From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 Description of problem: For AMD64 systems, concurrent installation of pango for i386 and x86_64 leaves a conflict at /usr/bin/pango-querymodules. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. install pango for x86_64 2. install pango for i386 (or reverse the order) Additional info:
As discussed, we should move the binaries, and make /usr/bin/pango-querymodules a script that calls the right binary. We'll also need to make sure other files don't conflict, such as the contents of /etc/pango
Fix seems to work well for general use, but breaks Anaconda for buildinstall. Jeremy has created a work around for AMD64 test1 but a more long term solution needs to be reached.
To be a little bit more verbose, the problem is that since anaconda doesn't actually install the package but instead does rpm2cpio on it, there ends up having to be special cased fun to figure out what rpm decided various macros were at build time. Unfortunately, I haven't thought of anything better yet (well, beyond the "make it smart like ldconfig so that the 64bit one can handle both 32bit and 64bit modules and do the right thing" :-) For FC1, what's there will be fine, but I'd rather not carry the hack forward for all time if we can avoid it
http://bugzilla.gnome.org/show_bug.cgi?id=129540 as a sketch of the general solution. I believe that Jeremy has hacked together something for Anaconda, I'm going to close this one and leave further refinements tracked upstream.