Math.h error "long double" compile flag?

[nirav dot dalal at teradyne dot com]

Hi,

I'm having issues with compiling gcc when math.h is included.

I have tried many switches to include math.h, like
-lm, -ffast-math, -m128bit-long-double, -m96bit-long-double

but everytime it gives me the following errors:
----------------------------------------------------------------------------
c:/mingw/include/math.h:276: error: short, signed or unsigned invalid for
`x'
c:/mingw/include/math.h:315: error: short, signed or unsigned invalid for
`_x'
c:/mingw/include/math.h:344: error: short, signed or unsigned invalid for
`x'
c:/mingw/include/math.h:356: error: short, signed or unsigned invalid for
`sinl
   '
c:/mingw/include/math.h:356: error: short, signed or unsigned invalid for `
   parameter'
c:/mingw/include/math.h:359: error: short, signed or unsigned invalid for
`cosl
   '
c:/mingw/include/math.h:359: error: short, signed or unsigned invalid for `
   parameter'
c:/mingw/include/math.h:362: error: short, signed or unsigned invalid for
`tanl
   '
c:/mingw/include/math.h:362: error: short, signed or unsigned invalid for `
   parameter'
c:/mingw/include/math.h:365: error: short, signed or unsigned invalid for `
   asinl'
c:/mingw/include/math.h:365: error: short, signed or unsigned invalid for `
   parameter'
..........
---------------------------------------------------------------------------

The issue seems to be not accepting "long double" format somehow in math.h
file.

The gcc compile command used for x86 target is as under:
>gcc -I "c:\program files\teradyne\ig-xl\3.64.50\inc" -mdll -DBUILD_DLL -O3
-c
-fno-exceptions -lm -ffast-math -m128bit-long-double -m96bit-long-double
*.cpp

Any help on this problem is appreicated.

Thanks,
- Nirav