Attempt to reduce Dartium debug bot redness.

tests/co19/co19-dartium.status

Index: tests/co19/co19-dartium.status
diff --git a/tests/co19/co19-dartium.status b/tests/co19/co19-dartium.status
index 71561d7d98e9a9392bc819fb14a79073e94f2879..096061d943137439485d09fee25a99a868512725 100644
--- a/tests/co19/co19-dartium.status
+++ b/tests/co19/co19-dartium.status
@@ -637,7 +637,7 @@ LayoutTests/fast/css/counters/complex-before_t01: RuntimeError, Pass # co19-roll
 LayoutTests/fast/css/getComputedStyle/computed-style-select-overflow_t01: RuntimeError # co19-roll r761: Please triage this failure.
 LayoutTests/fast/css/vertical-align-length-copy-bug_t01: RuntimeError # co19-roll r761: Please triage this failure.
 
-[ $compiler == none && $runtime == dartium ]
+[ $compiler == none && $runtime == dartium && ($mode != debug || $system != macos) ]

[Bill Hesse, 2014/09/01 14:54:42]

I really think it would be better to make a section with mode == debug and
system == macos, and put the exceptions in there as Pass or Pass,Fail.  This
file is hard to maintain, so keeping the big sections as general and uniform as
possible is necessary.

I see that about 2/3 of these tests pass, and 1/3 fail, on macos debug.  I'm not
sure we will make much progress by trying to pin down what fails on what
platform, and whether it is flaky.  I think that only by fixing most of these
failures will we turn all the bots green.  The failures are either a bug in
dartium or a problem in the test, so they need to be fixed either way.  I can't
imagine trying to maintain 6 or 8 sections, one for each platform, when adding
more co19 tests from an update.

 LayoutTests/fast/canvas/webgl/WebGLContextEvent_t01: RuntimeError # co19-roll r761: Please triage this failure.
 LayoutTests/fast/canvas/webgl/array-bounds-clamping_t01: RuntimeError # co19-roll r761: Please triage this failure.
 LayoutTests/fast/canvas/webgl/attrib-location-length-limits_t01: RuntimeError # co19-roll r761: Please triage this failure.