Collecting LatencyInfo in telemetry

tools/telemetry/telemetry/core/browser.py

Index: tools/telemetry/telemetry/core/browser.py
diff --git a/tools/telemetry/telemetry/core/browser.py b/tools/telemetry/telemetry/core/browser.py
index 3900c45f09f1cd2360e75b16b2bb99f45f5ece03..f3c6c7f1d4a2077af628c7a7b6e605e14a25660d 100644
--- a/tools/telemetry/telemetry/core/browser.py
+++ b/tools/telemetry/telemetry/core/browser.py
@@ -85,6 +85,10 @@ class Browser(object):
   def supports_tracing(self):
     return self._browser_backend.supports_tracing
 
+  @property

[nduca, 2014/01/13 18:49:32]

we never expose the browser pid. there are good reasons for this, even though it
seems innocent.

talk to @epenenr about how to get the browser process in telemetry. he went
through this exercise recently and probalby can advise you on how to do it.

[Yufeng Shen (Slow to review), 2014/01/14 00:05:47]

Talked to Eric and he suggests to search the desired thread by name instead of
acting on process. Since all the input event LatencyInfo starts on browser main
thread, so I change to add
TimelineModel.GetBrowserMainThread() to return the thread with name
'CrBrowserMain' and the async slices of that thread has all the dumped
LatencyInfo.

+  def browser_pid(self):
+    return self._browser_backend.pid
+
   def is_profiler_active(self, profiler_name):
     return profiler_name in [profiler.name() for
                              profiler in self._active_profilers]