Serpent OS Build Infrastructure Launched

[skeevy420]

With that infrastructure work out of the way, they are now pursuing proper GNOME packaging and then working towards a new spin as their first "real" ISO release.

The only way to properly package GNOME.

I can dream, dammit.

[matthewtrescott]

It's true that distro packaging isn't perfect, has legacy cruft, etc... but the most important features mentioned about Serpent OS are already available in the Open Build Service (from SUSE) for years. Sometimes it's easier to start from scratch than to clean something up afterward but I don't think this is one of those cases.

OBS supports parallel builds, signing, repository snapshotting, creating installable images, deb packages, pacman packages, ... there's no magic; it just works. It would probably be relatively easy generate ostree filesystems with it, too. It's too bad to see so much effort that could have been directed to improving the existing tools. Is the fact that it's written in Perl really a good reason to reject an actively-developed and successful tool like OBS?

[skeevy420]

Is the fact that it's written in Perl really a good reason to reject an actively-developed and successful tool like OBS?

That's like asking: "Is the fact that they're built out of wood and fabric really a good reason to abandon actively-developed and successful aircraft like the Fokker M.7?"

[kn00tcn]

It's true that distro packaging isn't perfect, has legacy cruft, etc... but the most important features mentioned about Serpent OS are already available in the Open Build Service (from SUSE) for years. Sometimes it's easier to start from scratch than to clean something up afterward but I don't think this is one of those cases.

OBS supports parallel builds, signing, repository snapshotting, creating installable images, deb packages, pacman packages, ... there's no magic; it just works. It would probably be relatively easy generate ostree filesystems with it, too. It's too bad to see so much effort that could have been directed to improving the existing tools. Is the fact that it's written in Perl really a good reason to reject an actively-developed and successful tool like OBS?

sometimes isolated experiments can later be cherrypicked back to the established projects, especially after multiple revisions, rather than trying to heavily alter the existing project