If the same test class is run over and over, we assume the existing browser session will suffice to handle it.  However, if the test fails to even get launched, we end up just sitting around waiting for nothing until timeouts occur for all the rest of the tests.  This patch records whenever the last run failed to start, so that it won't even try to run subsequent tests.  This particularly improves the failure mode when the test class module fails to compile.  Also bumped the timeout to 30 seconds instead of 20.

Review by: tobyr
           mmendez

git-svn-id: https://google-web-toolkit.googlecode.com/svn/trunk@724 8db76d5a-ed1c-0410-87a9-c151d255dfc7
1 file changed