https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126111
--- Comment #10 from Sebastian Jude <sebastian.jude at st dot ovgu.de> --- (In reply to Jonathan Wakely from comment #9) > Ah, so the problem is HAVE_STRTOLD not HAVE_BROKEN_STRTOLD. I assumed the > latter (which is only relevant for HP-UX, I think). > > Not having long double support on microcontrollers is more common, and > probably means that std.cc can't be compiled on a lot of small systems. > Ideally we would support that. Understand. TBH, that "#if (_GLIBCXX_HAVE_STRTOLD && ! _GLIBCXX_HAVE_BROKEN_STRTOLD)" was just a copy+paste from what i saw around the definition of std::stold, so i thought it makes sense to use the exact same guard here as well. I didn't even know what _GLIBCXX_HAVE_BROKEN_STRTOLD was good for ;) > Is strtold the only problem for your target? std::sinl etc. are present? Those are the only problems with std module exports i encountered so far and i got it to build for my target after applying the suggested fixes. But i did not look into it in depth, its just a "does it compile and can i use it?" test.
