https://bugzilla.redhat.com/show_bug.cgi?id=1083962 --- Comment #9 from Oden Eriksson <oe@xxxxxx> --- (In reply to Paulo Andrade from comment #8) > On a quick look, I can say that I do not like when > one does not run a proper "make install" in %install, > this can easily break in subtle ways, if not in the > first spec write, whenever the package is updated. > I see the Makefile at least appears to respect > $DESTDIR for the install target. I know. The problem is that libtool for some reason swaps linking order when relinking while doing make install. I'm not a autopoo or libtool guru enough to fix that. Upstream knows that autopoo should be implemented and this might happen in a future release. > You should run "make testall" in %check and give > good reasons if it does not work. This way both you > and reviewer will at least have a good hint the > package is functional if it pass %check. Same here. Poor autopoo support. > Please post a link to the failed arm build, to > have an idea of why it fails. http://koji.fedoraproject.org/koji/taskinfo?taskID=6713406 > I suggest renaming the patches to maxscale-xyz.patch, > and use "PatchN: %{name}-xyz.patch", but this is a > cosmetic change, just a common pattern in fedora > packages. > > Instead of using CFLAGS="%{optflags} -fpie" you > should use https://fedoraproject.org/wiki/Packaging:Guidelines#PIE > that automatically does it for you; could then just need > to force CFLAGS="%{optflags} and LDFLAGS="%{__global_ldflags}" > But I am not sure if it is handled the same way for rhel. No it doesn't. Sure, I could use conditionals here but... -- You are receiving this mail because: You are on the CC list for the bug. You are always notified about changes to this product and component _______________________________________________ package-review mailing list package-review@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/package-review