On Wed, 3 Apr 2024 at 15:34, Martin Uecker via Gcc <gcc@gcc.gnu.org> wrote: > I ask a very specific question: To what extend is testing > for features instead of semantic versions and/or supported > standards still necessary? This seems like a problematic approach > that may have been necessary decades ago, but it seems it may be > time to move on.
What standard or version should I check for to detect nl_langinfo_l support? It's in POSIX 2017 but macOS supports it in <xlocale.h> despite not being POSIX 2017 compliant. What about _get_osfhandle? Or fwrite_unlocked? What about whether the platform ABI allows aligning global objects to the cacheline size? Those are just a few of the most recent autoconf checks I've written for libstdc++ in the past few months, not decades ago.
