Hi, On Thu, Apr 17, 2025 at 08:41:35AM +0200, Sebastian Ramacher wrote: > On 2025-04-15 17:10:44 +0200, Michael Banck wrote: > > On Tue, Apr 15, 2025 at 08:40:59AM +0200, Sebastian Ramacher wrote: > > > Control: tags -1 moreinfo > > > > > > On 2025-04-14 19:49:00 +0200, Michael Banck wrote: > > > > Package: release.debian.org > > > > Severity: normal > > > > User: release.debian....@packages.debian.org > > > > Usertags: binnmu > > > > > > > > nmu coot_1.1.15+dfsg-1 . ANY . unstable . -m "Rebuild against > > > > librdkit-dev" > > > > > > Why is this needed? > > > > The autopkgtests fail: > > > > https://ci.debian.net/packages/c/coot/testing/amd64/ > > https://ci.debian.net/packages/c/coot/testing/amd64/59906919/ > > If symbols go missing from librdkit, this smells like a SONAME bump that > is missing from librdkit.
Right. But coot is the only reverse-depdency, and I think it is ok to rebuild it for trixie and then re-assess the situation post-trixie. E.g. we could remove the shared library. > Why did this symbol disappear? Because upstream is playing loose with the library API I guess and/or do not care all that much. It is nothing new with scientific packages and a reason I mostly only ship -dev libraries anymore. Should we rename the shared library package to librdkit1d or something instead? Michael
