[llvm-commits] CVS: llvm/lib/Support/APInt.cpp

[Evan Cheng]

Hi Reid,

This breaks the build for me.

$ make ENABLE_OPTIMIZED=1 -j2
make[1]: Nothing to be done for `all'.
llvm[1]: Compiling Annotation.cpp for Release build
llvm[1]: Compiling APInt.cpp for Release build
APInt.cpp:841: error: floating constant exceeds range of 'float'
APInt.cpp:843: error: floating constant exceeds range of 'float'

Evan

On Feb 20, 2007, at 10:29 AM, Reid Spencer wrote:

>
>
> Changes in directory llvm/lib/Support:
>
> APInt.cpp updated: 1.22 -> 1.23
> ---
> Log message:
>
> Use INFINITY macro from math.h instead of constructing hex floating  
> point
> constants (avoids warnings).
>
>
> ---
> Diffs of the changes:  (+3 -2)
>
>  APInt.cpp |    5 +++--
>  1 files changed, 3 insertions(+), 2 deletions(-)
>
>
> Index: llvm/lib/Support/APInt.cpp
> diff -u llvm/lib/Support/APInt.cpp:1.22 llvm/lib/Support/APInt.cpp: 
> 1.23
> --- llvm/lib/Support/APInt.cpp:1.22	Tue Feb 20 02:51:03 2007
> +++ llvm/lib/Support/APInt.cpp	Tue Feb 20 12:29:12 2007
> @@ -17,6 +17,7 @@
>  #include "llvm/Support/MathExtras.h"
>  #include <cstring>
>  #include <cstdlib>
> +#include <cmath>
>  using namespace llvm;
>
>  // A utility function for allocating memory, checking for  
> allocation failures,
> @@ -837,9 +838,9 @@
>    // Return infinity for exponent overflow
>    if (exp > 1023) {
>      if (!isSigned || !isNeg)
> -      return double(0x0.0p2047L); // positive infinity
> +      return double(INFINITY); // positive infinity
>      else
> -      return double(-0x0.0p2047L); // negative infinity
> +      return double(-INFINITY); // negative infinity
>    }
>    exp += 1023; // Increment for 1023 bias
>
>
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits