Frank Lichtenheld wrote: > On Wed, May 04, 2005 at 12:27:44PM +0200, Marco van Zwetselaar wrote: > >> qtstalker (0.26-5) unstable; urgency=low >> >> * Bumped libmysqlclient deps to fix FTBFS in sarge, closes: #306240 >> * Retrofitted workaround for another qmake bug in debian/rules > > I wonder why you didn't used a Build-Conflicts against qtstalker-doc > instead of the rather lengthy and clumsy > > + # Prevent another qmake bug, see > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=259081 > + @ [ ! -L /usr/share/doc/qtstalker/html ] || ( \ > + echo "It seems qtstalker-doc is installed in your build > environment; this will trigger a qmake bug (see Debian bug #259081)." && \ > + echo "Please uninstall qtstalker-doc from your build > environment." && false )
Good point. I guess that didn't come to mind because there isn't a build conflict with qtstalker-doc per se. The problem is in qmake, which generates faulty makefiles whenever the local file system contains symlinks somewhere on an install path (bug #259081). The -L was primarily intended to prevent the most likely occurrence of that: when rebuilding the package after having done a debi and test run. I agree that a Build-Conflicts would work equally well in that situation. Good thing about the -L test is that it will also work for people with a home-installed Qtstalker version on their system. What about adding a Build-Conflicts /and/ keeping the -L test in? Best regards, Marco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
