> > But the caller can do the same, right now. When the caller gets the > PCRE_PARTIAL return, a temporary buffer starting "123" can be > constructed, joined onto more data from the second buffer, and the > entire match can be retried, *not* using PCRE_DFA_RESTART. >
True, but the caller will have to do that each time PCRE_PARTIAL is returned rather then only when there is a chance of losing matches. That was exactly the point I was trying to make in my previous post. In the first scenario PCRE_PARTIAL_ALTERNATIVE could be returned (instead of PCRE_PARTIAL) to indicate that if the caller wants to make sure that no matches are lost, the buffers must be linearized. In the second scenario, there is no such issue and thus PCRE_PARTIAL can be returned and the second buffer can be scanned with PCRE_DFA_RESTART with no worry of losing matches. Thanks, E. -- ## List details at http://lists.exim.org/mailman/listinfo/pcre-dev
