https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120993
--- Comment #3 from Paul Eggert <eggert at cs dot ucla.edu> --- Thanks for following up. I understand the backward-compatibility argument for leaving things alone in this messy area, even if the resulting macros don't conform to C23. Should this conformance issue be documented in the GCC manual? I didn't see any obvious place to do that. Since Gnulib already overrides LDBL_MAX on this platform, I patched Gnulib to override LDBL_NORM_MAX as well[1]. Admittedly a lot of this is notional, as this platform's long double operations are pretty flaky near LDBL_MAX. [1]: https://lists.gnu.org/r/bug-gnulib/2025-07/msg00036.html
