Future of scScript Patterns

The current Match algorithm has a very hard time going backwards in the text buffer. If a particular Pat starts at the end of the buffer and jumps (perhaps N chars/lines) backwards, it will eventually go past the beginning of the buffer (or permissible area) and fail. But Match is not smart enough to just stop, instead it advances +1 position from the previous start and tries again, and again.... just as it would for a forward match going through a gap area.

While not quite as dramatic, mindlessly advancing +1 position can also be inefficient for forward matches. If the particular word does not match, it makes no sense to advance +1 char into the word and try again. It would be much better to jump directly to the next word instead. The same logic may apply to lines or paragraphs, depending on the particular pattern definitions.

There are a few possible solutions, but the most appealing general one is to also provide a FailMFunc. The regular MFunc (if provided) is normally called only when the Pat matches. It tells the Match apparatus whether to continue and where to go next. A similar FailMFunc could perform that function when the Pat does not match. So instead of mindlessly advancing +1 in search of the next match, it could advance backwards or even jump to the next word or line. This mechanism can enable proper backwards traversals, but also make forward matching more efficient. Given the more advanced functionality, it might be best to limit its use to a new RMatch function, one that takes an extra FailMFunc arg right after MFunc.

(An expedient step would be to supply a #SkipWord option for the current Match!)