[Bug 1868] New: ftol function from msvcrt emulation incorporates GNU-C specific artefact

[Wine Bugs]

http://bugs.winehq.com/show_bug.cgi?id=1868

Summary: ftol function from msvcrt emulation incorporates GNU-C
                    specific artefact
           Product: Wine
           Version: 20030813
          Platform: PC
        OS/Version: FreeBSD
            Status: UNCONFIRMED
          Severity: minor
          Priority: P2
         Component: wine-kernel
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

There is an feature in way how GNU-C generated code handles double to int 
conversion, i.e. the following code:
-----
#include <stdio.h>
int main() {double t[]= {
        2777777777.0, 
        1777777777.0, 
        };int i;

for(i=0;i<sizeof(t)/sizeof(*t);i++) {
                printf("%lf -> %#x\n", t[i], (long)t[i]);
        }
        return 1;
}
---
When compiled with GCC prints out 
2777777777.000000 -> 0x80000000
1777777777.000000 -> 0x69f6bc71
And when compiled with MSVC++ (use /MD and save as ftol.exe for the next test), 
prints out something better.
2777777777.000000 -> 0xa5918671
1777777777.000000 -> 0x69f6bc71

Unsuprisingly if ftol.exe launched under wine, it  prints out 0x80000000 as
result of first conversion. Using original msvcrt.dll with wine solves the 
problem. However problem could be solved in msvcr emulation layer by applying 
the following ha^H^Hpatch:
--- ./work/wine-20030813/dlls/ntdll/misc.c.orig Sun Dec  7 21:57:45 2003
+++ ./work/wine-20030813/dlls/ntdll/misc.c      Mon Dec  8 00:14:04 2003
@@ -59,8 +59,10 @@
        /* don't just do DO_FPU("fistp",retval), because the rounding
         * mode must also be set to "round towards zero"... */
        double fl;
+       long long r;
        POP_FPU(fl);
-       return (LONG)fl;
+       r = (long long)fl;
+       return (long)r;
 }
 #endif /* defined(__GNUC__) && defined(__i386__) */