GTK2 revival

[jarsealer]

I saw a YT video talking about the GTK2-ng fork:

https://inv.nadeko.net/watch?v=7S2qf8i5nWk

[Moonchild]

I saw a YT video talking about the GTK2-ng fork:

https://inv.nadeko.net/watch?v=7S2qf8i5nWk

Hate to say it but Brodie is one of those people who can never appreciate genuine effort being put into something. Here too he goes on talking down on making all the tools specifically available as forks, etc. But he forgets that without anyone to organise it, it will become the chaos it already was: individual projects and spin-offs that go nowhere because they were made by someone for their own use and that use might end in 2 years when they find another fav flavour of Linux or DE for the next few.

So, take it with a few kg of salt if you end up watching that video. For me, it feels a bit sour, but maybe I just don't like the way he tries to do a reality check without actually respecting people's efforts.

[jarsealer]

I saw a YT video talking about the GTK2-ng fork:

https://inv.nadeko.net/watch?v=7S2qf8i5nWk

Hate to say it but Brodie is one of those people who can never appreciate genuine effort being put into something. Here too he goes on talking down on making all the tools specifically available as forks, etc. But he forgets that without anyone to organise it, it will become the chaos it already was: individual projects and spin-offs that go nowhere because they were made by someone for their own use and that use might end in 2 years when they find another fav flavour of Linux or DE for the next few.

So, take it with a few kg of salt if you end up watching that video. For me, it feels a bit sour, but maybe I just don't like the way he tries to do a reality check without actually respecting people's efforts.

Don't have much to say, but he does make (daily) videos on niche topics or projects that I'm also interested / curious in, which is why I watch him some times.

[ownedbywuigi]

I saw a YT video talking about the GTK2-ng fork:

https://inv.nadeko.net/watch?v=7S2qf8i5nWk

Hate to say it but Brodie is one of those people who can never appreciate genuine effort being put into something. Here too he goes on talking down on making all the tools specifically available as forks, etc. But he forgets that without anyone to organise it, it will become the chaos it already was: individual projects and spin-offs that go nowhere because they were made by someone for their own use and that use might end in 2 years when they find another fav flavour of Linux or DE for the next few.

So, take it with a few kg of salt if you end up watching that video. For me, it feels a bit sour, but maybe I just don't like the way he tries to do a reality check without actually respecting people's efforts.

I hate, and I mean HATE people like Brodie (people who pretend to know what they are doing but in reality have a room-temperature IQ). They always fear-monger and don't appreciate effort being made into projects, so his lazy ass can just go
'OOOO LOOK FUNNY PROJECT, I HATE IT GRRRR LOOK GUYS MY OPINION ISN'T POPULAR'

[andyprough]

I hate, and I mean HATE people like Brodie (people who pretend to know what they are doing but in reality have a room-temperature IQ). They always fear-monger and don't appreciate effort being made into projects, so his lazy ass can just go
'OOOO LOOK FUNNY PROJECT, I HATE IT GRRRR LOOK GUYS MY OPINION ISN'T POPULAR'

When he first came on the scene, and I used to watch him almost from day 1, Brodie was one of the first of the people who wanted to be Luke Smith clones and grab a bit of that Luke Smith audience. Of course, Brodie has since gone very, very far away from Luke's ideal of computing being super-simplistic, minimalistic, keyboard-centric. Brodie's still a pretty funny watch every once in awhile though. He makes fun of himself as much as poking fun at projects. His best videos are when he tries to explain fights and drama within open source communities. Those can be pretty hilarious.

[Daemonratte]

...

Sorry for the late reply. I'm not getting any email notifications and just noticed there are many new comments.
We still didn't tidy up or test all the engines and themes, but we started collecting them here:
https://git.devuan.org/Daemonratte/gtk2-ng-themes
https://git.devuan.org/Daemonratte/gtk2-ng-engines

There are binaries available now too:
https://git.devuan.org/SteveM/DTK_debs

And greenjeans wrote some scripts and isos that make it way easier to test it
https://git.devuan.org/greenjeans/gtk2-ng-test-scripts
https://sourceforge.net/projects/vuu-do ... g-testing/

About the warnings: I fixed MANY more and not there are no introspection errors anymore. I wonder if the original gtk2 developers broke bindings for other languages and just never fixed it.

...

When I watched that video, I was really mad

"You preserved something. Meaning you didn't change anything."
"there's a bunch of random little [fixes] out there"
"a lot of the changes being done are either merging existing work that has been done"

First of all, when I started compiling gtk2 gave me 2mb of compiler warnings. Now I only get around 130 depending on how many threads I use. Fixing some warnings was easy. Others required me to rewrite code segments and then fix errors I got from the rewrites.
I also noticed that aparently they just let this toolkit ROT. There were bugs in the demos, in the build script, in the code, etc and a couple days ago I merged over 100 commits that fixed introspection errors. They literally messed up GIR metadata (you need that for bindings for Python, JavaScript, C#, Perl, Ruby, etc) and I digged through the code to fix it. I could have chosen to simply mute all the compiler warnings, but I fixed them.
So yes! The fact that we didn't break anything is something we can be proud of.
I also included fallbacks so you can compile it on anything from Debian 8 to bleeding edge distros like Arch and Gentoo.

Him saying that all we did was merging existing patches also really pissed me off. Just look at how many commits I added. Hundreds. There aren't hundreds of patches out there that I could have applied.

Not trying to be a kissass, but all of this let's me respect your work even more, because now I know exactly what you had to endure for years.

...

That's also how I found out about him. In the beginning he really just copied Luke Smith.

[athenian200]

If I seem less than enthusiastic about this, I should clarify that it's not because I have the mentality of Brodie or those who don't think this is a lot of work. It's actually a bit more personal, and I want to clarify before people start assuming I'm as bad as him.

It's... basically, I never shipped my SunOS/Solaris builds with GTK2. I used GTK3 from the beginning, used to complain as early as 2019 that it was nonsense that we had to have two Linux builds just because some people prefer an old toolkit because it looks better with themes or something (before we dropped 32-bit builds, I think it was 4 Linux builds). Always thought it was so weird we dropped 32-bit builds in favor of keeping GTK2 because the community here was so obsessed with it. I never shipped Epyrus with GTK2, and got annoyed with people who kept asking for it. Even did a Python 3 port of the Python 2 codebase in part so people self-building Epyrus to have a GTK2 version couldn't throw it in my face that building UXP was "too hard" because tracking down Python 2 or Tauthon is difficult, SSL doesn't work, they don't know which branch to build against, they don't know how to use a Docker container, etc. Plus, I was hoping supporting Python 3 might draw more modern Linux users whose distros don't ship Python 2 anymore into contributing. I always wanted people to fix themes so they look better on GTK3, or just... do something more reasonable than constantly repeat their old saw about GTK2 being "the best version" of our application and everything new being bad like a bunch of people at a high school reunion who feel like they peaked in high school and have nothing good to say about their lives since then.

And now... instead of newer contributors or people who will help us target modern Linux... we get, well... this. Pulled right into the firestorm of GTK2 revival energy, right when I thought people would start taking GTK3 more seriously and finally realize we can't do GTK2 forever, and I have to get demoralized seeing just how many people on this forum have more in common with the values of the types on VOGONS or MSFN or whatever than I was hoping. The timing just made it feel like a giant middle finger from the universe, an admonishment, "No, sorry, sir, you are working on a project for old people that like old stuff, you don't get to move it forward, you don't get to do anything except watch them worship their idol called GTK2, and scoff at your GTK2-atheism, at your non-belief. The GTK3 version won't be made better, there won't be a port to anything newer, this new fork will be used as an excuse to ignore modern Linux, all your work is going in the retro Linux bin as a toy for Slackware and Devuan users to play with. Sorry, that's what you get for working on UXP and thinking it was worthy of a serious investment of time! The kind of people that surround your project won't ever do anything but drag you backwards and scold you for wanting to support modern platforms, getting excited about every retro fork and lashing out at every deprecation. They'll keep pining for GTK2, for SSE2-only builds, and Windows XP, and never forgive you for moving forward. That's your potential userbase, and it's not the one you wanted, not the one you were working this hard for. Deal with it."

But... being honest, I guess GTK2 existing is a good thing. There was nothing wrong with GIMP's old interface and I'm sure a lot of old games depend on it. I guess as selfish as this sounds, I only care about what it means for us, and I just don't like the fact that we're being dragged backwards and told it's okay to never change, never adapt, become irrelevant to anyone except people who like retro desktops as novelties. If it was only happening to other projects I hadn't invested in and didn't care about futureproofing? I'd probably think it was kind of funny and not care. And it's not Daemonratte's fault... but unfortunately he's opened a window into something I didn't want to look at, and now I'm seeing how far apart my values and goals are from a lot of people in this community, and feeling like I'm wasting my time trying to bring anything forwards. Because the people here just want the past preserved pixel-for-pixel, bug-for-bug, they don't want the spirit of the past carried forward and kept relevant on newer platforms. I don't know what I was expecting... I guess I liked to think other people had a similar vision, that they understood, but instead I'm surrounded by backwards-looking mechanics who remind me too much of the type of man who refuses to buy a car made after 1970 on principle, and rebuilds the carburetor every two weeks because he can't find replacement parts but is just that stubborn. He has something bad to say about no 8-track tape player, about air bags, about power steering, about automatic transmissions, about power windows he can't roll down... can't see any good in anything after his era. That man is not a visionary... he's just stubborn and likes old things and can't see any good in anything new. Maybe because of trauma, maybe because of nostalgia, I don't know, but it's not rational, it's not healthy, and it's not something I want to spend my life catering to.

If this was a fork of GTK3 or something still in wide use, I would get it... but this is something that was deprecated 6 years ago that's been on the chopping block of a lot of distros for a long time. I just... I don't know anymore. I was already sick to death of the GTK2 forever brigrade, and now they have a huge advocate giving them a shot in the arm. Again, Daemonratte didn't do anything wrong... he just showed me something that's really hard to look away from about this community and took a lot of the wind out of my sails.

[greenjeans]

Hi athenian200, i'm greenjeans, one of the folks helping Daemonratte with the project, I read your post and was very moved by it, enough to want to register here so I could offer a message of hope.

For myself when we started our new forum and began these projects, I had been working solely with gtk3 in my app development, so gtk2 wasn't my first priority and I had lobbied some for forking it instead. But I do love choices, so when Daemonratte asked me to help out I was honored and happy to help in my limited capacity and skill level.

But see gtk2 isn't the only thing our team is thinking of forking/reviving, we have a LOT of things on the table and at some point when this current project has gone stable I want to actively work on doing the same for gtk3. In a perfect world they'd eventually meet in the middle and be a nice universal toolkit, but these ideas and projects are rapidly evolving, so we'll see what happens. But let me assure you as we did for gtk2: GTK3 is not dead yet, we will not let it go gentle into that good night.

So first post and all, sorry I didn't do an introduction, but I also want to say kudos to the whole Palemoon team, I spent a great deal of time yesterday running a slightly older gtk2 version on gtk2-ng, and it was awesome! I used it in fact for the test iso, and plan on replacing chromium with it in my distro project.

[andyprough]

Again, Daemonratte didn't do anything wrong... he just showed me something that's really hard to look away from about this community and took a lot of the wind out of my sails.

I was reading through all of Daemonratte's posts in this and the Qt6 threads last night and getting the opposite impression. Seems to me that part of what Daemonratte is saying is, "you should go forward with Pale Moon development and exploring new toolkits, I've got the gtk2 tail covered - I'll make my own gtk2-ng builds and add things into them that weren't previously available, I'll revive Galeon and work on other gtk2-ng UXP projects and gtk2-ng Desktop environments myself" (which, if you look at the original Galeon objectives is a pretty cool idea). gtk2-ng, if it is going to be successful, is going to need a browser to showcase what it can do. You should definitely be focusing on future technology, because UXP still has so much untapped potential. Let the people who are passionate about the older technologies keep UXP alive in those arenas. Not your problem I don't think.

Doing anything else would be like worrying that you have to support KDE3 because the Trinity Desktop Environment exists, but that's never been the case. Let the TDE folks deal with keeping KDE3 alive and in use.

[Drugwash]

Sorry for the late reply. I'm not getting any email notifications [...]

Don't worry, it happens, no biggie. Happened to me too in the past especially with forums of different formats than I knew, or with different defaults.

We still didn't tidy up or test all the engines and themes, but we started collecting them here:

Yeah, I noticed the new themes and engines repos incidentally a few days ago. Cloned them then but didn't have time to look through the code or try to compile them. Will do at some point.

greenjeans wrote some scripts and isos that make it way easier to test it

Will look at that too, soon. Just woke up and there's e-mail to take care first, forum posts to read (and reply if needed).

About the warnings: I fixed MANY more and no there are no introspection errors anymore.

That's great! I'll compare to my own previous fixes, see what it's been done and what I could've done [better].
By the way, your editor/IDE doesn't seem to automatically strip whitespace, or maybe you didn't get to go through all files. It's annoying sometimes - OK, everytime - to make a small change to a file and upon comparison with the original to get dozens or hundreds of mismatches just because the original had a truckload of whitespace. I wonder, were all those empty lines containing only spaces or tabs counted in as lines of code, back then...?

I wonder if the original gtk2 developers broke bindings for other languages and just never fixed it.

Could be, anything's possible. Maybe it's my paranoia but I always thought programmers intentionally leave their previous project in a bad state, maybe adding some bad commits right before throwing it aside, in order to justify the existence of the new and shiny version of it. Although in all fairness it's also possible that the old concept was so bad and/or the code was so screwed up that it simply wasn't worth the effort of fixing it for whatever retro fans might've had at some point in the future. Speaking generally, not particularly about Gtk2.

When I watched that video, I was really mad

Me too, and I don't even have a dog in this fight. Dunno who that guy is, never watched him before, but it stroke me right off the bat as a douchebag that speaks a lot without saying anything [of value]. He can't even read or spell correctly FFS! Don't mind little dogs barking, go about your business!

[...]

I've been where you are, in a way, many years ago. Always looking forward to the future. That's because back then the future genuinely seemed to bring good things. Useful things. Fixes and improvements. New features I wanted/needed. And for a while - maybe too short a while - it was true. I was happy. And then at some point things started going backwards. Those good things started to be hidden behind [invisible] guards, or were even removed completely. At best they were crippled enough to make them undesirable or unusable.

At that point I said to myself: «fuck the future, it's gonna destroy everything that's been good so far; gotta stick with whatever still works as I want/need». And I managed to do that for some time. Put in some effort, sometimes a great deal of effort just to stay afloat in a sea of "modernism" that was trying to take away all the good things the past had given to me/us. But slowly I had to give in. First to one thing, then to another, and another, and another until I got to the point that I had to completely shift the paradigm. It was sad and dire. But I survived. Reborn from the ashes I adapted to something new and different. I worked hard to get things back to the level I wanted/needed. It was fine for a while - too short a while. Then change came again upon me.

You know how it felt to me? As if the builders who built my house came to my door and said: «get out, we gotta tear down your house and build a newer, better one». And I was: «why? I'm just so happy here, finally adjusted [to] it, and it became all I was dreaming of. why would you destroy the house I've been working so hard to fix and decorate to my liking?!» And they said: «oh, that's just how things work. your house is obsolete. move on.»

I understand why this is happening. Not gonna say it out loud but I guess/hope people would suspect the real reason. Do I have to agree and comply? Hell no. But that's gonna cost me. Dearly at some point. Or maybe I won't live enough to get there.

Dunno how old you are, and would hate to make assumptions. Maybe you're still in that phase where the future seems bright and brings a lot of good things so you'd gladly let go of the past. But I believe there will come a day when you realize this future was deceiving, and it took away much more than it gave. And you'll want to stick with the little good things you still have. You'll wanna live, or at least to survive, to be left in peace in your old dear house. Just like I (and many others like me) want[ed] to. But you won't be able to - not for too long. Because... oh, that's just how things work. And you'll remember: what goes around comes around.

I hope we all come to peace with all these changes and requirements and whatnot. The last thing we need is to get more fragmented than we already are. As a community, as a society.

[gabrgv]

There are binaries available now too:
https://git.devuan.org/SteveM/DTK_debs

Any changes we'll get binaries for Daedalus? I didn't/won't move to Excalibur due to usrMerge and... GIMP 3.

Also, I though it was a really wise move you did when rejected Wayland support.


And, @athenian200, relax, at this point, "deprecation" is just a buzzword made up by corporate big-shots and marketeers to fright people off stable paths they don't want them to follow, because being stable is bad for their business. The men that follow this trend are either unaware of better options (but being a hardcore smartphone user helps here, too) or neophiles/hipsters anyway (the devs, mostly). That's what "modern Linux" is for. But don't forget that, while we cheer for this most-welcomed non-deprecation of GTK2, a Qt6 port of UXP looks to be on its way too.

It really looks like two clashing views here: the "community", the hackers, the men, against the companies who need to make their money one way of another, and do it by, on one hand, pushing half-baked shit into everyone's face, and, on the other, deprecating what currently works just to be sure people won't stick to the "old and insecure". To me, UXP aligns itself better to the former.

Well, if it's...

on principle

, why it cannot be rational?

By the way, how do you think NPAPI figures in here, Flash Player being deprecated about the same time as GTK2?

[athenian200]

And, @athenian200, relax, at this point, "deprecation" is just a buzzword made up by corporate big-shots and marketeers to fright people off stable paths they don't want them to follow, because being stable is bad for their business. The men that follow this trend are either unaware of better options (but being a hardcore smartphone user helps here, too) or neophiles/hipsters anyway (the devs, mostly). That's what "modern Linux" is for. But don't forget that, while we cheer for this most-welcomed non-deprecation of GTK2, a Qt6 port of UXP looks to be on its way too.

It really looks like two clashing views here: the "community", the hackers, the men, against the companies who need to make their money one way of another, and do it by, on one hand, pushing half-baked shit into everyone's face, and, on the other, deprecating what currently works just to be sure people won't stick to the "old and insecure". To me, UXP aligns itself better to the former.

Yeah, honestly I hope the Qt6 port (or similar) happens, it would be a bit of a relief for me... getting a Qt version working would likely buy us more time as far as working on modern distros, and there's a chance people would like it better than newer GTK.

I don't think it's really as simple as you make it out to be, and people have been saying that for years. To me it seems like the "community" just maintains the long tail and keeps the old, stable stuff going a little longer while the less stable stuff is adopted at the fast-moving, corporate end first. There's an element of truth, but the world is not as simple as "corporate bad and fast moving, community good and stable," at least not in my book. If we disagree, we'll just have to agree to disagree. But what I usually see is, eventually the newer stuff stabilizes and comes downstream, so it's not like you avoid the future forever... you just delay having to deal with it and maybe get a better version of it than if you'd been forced to adopt earlier, a little sugar to make the medicine go down. I also feel like this glosses over very real security implications in a way that I have repeatedly seen people get burned by because of this very way of thinking.

Like, it's not that we intend to drop support for the old stuff (unless it genuinely becomes a burden or gets in the way of something), but I think ignoring where things seem to be headed and the mainstream while refusing to prepare for them just because they seem to suck isn't a very smart way to do things. It's never been a winning strategy in my experience, and I'm not keen on the whole "give up on modern Linux because it's too hipster, just trust the community hacker rebels bro," mentality, because I've never seen anything good come out of thinking that way and betting the farm on that sort of thing.

By the way, how do you think NPAPI figures in here, Flash Player being deprecated about the same time as GTK2?

Off-topic:
Honestly, I find it rather annoying because GTK2 is not a dependency I want us to have on Linux. I've actually been experimenting with trying to get NPAPI plugins to work with raw Xlib/XEmbed in situations where GTK2 isn't available. I have had some success, but Flash Player is among the more stubborn ones. I know this a little off-topic, but I'm even looking into stuff like Pipelight (de-GTKed of course) or whatever to get the Windows version working on Linux in the future because I don't trust GTK2 to remain available on most distros and really don't see being relegated to retro-friendly distros as a future I want, even if I'm grudgingly appreciating the stopgap existing if mainstream distros pull the plug too fast, I still want the long-term strategy to be supporting both modern and retro Linux rather than being chained to the latter and having to put all our faith in it. The nuclear option (which I've already added as a build flag and which has been used on some distros already), is just disable NPAPI on Linux to build with pure GTK3 and tell Linux users to suck it up and it's not our fault they don't have a stable ABI, though we would still have NPAPI on Windows and Mac even in this case. Truth be told, Flash Player is sometimes half-busted even with GTK2 available due to modern glibc issues, so you'd need an old glibc to run it reliably in some cases as well. Linux is just... not friendly to its own native binaries, and honestly I'm not sure even GTK2 forever would save existing NPAPI plugins that can't be recompiled on Linux long-term given what I've been seeing with glibc lately.

[athenian200]

Hi athenian200, i'm greenjeans, one of the folks helping Daemonratte with the project, I read your post and was very moved by it, enough to want to register here so I could offer a message of hope.

For myself when we started our new forum and began these projects, I had been working solely with gtk3 in my app development, so gtk2 wasn't my first priority and I had lobbied some for forking it instead. But I do love choices, so when Daemonratte asked me to help out I was honored and happy to help in my limited capacity and skill level.

But see gtk2 isn't the only thing our team is thinking of forking/reviving, we have a LOT of things on the table and at some point when this current project has gone stable I want to actively work on doing the same for gtk3. In a perfect world they'd eventually meet in the middle and be a nice universal toolkit, but these ideas and projects are rapidly evolving, so we'll see what happens. But let me assure you as we did for gtk2: GTK3 is not dead yet, we will not let it go gentle into that good night.

So first post and all, sorry I didn't do an introduction, but I also want to say kudos to the whole Palemoon team, I spent a great deal of time yesterday running a slightly older gtk2 version on gtk2-ng, and it was awesome! I used it in fact for the test iso, and plan on replacing chromium with it in my distro project.

Just saw your message. I'm glad you guys aren't upset with me for not being as enthusiastic about the GTK2 revival as everyone else, and seem to get where I'm coming from. Yeah, I really am way more worried about GTK3 deprecation than anything else. Because we already have a strategy to survive GTK2 going away (I seem to recall Arch dropped it years ago and that's about when I implemented the flag to disable NPAPI for self-builders who don't want to fight with GTK2 dependencies), but knowing how these things work, that means GTK3 is next on the chopping block and that means I have to worry about possibly Debian 16 or 17 not having it? It was stressful because it meant we went from having one legacy toolkit and one supported toolkit... to a dead toolkit and a legacy one with no further fallbacks unless we get a new one working ourselves.

My discomfort was actually more based on the way people who like GTK2 tend to act in our community at times, being a little exasperated with that, and that biasing my perception of what you guys are doing. Though I will say, there was actually a guy a long time ago who wanted to do something like what you're talking about... he was called The_Squash, and it seemed like he tried to do a sort of merged GTK2/GTK3 thing, but it wound up being vaporware and not really getting off the ground. There was also a guy named mrnhmath who I think wanted to fork GNOME 2 at one point, though I don't know if that went anywhere (don't think it did). It sounds like you guys are a bit more serious about this, and are possibly creating a whole distro around the idea.

[andyprough]

There's an element of truth, but the world is not as simple as "corporate bad and fast moving, community good and stable," at least not in my book.

The development money is almost all being poured in by multi-trillion dollar corporations whose customers are data centers and huge corporate server farms. IBM and Google and Facebook and whoever else is paying the salaries for Gnome and systemd and Wayland engineers do not give a single thought to the home desktop user experience, they just want a simple Gnome desktop environment that's easy to roll out across thousands of seats at a time, is rarely used by mostly headless data center installers, and is easy to support. And hardly anyone uses the standard Gnome corporate desktop as a home desktop. Ubuntu heavily modifies it, and supports many alternative desktops. Debian's popcon shows that most of their users don't use the Gnome desktop. Mint and MX, two of the more popular home user distros with installed bases in the millions don't even offer a Gnome version, because none of their users are asking for it. So there's a huge disconnect between what gets corporate development, and what home users ultimately end up using.

I think the gulf between what the trillion dollar corporations want on the desktop and what the home users want is widening. So rather than a convergence on standards, the next decade is likely instead to see un-healable rifts. Following the Qt path may be a good middle ground, as it's always been the primary alternative and is widely supported in every popular distro.

[athenian200]

The development money is almost all being poured in by multi-trillion dollar corporations whose customers are data centers and huge corporate server farms. IBM and Google and Facebook and whoever else is paying the salaries for Gnome and systemd and Wayland engineers do not give a single thought to the home desktop user experience, they just want a simple Gnome desktop environment that's easy to roll out across thousands of seats at a time, is rarely used by mostly headless data center installers, and is easy to support. And hardly anyone uses the standard Gnome corporate desktop as a home desktop. Ubuntu heavily modifies it, and supports many alternative desktops. Debian's popcon shows that most of their users don't use the Gnome desktop. Mint and MX, two of the more popular home user distros with installed bases in the millions don't even offer a Gnome version, because none of their users are asking for it. So there's a huge disconnect between what gets corporate development, and what home users ultimately end up using.

I think the gulf between what the trillion dollar corporations want on the desktop and what the home users want is widening. So rather than a convergence on standards, the next decade is likely instead to see un-healable rifts. Following the Qt path may be a good middle ground, as it's always been the primary alternative and is widely supported in every popular distro.

Yeah, I'm not saying you're wrong per se, but here's the thing... while home users may not be running GNOME as it's shipped upstream, they are definitely running MATE and Cinnamon, which are downstream of GNOME. MATE at least, was originally a fork of GNOME 2, but eventually adapted to GTK3. And from what I have seen of GTK4 (I studied it a bit), they WILL be able to adopt it after GTK3 is deprecated because it is designed to work for both GNOME and what GNOME forks want (see the other thread about the Qt port, the common assumptions about GTK4 being GTK3 but worse and a GNOME-only toolkit are slightly wrong), but it will be work for them. Because of that, they probably won't be able to justify a GTK3 fork long-term, however there is the hope that if someone steps up to maintain GTK3 before it's deprecated, they might decide to stick with it for longer. MATE went GTK3 well before GTK2 was deprecated upstream at GNOME in 2020. However, I do hope that MATE and Cinnamon can converge on a shared "libclassic" as a libadwaita alternative for their GTK4 version rather than both developing independent alternatives.

So the point isn't so much, "I want UXP to target modern GNOME" or anything like that, but rather, "what GNOME does eventually impacts downstream forks of GNOME which are much more popular." Basically, it's precisely because the development dollars come in from upstream and are poured into server distros that most of the downstream distros (and downstream forks of GNOME) eventually have to use the pieces they are given to build on top of RHEL's foundation or else become increasingly broken, fiddly, and left behind with the amount of tech debt from staying behind increasing year on year. Like, we're seeing that already with MATE, Cinnamon, XFCE looking into Wayland, wlroots and similar trying to give people some of what they had on X11 with Wayland, etc. wlroots is the community's "spoonful of sugar" to make the Wayland medicine go down. And it's like... as much as it may annoy users for me to say this, from a developer standpoint, targeting all the GTK3 desktops is similar to targeting GNOME 3. Users may like MATE and Cinnamon better because of creature comforts and classic feel, but under the hood from the perspective of an application developer, targeting them feels a lot like targeting GNOME 3, especially in the case of Cinnamon. So really we just support what we internally think of as "GNOME family of desktops" with GTK3, which includes MATE, Cinnamon, XFCE, and GNOME 3 itself of course. From our perspective, these are all GTK3 desktops that are all "spiced/riced" a little differently for the taste of users, but that customization doesn't really change the API or what we have to do much in the end. And if they go GTK4, then that still requires a lot of the same work for us as if we were targeting modern GNOME, even if from the user perspective they manage to put in the "sugar" so that everything looks similar or maybe even better for them than it did with GTK3.

My point is, what the user sees on the surface from "cooler" distros/DEs that look radically different from GNOME is often quietly built on that very same RHEL stack with several coats of paint added, not usually or always a fundamentally different architecture. Unless they use KDE, which really, truly is its own stack. But most of the others are just GNOME forks trying to provide a cooler experience for users and buffer them from upstream a bit until they can adapt their stuff. That probably feels dismissive of their efforts, but that is honestly how the average developer feels about it.

[andyprough]

And it's like... as much as it may annoy users for me to say this, from a developer standpoint, targeting all the GTK3 desktops is similar to targeting GNOME 3. Users may like MATE and Cinnamon better because of creature comforts and classic feel, but under the hood from the perspective of an application developer, targeting them feels a lot like targeting GNOME 3, especially in the case of Cinnamon.

That's all well and good, but we're on Gnome version 50 now, and anything on GTK3 is already ancient history if you are talking in terms of future-proofing your project. As i said, the disconnect between Gnome development and actual home desktop use is widening so much that at some point something is likely to break badly. Anyway, Qt looks good from that perspective, but I'm not the one that's going to do all the hard work, only you can decide what's worthwhile for you to do.

[athenian200]

That's all well and good, but we're on Gnome version 50 now, and anything on GTK3 is already ancient history if you are talking in terms of future-proofing your project. As i said, the disconnect between Gnome development and actual home desktop use is widening so much that at some point something is likely to break badly. Anyway, Qt looks good from that perspective, but I'm not the one that's going to do all the hard work, only you can decide what's worthwhile for you to do.

Yeah, that sounds like we pretty much agree, we're just coming at this from different angles.

But yeah, I am definitely well aware (painfully aware and more so than I've wanted to let on for years) that MATE and Cinnamon are buying us time on Linux already... if it weren't for those, distros might have dropped GTK3 already by now and we would have been left with no way to run on modern Linux period. I was honestly a little worried while I was in college that when I was done and able to get back to UXP development, I'd come back to see that everything was basically unusable for end users on Linux.

[mr tribute]

That's all well and good, but we're on Gnome version 50 now, and anything on GTK3 is already ancient history if you are talking in terms of future-proofing your project. As i said, the disconnect between Gnome development and actual home desktop use is widening so much that at some point something is likely to break badly. Anyway, Qt looks good from that perspective, but I'm not the one that's going to do all the hard work, only you can decide what's worthwhile for you to do.

Yeah, that sounds like we pretty much agree, we're just coming at this from different angles.

But yeah, I am definitely well aware (painfully aware and more so than I've wanted to let on for years) that MATE and Cinnamon are buying us time on Linux already... if it weren't for those, distros might have dropped GTK3 already by now and we would have been left with no way to run on modern Linux period.

I like reading your posts, but I think you worry about these things about a decade too soon. Cinnamon, MATE and XFCE have no stated goals or road-maps for a gtk4 transition. They have enough trouble with Wayland. IMO the really important thing concerning Pale Moon is web compatibility. Everything else pales in comparison.

Ubuntu 26.04 is out with a fully intact gtk2 stack. Ubuntu 26.04 is supported 10 years for free (maybe even 15 years). So Ubuntu users will have gtk2 intact at least until 2036. What that means is that 16 years after being declared dead by Gnome, gtk2 will still be supported by Ubuntu. Ubuntu is by far the largest distro by user base - Ubuntu and official flavors, Linux Mint and Zorin OS. Fedora is very small compared to that. Arch will likely always have gtk2 in AUR. The only question mark is Debian which is still discussing the fate of gtk2 in official Debian 14 repos (which will be released summer/autumn 2027).

So in essence gtk2 will be supported until 2036 (the only possible big exception being official Debian repos). Gtk3 isn't declared dead yet. Gtk5 is expected in the 2028 - 2030 time frame. Add 16 years to 2028 and gtk3 will enjoy major distro support until 2044.
Gtk3 is a great toolkit because it is mature. I don't think there will ever be better toolkit-integration for Pale Moon on Linux desktops. If Pale Moon is still developed in 2044 I think you have won either way and possibly the toolkit problem has solved itself. It is possible that both gtk and Qt are dead-ends as universal Linux/nix toolkits.

[andyprough]

So Ubuntu users will have gtk2 intact at least until 2036. What that means is that 16 years after being declared dead by Gnome, gtk2 will still be supported by Ubuntu

No, gtk2 already does not get security or any other updates, and by 2036 Ubuntu 26.04 will only be getting critical security updates to things like the kernel. Also, you are missing @athenian200's main point - he doesn't want to be the guy working on the browser whose users have to run 5 or 10 year old versions of Ubuntu to run it. And if your developers don't want to develop it then you can either fork it or not have it at all.

[athenian200]

I like reading your posts, but I think you worry about these things about a decade too soon. Cinnamon, MATE and XFCE have no stated goals or road-maps for a gtk4 transition. They have enough trouble with Wayland. IMO the really important thing concerning Pale Moon is web compatibility. Everything else pales in comparison.

I tend to agree, in fact that's why this hasn't come up until recently. Debian talking about dropping GTK2 created quite a stir and got everyone in the Linux community started worrying if GTK3 is next or not, with the nearby GNOME 50 release timing (the one that finally drops the XOrg session) not helping matters. Add to that this thread existing and people talking about GTK2 all over again, and now it's on my mind because it was brought to the foreground. LOL.

So in essence gtk2 will be supported until 2036 (the only possible big exception being official Debian repos). Gtk3 isn't declared dead yet. Gtk5 is expected in the 2028 - 2030 time frame. Add 16 years to 2028 and gtk3 will enjoy major distro support until 2044.
Gtk3 is a great toolkit because it is mature. I don't think there will ever be better toolkit-integration for Pale Moon on Linux desktops. If Pale Moon is still developed in 2044 I think you have won either way and possibly the toolkit problem has solved itself. It is possible that both gtk and Qt are dead-ends as universal Linux/nix toolkits.

Yes, I would actually say GTK3 has a better chance of surviving as something of a "near-permanent toolkit" than GTK2 does. Most applications have already bitten the bullet and done a GTK3 migration or gone Qt if they couldn't stand it. MATE, Cinnamon, XFCE, and even GNOME's own GNOME Flashback all rely on GTK3 currently, putting aside a small but vocal contingent of actual GNOME 3 fans who still prefer it over modern GNOME. That's 5 desktops using it as their main toolkit. There was maybe GNOME 2 and MATE briefly using GTK2. Also, GTK3 supports Wayland (which doesn't help us, but does help GTK3 as a platform). GTK2 was X11 only pretty much. That's another point in favor of GTK3. It supports X11 and Wayland reasonably well, while GTK2 is X11 only, and GTK4 is significantly better at supporting Wayland than X11. So the ecosystem being in flux between X11 and Wayland, and needing something that supports both reasonably well the thing that could keep it alive longer than GTK2, if distros care enough and are sensible. Distros might decide to give users some mercy on the GTK3 front given how slow adoption of GTK3 was to begin with and how hard they had to fight to get people off GTK2, and how chaotic the Wayland transition has been (and frankly I hope they do).

Just bear in mind that I think like a long term planner (in case it wasn't already obvious). Like, in MBTI terms my dominant psychological function is Ni (introverted intuition)... if you look that one up, it will explain everything as to why my brain works like this, but for this context, all you have to know is it means my natural tendency is to look at long-term ripple effects, patterns, and anxiously try to make future plans based on them rather than live in the past or the present. It's probably why GTK2 sticks in my craw so much... for a person who thinks the way I do and tries to prepare for worst-case scenarios and deprecation hits before they happen, discussion of GTK2 feels like "noise in the signal" from people who don't see the bigger picture and think primarily of their own comfort, it doesn't give me useful information about the future or help me adapt to a chaotic world. Which is probably why I wound up overreacting to this thread. My mind was screaming at me from the moment it was posted to push in the opposite direction and seeing this as danger, which makes a lot of sense once you know how my mind works.

No, gtk2 already does not get security or any other updates, and by 2036 Ubuntu 26.04 will only be getting critical security updates to things like the kernel. Also, you are missing @athenian200's main point - he doesn't want to be the guy working on the browser whose users have to run 5 or 10 year old versions of Ubuntu to run it. And if your developers don't want to develop it then you can either fork it or not have it at all.

Yeah, you basically get it. It's like... on Linux, I'm sure people who run LTS distros or know how to self-compile stuff can run much, much older stuff than anyone expects. But like... to put this in context. There are niche CDE revival projects out there, and a lot of people who really want that find ways to get it working. But few or no distros actually ship this. Same with TDE and MiDesktop's forks of older Qt they use for a retro KDE environment.

So from my perspective, it wouldn't make a lot of sense to target a niche like CDE revival forks, or an older KDE fork like TDE, and say you "support Linux" just because someone who tries hard enough could get your application working on Linux under some circumstances. And I guess I just don't really see why a fork of GTK2, or even a whole fork of GNOME 2, would move the needle much. TDE is cool, CDE revival forks are cool... don't get me wrong, but you don't exactly see application developers with cross-platform applications still in development targeting them and then saying that's how they want to support Linux as a whole.

My skepticism isn't that they haven't put in the work or what they're making isn't usable... it's more that it's going to be very niche and won't see mainstream adoption like MATE or Cinnamon did, because those projects took a very careful middle ground and supported a GTK3 platform just new enough that they seemed to have a future and could spin it out for a long time. I guess I really should have just been blunter from the start and said... I think this GTK2 revival will end up like the GNOME equivalent of TDE or MiDesktop, which are retro KDE forks based on hard forks of older Qt. Sure, they work. Sure they have their fans. Yeah, they're fun to play around with and take screenshots of. Maybe if you're a power user, you can even make it work as your environment for a time. But no, they didn't move the needle as far as what most Linux distros ship and what toolkits people expect applications to target. I feel like that's almost rude to say given how hard I'm sure they worked (and believe me, I respect the work), but it's hard to see how that wouldn't be the case.