Hi, >thanks for sharing these improvements!
you are welcome. I think it is important to share any major progress with the community. >That is, it can detect that anything coming after the \s* won't >include a \s, so it can possessify the quantifier? Exactly. >Are we going to see these optimizations as part of PCRE 9 (together >with the new API) or do you plan to roll them in the 8.xx series too? Currently working on the new API is a bit postponed, so this will be part of 8.x I think the best way to improve an engine is finding performance bottlenecks. Especially when a large number of patterns from different sources has the same bottleneck. If anybody has pattern sets he/she can share with us, we can use them to improve the engine on the long run. Regards, Zoltan -- ## List details at https://lists.exim.org/mailman/listinfo/pcre-dev
