riccibruno added a comment. In D57914#1416215 <https://reviews.llvm.org/D57914#1416215>, @jyknight wrote:
> The intricate initialization-order workarounds apparently don't work in all > build modes, so I've updated this code to have constexpr functions and > initializations in rL355278 <https://reviews.llvm.org/rL355278>. Looking at what guarantees are given for dynamic initialization (in [basic.start.dynamic]), it seems that no guarantees are given if the variable is an explicit or implicit instantiation. So unless I am mistaken the original code was not guaranteed to work indeed. Apologies for missing that! Repository: rL LLVM CHANGES SINCE LAST ACTION https://reviews.llvm.org/D57914/new/ https://reviews.llvm.org/D57914 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits