On Tue, May 09, 2023 at 01:26:10PM -0700, Sean Whitton wrote: > I call for votes on the following resolution. > Voting lasts for one week or until the outcome is no longer in doubt. > Let me take this opportunity to thank Helmut for all his recent work on > this topic. > > === BEGIN > > OPTION A: > > Under Constitution 6.1.5, the Technical Committee recommends that the > maintainers of individual packages should not proactively move files > from the root filesystem to corresponding locations under /usr in the > data.tar.* of packages. So, /foo/bar should not move to /usr/foo/bar. > > Files that are in /usr in the Debian 12 release should remain in /usr, > while files that are in /bin, /lib* or /sbin in the Debian 12 release > should remain in those directories. If any files are moved from /bin, > /lib* or /sbin into /usr after the Debian 12 release, they should be > moved back to their Debian 12 locations. > > This moratorium lasts until we vote to repeal it. We expect to do that > during the trixie development cycle, and sooner rather than later. > We will continue to facilitate efforts to resolve the remaining issues > that stand in the way of safely repealing the moratorium. > > OPTION B: > > As option A, except that only maintainers of essential and transitively > essential packages should refrain from proactively moving files from the > root filesystem to corresponding locations under /usr in the data.tar.* > of packages. > > OPTION N: > > None of the above. > > === END I vote A > B > N. Rationale I was initially very unhappy about extending the moratorium, because it poses a significant mental cost to the entire project. I subsequently proposed to limit it to essential packages to reduce the impact. However, that would increase the complexity of the rule to follow and the essential set isn't static either. If we were to move files in bookworm already (which has not been proposed), we could easily break systemd as systemd-resolved and systemd-boot both replace systemd and that would become exactly a situation as the one to be prevented by the moratorium. In the bullseye -> bookworm upgrade, we have about 13 situations that would have been broken (in a /usr-merged bullseye upgraded to bookworm) in the absence of the moratorium. I think it is reasonable to expect that we'd encounter a similar amount in trixie. Helmut
Attachment:
signature.asc
Description: PGP signature