Pass through in dual_browser_story for disabled benchmarks.

tools/perf/page_sets/dual_browser_story.py

Index: tools/perf/page_sets/dual_browser_story.py
diff --git a/tools/perf/page_sets/dual_browser_story.py b/tools/perf/page_sets/dual_browser_story.py
index d21e0e666782eed1804a9ac989b92c4e5ca62e24..18ae6ad03a61c0a5ecdb84ad04e8c2915dc687e4 100644
--- a/tools/perf/page_sets/dual_browser_story.py
+++ b/tools/perf/page_sets/dual_browser_story.py
@@ -117,9 +117,7 @@ class MultiBrowserSharedState(story_module.SharedState):
     if should_skip:
       logging.warning(msg)
       logging.warning('You are trying to run a disabled test.')
-      logging.warning(
-          'Pass --also-run-disabled-tests to squelch this message.')
-      sys.exit(0)
+      logging.warning('Valid chartjson ouput will not be generated.')

[perezju, 2016/09/29 13:01:45]

This will effectively cause the benchmark to run even if it's disabled, no?
Causing it to fail on devices where it cannot run.

Is there an alternative to make this produce a valid disabled chartjson?

[eyaich1, 2016/09/29 13:05:30]

Well it won't fail, it will still technically succeed it just wont upload
results and the json will return as results['enabled'] = False.  

As I mentioned in the team meeting yesterday, this is the only difference people
should see with this new logic and we may want to do something like print
DISABLED in the step or something since it will appear green even though it
didn't actually run.

I don't see an alternative to this as we made a conscious decision not to
dynamically generate the set of perf benchmarks to run but instead run them all
on each bot and just ignore the disabled ones.

[nednguyen, 2016/09/29 13:08:40]

Hmhh, why not define ShouldDisable(..) for the benchmarks that are using this
story?

 
   def _PrepareBrowser(self, browser_type, options):
     """Add a browser to the dict of possible browsers.