers/compositors/apps/etc depending on what sort of software we're talking about. You should also try and this through that thick skull of yours: Not everyone likes Gnome. It's simple as that. Even if Gnome where to be the only project working on Wayland, daily reminder for you: it's not, and then after years people came along and used their work for their own compositors that would still be fine and acceptable. Isn't this how foss works? Isn't this what happened to Gtk and Gimp?

It's also funny how you keep blaming people for not integrating Mutter and standardizing Gnome stuff when Gnome devs themselves are rejecting every patch and suggestion coming from outsiders. How do you expect anyone to integrate their stuff and deduplicate effort if Gnome won't ever accept anything upstream? Do you seriously believe this has anything to do with code quality? I'm sure people would be more willing to work with them if they ever wanted to work with anyone but themselves.

I'm sad for all of you Gnome fanboys trying to force your desktop on everyone by devalueing other projects just cause people abandoned Gnome for constantly not giving a hot damn about their users opinions in regards to their decisions. I'm sure lots of Gnome devs are better than this and I'm sorry if I have spoken ill of them, but there's a great deal to say about users like you and this muddies the waters as to who is it that's actually guiding and supporting this hatred towards everybody.

(It seems like I can't even add attachments. Here's one for you)