Gnome Shell Extensions are powerful because they are allowed to hook into everything the gnome-shell process does. If the extension API would be changed so they couldn’t crash our shell session, extensions would become way less powerful and be mostly useless.
Nevertheless, it would be great if Gnome Shell could keep/recover your application state after restarting (like KWin).
If the extension API would be changed so they couldn’t crash our shell session, extensions would become way less powerful and be mostly useless.
That’s not true, a proper API would enable probably 90% of existing extensions to be portable since they more often than not “just” add stuff into different places without really modifying anything. And if they really are worried so much about developer freedom they could still allow for the current monkey-patching approach to exist for those extensions that would otherwise not work, so the user ends up with more options than just “no extension at all” or “from now on shit may crash”. To allow random code to run right in the shell (which is one of the most important pieces of software in an end-user facing system) and/or actively manipulate its code without sufficient measures to ensure stability or at least recoverability is just not an acceptable status Quo in my opinion.
I’m sad and annoyed about this whole situation (therefore memes) ever since a Gnome dev confirmed to me they (meaning some board; their orga structure is rather stiff) actively decided against doing anything about this problem. I love Gnome, but I had to move to KDE simply to have a modern, stable desktop that I could trust in a production environment without feeling so barebone.
The thing is that the shell provides so little innate functionality and delegates anything more to an ecosystem of random quality, and then subjects those authors to a pretty capricious interface that breaks random extensions every six months generally driving a lot of the authors to throw up their hands and give up.
So the native functionality is solid, though even lower features than Microsoft windows window management, and then have to apply dodgy extensions to get features that other solutions just have as a matter of course.
If I didn’t know any better, I would have assumed that gnome shell was some small demonstrator project to serve as a reference implementation (e.g Weston) rather than intended for serious use. I came over from gnome 2 thinking things went pretty far backwards, but the extensions are going to be stop gaps while they build back up to a balance desktop. But they never seemed to do that.
Ultimately, I landed on Plasma and that’s been pretty good. Have some embedded/kiosk stuff using sway thanks to the very nice scriptable facilities there, but still sticking with kwin as a daily driver for now.
Gnome Shell Extensions are powerful because they are allowed to hook into everything the gnome-shell process does. If the extension API would be changed so they couldn’t crash our shell session, extensions would become way less powerful and be mostly useless.
Nevertheless, it would be great if Gnome Shell could keep/recover your application state after restarting (like KWin).
That’s not true, a proper API would enable probably 90% of existing extensions to be portable since they more often than not “just” add stuff into different places without really modifying anything. And if they really are worried so much about developer freedom they could still allow for the current monkey-patching approach to exist for those extensions that would otherwise not work, so the user ends up with more options than just “no extension at all” or “from now on shit may crash”. To allow random code to run right in the shell (which is one of the most important pieces of software in an end-user facing system) and/or actively manipulate its code without sufficient measures to ensure stability or at least recoverability is just not an acceptable status Quo in my opinion.
I’m sad and annoyed about this whole situation (therefore memes) ever since a Gnome dev confirmed to me they (meaning some board; their orga structure is rather stiff) actively decided against doing anything about this problem. I love Gnome, but I had to move to KDE simply to have a modern, stable desktop that I could trust in a production environment without feeling so barebone.
The thing is that the shell provides so little innate functionality and delegates anything more to an ecosystem of random quality, and then subjects those authors to a pretty capricious interface that breaks random extensions every six months generally driving a lot of the authors to throw up their hands and give up.
So the native functionality is solid, though even lower features than Microsoft windows window management, and then have to apply dodgy extensions to get features that other solutions just have as a matter of course.
If I didn’t know any better, I would have assumed that gnome shell was some small demonstrator project to serve as a reference implementation (e.g Weston) rather than intended for serious use. I came over from gnome 2 thinking things went pretty far backwards, but the extensions are going to be stop gaps while they build back up to a balance desktop. But they never seemed to do that.
Ultimately, I landed on Plasma and that’s been pretty good. Have some embedded/kiosk stuff using sway thanks to the very nice scriptable facilities there, but still sticking with kwin as a daily driver for now.
Gnome’s extension power is true, but it’s also because they don’t give a toot about supporting them natively