If gnome-shell extensions didn't break with every release, forks would not be needed
Since I treat Ubuntu alphas as a rolling "upstream" release, extension breakage was a major problem until Cinnamon came along. I hope they do refork from 3.12 to bring in Wayland support. If they do that, debug it, then "park" the code, that's good enough. IceWM hasn't had new development in many years but is still available and works damned well for low-resource machines like my netbook. No reason Cinnamon can be made to work well with Wayland, then parked with what is not broken not needing to be fixed.
Wayland will matter on big machines like I use when games and other demanding applications are ported to it or otherwise run faster than they do on X. Until then X will be superior for end users, but not after that point. If Cinnamon dies and I want a stable DE for my own use, I can do a private fork of either Cinnamon or g-s for my own use, worrying only about the featuures I use and ignoring all those I do not. Just changing gnome-shell's name at compile time would be enough to install it alongside vanilla GNOME for comparison while having Frippery/Cinnamon style extensions always work.
One shortcoming in Frippery is not including a traditional system tray. Extensions written by others that provide it come, die when g-s changes, then sometimes another comes into existance. Nobody can begin writing these until the GNOME changes for a generation are finalized, however! Cinnamon always has a traditional system tray, essential for running Volti's volume control when Pulseaudio is not being used.
A private, post-Cinnamon fork of g-s for my own use in a worst-case scenario would require a version of GNOME for which both Frippery and something like Topicons exist or for which I can port them over, and which supports Wayland. Once I have that working, I am free to install upstream GNOME alongside it and not worry about extensions always being broken in a DE used only for testing and not daily use. I would attempt to combine the bottom and top panels like in Cinnamon, would look at Cinnamon code for how to do that, but that will surely require figuring out more Javascript than I now know. If nobody else has a use for it, fine-I would be writing it to use in my own machines, then probably offering it to anyone else who wants it with no guarantees.
Hell, the whole OS in my machines is effectively a private fork and blend of both UbuntuStudio and Mint, but has diverged quite far from either one over the years.
Originally posted by Honton
View Post
Wayland will matter on big machines like I use when games and other demanding applications are ported to it or otherwise run faster than they do on X. Until then X will be superior for end users, but not after that point. If Cinnamon dies and I want a stable DE for my own use, I can do a private fork of either Cinnamon or g-s for my own use, worrying only about the featuures I use and ignoring all those I do not. Just changing gnome-shell's name at compile time would be enough to install it alongside vanilla GNOME for comparison while having Frippery/Cinnamon style extensions always work.
One shortcoming in Frippery is not including a traditional system tray. Extensions written by others that provide it come, die when g-s changes, then sometimes another comes into existance. Nobody can begin writing these until the GNOME changes for a generation are finalized, however! Cinnamon always has a traditional system tray, essential for running Volti's volume control when Pulseaudio is not being used.
A private, post-Cinnamon fork of g-s for my own use in a worst-case scenario would require a version of GNOME for which both Frippery and something like Topicons exist or for which I can port them over, and which supports Wayland. Once I have that working, I am free to install upstream GNOME alongside it and not worry about extensions always being broken in a DE used only for testing and not daily use. I would attempt to combine the bottom and top panels like in Cinnamon, would look at Cinnamon code for how to do that, but that will surely require figuring out more Javascript than I now know. If nobody else has a use for it, fine-I would be writing it to use in my own machines, then probably offering it to anyone else who wants it with no guarantees.
Hell, the whole OS in my machines is effectively a private fork and blend of both UbuntuStudio and Mint, but has diverged quite far from either one over the years.
Comment