I’m curious why Gnome OS is built in a custom way, as opposed to for example starting with a traditional package manager (f.e. apt, pacman, rpm, etc) and then tracking the results with OSTree for distributing to users.
I’m not criticizing or doubting the current approach, but only curious, as I don’t know why one is better than the other (except that I do know building from source is more work, at least for me personally, but maybe it has advantages).
ostree is the future as fedora now switching to bootc container.
aeon to kalpa and even vanilla os all are making os with immutable variant which make more sense as snap and flatpak are better way to install apps.
Taking something that is already distribution as a base, and then removing any part of it that is a distribution just to get the resulting file system tree, is just a waste of time. You end up following the decisions of another project—downstream patches, pre/post-install directives, configuration files, file system layout—which means you also lose any sort of neutrality.
Also: GNOME OS does not use OSTree any more.
Oh that’s interesting, so the OSTree docs that mentions projects is outdated, that’s where I saw it. What is Gnome OS using now?
@dnome I was thinking that, for example, using a distro with a package manager, would make it easier to install stuff and then test updates, but then to ship the update to users the distro would ship the OSTree snapshot/commit (rather than having the users using the package manager for updates).
Oh that’s interesting, so the OSTree docs that mentions projects is outdated, that’s where I saw it. What is Gnome OS using now?
I believe it’s using an A/B partition scheme instead.
GNOME OS has switched to systemd-sysupdate.
will gnome os release without a beta tag
Out of interest to gain insight, why did it switch to that? EDIT: Nevermind, you didn’t link to systemd-sysupdate, but the actual article with the reasoning behind the switch. Thanks!