sebpop added a comment. In https://reviews.llvm.org/D22782#495436, @mclow.lists wrote:
> Do we have a test for the problem that this is solving? I think we can write a testcase that shows that copy constructors are not optimized away unless the string constructor is inlined. This patch fixes the performance of a proprietary benchmark when compiled with libc++, closing the performance gap with the same benchmark compiled with the gnu libstdc++. Overall with this patch we have half a billion fewer instructions out of about 10 billion. https://reviews.llvm.org/D22782 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
