Clobber layout test results at the beginning of each run....

Clobber layout test results at the beginning of each run.
This should fix issues with the layout_test_results.zip files
on the bots have results from other runs, which makes debugging
flakiness and other failures harder.

Add a flag, clobber-old-results that defaults to false.
We'll need to add the flag to the bots, but this will
maintain the old behavior for local runs of run_webkit_tests.

I'd also be fine not having a flag and always clobbering,
but I don't know if people rely on the current behavior
for their workflows.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=30736

[tony, 2009-11-02 19:36:54]

LGTM.  The other option would be to just have the buildbot do the clobbering
rather than this script.  I don't have a strong opinion so this seems fine to
me.

http://codereview.chromium.org/340064/diff/1/2
File webkit/tools/layout_tests/run_webkit_tests.py (right):

http://codereview.chromium.org/340064/diff/1/2#newcode1037
Line 1037: test_dirs = ["LayoutTests", "chrome", "pending"]
Nit: Use () instead of [] to make it immutable.