Hi,In my previous reply I may have gotten some details wrong, as I hadn't read the bug about dcmtk yet.
On 19-04-2025 20:01, Sébastien Jodogne wrote:
If I read that other bug correctly, the changed path in dcmtk broke orthanc. So, the right binary package of src:dcmtk should declare that with a *versioned* Breaks on the right binary package of src:orthanc. It seems that orthanc (and thus orthanc-wsi) from unstable works in unstable (at least on amd64 and some other architectures, it seems broken on arm64, loong64, ppc64el and s390x), but not with the older src:dcmtk in testing. So the right package from src:orthanc should have a *versioned* Depends on the right package from src:dcmtk.
Unfortunately, I am totally lost here. Even though I understand the general philosophy, I still do not see what I must do. Should I upload to testing?
No, not at all. All changes should go via unstable.
Must also orthanc-wsi and dcmtk be updated, given that the issue seems to lie in liborthancframework (part of src:orthanc)?
I *suspect* that only orthanc and dcmtk need changes.
And what about the fact that orthanc-wsi_3.2+dfsg-1 is already there, while it seems like that it is orthanc-wsi_3.0+dfsg-1 for which a fix is expected?
The fix for orthanc-wsi/testing is for orthanc-wsi/unstable to migrate. So you need to remove the blockers for migration (which may lay outside of orthanc-wsi).
Sorry, I am unable to solve this without further instructions and a full example...
I suggest asking debian-ment...@lists.debian.org for help then. Paul
OpenPGP_signature.asc
Description: OpenPGP digital signature
