Reviewers: Yang,
Message:
Original CL https://chromiumcodereview.appspot.com/9595012/ by Jonathan Liu
<[email protected]>. Added cosmetic changes.
LGTM and landing.
Description:
Fix double-rounding in strtod for MinGW.
BUG=v8:1062
TEST=
Please review this at http://codereview.chromium.org/9578011/
SVN Base: https://v8.googlecode.com/svn/branches/bleeding_edge
Affected files:
M src/strtod.cc
Index: src/strtod.cc
diff --git a/src/strtod.cc b/src/strtod.cc
index
be79c80085200dc8015e53687e6e3e9c80f38418..0dc618a399f9df8b0ecb9c8b3039bd4edb158727
100644
--- a/src/strtod.cc
+++ b/src/strtod.cc
@@ -1,4 +1,4 @@
-// Copyright 2011 the V8 project authors. All rights reserved.
+// Copyright 2012 the V8 project authors. All rights reserved.
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
@@ -175,13 +175,15 @@ static void ReadDiyFp(Vector<const char> buffer,
static bool DoubleStrtod(Vector<const char> trimmed,
int exponent,
double* result) {
-#if (defined(V8_TARGET_ARCH_IA32) || defined(USE_SIMULATOR))
&& !defined(WIN32)
+#if (defined(V8_TARGET_ARCH_IA32) || defined(USE_SIMULATOR)) \
+ && !defined(_MSC_VER)
// On x86 the floating-point stack can be 64 or 80 bits wide. If it is
// 80 bits wide (as is the case on Linux) then double-rounding occurs
and the
// result is not accurate.
- // We know that Windows32 uses 64 bits and is therefore accurate.
- // Note that the ARM simulator is compiled for 32bits. It therefore
exhibits
- // the same problem.
+ // We know that Windows32 with MSVC, unlike with MinGW32, uses 64 bits
and is
+ // therefore accurate.
+ // Note that the ARM and MIPS simulators are compiled for 32bits. They
+ // therefore exhibit the same problem.
return false;
#endif
if (trimmed.length() <= kMaxExactDoubleIntegerDecimalDigits) {
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
