This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: basic_istream error building SPEC eon on AIX
- From: Gabriel Dos Reis <gdr at integrable-solutions dot net>
- To: David Edelsohn <dje at watson dot ibm dot com>
- Cc: libstdc++ at gcc dot gnu dot org
- Date: 12 Aug 2003 03:59:57 +0200
- Subject: Re: basic_istream error building SPEC eon on AIX
- Organization: Integrable Solutions
- References: <200308112255.SAA28402@makai.watson.ibm.com>
David Edelsohn <dje@watson.ibm.com> writes:
| >>>>> Gabriel Dos Reis writes:
|
| Gaby> | std::basic_istream<char, std::char_traits<char> >::get(unsigned char&)'
|
| Gaby> This does not exist.
|
| 252.eon includes the following code:
|
| ggRasterSurfaceTexture::ggRasterSurfaceTexture( istream& in) {
| ...
| unsigned char red, green, blue;
| in.get(red); // eat a byte
Ah, then, that is bogus.
| ...
| in.get(red);
| in.get(green);
| in.get(blue);
|
|
| Why did this work before?
I have no theory for the moment. Do they still have a working SPEC on eon?
If GCC accepted the above in the past, it was simply by mistake.
| Why do the GCC SPEC testers runs by Andreas and Diego compile this?
|
| Is this a case of G++ being too pedantic in non-pedantic mode?
No, I don't beleive so. You cannot pass a T& when a U& is expected
unless U and T are related. And the case above is explicitly invalidated.
-- Gaby