[Bug middle-end/17603] [4.0 Regression] Cpowf and cpowl builtins broken

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-22 12:57 -------
Hmm this works for me even with the float version, I really suspect that this glibc bug.

The tree level looks right:
  REALPART_EXPR <x> = -1.0e+0;
  IMAGPART_EXPR <x> = 0.0;
  REALPART_EXPR <y> = 5.0e-1;
  IMAGPART_EXPR <y> = 0.0;
  z = __builtin_cpow (x, y);
  D.1119 = IMAGPART_EXPR <z>;

So does the Assembly on powerpc-darwin:
        li r5,0
        mflr r31
        li r6,0
        stw r0,8(r1)
        lis r3,0xbff0
        stwu r1,-96(r1)
        li r4,0
        lis r7,0x3fe0
        li r8,0
        mr r9,r5
        mr r10,r6

I wonder if someone broke the ABI.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17603