On Sun, 2014-09-14 at 06:14 +0200, Ralf Corsepius wrote: > On 09/13/2014 01:52 PM, Florian Weimer wrote: > > On 09/12/2014 06:25 PM, Michael Cronenworth wrote: > >> You'll have to download all the source and perform a massive grep. > > > > You have to unpack the sources before the grep, which can be quite a > > challenge because SPEC files aren't declarative. The %prep stage > > sometimes even needs build dependencies already installed. :-( > > This is the list, I got from doing something similar as you outlined > (unpackaging src.rpm, running prep, and find -name "valgrind.h") > > 0ad-0.0.16-10.fc22 > cockpit-0.23-1.fc22 > condor-8.1.4-7.a1a7df5.fc22 > exim-4.84-3.fc22 > fwbuilder-5.1.0.3599-5.fc20 > gcr-3.13.91-1.fc22 > gearmand-1.1.12-8.fc22 > ghostscript-9.14-6.fc22 > glib2-2.41.4-3.fc22 > gperftools-2.2.1-2.fc22 > ipxe-20140303-3.gitff1e7fc7.fc22 > libgnome-keyring-3.12.0-4.fc22 > libmemcached-1.0.18-4.fc22 > libsecret-0.18-4.fc22 > lwp-2.6-10.fc22 > mingw-glib2-2.41.2-1.fc22 > mingw-qt5-qtjsbackend-5.1.1-4.fc21 > mongodb-2.4.9-7.fc22 > mono-2.10.8-7.fc21 > openvswitch-2.3.0-1.fc22 > pidgin-2.10.9-5.fc22 > planner-0.14.6-17.fc22 > qemu-2.1.1-1.fc22 > R-3.1.1-5.fc22 > re2-20131024-3.fc22 > realmd-0.15.1-4.fc22 > rubygem-passenger-4.0.18-9.fc22 > squid-3.4.7-1.fc22 > v8-3.14.5.10-13.fc22 > valgrind-3.10.0-1.fc22 > wine-mono-4.5.2-4.fc21 Thanks. That is a much bigger list than the packages I already filed bugs for based on the reproquery against the debuginfo packages. https://bugzilla.redhat.com/show_bug.cgi?id=1141461 Specifically missing are: cockpit, exim, gearmand, ghostscript, ipxe, libmemcached, mingw-glib2, mingw-qt5-qtjsbackend, mongodb, openvswitch, qemu, R, realmd, rubygem-passenger, squid, wine-mono. I think that means they either don't enable valgrind support in the binary package or they don't generate proper debuginfo. I assume it still makes sense to file a bug report against these packages so the maintainer can investigate. If it turns out the package source does include a private copy of valgrind.h, but they don't actually use/activate support for it in the binary package, how should the package be marked? Thanks, Mark BTW. Sorry for the slow responses and for not adding CCs to my messages. Apparently the fedora-devel-list mangles the Reply-To header somehow so replies are sometimes lost (or at least don't go to individuals, but only to the list). -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
