Dear DSA, ftp-master, release team,As you are probably aware, architecture qualification is a notorious hard problem for the Release Team and the project. This year at DebConf, we have been discussing [1,2] a potential change in the way we look at the (release) architectures. Although my original intent of that BoF was broader, I'd like to discuss the option to split the group of official architectures into two groups.
There are a couple of architectures (armel, armhf, i386, mips64el and mipsel) that we have serious concerns about, but not so much that we have already decided to kick them out from the release architectures. I am under the impression (acknowledged by the big crowd at the BoF) that it would help the project if we would reduce the "Fully Supported architecture" (FS) without demoting the other architectures to ports (on ports.debian.org). The idea that I have is that we don't change what happens on the DSA and ftp-master side of the archive, but that we change the support level for some of the current release architectures. For the scope of this e-mail, lets call the new group the "Best Effort architectures" (BE). The main idea is that (most) problems on those architectures are not blocking migration from unstable to testing on the FS archs.
Of course there are details to figure out and agree on, but before diving into those I'd like to hear if you are open to support the idea (hopefully even in time for bookworm) or if there are already deep concerns (that would take long to resolve if at all).
Paul[1] https://debconf22.debconf.org/talks/47-investigating-a-tier-system-for-release-architectures/
[2] https://gobby.debian.org/export/debconf22/bof/tier-architecture-system
Attachment:
OpenPGP_signature
Description: OpenPGP digital signature