Hi, Am Wed, May 04, 2022 at 10:56:57AM +0200 schrieb Johannes Schauer Marin Rodrigues: > > I'm going to rebuild all reverse dependencies and see if anything breaks > > and report back to you in case I find any FTBFS caused by the new dcmtk > > version. > > the following source package build depend on libdcmtk-dev: > > aeskulap, amide, ants, biosig, cmtk, dicomscope, elastix, insighttoolkit4, > insighttoolkit5, itksnap, mia, odil, odin, openimageio, orthanc, orthanc-wsi, > plastimatch > > I cannot test insighttoolkit4 or insighttoolkit5 because my system lacks the > resources to successfully build either source package (No space left on > device).
This is really a hard one. > ants FTBFS but is broken beyond repair and hasn't been in testing since 2017. I'd vote for a removal of ants. > itksnap FTBFS for for an unrelated reason (#1010549). > > plastimatch FTBFS because of a missing build dependency on > libinsighttoolkit5-dev: > https://buildd.debian.org/status/package.php?p=plastimatch > > It seems the new dcmtk version did not just bump ABI but also changed its API > (the DcmTransportLayerStatus enum including TCS_ok was removed from dcmlayer.h > and defining INCLUDE_{CSTRING,CSTDLIB,CSTDIO} now raises an error), so some > patches were necessary: > > biosig: #1010545 > orthanc: #1010554 > > Mathieu, since you filed #1010474 (upgrading dcmtk to 3.6.7) could you help > clean this up? I wished some warning would have been added. I'm not that involved into those medical imaging tools and a deeply regret that I have messed up things that heavily. > For example maybe you find a solution to get orthanc to > successfully compile again (I X-Debbugs-Cc-ed you on the last bug). Currently, > the testsuite fails with: > > /usr/bin/ld: > /usr/lib/gcc/x86_64-linux-gnu/11/../../../x86_64-linux-gnu/libdcmdata.so: > undefined reference to symbol 'inflateEnd' > /usr/bin/ld: /lib/x86_64-linux-gnu/libz.so.1: error adding symbols: DSO > missing from command line > collect2: error: ld returned 1 exit status > > Which may be something that we have to fix in dcmtk? Mathieu, could you say something about this? Wouldn't it be better to upload dcmtk-3.6.7+really3.6.6 and go a more sensible route? > Note, that I'm not a Debian Med team member. I'm just putting my time here, > because the last dcmtk upload broke blender (because it depends on > openimageio) > which in turn hampers my work on the MNT Reform system image. So for me this > is > just one big yak shave... I can absolutely understand your situation. Medical imaging is also kind of yak shaving for me - well, may be re-shaving by updating packages others prepared. But obviously I'm trapping fully into wide open pitfalls. Sorry again to pull several people into this Andreas. > _______________________________________________ > Debian-med-packaging mailing list > [email protected] > https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/debian-med-packaging -- http://fam-tille.de
