This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug math/21930] C-only gcc builtins used in <math.h> isinf
- From: "joseph at codesourcery dot com" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sourceware dot org
- Date: Mon, 21 Aug 2017 13:37:49 +0000
- Subject: [Bug math/21930] C-only gcc builtins used in <math.h> isinf
- Auto-submitted: auto-generated
- References: <bug-21930-131@http.sourceware.org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=21930
--- Comment #9 from joseph at codesourcery dot com <joseph at codesourcery dot com> ---
Most of the macros, including fpclassify, should be irrelevant because
libstdc++ headers undefine the macros and replace them with function
overloads (as required by the C++ standard). The problem with iszero is
specifically that the C++ iszero definition in the C math.h is using the
macro fpclassify definition (this is before libstdc++ gets a chance to
undefine it), and, in the case (__SUPPORT_SNAN__ and float128 supported),
that definition is using __builtin_types_compatible_p. So some approach
is needed for iszero to work in that place - without conflicting with
libstdc++'s own fpclassify overloads.
--
You are receiving this mail because:
You are on the CC list for the bug.