Added a test to demonstrate problem tokenizing the equals comparison.

dokai · dokai · commit d7cdae93de92 · 2016-05-17T14:09:34.000-07:00
The equals comparison, `==`, is tokenized into two consecutive
assignments instead of the comparison.

One effect of this is that if you're using a font with programming
ligatures, e.g. FireCode, then the correct ligature will not be used for
the equals operator.
diff --git a/test/expressions/expr20.py b/test/expressions/expr20.py
@@ -0,0 +1,39 @@
+a == b
+a != b
+a < b
+a <= b
+a > b
+a >= b
+
+
+
+a             : source.python
+              : source.python
+==            : keyword.operator.comparison.python, source.python
+              : source.python
+b             : source.python
+a             : source.python
+              : source.python
+!=            : keyword.operator.comparison.python, source.python
+              : source.python
+b             : source.python
+a             : source.python
+              : source.python
+<             : keyword.operator.comparison.python, source.python
+              : source.python
+b             : source.python
+a             : source.python
+              : source.python
+<=            : keyword.operator.comparison.python, source.python
+              : source.python
+b             : source.python
+a             : source.python
+              : source.python
+>             : keyword.operator.comparison.python, source.python
+              : source.python
+b             : source.python
+a             : source.python
+              : source.python
+>=            : keyword.operator.comparison.python, source.python
+              : source.python
+b             : source.python
