https://bugzilla.redhat.com/show_bug.cgi?id=833622 --- Comment #13 from Ralf Corsepius <rc040203@xxxxxxxxxx> --- (In reply to comment #12) > (In reply to comment #11) > > * What is reason to run the autotools while building? > > gmp is supposed to build fine for mingw without it (and it actually does). > > I was following the native package spec. I have removed this. Thanks. > > * What is the reason to ship and use gmp.h and gmp-mparam.h? > > See comment 10 or the comment I left in the spec. I followed the native > package in shipping these wrappers. I will copy it again here: > # Some apps seem to assume that they are building against the > # gmp source tree and require the source versions of the gmp.h > # and gmp-mparam.h files. Well, * the gmp.h-wrapper (gmp.h) is a (RH/Fedora-specific) cludge to work-around the original gmp.h not being multilib-capable. I.e. this wrapper is not required on single-arched/lib'ed systems, such as mingw. * The gmp-mparam.h-wrapper is a similar cludge/hack aiming at gmp-mparam.h not being multlib-capable, with similar considerations applying to it. The delicacy behind this: gmp does not export the gmp-mparam.h header. => No gmp package should ship it. Packages expecting it should be considered broken (I guess, this is what is meant by "some apps seem to .. gmp source-tree" in the comment above.) In other words, both wrappers are not necessary for mingw, shipping the gmp.h wrapper makes some (limited) sense, but shipping gmp-mparam.h doesn't. -- You are receiving this mail because: You are on the CC list for the bug. _______________________________________________ package-review mailing list package-review@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/package-review