On Fri, Sep 28, 2007 at 02:32:16PM +0200, Stefano Zacchiroli wrote: > On Fri, Sep 28, 2007 at 01:17:00PM +0100, Richard Jones wrote: > > No I don't actually. It's strange because I didn't hit this bug at > > all when compiling ocamlnet for Fedora. > > Are you building the Apache connector? The bug manifests itself only if > you're building that, and only on architecture in which PIC code is > actually different than non-PIC code.
I thought we were but I just checked the specfile and it turns out we aren't. > > Unfortunately Fedora policy stops me from incorporating this fix: > > http://caml.inria.fr/mantis/view.php?id=3866 > > into our OCaml. It would have to be accepted into upstream (INRIA's) > > OCaml first. > > Why? Because of our policy ... http://fedoraproject.org/wiki/PackageMaintainers/WhyUpstream In particular I could do it if INRIA said that they would support the change in some future release (see the exception "Patches Heading Upstream"). But otherwise this is quite a large ABI change -- if Fedora users started to build lots of 64 bit shared libraries linked with -lcamlrun I could end up maintaining it separately forever. Rich. -- Richard Jones Red Hat -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]