On Nov 02, 2011, at 17:07, Matthias Klose wrote: > On 11/02/2011 04:09 PM, Moffett, Kyle D wrote: >> Hello, >> >> I have verified that gcc-4.6 and gcj-4.6 both build successfully >> on "powerpcspe" with that patch. It turns out gcj needs to build >> an embedded copy of libffi (is that OK under Debian policy?), so >> I also had to apply the new libffi patch in #647288 [1]. >> >> This patch modifies a case conditional that only applies for SPE, >> so it should be safe to merge into the existing multiarch patch. > > the multiarch dir is already encoded in the rs600/t-spe file; can we move the > MULTIARCH_DIRNAME macro to this file as well? Or the other way around?
Well, that "t-spe" file is not being included at all for our target triplet right now; it only gets referenced for "powerpc-linux-eabispe" targets as per "src/gcc/config.gcc". Furthermore, the Debian PowerPCSPE port does not want any of the stuff in that file, as it's all multilib stuff that doesn't apply to e500 anyways. So the MULTILIB_OSDIRNAMES value in "t-spe" needs to be changed to "eabispe" upstream as a bugfix, but that is not applicable to the Debian e500 port. I suppose in theory PPC EABI is better than PPC SYSV and if I'd have given it a bit more thought I might have tried to build the port based on "eabispe" in the first place, but that's water under the bridge at this point. Even still, we wouldn't want the multilib stuff anyways. Cheers, Kyle Moffett -- Curious about my work on the Debian powerpcspe port? I'm keeping a blog here: http://pureperl.blogspot.com/ -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org