bug: ambiguous fabs/sqrt/isdigit/etc

[Geoffrey Furnish]

Found also in the 20000821 snapshot, gcc claims this code is
ambiguous.

valinux[39] cat z.cc
#include <cmath>
using namespace std;

void f()
{
    double x = 1, y;
    y = fabs(x);
}
valinux[40] c++ -c z.cc
z.cc: In function `void f ()':
z.cc:7: call of overloaded `fabs (double &)' is ambiguous
/usr/include/bits/mathcalls.h:165: candidates are: double fabs (double)
/va1/furnish/gcc/include/g++-v3/bits/std_cmath.h:423:                 long double std::fabs (long double)
/va1/furnish/gcc/include/g++-v3/bits/std_cmath.h:310:                 double std::fabs (double)
/va1/furnish/gcc/include/g++-v3/bits/std_cmath.h:140:                 float std::fabs (float)
valinux[41] 

I think this must be a bug.  

This happens with other math ops mirrored by compiler intrinsics as
well, for example sqrt.

BTW, the same sort of problem also arises when including <cctype> and
using isdigit, etc.

-- 
Geoffrey Furnish            Actel Corporation        furnish@actel.com
Senior Staff Engineer      955 East Arques Ave       voice: 408-522-7528
Placement & Routing     Sunnyvale, CA   94086-4533   fax:   408-522-8041