[tracker/class-signal] Added a comment, squash

[Philip Van Hoof <pvanhoof src gnome org>]

commit 549d00372ac199700b533c5b2aff3c621eade65e
Author: Philip Van Hoof <philip codeminded be>
Date:   Wed Aug 25 15:22:43 2010 +0200

    Added a comment, squash

 .../class-signal-performance-test.vala             |    7 +++++++
 1 files changed, 7 insertions(+), 0 deletions(-)
---
diff --git a/tests/functional-tests/class-signal-performance-test.vala b/tests/functional-tests/class-signal-performance-test.vala
index b520271..b323c57 100644
--- a/tests/functional-tests/class-signal-performance-test.vala
+++ b/tests/functional-tests/class-signal-performance-test.vala
@@ -28,6 +28,13 @@ const string title_data = "title";
 // The test IS NOT deleting existing resources, so you CAN'T RUN IT TWICE unless
 // you clear the database before starting it the second time.
 
+// On Aug 25 the difference between using tracker-store on master and the
+// tracker-store of class-signal, and then letting this wait until all 10000
+// (in case of max_signals = 10000) insert queries' signals arrived (you'll
+// have in total 20002 events in the signals in both tracker-store versions)
+// was: 20s for class-signals (new class signal) and 23s for master (old class
+// signals). Measured using this performance test.
+
 struct Event {
 	int subject_id;
 	int pred_id;