Rewrite double.parse.

sdk/lib/core/double.dart

Index: sdk/lib/core/double.dart
diff --git a/sdk/lib/core/double.dart b/sdk/lib/core/double.dart
index 65b9c3096ed87dacce8d40d51f29c4892a466eca..ef0953f31d9fdb8770ab2440595e76916dfa6f80 100644
--- a/sdk/lib/core/double.dart
+++ b/sdk/lib/core/double.dart
@@ -133,16 +133,31 @@ abstract class double extends num {
   /**
    * Parse [source] as an double literal and return its value.
    *
-   * Accepts the same format as double literals:
-   *   [: ['+'|'-'] [digit* '.'] digit+ [('e'|'E') ['+'|'-'] digit+] :]
+   * Accepts an optional sign followed by either "Infinity",

[Lasse Reichstein Nielsen, 2013/05/22 08:47:26]

.. optional sign (`+` or `-`) immediately followed by either the characters
"Infinity", the characters "NaN" or a ...

[floitsch, 2013/05/23 02:30:03]

Done.

+   * "NaN" or a floating-point representation. A floating-point number is

[Lasse Reichstein Nielsen, 2013/05/22 08:47:26]

"floating-point representation", then "floating point number". Probably change
the "number" to "representation".

[floitsch, 2013/05/23 02:30:03]

Done.

+   * composed of the mantissa and and optional exponent part. The mantissa is

[Lasse Reichstein Nielsen, 2013/05/22 08:47:26]

"the mantissa" -> "a mantissa"
"and and" -> "and an".

[floitsch, 2013/05/23 02:30:03]

Done.

+   * either a dot followed by a sequence of digits, or a sequence of digits

[Lasse Reichstein Nielsen, 2013/05/22 08:47:26]

"dot" -> "decimal point (`.`)"
"digts" -> "(decimal) digits"

[floitsch, 2013/05/23 02:30:03]

Done.

+   * optionally followed by a dot and more digits. The (optional) exponent part

[Lasse Reichstein Nielsen, 2013/05/22 08:47:26]

This disallows "4." (which I approve of!) but probably isn't what you want.
"dot" -> "decimal point".

[floitsch, 2013/05/23 02:30:03]

Changed to "and optionally more digits. Did not repeat the "decimal".

+   * consists of the character "e" or "E", an optional sign, and the exponent

[Lasse Reichstein Nielsen, 2013/05/22 08:47:26]

"and the exponent digit(s)" -> "and one or more digits"

[floitsch, 2013/05/23 02:30:03]

Done.

+   * digit(s).
    *
-   * Also recognizes "NaN", "Infinity" and "-Infinity" as inputs and
-   * returns the corresponding double value.
+   * The whole string may be surrounded by whitespace.

[Lasse Reichstein Nielsen, 2013/05/22 08:47:26]

Refer to [String.trim] for a definition of "whitepsace", if possible.

[floitsch, 2013/05/23 02:30:03]

Changed to: "The input string is trimmed (see [String.trim]) before conversion.

    *
-   * If the [soure] is not a valid double literal, the [handleError]
+   * If the [source] is not a valid double literal, the [handleError]
    * is called with the [source] as argument, and its return value is
    * used instead. If no handleError is provided, a [FormatException]
    * is thrown.
+   *
+   * Examples of accepted strings:
+   *
+   *     "3.14"
+   *     "  3.14 \xA0"
+   *     "0."
+   *     ".0"
+   *     "-1.e3"
+   *     "1234E+7"
+   *     "+.12e-9"
+   *     "-NaN"
    */
   external static double parse(String source,
                                [double handleError(String source)]);