Richard W.M. Jones wrote: > This question has come up a lot: Whether we should backport changes > from Rawhide to EL-5/F-10. A corollary is whether we should keep > certain packages in synch across the different branches (particularly > mingw32-filesystem). This is my opinion on this subject. Comments & > follow-ups welcome. > > No we should not do this, except in the following cases: > > (1) When a package is approved after a Fedora review, we should create > new packages in EL-5/F-10 as long as this doesn't require a huge > amount of effort. > > (2) We should upgrade those packages to track the native versions of > packages[1] and where there are security updates. > > (3) If someone asks *and* is willing to do all the required work > (provide a sensible, working patch), then we or they can upgrade the > package. > > But apart from those reasons, we should leave working packages in > EL-5/F-10 untouched. > > In particular there are two outstanding patches from mingw32- > filesystem-44 in Rawhide which haven't been backported. The C++ > compiler detection stuff isn't necessary for F-10 or EL-5 - it appears > that the brokenness comes about as a side-effect of something in > libtool-2. The rpmlint message suppression is cosmetic. So both > patches fall under clause (3) above. > > Rich. > > [1] We need to get Dan's compare stuff working ...
mostly agree, but eg. for filesystem these patches usually happened because of some new RR required it. imho it's currently enough, when all mingw32* packages pushed into fedora then may be a big rebuild would be useful on all distro. -- Levente "Si vis pacem para bellum!" _______________________________________________ fedora-mingw mailing list [email protected] https://admin.fedoraproject.org/mailman/listinfo/fedora-mingw
