Fixing packages for broken Gtk3
As mentioned on sunweaver’s blog Debian’s GTK-3+ v3.21 breaks Debian MATE 1.14, Gtk3 is breaking apps all around. But not only Mate, probably many other apps are broken, too, in particular Nemo (the file manager of Cinnamon desktop) has redraw issues (bug 836908), and regular crashes (bug 835043).
I have prepared packages for mate-terminal and nemo built from the most recent git sources. The new mate-terminal now does not crash anymore on profile changes (bug 835188), and the nemo redraw issues are gone. Unfortunately, the other crashes of nemo are still there. The apt-gettable repository with sources and amd64 binaries are here:
deb http://www.preining.info/debian/ gtk3fixes main
deb-src http://www.preining.info/debian/ gtk3fixes main
and are signed with my usual GPG key.
Last but not least, I quote from sunweaver’s blog:
Questions
- Isn’t GTK-3+ a shared library? This one was rhetorical… Yes, it is.
- One that breaks other application with every point release? Well, unfortunately, as experience over the past years has shown: Yes, this has happened several times, so far — and it happened again.
- Why is it that GTK-3+ uploads appear in Debian without going through a proper transition? This question is not rhetorical. If someone has an answer, please enlighten me.
(end of quote)
<rant>
My personal answer to this is: Gtk is strongly related to Gnome, Gnome is strongly related to SystemD, all this is pushed onto Debian users in the usual way of “we don’t care for breaking non-XXX apps” (for XXX in Gnome, SystemD). It is very sad to see this recklessness taking more and more space all over Debian.
</rant>
I finish with another quote from sunweaver’s blog:
already scared of the 3.22 GTK+ release, luckily the last development release of the GTK+ 3-series
big mistake mate switched to gtk3 !
I still prefer gtk2
Nope, GTK3 breaking the desktops is not that bad. It is one of those issues everybody sees and that get fixed soon or later.
The worst issues with updates that break the “userspace” are those related to the hardware. When SystemD/NetworkManager/what else gets updated and your wireless card/adapter becomes unusable, WHO CARES? You can report the issue but it is usually difficult to locate the cause. In the end you know the issue “won’tfix”, EVER, unless it goes away by accident like it came by accident. In my case I “fixed” it by try&error, in the end using Wicd.
This kind of hardware-relate problems PILE UP so that instead of getting a system that becomes better as time passes, we get a system that is increasingly worse, but hey, there are all those “exciting new features” that are so much fun on the brand new hardware the developers are using.
1) No, GTK+ has absolutely nothing to do with systemd in practice.
2) It’s “systemd”, not “SystemD”.
3) systemd cares a great deal about not breaking applications. Many applications do cause breakage by depending on systemd and not working without it, but that’s not a systemd issue, that’s an application issue.
A question I’d like to see the answer to: why are various applications being “fixed” to work with GTK+ 3.21, rather than GTK+ being fixed to not break applications that worked with GTK+ 3.20? A shared library breaking backward compatibility without changing SONAME, without an RC bug?
Yes but systemd (right?) comes from the same direction as lots of other stuff that is becoming the “de facto standard”.
I did not say systemd breaks “applications”, I said because of updates in key components distros are increasingly inconsistent and regressions aren’t getting fixed because 1. nobody cares of boring stuff hardware related, even less if it is old hardware, 2. new things must be pushed forward regardless what thy leave behind.
See, if it was about “applications”, you could just use another application, unless there isn’t any alternative, which in the long term is another bad consequence of the systemd/gnome development. Mate broken? You can use XFCE or LXDE until they still work with systemd udev etc and they don’t depend on GTK3.
But when some hardware stops working after an update, you can’t do much about it, besides trying all the distros out there, hoping for some untold reason there is one that don’t have the said regression. Because, please note, there isn’t anything like proper testing and/or documentation about what works, what doesn’t work , where, when and why. Shit happens out of the blue, nobody knows why and nobody cares.
It is not Debian, I got the same kind of issues on Ubuntu. You update and some piece of hardware is broken. You report the issue and you are basically told to debug it yourself, lets try vanilla kernel and so on. Another update, another bag of regressions which pile up with the unfixed old ones.
Maybe interesting for reading:
https://blogs.gnome.org/desrt/2016/06/13/gtk-4-0-is-not-gtk-4/
Hi Tanty,
thanks for the link. This is an excellent example of stupidity. Let us break *every* expectation *everyone* has out there just to be different. It is very disappointing, but honestly falls into my expectation of Gnome/Gtk developers. Sad to see such a good toolkit run down the gutter.
Go away gnome, go away systemd, Welcome Devuan, Welcome Void.