That's a good question. For reference, CentOS7 is expected to be maintained until 2024. Now for the future ?

First, we will have to provide RHEL8 support because this is the platform the most deployed by our customers. That means that those package will be installable on CentOS8 even if, given the timeframe, that would be an odd move. This also means that other RHEL8 derivation (like Oracle Linux or the ones announced after RedHat move on Stream) will also be an option.

It's also likely that we will support Stream because, at the end of the day, a rolling release is way more interesting to a tool like Tuleap than a fixed OS like RHEL is / CentOS used to be. We historically depend on RHEL because it was the most deployed OS in Big Corp ITs but the buisness value of RHEL (10 years without modifications) is a PITA. We need up to date versions of the tooling PHP, Node, Git, SVN, and all related libs, it slows us down to have to maintain a backward compatiblity with stuff that was release a decade ago.

At this time we do not envision to provide package outside "RedHat world" though (ubuntu, debian, suse).

Not that we don't like those platform but mainly because:

• it means to have a not null effort to build the packages (.deb for instance) and to maintain them in time
• completely different stacks means completely different version of everything and then a new source of bugs. The worst of all bugs: environment bugs, that can be reproduced only with specific conditions, that takes ages to spot.

One of the strength of Tuleap is that when we fix a bug it's likely to be fixed for everybody (because everybody is supposed to run the same environment) and we are pretty much sure that it doesn't introduce regressions because the environment is under control. With more moving parts, this promise would no longer be possible (or would cost a lot in term of tests, CI, slowed down developments, etc).