Collecting LatencyInfo in telemetry

tools/telemetry/telemetry/core/timeline/model.py

Index: tools/telemetry/telemetry/core/timeline/model.py
diff --git a/tools/telemetry/telemetry/core/timeline/model.py b/tools/telemetry/telemetry/core/timeline/model.py
index ec8dccb4b786b19c605dbbac5dad2898bb4e9fb1..521c1f6057ff100a6823aa9d1ecabf51657fe920 100644
--- a/tools/telemetry/telemetry/core/timeline/model.py
+++ b/tools/telemetry/telemetry/core/timeline/model.py
@@ -150,6 +150,12 @@ class TimelineModel(object):
       threads.extend(process.threads.values())
     return threads
 
+  def GetBrowserMainThread(self):

[nduca, 2014/01/23 18:34:45]

I'd like to remove this from this CL and have a CL that introduces just this. I
think you're going to need to do something slightly different but we can do the
review separate from this patch --- epenner and i need to review it.


namely, the model class is browser agnostic. It doesn't know if its a model that
came from chrome, or firefox. it might even be a model that came from a devtools
timeline recording, which doesn't have a browser main.

i think what you want to do is telemetry.core.backends.chrome_trace_result to
find the CrRendererMain, and then add AddCoreObjectToContainerMapping for the
browser object to the renderer main thread.

Then, on this object, have a GetBrowserProcesssFromTab(tab) could look into the
mapping just as GetRendererProcessFromTab does.

[Yufeng Shen (Slow to review), 2014/01/30 01:16:05]

created a separate CL https://codereview.chromium.org/144423003/  for this.

+    for thread in self.GetAllThreads():
+      if thread.name == 'CrBrowserMain':
+        return thread
+    return None
+
   def GetAllEvents(self):
     return list(self.IterAllEvents())