WIP: Add support for additional flag-specified tracing categories.

telemetry/examples/benchmarks/tbm_benchmark.py

Index: telemetry/examples/benchmarks/tbm_benchmark.py
diff --git a/telemetry/examples/benchmarks/tbm_benchmark.py b/telemetry/examples/benchmarks/tbm_benchmark.py
index a335805b558a70168517cd8aaf78649c46e4c42a..ea8f1cd0a82a2628f5e462456d62ac09af79915f 100644
--- a/telemetry/examples/benchmarks/tbm_benchmark.py
+++ b/telemetry/examples/benchmarks/tbm_benchmark.py
@@ -11,7 +11,7 @@ class TBMSample(benchmark.Benchmark):
   def CreateStorySet(self, options):
     return simple_story_set.SimpleStorySet()
 
-  def CreateTimelineBasedMeasurementOptions(self):
+  def CreateBaseTimelineBasedMeasurementOptions(self):

[charliea (OOO until 10-5), 2017/07/07 20:06:59]

Maybe it's just me, but it feels like this name is getting long and hard to
understand for how easy a concept it is. I especially don't like the fact that
we have both "Base" and "Based" in the name and that they carry two different
meanings in it.

What about something like "CreateBaseTBMOptions"? Is this too acronym-y?

nednguyen@google.com, WDYT?

     options = timeline_based_measurement.Options()
     options.SetTimelineBasedMetrics(['sampleMetric'])
     return options