> On 2016-Feb-06, at 13:28, Eric Fiselier <e...@efcs.ca> wrote: > > EricWF added a comment. > > This is very subtly broken. The requirements in the standard for emplace are > > [unord.req] Table 102 > >> a_eq.emplace(args) > >> Requires: : value_type shall be EmplaceConstructible into X from args. > > > Unfortunately __hash_table doesn't know that unordered_map has this "special" > value_type, so it can't actually extract and construct the correct value_type > from it. I'm going to change > this over the weekend. Once I've done some prep work this patch should be > ready to go.
Nice catch. Even though this "special" value type is what makes the malloc-optimization so difficult, I missed that behaviour change here. Let me know if you run out of time to help with the prep, and I can take over. Thanks for helping. > > > http://reviews.llvm.org/D16792 > > > _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
