GNOME's Platform Design Continues Evolving From Dark Mode To Toasts

[blacknova]

I only have one question: which one of them, GTK or Qt is truly free. Looks like neither, only once you buy a commercial license for Qt it actually makes a lot more sense.

AFAIK the main problem with QT is that you actually need to buy license before you develop your application. You cannot switch from free QT to commercial mid-cycle.
Obviously that is not a problem if your project is open-source only.

[You-]

I was under the impression that GTK was a multiplatform library not bound to Gnome and not requiring independent developers to cater to it but it looks like I was wrong...

I think I said precisely the opposite - Libadwaita was created as a platform library to decouple the gnome inspired widgets from GTK. Instead of forcing GTK developers to use Gnome patterns, they are out elsewhere.

As for GTK, its theming abilities have been unchanged over the last 10 years and will remain so going forward. The only change here for GTK4 is that it is so so much faster.

Only the gnome platform library - Libadwaita - will change how its apps are themed. By keeping the css framework in place and giving other options that stop developers shooting their own feet.

You should be jumping for joy - Gnome has decoupled from GTK, GTK theming remains the most powerful out there (far more capable that Qt, though I dont know how it compares to EFL) and GTK is no longer beholden to any specific platform. Precisely everything you wanted.

It should be green grass and unicorns for everyone.

I do wonder though why you ignored the line from my previous post that was in bold. It was a pretty strange stance from the blog author considering his hampering for freedom.

[White Wolf]

IMO that is actually a surprisingly ignorant take by someone who leads the efforts to make a major DE and distribution.

I am not a proper programmer but I follow a lot of developer discussions and so far it seems everyone find gtk4 much easier to use that gtk3 and earlier - there is early frustration sometimes where the change in focus is not fully understood and the developers have attempted the old way.

An example of this is the complaints about subclassing in that blog. In GTK4 subclassing is less important as making your own widgets has been made easier. however if you want to subclass like before instead of composing your own widget, there will be frustrations due to the changes made to the underlying platform.

He wants both a GTK that isnt limited to gnome, but at the same time he does not want a platform library that separates the components that are specific to gnome.

His distro is also the one that has frozen many gnome apps to ancient released because he takes issue with the name and origins of libhandy. The links trying to discuss this with Solus developers are hilarious for the wrong reasons. A developer even got banned for life for trying to explain the need to allow libhandy to be used by apps that choose to do so.

I will ignore the previous lead's previous plan to migrate to Qt, which has failed. The current lead's major frustration is that he wants to be just a consumer of GTK without being involved and then is frustrated that not all decisions made by others are what he would make.

I look forward to see how far they get with their new EFL based foundation.

Not only him but sytem76 devs also make a statement that GTK and Gnome is going against free choose. So Solus and Pop devs are wrong..... uhhh yeah. The only ignorant ones are Gnome devs.

[You-]

Not only him but sytem76 devs also make a statement that GTK and Gnome is going against free choose. So Solus and Pop devs are wrong..... uhhh yeah. The only ignorant ones are Gnome devs.

The system76 dev was mistaken and (potentially accidentally) made false allegations.

He:

1. Claimed that Gnome was breaking dark mode. When corrected that that wasnt the case and that gnome adding a non-hacky way for apps etc to query the prefer mode, changed his view to saying that they had been pioneers.
2. Thought Libadwaita was both complete and being used for all apps in Gnome 41. He was simply wrong - it is not complete and precisely zero core apps are using it in gnome 41.

I will blame it on not paying attention since they have copied Ubuntu and using a version of gnome over 6 months after its feature development has compeleted and released.

He also thought that Ubuntu held the same belief on these issues as he did, but their developers have confirmed that that isnt the case.
Tldr; threw mud and got it all wrong.

What the interesting bit though was that the volunteers and companies that had worked on the changes were small, maybe even smaller than system76. They had done work and others wanted to use it. System76 had hinted that (after a 2019 theming BoF) it would be going back and giving its needs for a styling API or even developing it, but it never did so and was suddenly found to be an emperor with no clothes.

System76 do great products, so it is unfortunate that on situations like this (and their former war against LVFS) take place. I was also looking forward to their merge requests towards mutter to improve tiling (something they had promised when announcing Pop-Shell - that once the experimentation was complete, they will upstream the non controversial bits), but they have now said they want to make a se[arate framework a layer above mutter.

I suspect the main frustration by System76 is that they are using an old version of gnome. When working towards their next release, they found that they needed to make changes to gnome 40. However that was already stable. Like sane adults, they upstreamed the changes they needed to the now released gnome 41. However their next release will be on gnome 40, without full access to those fixes (because they can break extensions). They will start using gnome 41 in April next year, when the gnome community will have already moved on to gnome 42+.

On the other hand, there was previously a tussle with how things should be done between gnome and Elementary and Ubuntu. Following the 2019 BoF they have formulated plans and sing from the same hymn sheet. You will see System76 being called out by prominent developers from both Ubuntu and Elementary.

The funny thing is that the majority of support for System76 is

[Teggs]

- Ongoing improvements to the GNOME/GTK application styling, especially around libadwaita.

I would hardly name changes that drive entire distros away from GTK as 'improvements', but maybe I misunderstood the reason GTK exists. Or maybe some people changed that reason over time.

[Shiba]

Limited theming is not actually that bad - better to have sane default themes, than a lot of installable eye tearing ones

There is absolutely no reason on earth why having "eye tearing themes" should block you from making your default "sane" one.

[pal666]

"We are unable to come up with a good theming support/API, so let's remove the old CSS based theming altogether and make it inaccessible for most people out there"

you are saying like it's a bad thing. in the world of grownups it's the right thing: if nobody volunteered effort to make a good theming api, then it's not really needed. if you can't design it yourself, hire someone who can. and all whine comes out of distros who want to rebrand gnome without contributing api for it

[pal666]

So Solus and Pop devs are wrong..... uhhh yeah. The only ignorant ones are Gnome devs.

lol? obviously gnome devs are smarter than ubuntu reskin devs(and than random phoronix commenters btw)

[pal666]

drive entire distros

every schoolkid can have its own entire distro

[arQon]

That's a lot of

truth that GNOME has made not just blatantly clear, but explicitly so, ever since GTK3 was introduced and big chunks of theming support were dropped specifically because GNOME devs were butthurt that nobody liked the direction they were taking GNOME 3 in.

When even they admit that removing theme support is one of the *goals* for GNOME, you claiming that others saying so is making "bad assumptions" isn't exactly credible.