On Thu, Oct 20, 2011 at 08:18:10PM +0200, Andreas Metzler wrote: > a transition using Breaks looks too heavy-handed to me. Installing > multi-arched p11-kit does not break other software's use of PKCS#11 > modules. Also at least on Debian using relative paths and relying on > the module search path simply has never worked, the initial p11-kit > (0.3-1) upload already used the multi-arch path. Ubuntu was quicker > (first upload July 28) but I doubt there are many active users yet. If > the currently available modules moved, p11-kit users could make use of > them without specifying the full path in the configuration file.
> I guess the critical part of the transition is to not break other > PKCS#11 users with the new path. Yes, given that there are other consumers of these modules, that is certainly a concern. > > Or should libp11-kit0 be patched to support looking up modules in > > both locations? > [...] > Won't this introduce problems on partial upgrades to multiarch? Evil > things might happen if a i386 p11-kit tried to load a amd64 module > from /usr/lib/. "Evil things" - the only thing that happens is that dlopen() will return NULL. Which is, after all, the same result you would get if you tried to open a non-existent file, so I don't think this is a major problem provided that the code is written to allow for this. Cheers, -- 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/ slanga...@ubuntu.com vor...@debian.org
signature.asc
Description: Digital signature
