On 01/05/20 15:21 +0200, Jakub Jelinek via Libstdc++ wrote:
On Fri, May 01, 2020 at 02:17:54PM +0100, Jonathan Wakely via Gcc-patches wrote:
The libstdc++ manual documents that _T can not be used, because it's a
macro in system headers on some targets.
PR libstdc++/94901
* include/std/type_traits (__is_complete_or_unbounded): Replace
BADNAME _T with _Tp.
* testsuite/17_intro/badnames.cc: New test.
Tested powerpc64le-linux and x86_64-linux, committed to master.
This should be backported to gcc-10 too.
Ok for 10 branch, thanks.
+#define _E9 _E9 is a BADNAME
+ // ..
+#define _E24 _E24 is a BADNAME
Just wonder about the above, IMHO it is worth spelling them all in between
too, otherwise we don't catch _E10 or _E13 or _E23 in libstdc++ headers.
OK, can do.
I decided it was pretty unlikely we'd have _E10 without hitting one of
the earlier ones, but I guess it's possible somebody would define _E12
to mean 10e12 or something.
I'll add that to master and then backport the enhanced version.
