7 Apr
2006
7 Apr
'06
9:48 a.m.
Obviously easier said than done, but if the library can turn the appropriate m[index].matched to "true", then perhaps there could be another variable in the class to reflect the lowest index of the first group that is true. I realize this gets more complicated because there can be multiple groups.
It's backtracking that gets you: each time the "lowest sub-expression set" gets updated you'd need to store a complete history of previous settings so you can backtrack out again. However, now I think about that some more, I actually think it might be possible after all :-) John.