GTK4's Vulkan Support Continues Maturing

[phoronix]

Phoronix: GTK4's Vulkan Support Continues Maturing

One of the questions that came up following our GNOME 3.26 feature overview was how GTK4's Vulkan renderer is coming along...

http://www.phoronix.com/scan.php?page=news_item&px=GTK4-Vulkan-Maturing

[doom_Oo7]

Yeah a proof of concept for the Standard Desktop.

Which is KDE Plasma, of course ; though I'd say "shitty imitation" rather than "proof of concept"

[Kemosabe]

Which is KDE Plasma, of course ; though I'd say "shitty imitation" rather than "proof of concept"

Plasma is sadly a huge bugparty with some tech demo elements which is apparently created by developers for developers without a design strategy for a efficient user interface. I'm not even talking about a modern one but an efficient and consistent one.

At the beginning I thought it was stupid from Valve to not choose Plasma but later on I realized everything would simply be worse. Performance, Usability and the aged UI.

That's all a real pitty since Qt should be the leading and only GUI framework in the *nix desktop world (ignoring the moc mess for a moment ).

[duby229]

All large scale Qt deployments have failed on the free desktop. kde, Nokia, Unity8, Jolla. It is about time to let go of "Qt save Us All" narrative. Who would think that handing over complete API control to a commercial entity would be smart in the first place?!?

GTK is tailored to the Standard Desktop. They design and decide the API which is an absolute necessity if you want something like a Wayland desktop.

Well, I have to say I do like Redhat's financial model of supporting itself. Because in supporting itself it supports a broad portion of the open source projects it uses. It's a good thing that even GTK falls under to some degree. But that still doesn't excuse the GTK devs from breaking shit constantly. And then bragging about it. It's definitely the worst widget set to choose if you want a stable interface between versions. GTK -will- break between versions. It's already happened again and again and again.

[Kemosabe]

All large scale Qt deployments have failed on the free desktop. kde, Nokia, Unity8, Jolla. It is about time to let go of "Qt save Us All" narrative. Who would think that handing over complete API control to a commercial entity would be smart in the first place?!?

GTK is tailored to the Standard Desktop. They design and decide the API which is an absolute necessity if you want something like a Wayland desktop.

C'mon it's really not Qt's fault. And from a technological point of view: Gtk always has been and is inferior. Who actually wants to develop with Gtk? For me, it is a nightmare because I know the other side. Besides that the platform independence is kinda iffy <- And yes, this is a crucial feature these days.

You mean "tailored to the standard desktop" because it's being developed hand in hand with the one and only DE you can show the world without being completely embarrassed?

[DMJC]

Redhat has no clue what they are doing on desktop, as evidenced by anyone using a multi-screen setup with gnome. It runs terribly as it keeps throwing applets and things into corners where they are unwanted. It's not even aware of a second screen in so far as spreading applets and notifications across both monitors (it only throws them on the primary.) The Gnome Control Center is a poor ripoff of OSX's System Preferences.app. If you're going to try ripping off OSX at least use GNUstep so you get the same API. I'll be switching back to XFCE4 as soon as I have free time to organise it. My computer shouldn't be doing random actions to my apps every time my mouse moves to a corner of my screen. I am starting to understand why Motiff was designed the way it was. Actions shouldn't occur without conscious user intervention.

[starshipeleven]

GNOME is sadly a huge devparty with some tech demo elements which is apparently created by developers for developers without a design strategy for a desktop user interface. I'm not even talking about a modern one but an efficient and consistent one.

At the beginning I thought it was stupid from Valve to not choose GNOME but later on I realized everything would simply be worse. Performance, Usability and the touch UI. That's why they have SteamOS use their own UI.

That's all a real pitty since GTK should be the leading and only GUI framework in the *nix desktop world (ignoring the moc mess for a moment ).

fixed.

[starshipeleven]

Ding, ding, ding. We got a winner. As stated earlier API design is harder than surgery on a beating heart.

No, that's just GNOME devs that suck at it, so they make and remake and reamke them on each revision.

QT decided stuff like QNX is more important than kde. That's why we get a couple QT regressions, while bad release management is a distro issue, to read about on Phoronix.

Fixed.
QNX is used in markets where companies pay for Qt, KDE or linux desktop in general does not pay for Qt.

[starshipeleven]

Benjamin Otte made the vulkan renderer. That takes a decade of experience. You wont get there by Gsoc, ridiculous Randa meetings or weekend hacks. You need paid support from RH or similar.

You won't get a vulkan renderer in Kwin because it is plain pointless for a DE running on Wayland, as usual GNOME devs waste resources playing with frivolous things none in the userbase really need, just because of RedHat's deep pockets.

https://blog.martin-graesslin.com/blog/2015/08/thoughts-on-vulkan-in-kwin/

[Masush5]

You won't get a vulkan renderer in Kwin because it is plain pointless for a DE running on Wayland, as usual GNOME devs waste resources playing with frivolous things none in the userbase really need, just because of RedHat's deep pockets.
https://blog.martin-graesslin.com/bl...ulkan-in-kwin/

You realize this is about a vulkan renderer for gtk4 and not mutter right? Rendering all widgets of a gtk application with vulkan instead of OpenGL sounds like it has the potential to reduce overhead to me.