On Wed, Jun 08, 2011 at 06:11:08PM +0100, Simon McVittie wrote: > On Wed, 08 Jun 2011 at 09:24:44 -0700, Steve Langasek wrote: > > But, do we risk getting a false negative for cases where the real file has > > a weird filename whereas the soname (as provided by the symlink) follows one > > of the standard patterns? > Possibly... either way, lintian and debhelper should be consistent about > what they consider to be a "good enough" reason to pre-depend (this could > become a debhelper bug if you prefer?). right, so debhelper's check is: find $tmp -type f \( -name '*.so' -or -name '*.so.*' \) thus matching any "*.so*" real files at all in the multiarch dir. That would indeed be a better check to have in lintian also. I probably didn't think of it when writing the lintian check because I didn't see an obvious way to check if a file was a symlink. > See commit 6463abf33ad (on the experimental branch) in > git://anonscm.debian.org/pkg-utopia/dbus.git for an example of a package that > seems to work fine, but needed an explicit pre-dependency to avoid the > lintian error. Heh, I would've assumed you would use a lintian override instead of adding a pre-dependency... :) -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developer http://www.debian.org/ slangasek@ubuntu.com vorlon@debian.org
Attachment:
signature.asc
Description: Digital signature