} If it does a single replace, there is no issue. The problem is with
} finding *all* matches in the same string. As the current code does
} not mention PCRE_NOTEMPTY, I assume that it does not do the
} Perl-type trickery that is needed to emulate Perl's /g behaviour.
Yes, that's probably true. Though who knows, maybe we do something
like it -- the trickery is to deal with matches of the empty string?
We have this behavior:
// advance one character if we matched an empty string at the same
// place as the last match occurred
} But maybe it should be doing it?
Likely so. It sounds complicated, though. :-) It would be best if
someone who understood both what perl was doing, and C++, could make a
go at any change, assuming one is needed.
craig
