The GTK3 Port Of Firefox Is Making Progress, Firefox Can Run On Wayland

[Karl Napf]

What kind of design is this?

So you have a library with two users, one that needs UI elements and one that does not. In fact the second one needs a different UI library, causing the second use case to crash.

So the obvious choice is to add a new library that can be used to mask out the UI elements that the first and the second user needs, so that you can switch between those two at runtime. Wow, that is software design as it should be. Both process depend on a new library now and the second that used to load a lot of crap it did not ever need is still loading that.

The reason? "libxul was to intricate". So we add a new layer (a thin one, of course, has anybody ever added a thick one?), because the existing code is so interwoven that the developers do not dare to move one piece of functionality from one library into another. Great... so the code is bad, let's make it worth by adding more cruft that we actually know is unnecessary. That way we can get a new feature a bit faster, yeah!

That is exactly what I want in my browser. No wonder chrome is eating firefoxes lunch.

[curaga]

+1

The arrogance of the GTK developers, who can not even be bothered to implement support for other fileselection dialogs, makes every day an application switches to Qt a good day.

We are at a point where application developers are implementing support for different file dialogs, even though this clearly is the job of the toolkit. Qt apps integrate pretty well in GTK based desktops but the other way round it's a nightmare. The arrogance of the GTK developers harms the user experience of all users of Qt-based desktops! The sooner GTK dies the better, the Open Source community has no need for these kind of developers, who play their little power games on the back of the users.

Shame on you GTK devs.

Have they turned down a patch to pop up the Qt file selector?

[Karl Napf]

+1

The arrogance of the GTK developers, who can not even be bothered to implement support for other fileselection dialogs, makes every day an application switches to Qt a good day.

We are at a point where application developers are implementing support for different file dialogs, even though this clearly is the job of the toolkit. Qt apps integrate pretty well in GTK based desktops but the other way round it's a nightmare. The arrogance of the GTK developers harms the user experience of all users of Qt-based desktops! The sooner GTK dies the better, the Open Source community has no need for these kind of developers, who play their little power games on the back of the users.

Shame on you GTK devs.

Oh, come on, calm down.

There are not that many gtk devs: They need to concentrate on their priorities. There is no use calling any of them names, just because their priorities are different from yours: That will only discouraging people from working on the library.

Check the statistics yourself: 90% of the commits in the last 30 days where done by 5 people (check ohloh). Those 5 people do a really heroic job at improving GTK, so please be more respectful.

[randomizer]

I wish they'd use Qt. Then Firefox would finally honour mouse wheel scroll rate settings specified in KDE, instead of using the broken GTK+ black magic that requires messing with the bowels of about:config to work around.

[Luke_Wolf]

How is Mozilla Shumway progressing? Flash won't work on Wayland, and in order to switch to GTK3, Shumway has to be a drop in replacement. I tried it a while ago, and it didn't work for video for example (which is probably 90%+ use cases for Flash these days).

Or you know they could finally stop holding up the web and support PPAPI

[Karl Napf]

3) All APIs break upon major revisions, and Gtk+ and Qt are no exceptions. I've upgraded my Gtk+ apps a few times, and have found that the changes are usually quite small and make sense. Generally, Gtk+ adds features but doesn't remove them.

Morten Welinder is challenging that point in his recent blog post. The blog post was on planet gnome and I have not seen anybody challenge it there, so I assume there is some validity to the claims Morten makes.

Do you think he is over the top with his criticism?

[wargames]

CSS is awful, idiotic, and slow as shit. There shouldn't be any web technologies in anything other than a browser and that's only because we have no choice in the browsers at this point. JS, CSS, HTML -- all that crap is absolutely awful means of programming anything. They are some of the worst software technologies in the world, completely made by the limitations of the browser. All those things do is drive developers to alcohol.

Wow, those are exactly my thoughts... You just forgot PHP, which is usually used with that crap

[Temar]

There are not that many gtk devs: They need to concentrate on their priorities.

Is this sarcasm or are you serious? Qt 1.0 was released in 1996, which gave them about 18 years to implement support for that file selector.

This is not an issue of not having enough time, but it is pure arrogance. Their arrogant attitude not only mocks their user base but also all GTK application developers who have to manually implement support for other file selectors to get an application which integrates well on all desktop environments.

[emblemparade]

You made these exact same false claims in the thread about audacious, then immediately ducked out of the thread when people corrected you. I, for one, responded in detail here, which you completely ignored. I am not going to bother rewriting the same thing again if you couldn't be bothered to read it last time. If you actually care about having a discussion on the subject, you can respond to the people who addressed this post in the last thread.

It's simply not always worth the effort to discuss things on the Internet. Sometimes it's best just to leave the conversation as is, and to let other readers judge.

You made a lot of statements there about "Gtk+ devs have said" without any citations, so I can't respond to those.

You also make the odd conclusion that "The very fact that so many projects are switching, and the switch is pretty much entirely one-directional, should tell you that there is something different." Corellation doesn't equal causation. That *some* projects are switching can say many things, but that doesn't lead to the conclusion that Qt is generally better than Gtk+ and that all projects should switch. The reason can be very project-specific reason, for example the choice of programming language (C or C++), or licensing ideology. It could also be jumping on the bandwagon. Trends are funny: in 5 years, there might be something really annoying happening with Qt (too much focus on mobile or something) and then "so many projects" will want to switch back. We've seen these changing fashions a lot.

In any case, it's ridiculous and annoying for people to want Firefox to switch from Gtk+ to Qt without any understanding as to what this would mean for the Firefox devs. For sure, the end-user experience would be pretty much identical: there is a slight advantage for one or the other depending on whether you are using a Gtk+-based desktop versus a Qt-based desktop. It could mean fewer packages to install. However, the fact that so many popular software packages use one or the other means that you probably have both sets of packages installed anyway. It's really not such a terrible situation as long as the end-user experience is consistent.

I agree with one thing you said: it is indeed easier to create new widgets in Qt. But whether or not it's this particular reason that is causing some projects to switch needs to be looked at project by project. How many projects need to create custom widgets? And is that generally a good idea?

[psychoticmeow]

CSS is awful, idiotic, and slow as shit. There shouldn't be any web technologies in anything other than a browser and that's only because we have no choice in the browsers at this point. JS, CSS, HTML -- all that crap is absolutely awful means of programming anything. They are some of the worst software technologies in the world, completely made by the limitations of the browser. All those things do is drive developers to alcohol.

And yet it's understood very widely, and despite it's major flaws is still better than the alternatives. Do you really want to go back to the days where the theme engine had final say over what it was possible for designers to implement?