Hi, On 3/17/26 08:12, Sébastien Jodogne wrote:
I have just released a new version of the liborthancframework-dev package (1.12.10+dfsg-2), which notably includes the following changeset:https://salsa.debian.org/med-team/orthanc/-/ commit/43304015d951fc6701fe1329769943c10d3301c5
I had a look when the bug was closed :).
Could you please confirm that this is what you expected? In particular, I was wondering whether the ">=" should be replaced with "=".
Well, the first answer is no. But your second suggestion is less good. Let me explain. What I wanted you to do is think about what the expectations are for liborthancframework-dev. And *if* a versioned dependency was necessary, I was expecting something like >= major-version-of-boost-during-build, << next-major-version-of-boost~. Such that during the *next* boost transition, it will be lock stepping, but not during regular Debian revisions. If a versioned dependency is *not* necessary I'd like to understand what's the missing piece of the puzzle then.
I also wonder why explicit versioning is required only for libboost-all- dev in liborthancframework-dev, but not for the other "-dev" package dependencies on C++ libraries (i.e. libdcmtk-dev, libjsoncpp-dev, and libpugixml-dev).
That I don't know, I'm not very versed in c libraries. But what I'm guessing is that liborthancframework encodes some of the ABI (or is it API, I always mix those up) of boost in their own ABI and hence when boost transitions, liborthancframework also transitions. Now that I write this down, I wonder if technically liborthancframework now should have changed SONAME too ...
Paul
OpenPGP_signature.asc
Description: OpenPGP digital signature
