Update record_controller to pass memoryDumpConfig to all controller clients.

tracing/tracing/ui/extras/about_tracing/record_controller.html

Index: tracing/tracing/ui/extras/about_tracing/record_controller.html
diff --git a/tracing/tracing/ui/extras/about_tracing/record_controller.html b/tracing/tracing/ui/extras/about_tracing/record_controller.html
index 8883da7a89b2b63fe4aa0f326ac9bce5cfcb2845..a9b42b589d884ba63dc9f05ca63b53c4a9f553f7 100644
--- a/tracing/tracing/ui/extras/about_tracing/record_controller.html
+++ b/tracing/tracing/ui/extras/about_tracing/record_controller.html
@@ -92,6 +92,13 @@ tr.exportTo('tr.ui.e.about_tracing', function() {
         enable_systrace: selectionDlg.useSystemTracing,
         record_mode: selectionDlg.tracingRecordMode,
       };
+      if (categories.included.indexOf(
+          'disabled-by-default-memory-infra') !== -1) {
+        const memoryConfig = { triggers: [] };
+        memoryConfig.triggers.push(
+            {'mode': 'detailed', 'periodic_interval_ms': 10000});

[fmeawad, 2017/06/23 17:38:09]

I lost track of the original interval, but as I was testing this CL I noticed
that 10s in practice might be too long (specially for the about:tracing use
case). We should plumb it from the UI with a default value. (I will try to do it
in a follow up CL)

+        recordingOptions.memory_dump_config = memoryConfig;
+      }
 
       const requestPromise = tracingControllerClient.beginRecording(
           recordingOptions);