numberOfErrors are not defined in should().beEqualToArray()

third_party/WebKit/LayoutTests/webaudio/unit-tests/audit-failures.html

Index: third_party/WebKit/LayoutTests/webaudio/unit-tests/audit-failures.html
diff --git a/third_party/WebKit/LayoutTests/webaudio/unit-tests/audit-failures.html b/third_party/WebKit/LayoutTests/webaudio/unit-tests/audit-failures.html
index 7e6da6317188de7a5994f1ff88f5ddcb850d9308..2ce7c1ec40d6662cf47fbce83fd8ab0611bf25c3 100644
--- a/third_party/WebKit/LayoutTests/webaudio/unit-tests/audit-failures.html
+++ b/third_party/WebKit/LayoutTests/webaudio/unit-tests/audit-failures.html
@@ -54,8 +54,8 @@
       should(59, 'The measured decibel').beCloseTo(62, { threshold: 0.01 });
       should([1, 8, 11, 18, 6, 5, 5, 5, 123, 49]).beConstantValueOf(5);
       should([0, 0, 0]).notBeConstantValueOf(0);
-      should([1, 8, 11, 18, 6, 5, 5, 5, 123, 49])
-        .beEqualToArray([1, 8, 10, 18, 6, 5, 6, 5, 123, 48]);
+      should([1, 8, 11, 17, 6, 6, 5, 6, 123, 49])
+        .beEqualToArray([1, 8, 9, 18, 6, 5, 6, 5, 123, 48]);

[Raymond Toy, 2017/01/25 19:42:07]

Although there's nothing wrong with this change, but why not just keep the
change smaller and only change one of the arrays?

And actually since the contents don't really matter, why not make the reference
a simple linear ramp?  That makes it obvious what's wrong with the actual data
compared to the reference.

[hongchan, 2017/01/25 22:20:33]

Done.

       should([1, 1, 1, 1, 2, 2, 3, 3, 3]).containValues([1, 3, 2]);
       should([0.5, 0.5, 0.55, 0.5, 0.45, 0.5]).notGlitch(0.04);
       should([0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.9])