https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70095
Jason Merrill <jason at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution|--- |FIXED Target Milestone|--- |6.0 --- Comment #4 from Jason Merrill <jason at gcc dot gnu.org> --- (In reply to Tom Honermann from comment #2) > The change in comment 1 introduced a regression. The following test passes > with r234230, but fails with r234231: > > $ cat t.cpp > template <class> > constexpr bool b = false; > template<typename T> > constexpr bool b<T*> = true; > int main() { > b<int*>; > b<double*>; > } > > $ g++ -std=c++14 t.cpp -o t > /tmp/ccJTAQId.s: Assembler messages: > /tmp/ccJTAQId.s:27: Error: symbol `_ZL1b' is already defined I think this was fixed by the patch for 69515. In any case, it works now.