Contribution of PowerPC port (continuation of 422063005) - AIX Common1

src/assembler.cc

Index: src/assembler.cc
diff --git a/src/assembler.cc b/src/assembler.cc
index 21f1715859859fb5cd54107a26e06b218c43a3cd..dba577266948f1d78d043a4aee8355f09213a1f7 100644
--- a/src/assembler.cc
+++ b/src/assembler.cc
@@ -1458,6 +1458,22 @@ double power_double_double(double x, double y) {
       return std::ldexp(1.0, y_int);
     }
   }
+#elif V8_OS_AIX

[Sven Panne, 2015/01/27 11:47:05]

Can we merge this with the case above? If I see this correctly, only the
copysign is new, and this should be OK above, too (I hope :-).

[michael_dawson, 2015/01/29 00:08:29]

It looks like there are a few other differences in that in the AIX case it
always returns while in the MinGW case it may not.  I created a merged version
which I think preserves the behaviour of both.  I validated the AIX case but we
compile with MSVC so I don't have a MinGW environment to validate that case.

[Sven Panne, 2015/01/29 09:54:33]

We don't test MINGW builds, either, it was just part of an external contribution
IIRC. Just use your AIX code for the MINGW stuff above, too, and let's see if
somebody complains. >:-) Having weird conditional untested code around is
definitely worse.

[michael_dawson, 2015/01/29 18:21:58]

ok will do

+  // AIX has a custom implementation for pow.  This handles certain
+  // special cases that are different.
+  if ((x == 0.0 || std::isinf(x)) && y != 0.0 && std::isfinite(y)) {
+    double f;
+    double result = ((x == 0.0) ^ (y > 0)) ? V8_INFINITY : 0;
+    /* retain sign if odd integer exponent */
+    return ((modf(y, &f) == 0.0) && (static_cast<int64_t>(y) & 1))
+               ? copysign(result, x)
+               : result;
+  }
+
+  if (x == 2.0) {
+    int y_int = static_cast<int>(y);
+    if (y == y_int) return ldexp(1.0, y_int);
+  }
 #endif
 
   // The checks for special cases can be dropped in ia32 because it has already