Hi Thorsten, Thorsten Alteholz <debian-med@alteholz.de> schrieb : > Hi Ivo, > > On Wed, 28 Mar 2012, Ivo Maintz wrote: > > But now it should be ready for upload (also to close bug #665837). > > after building the package, I got some I:-lines from lintian: > > I: libsbml5-dev: package-contains-empty-directory usr/lib/jni/ > I: libsbml5-dev: package-contains-empty-directory usr/share/java/ > I: libsbml5-dev: package-contains-empty-directory usr/share/man/ > > Did something strange happen during my build or are these empty dirs > intentional? Oh no, there was an old (and now unused) debian/dirs file. > I: libsbml5-octave: binary-has-unneeded-section > usr/lib/x86_64-linux-gnu/octave/site/oct/x86_64-pc-linux-gnu/OutputSBML.mex > .comment > I: libsbml5-octave: binary-has-unneeded-section > usr/lib/x86_64-linux-gnu/octave/site/oct/x86_64-pc-linux-gnu/TranslateSBML.mex > .comment I have no idea, how to fix this. But it's not critical, right? > In libsbml5-octave.install you added x86_64-linux-gnu to the path. Is > this correct in case the package is built on some other architecture? I see, there is some more work to do. Up to now, also on amd64 the octave site packages were situated directly under /usr/lib; now it's usr/lib/x86-64-linux-gnu. I can only build for amd64 and i386; how to get infos about the right path for the octave site packages on other architectures? > Also there are some W:-lines: > W: libsbml source: package-depends-on-hardcoded-libc libsbml5-matlab > depends > W: libsbml source: package-depends-on-hardcoded-libc libsbml5-octave > depends > > Is there really any reason to depend on libc6? Yes. And no. dh_shlibdeps does not examine the *.mex and *.mexa64 files (octave and matlab bindings), but lintian gave an error: "missing-dependency-on-libc" So I hardcoded it, but now I added two lines in debian/rules: dpkg-shlibdeps \ debian/libsbml5-octave/usr/lib/x86_64-linux-gnu/octave/site/oct/x86_64-pc-linux-gnu/* -Tdebian/libsbml5-octave.substvars dpkg-shlibdeps debian/libsbml5-matlab/usr/lib/* -Tdebian/libsbml5-matlab.substvars Is this still the right way? > Thorsten > > PS. There are also lots of P:-lines :-). I know. But as far it was such a hard thing... Ivo
Attachment:
signature.asc
Description: PGP signature