* 1) named definitions, e.g. digit = [0-9];
* 2) counted repetition, e.g. "a"{3}, "a"{3,}, "a"{3,5}
* 3) multiple DFA's sharing the same regexp , e.g. <c1, c2> "abc" { }
- * 3) common suffixes, e.g. suffix [\x80-\xBF] in UTF-8 ranges
+ * 4) common suffixes, e.g. suffix [\x80-\xBF] in UTF-8 ranges
* In cases 1-3, regexp must be recompiled each time it's reused.
* In case 4, regexp should be compiled only once, and instructions
* should be shared in order to reduce space.