Re: itk-4.5 and python
On 03/18/14 02:19, Steve M. Robbins wrote:
On March 12, 2014 11:16:05 PM Steve M. Robbins wrote:
On March 12, 2014 04:40:35 PM Gert Wollny wrote:
Sorry, it seems I didn't run lintian at all - spoiled from
git-buildpackage which does it always I forgot to run it manually. Now
I've configured svn-buildpackage to run lintian as well.
E: insighttoolkit4-python: arch-independent-package-contains-binary-or-object
usr/lib/ITK-4.5/Python/_ITKAnisotropicSmoothingPython.so
[... and all other ITK*Python.so...]
My fault, because I changed d/control first, and did not copy all the
changes to d/control.in, specifically, the line
Architecture: amd64 i386
didn't make it into control.in, so that the arch was set to "all" since
I copy-pasted that section from insighttoolkit4-examples.
When you regenerated the control file this information got lost.
W: libinsighttoolkit4.5: shared-lib-without-dependency-information
usr/lib/libITKGDCM-4.5.so.1
I guess that's because of the --as-needed flag I added. If the system
GDCM is used libITKGDCM is only a stub to have a target for the GDCM IO.
It would be better that when using the system GDCM, this library
wouldn't be created at all, and the GDCM IO module would reference the
GDCM library directly. Alternatively, we could make libITKGDCM static to
still have the cmake target.
W: libinsighttoolkit4-dev: extra-license-file
usr/lib/cmake/ITK-4.5/WrapITK/Configuration/Generators/SwigInterface/pygccxml-1.0.0/LICENSE_1_0.txt
E: libinsighttoolkit4-dev: package-installs-python-bytecode
usr/lib/cmake/ITK-4.5/WrapITK/Configuration/Generators/SwigInterface/pygccxml-1.0.0/pygccxml/__init__.pyc
This is obviously doing too much:
dh_install -p$(pkg_dev) --autodest debian/tmp/usr/lib/cmake
I wonder if the *.cmake inside WrapITK are really needed for development
with ITK.
I'll see what I can do in the next few days to fix these.
Best regards,
Gert
Reply to: