andyprough wrote: ↑2026-05-06, 20:19
Gimp started the transition process in 2012, and
by 2015 had a port that could be built showing their gtk3 progress. The final gtk3 pieces weren't in place
until the release of gimp 3.0 in March 2025. The long, laborious process of transitioning from gtk2 to gtk3 by gimp was well documented by tech outlets like Phoronix, and
gimp's own historical chronology page notes the tremendous amount of work required to transition [the 2024 date may be referring to the 3.0 RC1 release]:
Notice the long gap between 2.10 (2018) and 3.0 (2024). That span was consumed by the enormous effort of porting GIMP from the GTK2 toolkit to GTK3 - A rewrite of virtually every widget in the interface. It was unglamorous, difficult work that yielded no new user-visible features for years, but it was necessary to modernize the codebase and restore HiDPI display support on modern hardware.
Not that that's good, it actually shows evidence of what a giant pain in the butt the transitions can be, which is rightfully part of your concern. Also I am sure that you are right that their efforts got a lot more urgent after 2018, when the axe was about to fall on gtk2. Fortunately, you already have a lot of code on the gtk3 side for Pale Moon.
This is good information to hate. But... it almost sounds like you're trying to say, diplomatically, that we should be grateful we have a GTK3 port ready to go, but can't realistically support a new toolkit and will
have to accept the legacy bin on Linux because we are a small project and haven't done the groundwork for a newer toolkit, and you think it will take years and isn't worth pursuing, so we should just "let it happen," and these people offering us "deals" to be part of their fork ecosystem do indeed have us over a barrel and right where they want us because it takes big teams and years of planning ahead to support mainstream Linux? Because I mean... if people seriously think it takes 8 years to get a new toolkit port off the ground, then that explains why they are all suggesting subtly that they think we should just accept our fate and play around in the legacy bin because we don't have the engineering chops to stave off something like this, and they're basing it on the fact that GIMP couldn't and they planned well? I mean... it is also worth noting I did the Python 3 port in a lot less time than Mozilla, and without even referencing their code that much. That took them years as well. And Basilisk-Dev did get Qt working, and someone else got SDL working, so I'm not really convinced we'd need as much time as GIMP or should "just give up" on that basis.
So, anyway, here's what I was thinking... this part isn't really directed to anyone in particular, it's just what I was writing up before I saw the above response to the thread.
It is worth pointing out that there is no GTK5 yet. GNOME 50 dropped X11 support, but GTK5 has not been announced yet. People have been talking about GTK5 doing things like it's already out (I'm thinking they saw the guy saying that in the other thread), and it's not. There is no GTK5 yet. People have been saying Fedora will drop XWayland, when there is no indication of that and many think it will become a compatibility layer like Wine. It would be kind of stupid if they stopped letting X11 apps run on Wayland just to make a statement when people can run Win32 apps on Wayland through Wine, wouldn't it? I mean, I'm not saying that I'm
sure they wouldn't, but there's a very good chance that building and maintaining XWayland was the price RHEL had to pay to get their enterprise customers to accept Wayland in the first place. Also, while GTK3 is still getting critical bugfixes, it is definitely in maintenance mode, so it might be a bit of a stretch to say it's "safe forever." I think if anything it may be less safe than XWayland, but I'm willing to be proven wrong.
Now, what I'm really worried at the moment about is not XWayland going away, but rather dependency on specific toolkits. Things like GTK and Qt versions definitely have expiration dates, and at least with GTK, major versions are different enough from each other to cause problems. There is absolutely an expectation that eventually, you have to move to the next version or get consigned to the legacy bin. There is nothing you can target and have it be safe to run on most distros for a long time. There is nothing like Win32 you can target and be safe here. That breathing room between riding the upgrade treadmill hard and being consigned to the legacy bin is a pretty narrow space on Linux sometimes.
What I am hearing though, is that it seems like the Linux community would rather we go with Qt than adopt GTK4 if push comes to shove, and that if we stay on GTK they want us to stay on old versions of it (or forks of it) even if they no longer work on modern distros. Like, they're not even remotely interested in seeing what can be done with it, they aren't waiting to see what Cinnamon and MATE do with it, they've looked at modern GNOME and pretty much made up their minds. And staying on old GTK versions or forks doesn't sound so much like a long-term Linux strategy as it sounds like the path to becoming a toy for the GNOME counterparts of the kind of people that might want to do a CDE/Motif revival for fun in their own environment from time to time (yes that exists).
And from what I can tell, the reason they want us to go Qt is mostly because it means they get desktop integration in some form through Qt/GTK theme compatibility layers, whereas if we adopt anything that's not GTK or Qt they don't get that. And since GTK4 is hated, the gravity is towards the "other" Linux toolkit, Qt. And I can confirm there is a real trend of applications porting to Qt right now as GTK3 ages and people don't bite on GTK4... not all of them, but enough of them that this isn't coming out of nowhere. And from what I've heard, it sounds like Qt doesn't break stuff between major versions as much as it used to, which means porting to Qt6 could leave a path open to Qt7. And... there is such a thing as long-term support for older versions of Qt... but you have to pay for a commercial license to get it and it only lasts two years longer than regular support.
Honestly, I find the Linux community a bit... demanding at times. They want desktop theme integration, and to be treated like first-class users like Mac and Windows users are, which I find a bit absurd given how Linux is still a niche platform that's hard to support and develop for in a lot of ways. But they don't like GTK4/GNOME which is basically the descendant of what most applications have been using to get that sort of integration for a long time. Yet they also aren't all on board with KDE and use a lot of random stuff, including forks of older GNOME that thus far haven't shown a willingness to maintain their own toolkits and keep adopting newer GTK when the chips are down, meaning the direction is still the one defined by GNOME but slower and more fragmented? Basically, they have us in a frustrating corner. They want desktop theme integration, they don't like the modern versions of the tools that make that happen, and their solution is to have faith in legacy-friendly distros and hope everything works out somehow, without any clear path.
I guess from my perspective, it looks like this. Linux only worked as a psuedo-platform people could target because RHEL/GNOME was calling the shots enough that application developers could target something and know most distros would be able to run it. Not to say they were always fair or deserved their power, but they brought order. Now Linux users are in the process of rebelling against that order and asking people to plant their flag with their preferred faction and give up on supporting anyone outside their faction. And the major factions look like this:
1. Modern GNOME/Wayland advocates.
These are the people who are used to calling the shots and defining what Linux is. LSB was the modern POSIX and it was based on RHEL. They were trusted quite a bit up until GNOME 3 and later, started to lose influence slowly after that, but still have quite a huge following and a lot of software is built on foundations they ultimately control that no one else in the community has the manpower or resources to fork and maintain. XWayland is a compatibility layer they maintain for enterprise applications primarily, but many think they will drop it because they want to push Wayland... though I think they actually need to keep it in order to make Wayland viable. Anyone in this camp is probably on Fedora, latest RHEL, or at least a RHEL fork that they're happy with.
2. Cinnamon/MATE/XFCE group (GTK3-based desktop environments).
These people are not happy with GNOME, but seem willing to adapt to Wayland and ultimately have projects that are forks of GNOME and depend on GTK. They are the people who will "ride out" the deprecation wave and try to move towards Wayland and newer GTK at their own pace, probably writing a lot of custom widget code and their own versions of libadwaita like I observed I would have to do if I wanted a reasonable GTK4 port our users wouldn't hate too much. They are the people who might make GTK4 go from a total non-starter to something our users start begging us for 6 or 7 years from now because they want desktop integration on MATE/Cinnamon/XFCE who may well drop GTK3 at some point. Alternatively, if they can't or won't go GTK4, they may well be the force that keeps GTK3 on life support longer than a lot of people would like even on mainstream/modern distros. This is probably where Debian, Mint, Ubuntu, and a lot of mainstream Debian derivatives are living, and they probably expect GTK3 to have a fairly long life on their distros in some form, even if it's LTS or extended support or something.
3. Qt/KDE group.
These people have just never been part of the GNOME ecosystem, never used GTK by choice (web browsers often forced their hand), and Wayland was just another backend for them along with the dozens of others Qt supports. They pushed back on GNOME's vision for Wayland in some instances and insisted on pragmatism and it being a more neutral platform than it would have otherwise been. Their relationship to GNOME is more that of an old rival than someone who feels betrayed by them, they are pretty much in the same position in Linux they've always been in... the biggest GNOME alternative that isn't based on GNOME itself. These people are not that interested in preserving X11... because well, Qt doesn't need it, it's backend-agnostic from the ground up. They don't really cluster around specific distros, but they often gravitate towards the "KDE spins" offered by a variety of people as a GNOME alternative.
4. XLibre/XOrg fork group.
These people are actually somewhat outcasts in the Linux community and are viewed somewhat negatively by mainstream Linux, and some of them are straddling the Linux/BSD boundaries and wondering if they want to stay on Linux at all. These are the people noting that Wayland is Linux-first and leaves other Unix-like operating systems behind, and there's some overlap between them and the people who don't like SystemD (though it isn't universal). You see a lot of them on distros like Devuan, Artix, Gentoo, Hyperbola, etc. These people are halfway to going BSD and walking out the door, or at least building their own OS around the Linux kernel that will slowly and painfully diverge from mainstream Linux which is Wayland/SystemD/Linux as a base, and either GNOME or KDE on top. These people... are not
wrong per se, but they also are not the right people to ask for how to support Linux as a platform, because they reject the platform, want to build their own, and feel betrayed by mainstream Linux. These people will always use X11 no matter what, and are more likely to try and revive old toolkits, use forks, etc.
5. "Just use old Linux" group.
These people keep using LTS or Enterprise distros past their expiration date and always find someone willing to backport stuff to them as they need it. They don't understand the broader ecosystem or how anything is moving within it, they just... kinda like their old stuff, have it the way they want it, and know it broke on newer things, so they just keep using the old things without understanding. They are simple people and victims of a fast-moving ecosystem, but as bad as we might feel for them, we can't let them define what Pale Moon on Linux is. They have a lot in common with 4, except without the anger and the understanding of what's going on, maybe more just tired or apathetic than anything else.
The issue as I see it is... Linux as a platform is mostly defined by groups 1 and 2, but many in our userbase seems more aligned with 4 and 5, with the most realistic ones being people in 2 who sympathize with 4 and 5 and are trying to be diplomatic. Either people who hate mainstream Linux and are slowly moving towards their own platform, people who want it to move slower than it really does but adapt when needed, or people who just use old stuff and aren't worried about getting hacked because their setup working and feeling familiar is more important.
If it seems like the discussion steers towards Qt, it could well be in part because that group is the most "neutral" in all of this ironically, and thus the least offensive to all the other factions. Everyone else has an agenda that's diametrically opposed to everyone else's... ask a Wayland advocate who likes modern GNOME, they'll say the right time to go native Wayland was 10 years ago. Ask a grizzled old Devuan user, they might say no one should ever give in to Wayland because it's the same kind of force as SystemD, a control mechanism that must be resisted at all costs. Ask a person on Cinnamon or MATE who thinks they're headed in the right direction, they might well say to wait and keep to GTK3 and X11 until GTK4 and Wayland are stable for them, then move everything over after their preferred DE has figured out all the bugs. A KDE user? They'll just brag about how much easier their Wayland port was because Qt is good at supporting multiple backends, and how they managed to keep a more traditional desktop experience alive without giving up as much as GNOME did, while still getting the security benefits of Wayland. Will annoy GNOME fans a little, but no one else has a real problem with them because they don't have a dog in the same fight, and most can find some common ground with their perspectives.
So, that's the problem as I see it. Linux is less an operating system, and more just a divided community of opinionated computer nerds that hate each other and can't agree on how to do anything, but ultimately are either too lazy or too practical to follow though on most of their fork threats, and thus follow the money and the work being done by the majority eventually, just maybe 5 years slower than everyone else if they really don't like it and it causes drama for them with their users. RHEL/IBM/Fedora defines a platform that works for servers. All the other people involved complain about it, or start trying to make it work for them by building what they need to build on top of it to keep using that base, or try to drag out their resistance as long as possible before throwing in the towel. It seems like everything there is defined by drama, ideology, and trends more than sound engineering principles or a concrete plan.
"The Athenians, however, represent the unity of these opposites; in them, mind or spirit has emerged from the Theban subjectivity without losing itself in the Spartan objectivity of ethical life. With the Athenians, the rights of the State and of the individual found as perfect a union as was possible at all at the level of the Greek spirit." -- Hegel's philosophy of Mind
