rengolin added a comment. In http://reviews.llvm.org/D18035#440107, @rsmith wrote:
> Please try to appropriately apportion the responsibility here; if your > distribution opted into a non-standard ABI for their C++ standard library, > you should point out to them that they made a mistake. I think I have been clear enough on the bug and the list about responsibility... It is clear that the distros screwed that up on their own, and neither GCC not LLVM could do much to work around the complete lack of proper testing before such a big decision. In no way I think this is something we could have done anything to prevent from happening, nor I think that this patch is the best way forward for the future. I only asked this to be reviewed and accepted IFF it's the right (temporary) implementation, so that distros can pull those patches with a bit more confidence. If it makes to 3.8.1, that'd make things a lot easier for *them* (not us). If it doesn't, it wouldn't make it much harder (they can already pull odd patches). But since this has been tested by many people, including active usage in distros, and it doesn't touch anything outside the ABI tag issue, we could have it in (at least on trunk), so that we can test it better ourselves until 3.9 branches. I really don't want to pull this one in just before we branch. Nor just after. Regardless of who screwed up, having a fix would be good to LLVM as well, not just the distros. cheers, --renato http://reviews.llvm.org/D18035 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
