Language Selection

English French German Italian Portuguese Spanish

On Gentoo Releases and stuff

Filed under
Gentoo

When I got more involved with Gentoo, the release team was aiming to release the then-pending release (2007.1) in early December. Eventually, that date slipped time and again. It was decided to drop the entire release and do a new one (2008.0). Again, the dates kept slipping. Now, the question is why a release is slipping time and again.

* It's not quite as simple as it may seem. Building release media in itself isn't easy to begin with - catalyst is a powerful but complex (and complicated) tool. I got kind of dropped into cold water with release building for alpha (probably because I was the one who didn't have a chair when the music stopped). catalyst build errors can be notoriously hard to debug. On top of this, the central release coordinator has to keep in mind all of the gritty details of the arches that will see release media. There's arches like ppc which also have a differently-bitted cousin (ppc64); there are arches that are very, very slow when building stuff (MIPS). On top of that, some software just doesn't build on some arches (no Java on alpha, for example) which can make deciding what to put on the LiveCD very hairy. This also stresses one point: without the help and expertise of the arch teams, building a release is completely impossible. Some arches are more motivated to get a release out, some less so. Some are notoriously understaffed when it comes to release building, some have too few machines to test the releases (or no physical access, which is just as bad).

* People have lives. This is one that bit us this time: life struck at a very bad point (not that the event had been any better post-release). This occupied the time of a dev for a prolonged time. It made painfully obvious that in some spots, stand-in personnel wasn't there. Let me stress this lest I get misunderstood: such things happen and that the dev in question saw tending to the event as more important than get a release done is his prerogative and not to be questioned. That there wasn't a stand in to keep stuff going (even at a slower pace) is something that needs to be fixed.

More Here




More in Tux Machines

Android Leftovers

GNOME Shell vs. KDE Plasma Graphics Tests On Wayland vs. X.Org Server

A premium member this week had requested some benchmarks of openSUSE Tumbleweed when looking at the performance of KDE Plasma vs. GNOME Shell in some open-source graphics/gaming tests while also looking at the Wayland vs. X.Org Server performance. With KDE Plasma 5.12 that openSUSE Tumbleweed has picked up, there is much better Wayland session support compared to previous releases. While KDE developers aren't yet ready to declare their Wayland session the default, in my experience so far it's been working out very well but still routinely will find application crashes in Kate and the like when testing under the KWin's Wayland compositor. Read more

Stable kernels 4.15.6, 4.14.22, 4.9.84, 4.4.118 and 3.18.96

Android Leftovers