[Boost-users] Re: Re: 'long long' problem on VC++7.1

From: Nick Savoiu (savoiu_at_[hidden])
Date: 2003-11-13 13:02:37

Take the code at:

http://www.boost.org/libs/graph/example/strong_components.cpp

Compile it as is with /W4 and /Za. It compiles albeit with a slew of
warnings.

Now move

#include <iostream>
#include <vector>

before

#include <boost/config.hpp>

and recompile.

You should get:

...\boost\boost\concept_check.hpp(183) : error C2632: 'long' followed by
'long' is illegal

Nick

"John Maddock" <john_at_[hidden]> wrote in message
news:006d01c3a9d9$db7e7100$a3880252_at_fuji...
> > The problem seems to be that BOOST_HAS_LONG_LONG has two different
checks
> to
> > see if should be defined:
> >
> > 1 @ E:\users\savoiu\Projects\boost\boost\config\suffix.hpp(26):#include
> > <limits.h>
> >
> > # if !defined(BOOST_HAS_LONG_LONG)
> > \
> >    && !(defined(BOOST_MSVC) && BOOST_MSVC <=1300) &&
> !defined(__BORLANDC__)
> > \
> >    && (defined(ULLONG_MAX) || defined(ULONG_LONG_MAX) ||
> > defined(ULONGLONG_MAX))
> > #  define BOOST_HAS_LONG_LONG
> > #endif
> > Adding a '(...) || !defined(_MSC_EXTENSIONS)' to the MSVC test in (1)
> fixes
> > the problem. But it seems to me that BOOST_HAS_LONG_LONG should only be
> > defined in one place to reduce the possibility of things being out of
> sync.
>
> Can you let me have a test case please: (1) does *not* result in
> BOOST_HAS_LONG_LONG getting defined - well not for me anyway with vc7.1
and
> the /Za option.
>
> Thanks,
>
> John.