Frederic Crozat

Upgrading big components in Linux distributions is hard. But integrating them while minimizing regressions (for stable distributions) and not slowing down release pace (for rolling releases) requires a lot of process and tooling.

Let's deep dive in those.

Over the previous months, openSUSE and SUSE teams have worked together on integrating latest GNOME release (3.34) in their various distributions, while minimizing duplicated work and sharing as much code as possible.

We'll describe how it was done for the 3 differents flavors of distributions: - openSUSE Tumbleweed (rolling release) - SUSE Linux Enterprise 15 SP2 (Enterprise release) - openSUSE Leap 15.2 (stable release)

There has been several camps on distribution style: - Rolling style, where everything is constantly moving - Regular style, where everything get polished and fixed before some freeze - Stable style, where (almost) nothing move, fearing to break anything

openSUSE (and SUSE) have been releasing distributions in those various style and unlike common believe, they are not as different as you might thing

We will describe pro and cons for each style and look at how our tools (OBS / openQA) and our processes (Staging / Factory First) are being use, independently of the distribution style.

And how a "Stable" style can also be delivered like "Regular" style.

Starting with openSUSE Leap 42.2, a lot of cooperation has been done to bridge gaps between openSUSE and SUSE Linux Enterprise distributions. Things have been improving nicely with openSUSE Leap 42.3. We'll go into details on what this cooperation means for both openSUSE contributors and for SUSE and how we ensure it takes place.

We'll share a few statictics, discuss on the new SLE15 development and how is done in harmony with openSUSE Tumbleweed (rolling release).

This talk is to share with other distributions on our best practises, learned over time, and how we enforce them. We also want to share how beneficial this work is, for all parties.

We are also eager to learn from other distributions on how they work in similar environements and if we can learn from them.