Package: liborthancframework-dev
Version: 1.12.10+dfsg-1
Severity: important

Hi,

I've been checking on the autopkgtest regressions triggered by the binNMU of src:orthanc and noticed a pattern that's shared with bug 1129823 and it's clones. I have the impression (just a hunch) that the boost major version that gets installed by liborthancframework-dev via libboost-all-dev needs to be the same version as was used during building of liborthancframework1. If so, could you please generate a proper version for a versioned dependency during building? That way it's 1) clearer what's going on during the next transition during rebuilds 2) the migration software will know that the binNMU can't migrate until boost-defaults migrates and will ensure things are aligned. Currently the rebuild could migrate before boost-defaults and I'm betting things would break in testing if that happens. And vice versa, boost-defaults could migrate without the binNMU and I think things would be equally broken in testing.

Paul

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature

Reply via email to