Using the dense dfa.

Patterns:

• .+
• a+

Input:

• "aa"

Expected output:

• PatternID(0) range 0-2
• PatternID(1) range 0-2

My goal is to get all overlapping matches between patterns, but if from the same pattern then matches should report the longest match.

I'm currently using the method described here but this gets all overlapping matches including the shorter matches. Here's the fwd results:

HalfMatch { pattern: PatternID(0), offset: 1 }
HalfMatch { pattern: PatternID(1), offset: 1 }
HalfMatch { pattern: PatternID(0), offset: 2 }
HalfMatch { pattern: PatternID(1), offset: 2 }

Is there a better way of going about this? Other than:

• Separate DFA per pattern, would be slower (there will be a larger number of patterns
• Resorting to post-filtering to discard the smaller matches

@BurntSushi