Roll third_party/double_conversion

third_party/double-conversion/README.google

Index: third_party/double-conversion/README.google
diff --git a/third_party/double-conversion/README.google b/third_party/double-conversion/README.google
new file mode 100644
index 0000000000000000000000000000000000000000..322fe21b3c986f2d8c6bcc1c787c5308ac7af6b6
--- /dev/null
+++ b/third_party/double-conversion/README.google
@@ -0,0 +1,30 @@
+URL: https://github.com/google/double-conversion
+Version: 56a04575a474b929a4aa4fb72d83d46c8a93f294
+License: BSD
+License File: LICENSE
+
+Description:
+This project (double-conversion) provides binary-decimal and decimal-binary
+routines for IEEE doubles.
+
+Local Modifications:
+Only the source is included (no build files), and the following patch is
+applied to src/double-conversion.cc:
+
+34,35c34,40
+< #include "fast-dtoa.h"
+< #include "fixed-dtoa.h"
+---
+> 
+> // Remove the double-based fast cases and use big int operations always for
+> // Dartino. This cuts more than 8KB off the binary size.
+> //
+> // #include "fast-dtoa.h"
+> // #include "fixed-dtoa.h"
+> 
+387a393,395
+>   // Remove the double-based fast cases and use big int operations always for
+>   // Dartino. This cuts more than 8KB off the binary size.
+>   /*
+408a417
+>   */