[Not for landing - CL being split] Add Estimated Input Latency - EQT 90th Percentile definition

tracing/tracing/model/helpers/chrome_model_helper.html

Index: tracing/tracing/model/helpers/chrome_model_helper.html
diff --git a/tracing/tracing/model/helpers/chrome_model_helper.html b/tracing/tracing/model/helpers/chrome_model_helper.html
index 51b762abb3ca39b6423e2dfe3c3da2be5cb35a23..d974da7ad1ca7f266de4c1dde42aa586e9d1097d 100644
--- a/tracing/tracing/model/helpers/chrome_model_helper.html
+++ b/tracing/tracing/model/helpers/chrome_model_helper.html
@@ -114,6 +114,23 @@ tr.exportTo('tr.model.helpers', function() {
 
     get rendererHelpers() {
       return this.rendererHelpers_;
+    },
+
+    // This is often used to find the target renderer for metrics calculation

[charliea (OOO until 10-5), 2016/09/19 14:35:34]

Note: write jsdocs for function headers. Also, is there some guarantee that the
largets pid corresponds to the target renderer? If so, that's definitely worth
noting. In other words, this should probably look something like:

  /**
   * Returns the renderer with the largest PID.
   *
   * This is useful in metrics calculation because <reason>.
   */

[dproy, 2016/09/20 01:11:44]

I added more detailed comment in the latest patch. I really wish I didn't have
to use this hack, but I followed
https://github.com/catapult-project/catapult/blob/master/tracing/tracing/metr....


If you know of any way to avoid this approach all together I would be very very
happy to know.

+    get rendererWithLargestPid() {
+      var largestPid = -1;
+      for (var pid in this.rendererHelpers) {
+        var rendererHelper = this.rendererHelpers[pid];
+        if (rendererHelper.isChromeTracingUI)
+          continue;
+        if (pid > largestPid)
+          largestPid = pid;
+      }
+
+      if (largestPid === -1)
+        return undefined;
+
+      return this.rendererHelpers[largestPid];
     }
   };