Anyone interested in a patch to disable GTK2 on distros that don't support it?

[Moonchild]

Alternatively, make the IPDL as a whole a copy (i.e. have one for +npapi and one for -npapi scenarios) and do the conditional stuff in moz.build? That way you don't have to preprocess and it should Just Work™
You're looking at the hacky mess Mozilla had to do to prepare for multi-process that we unfortunately inherited a good chunk of (the whole parent/child split architecture)

[athenian200]

Alternatively, make the IPDL as a whole a copy (i.e. have one for +npapi and one for -npapi scenarios) and do the conditional stuff in moz.build? That way you don't have to preprocess and it should Just Work™
You're looking at the hacky mess Mozilla had to do to prepare for multi-process that we unfortunately inherited a good chunk of (the whole parent/child split architecture)

That seems like a much easier way, probably what I am going to do. It's so bizarre, though. I have now tested it on multiple Linux systems and even OpenIndiana/Solaris, and I can't reproduce dbsoft's build failure. Maybe there's some kind of race condition in the build system that he's hitting but I'm not, simply because our computers work at different speeds. Python 2 is old... maybe the lack of updates to the underlying thing is finally catching up with us. Sure wish I knew why, but it's probably easier to just do it without preprocessing rather than debug this.

[Moonchild]

remember I did run into something similar before not having it in the right moz.build category for preprocessing (I'm sure it's not that simple, right?)

[frostknight]

Linux is far from a safe haven from that sort of thing. They almost do the Chromium thing where they kill off smaller projects with constant API churn and breaking backwards compatibility, but with the whole OS rather than just the web.

You can thank redhat for that problem. They have been infecting linux for 15+ years possibly longer.

Their software additions including systemd, have been a really stupid unnecessary additions.

This is where BSD it least has some sense... if its OpenBSD for sure. Not sure about NetBSD and others though.

Well, I don't know specifics is what I mean. BSD supposedly is against constantly changing the code by adding over-engineered garbage. NetBSD is probably fine though.

Of course OpenBSD while minimal might have other problems such as scrapping code too often that is bloated. Which can mess stuff up potentially.

[athenian200]

You can thank redhat for that problem. They have been infecting linux for 15+ years possibly longer.

The truth is, in my view, all desktop Linux that isn't RHEL is basically just quirky RHEL anyway, and can only stray so far from it while still continuing to work. It seems like every other distro does what they do a couple years later, and at least with them you get a longer support window for the older stuff, whereas with most of the others you're expected to move to a newer base a lot sooner. If you want something that's really even close to being fundamentally different, you have to get away from Linux entirely. Even there, you already see Red Hat influence in the form of GTK dependency. X11 as a project is basically what has been keeping the other Unix systems on life support and letting them go their own way somewhat. It was an area of common standards from the last Unix wars that had survived for decades somehow. Wayland is basically going to kill a lot of them as anything other than terminal-mode server operating systems. But more realistically, I think we're going to see a lot of non-Linux systems stuck on X11 and GTK3 for a long time, even after those projects go unmaintained.

Their software additions including systemd, have been a really stupid unnecessary additions.

SystemD is a good example of what I mean. Fedora moved to SystemD, and every other distro followed shortly. I personally liked Upstart as an alternative to standard init, but it seems like the entire Linux community circles the wagons around Red Hat, always dutifully killing alternative ideas and shoving Red Hat's down everyone's throat out of some weird loyalty. Like... I do get why they don't like the idea of having to support multiple standards on Linux and fear fragmentation, but they effectively wind up serving the people forcing the change solely because pushing the smaller groups into submitting to Red Hat is easier than pushing Red Hat to listen to the community. Basically, when Red Hat says jump, every other mainstream Linux distro asks "how high?". Are there some other obscure distros that go their own way? Sure, but they can only go their own way so much before they start becoming incompatible with most Linux software. And Red Hat just said the clock is ticking for X11... that it will be dead with RHEL 9 in 2032. We all know what that means. And the next step after X11 is gone is killing off XWayland, you all know the aggressive Wayland people aren't going to stop with just ending standalone Xorg, they are going to argue that even using XWayland is insecure and we can't be allowed to have it. And then once we've gone too far down that path, it will be too late to argue with them.

This is where BSD it least has some sense... if its OpenBSD for sure. Not sure about NetBSD and others though.

It generally seems like BSD takes longer to get where Linux is, but when they get there it is a lot more stable and everything fits together well enough. I generally tend to think more highly of NetBSD than OpenBSD, though. Although really, there's a reason I prefer illumos distributions. FreeBSD is usable, but it feels like they just go in the same direction as Linux slower, and there are already Linux distros that do the same.

Of course OpenBSD while minimal might have other problems such as scrapping code too often that is bloated. Which can mess stuff up potentially.

Hmm... OpenBSD has its own problems and I don't really like their approach as a community, though. It's one of those operating systems that has Tor browser levels of paranoia and is too locked-down to work with. It's like, if Linux goes to one extreme and is too open and complicated, OpenBSD is too locked-down and simplified in a way that really doesn't suit our needs. I don't really care for the "aggressively cutting the fat" approach, because a lot of people would see Pale Moon itself as "unneeded pork" with no reason to exist that's overly complicated. Don't get me wrong, I feel like the OpenBSD kernel might have some decent ideas at the core, but someone would need to fork it to make into a really useful OS... the maintainers of it just don't have the right mentality and are impossible to work with.

As for me, I've thought about whether it would make more sense to roll my own Linux distro and ship Pale Moon and Epyrus with it than to keep trying to port UXP to different operating systems or keep up with mainstream Linux. I mean, suddenly the problem of upstream repo maintainers ripping out what we need goes away. All the stuff needed for Linux development could be shipped with it as a core package. I mean, it's been seeming like we're having to get pickier about glibc versions, GTK versions, Python versions, to the point that half the time we have to tell people on Linux that they are using versions of stuff that has known issues with us and is going to be a gimped experience, and we have no real way of keeping pace. The only problem is time... I just don't have time to maintain something like that. Plus, I don't like the direction the Linux kernel is going with Rust and all that... I don't want to work with Rust, and I would probably have to if I maintained a Linux distro thanks to that.

[frostknight]

Don't get me wrong, I feel like the OpenBSD kernel might have some decent ideas at the core, but someone would need to fork it to make into a really useful OS... the maintainers of it just don't have the right mentality and are impossible to work with.

Yeah they can be a toxic bunch I admit. Although the linux kernel devs can be also so... there's that

SystemD is a good example of what I mean. Fedora moved to SystemD, and every other distro followed shortly. I personally liked Upstart as an alternative to standard init, but it seems like the entire Linux community circles the wagons around Red Hat, always dutifully killing alternative ideas and shoving Red Hat's down everyone's throat out of some weird loyalty

Why do you think I dislike redhat so much lol. They make lots of bloated frameworks for linux and they are also infecting FreeBSD I heard as well, albeit not as much.

Btw, I think current firefox is ugly. Its like fisher price created a web browser... or like windows vista or newer as a web browser. Like really? So gross.

But more realistically, I think we're going to see a lot of non-Linux systems stuck on X11 and GTK3 for a long time, even after those projects go unmaintained.

I certainly hope so.

Hmm... OpenBSD has its own problems and I don't really like their approach as a community, though. It's one of those operating systems that has Tor browser levels of paranoia and is too locked-down to work with. It's like, if Linux goes to one extreme and is too open and complicated, OpenBSD is too locked-down and simplified in a way that really doesn't suit our needs. I don't really care for the "aggressively cutting the fat" approach, because a lot of people would see Pale Moon itself as "unneeded pork" with no reason to exist that's overly complicated.

Hyperbola devs have hard forked OpenBSD and intend to make their own version, I wouldn't doubt that they have better plans for their hard fork.

They most likely will go in their own direction more or less and I doubt it will be as locked down. Maybe they will still do some aggressive removals, but they are not as limited to older versions as OpenBSD devs are. Hyperbola devs will accept Copyleft code for anything. OpenBSD devs tend to usually only accept permissive code for most things.

ports are the sole exception to this.

Anywho, these are just some of my thoughts

[athenian200]

So, good news. It looks like the patch to conditionally disable NPAPI was accepted into the tree. So now it will be possible to build without GTK2 in the future, so long as you don't mind sacrificing plugin support...

Yeah they can be a toxic bunch I admit. Although the linux kernel devs can be also so... there's that

I suppose if I'm being honest, it seems like there's at least a few toxic people in every FOSS project I've ever interacted with... let's just say I always mentally prepare myself to be verbally abused whenever I report a bug in an open source project, and this is based on past experience. So it's not like there is some perfect project out there that doesn't have this problem, it's just a matter of degrees.

Btw, I think current firefox is ugly. Its like fisher price created a web browser... or like windows vista or newer as a web browser. Like really? So gross.

Yeah, that really goes for most things these days... the last version of Fedora that looked good to me was Fedora 14 from 2010. Every version since has looked awful. GTK4, Windows 11, etc. It all looks horrible aesthetically, with Touch UI and rounded corners.

Hyperbola devs have hard forked OpenBSD and intend to make their own version, I wouldn't doubt that they have better plans for their hard fork.

Yeah, I am looking forward to seeing what they do.

Anyway, we did get a little off topic with this, but I mostly enjoyed the conversation.

[Moonchild]

It all looks horrible aesthetically, with Touch UI and rounded corners.

The corners is up for debate but you made me think about "touch UI". How long has it been around now? and aside from ipads (not running windows) and the likes, when was the last time you saw anyone using touch on their laptop or desktop screen, even if it supports it? It's one of those failed paradigms that never caught on but is still pushed hard as being the next big thing - x years and counting and we're still waiting for it to become "big" to make all those compromises become useful, or have any real value, for that matter.