Run large html tests individually.

tools/testing/dart/status_file_parser.dart

Index: tools/testing/dart/status_file_parser.dart
===================================================================
--- tools/testing/dart/status_file_parser.dart	(revision 14350)
+++ tools/testing/dart/status_file_parser.dart	(working copy)
@@ -25,7 +25,7 @@
 
 const RegExp StripComment = const RegExp("^[^#]*");
 const RegExp HeaderPattern = const RegExp(r"^\[([^\]]+)\]");
-const RegExp RulePattern = const RegExp(r"\s*([^: ]*)\s*:(.*)");

[Emily Fortuna, 2012/11/01 05:19:11]

I changed this regexp so that test descriptions could be used to name the
"subtests", for example, 
element_test/.html has no parent: Fail
in the status file. (".html has no parent" is the description given in a test
case in element_test.dart)

whesse, is there a specific reason why this was previously disallowed?

[Bill Hesse, 2012/11/01 20:26:00]

This might lead to exponential backtracking, because (x|space)* (space)* has
many matches.  Also, all the space between the rule header and the : will be
grabbed, then.
A efficient regexp grabbing "I am the walrus" from "     I am the walrus       :
foo foo foo" may be hard to do in one step.

On 2012/11/01 05:19:11, Emily Fortuna wrote:

+const RegExp RulePattern = const RegExp(r"\s*([^:]*)\s*:(.*)");
 
 // TODO(whesse): Implement configuration_info library that contains data
 // structures for test configuration, including Section.