Bug 962488 - Update double-conversion for proper AArch64 support. r=froydnj

Fix imported from upstream repository:
https://code.google.com/p/double-conversion/source/detail?r=4e24bb31bcc76d6d218f3056b4c24a109d367561

mfbt/double-conversion/fix-aarch64-macro.patch

@@ -0,0 +1,23 @@
+Backport from upstream.
+
+https://code.google.com/p/double-conversion/source/detail?r=4e24bb31bcc76d6d218f3056b4c24a109d367561
+
+---
+ mfbt/double-conversion/utils.h |    2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+--- a/mfbt/double-conversion/utils.h
++++ b/mfbt/double-conversion/utils.h
+@@ -58,11 +58,11 @@
+     defined(__mips__) || \
+     defined(__powerpc__) || defined(__ppc__) || defined(__ppc64__) || \
+     defined(__sparc__) || defined(__sparc) || defined(__s390__) || \
+     defined(__SH4__) || defined(__alpha__) || \
+     defined(_MIPS_ARCH_MIPS32R2) || \
+-    defined(_AARCH64EL_)
++    defined(__AARCH64EL__)
+ #define DOUBLE_CONVERSION_CORRECT_DOUBLE_OPERATIONS 1
+ #elif defined(_M_IX86) || defined(__i386__) || defined(__i386)
+ #if defined(_WIN32)
+ // Windows uses a 64bit wide floating point stack.
+ #define DOUBLE_CONVERSION_CORRECT_DOUBLE_OPERATIONS 1

mfbt/double-conversion/update.sh

@@ -20,4 +20,7 @@ patch -p3 < add-mfbt-api-markers.patch
 patch -p3 < use-StandardInteger.patch
 patch -p3 < use-mozilla-assertions.patch
 patch -p3 < use-static_assert.patch
-patch -p3 < ToPrecision-exponential.patch
+patch -p3 < ToPrecision-exponential.patch
+
+# Merged upstream, part of 2.0.1 version
+patch -p3 < fix-aarch64-macro.patch

mfbt/double-conversion/utils.h

@@ -60,7 +60,7 @@
     defined(__sparc__) || defined(__sparc) || defined(__s390__) || \
     defined(__SH4__) || defined(__alpha__) || \
     defined(_MIPS_ARCH_MIPS32R2) || \
-    defined(_AARCH64EL_)
+    defined(__AARCH64EL__)
 #define DOUBLE_CONVERSION_CORRECT_DOUBLE_OPERATIONS 1
 #elif defined(_M_IX86) || defined(__i386__) || defined(__i386)
 #if defined(_WIN32)