This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Bootstrap failure in libstdc++-v3 on GNU/Linux x86
- From: Eric Botcazou <ebotcazou at libertysurf dot fr>
- To: Florian Weimer <fw at deneb dot enyo dot de>
- Cc: gcc at gcc dot gnu dot org,Mark Mitchell <mark at codesourcery dot com>
- Date: Mon, 28 Jun 2004 08:39:04 +0200
- Subject: Re: Bootstrap failure in libstdc++-v3 on GNU/Linux x86
- References: <87oen4l12v.fsf@deneb.enyo.de>
> I get a very obscure bootstrap failure for mainline:
>
> In file included from
> /home/fw/src/gnu/gcc/libstdc++-v3/src/complex_io.cc:30:
> /tmp/gcc/i686-pc-linux-gnu/libstdc++-v3/include/complex: In member
> function `long double& std::complex<long double>::real()':
> /tmp/gcc/i686-pc-linux-gnu/libstdc++-v3/include/complex:1296: error:
> invalid initialization of reference of type 'long double&' from expression
> of type 'long int'
> /tmp/gcc/i686-pc-linux-gnu/libstdc++-v3/include/complex: In member
> function `long double& std::complex<long double>::imag()':
> /tmp/gcc/i686-pc-linux-gnu/libstdc++-v3/include/complex:1304: error:
> invalid initialization of reference of type 'long double&' from expression
> of type 'long int'
Same on sparc-sun-solaris2.8.
> This doesn't make sense because the member function in question reads:
>
> inline long double&
> complex<long double>::real()
> { return __real__ _M_value; } // line 1296
>
> And _M_value does have type __complex__ long double. Should I start a
> binary search, or is this known to be some kind of recent breakage?
Could be Mark's latest patch to the parser.
--
Eric Botcazou