> Most GTK4 apps won't be usable outside the intended desktop environment for which they're built.
There have always been apps more heavily integrated with GNOME vs GTK apps.
Gtk4 is easier to port to other platforms, if anything it should be less painful than it has been before.
> Gtk4 is easier to port to other platforms, if anything it should be less painful than it has been before.
Oh, how desperately I wish this were true. My evidence is purely anecdotal, but I've found GTK4 to be almost completely unusable in my time working with it, so much so that I've basically decided to skip this release altogether and just stick with GTK3 for all intents and purposes. It started when I wanted to pull my Glade projects in to the new framework so I can build my apps with XML instead of writing it by hand... oops, that method has been unsupported. But don't worry! There's a replacement app right around the corner that's... not finished yet. Okay, I guess that's not a big deal, I'll just re-write everything by hand! Now it's compiling and doesn't look the least bit native on my system. And there are compositor issues. And the text looks ass-ugly too.
Asking around on IRC, I was told everything from "use Flatpak" to "stop trolling" to "your workflow is unhealthy/insane/unsupported". It would appear that GTK4 is only intended for GNOME extremists, unfortunately. If there are any portability improvements I won't likely see them.
If you depend on having a GUI builder then I agree it makes sense to wait for that to be stable before switching to GTK4. It's being worked on but progress is slow because of a severe lack of contributors. I don't know anything about compositor issues but it's aimed to give the text a substantial fix in GTK 4.6.
>I was told everything from "use Flatpak" to "stop trolling" to "your workflow is unhealthy/insane/unsupported". It would appear that GTK4 is only intended for GNOME extremists,
I don't see how that has anything to do with "extremists". The GNOME IRC is probably a bad place to ask for help with distro packaging, the focus is on Flatpak because that's realistically the only distro-agnostic packaging that upstream can support. For help with packaging apps for your distro you'd have better luck asking on your distro's IRC.
> If you depend on having a GUI builder then I agree it makes sense to wait for that to be stable before switching to GTK4.
Correction: traditional XML stylesheets were removed with the release of GTK4 so they could introduce a new paradigm, one that was not ready to replace it. I really don't know what to tell you here, because it's such a vast departure that it's going to make the porting process to GTK4 nearly impossible for most people. It's such a braindead, no-holds-barred move that it simply feels disgusting to me. Here we had this perfectly functional feature, and the GNOME devs, being the impractical idealists they are, said "let's get rid of it, offer a replacement and not develop the replacement". Same thing happened with LibAdwaita, GNOME 40 and pretty much everything else they seek to """improve""".
> It's being worked on but progress is slow because of a severe lack of contributors.
Who's fault is that? Is it the aggressive, unfriendly leadership? Or maybe it's the GNOME superstar celebrities, who make completely bonkers changes and expect the rest of the community to follow their lead without question? Maybe it's the lack of directive in their issue trackers, where their lead contributors decide to waffle on major issues instead of taking the directive to just shave the damn yak?
I don't care that they lack contributors. I really don't. That's a grave they've dug for themselves by ostracizing the rest of the community, and it's one they're going to have to deal with for years to come. They had opportunities to build bridges, and they tore them down. They harassed other distribution maintainers when they tried offering help of their own. They've shot themselves in the foot here, plain and simple. GTK2 and 3 weren't remotely this bad or unusable at launch, and it just makes me sad to see people clinging to elitism instead of pragmatism when building an inclusive, accessible desktop. If their ship is sinking, I'm perfectly contented to stay onboard GTK3 and watch as they sink beneath the surface. Several people said this was a bad idea, but it was their hubris that ultimately ruined GTK4. If they need "extra contributors" to fix that, then they should have thought about that before telling everyone else to screw off.
> The GNOME IRC is probably a bad place to ask for help with distro packaging
1. I was in freenode.
2. I wasn't asking about packaging.
My concern was getting my app to look right on every desktop that it ran on, which it most clearly did not. The text was distorted and completely unacceptable for distribution, there were black bars drawn around the window on half the desktops I tried it on, and the look-and-feel was inconsistent across systems. Utterly unbelievable for a toolkit that was trying to pride itself on accessibility.
> the focus is on Flatpak because that's realistically the only distro-agnostic packaging that upstream can support
The issue is that they don't even support that. Flatpak has over 600 open issues of it's own, and implementing it can oftentimes lead to more work, nonfunctional features (like distro-native features in Bottles, for example) and infinite bikeshedding. Every argument I've seen for using Flatpak is bogus. Every one. You can cry "it needs more maintainers!" and I'll probably agree with you, but it's become such a politicized and broken mess that, once again, very few people see the value to contributing to it anymore. Mark my words (and you're welcome to call me out when I'm inevitably wrong here): Flatpak will never get "finished". It has too many esoteric features, unfixable bugs and ideological missteps to compete with package managers that existed just fine 20 years ago.
>Correction: traditional XML stylesheets were removed with the release of GTK4 so they could introduce a new paradigm, one that was not ready to replace it.
Not really, the XML format is pretty similar. In my experience it works very well, even better than the GTK3 XML, but yes it's lacking a good GUI builder.
>I really don't know what to tell you here, because it's such a vast departure that it's going to make the porting process to GTK4 nearly impossible for most people.
I can't agree, I've talked to some GNOME developers and they don't have much of a problem with it. Glade on GTK3 is not very popular among GNOME developers anyway because it causes some other issues, hopefully those will be fixed with the new GUI builder too.
>said "let's get rid of it, offer a replacement and not develop the replacement". Same thing happened with LibAdwaita, GNOME 40 and pretty much everything else they seek to """improve""".
I'm very confused to what you're saying, there are replacements for all that being developed, but it takes time to do it.
>Who's fault is that? Is it the aggressive, unfriendly leadership? Or maybe it's the GNOME superstar celebrities, who make completely bonkers changes and expect the rest of the community to follow their lead without question?
I'm also very confused what you're talking about, I've talked to the Juan (the person working on the new GUI builder) and he's a pretty nice guy working as hard as he can. I've never seen him harass or ostracize anyone or refuse questions. You're generalizing and I wish you wouldn't do that, it's rude to the people who work hard on this and offer no ill will. I've never seen any aggressive leadership or superstar celebrities there either, this is mostly self-directed volunteers working in their area of interest.
>If their ship is sinking, I'm perfectly contented to stay onboard GTK3 and watch as they sink beneath the surface. Several people said this was a bad idea, but it was their hubris that ultimately ruined GTK4.
Well you lose some other improvements in GTK4 if you do this. Those improvements are very unlikely to be brought to GTK3 because they needed an API change. That's why it's a bad idea. Of course you can continue to use GTK3 if you want but you'll continue to miss out on those improvements. It's up to you to say how long you'll be OK with that, but you should know upstream has said that GTK3 will not gain any new features anymore.
>If they need "extra contributors" to fix that, then they should have thought about that before telling everyone else to screw off.
I don't know who was told to screw off, there are open calls for contributors in all these areas.
>I was in freenode.
Well Freenode is not the GNOME IRC so I have no idea what this has to do with GNOME or GTK. Yes there are trolls and unhelpful people in random freenode channels, I never used freenode much myself for that reason. GNOME has its own IRC server, and actually I think they switched to Matrix recently.
>The issue is that they don't even support that. Flatpak has over 600 open issues of it's own, and implementing it can oftentimes lead to more work, nonfunctional features (like distro-native features in Bottles, for example) and infinite bikeshedding.
I know about some of these issues and I'm personally affected by them but what you're missing is that there is actually no other option. Getting rid of Flatpak would leave them with no real way to deploy these apps at all. It's a choice between something buggy that's being worked on (slowly) and leaving app developers out in the cold with nothing. Distro package managers don't solve the problem, they created the whole problem that Flatpak exists to solve.
I'm not going to stop you from pretending like nothing is wrong if that's your prerogative. I'm letting you know here and now that there is an expanding sentiment among developers that GTK4 is a regression from GTK3, and if there are any improvements coming down the pipeline I have yet to see them. I'm not going to argue with you about this when my app is irreparably broken right now, though. Every attempt I've made to reach out to developers has either gone cold or been rejected. I fully intended to start contributing to GTK back in the v3 days, but now? The current leadership has made it clear that they don't have my best interests at mind. They'd rather scrap everything I used and promise a replacement instead of iterating on it or fixing it, which is unacceptable for a toolkit that is currently in active use. I refuse to donate my time or effort to people who disrespect me like that.
I'm a pragmatist, not an idealist. Promises mean very little, especially coming from the modern GTK team.
>I'm not going to stop you from pretending like nothing is wrong if that's your prerogative.
Please don't jump to this conclusion, I'm not doing that. If some GTK3 users want to continue developing GTK3 or use some other toolkit then that's absolutely fine with me. I said that before, I could even recommend some other toolkits for you to use. But I think it's a bad idea to view GTK3 as some kind of resistance against GTK4 because that's missing out on many years of improvements that would be very difficult to backport. Yes there are bugs in GTK4, but I've also talked to the developers and they are aware of them and are working on solving the issues even if progress seems slow. You can read some of the GTK4 blogs to see some of the enhancements that were made in GTK4 and decide if they're worth it to you or not: https://blog.gtk.org/2020/12/16/gtk-4-0/
>I'm letting you know here and now that there is an expanding sentiment among developers that GTK4 is a regression from GTK3
This is not really surprising or new or that much of a big deal, people complained when GTK2 was released because it changed stuff from GTK1, then they complained when GTK3 was released because it changed stuff from GTK2. It's totally understandable and normal for some to want to avoid or delay making the switch, but many others have been willing to make it for the improvements.
>Every attempt I've made to reach out to developers has either gone cold or been rejected. The current leadership has made it clear that they don't have my best interests at mind.
Well if you were on freenode then those probably weren't the upstream developers or any of the current project leadership so I think you might have a wrong impression; you're saying a whole lot of negative things but the underlying logic just doesn't really add up to me. I'm not interested to argue with you either, if you can just mention what the issue is with your app I might be able to help you.
>They'd rather scrap everything I used and promise a replacement instead of iterating on it or fixing it, which is unacceptable for a toolkit that is currently in active use.
This statement is also confusing to me, the replacement is the iteration and fixing. It's being worked on. It can't really be any other way, Glade is a very old application from the GNOME 1 days that's been continuously ported and it still inherits some old design flaws from that. To finally fix the design flaws the program needs some major breakage or a rewrite, so they decided to go with the rewrite because it's actually easier than hacking apart all the internals of Glade and breaking the whole program. And by "they" I mean the one person who has been working on Glade for the last few years and is now working on the replacement. I don't know what "promise" you're talking about either, I think it's a bad situation that only one person is working on this. He could stop working on it at any time and then you'd have no GUI builder at all, the "bus factor" on these projects is a concern right now.
>I refuse to donate my time or effort to people who disrespect me like that.
I'm not suggesting you do that, it seems more like you were asking about taking time or effort to fix your own application.
>There have always been apps more heavily integrated with GNOME vs GTK apps.
From my experience GTK2 apps were not "GNOME" apps, like there were many apps that did not installed the entire GNOME dependencies to work, I don't know if this changed lately since I so far only had the need for qt or GTK2 apps.
That's more because GNOME has completely dropped GTK2 while other projects continue to use it. In the past that wasn't true, GNOME 2 apps depended on a separate library called libgnomeui that had widgets specific to GNOME 2. All the actively-developed GNOME apps ported to GTK3 long ago and so they dropped that dependency which is why you don't see it anymore. The only apps left using GTK2 are non-GNOME apps and so they never depended on that in the first place.
There have always been apps more heavily integrated with GNOME vs GTK apps. Gtk4 is easier to port to other platforms, if anything it should be less painful than it has been before.