GCC build failed for native with your patch on 2003-11-20T22:26:08Z.
Geoff Keating
geoffk@geoffk.org
Fri Nov 21 01:09:00 GMT 2003
Benjamin Kosnik <bkoz@redhat.com> writes:
> >The build failures are new.
>
> the build log clearly implicates my check in in the bulid error.
> However, this builds on linux for me, with the same codepath.
> Furthermore, the error happens at this line:
>
> template<typename _T>
> void
> _Error_formatter::_M_format_word(char* __buf,
> int __n __attribute__((__unused__)), <-----------HERE
> const char* __fmt, _T __s) const
> {
> #ifdef _GLIBCXX_USE_C99
> std::snprintf(__buf, __n, __fmt, __s);
> #else
> std::sprintf(__buf, __fmt, __s);
> #endif
> }
>
>
> it's bailing at the __attribute__.
>
> What gives?? There's this code all over libstdc++-v3.
I see this error message:
/Users/regress/tbox/cvs-gcc/gcc/libstdc++-v3/src/debug.cc:499: error: expected nested-name-specifier
and line 499 is the one with 'template<typename _T>'.
It looks like _T was a poor choice of name, because in
/usr/include/ctype.h there is:
#define _T _CTYPE_T /* Special */
--
- Geoffrey Keating <geoffk@geoffk.org>
More information about the Gcc-regression
mailing list