Author: Philip Hazel Date: To: pcre-dev CC: 徐?z Subject: Re: [pcre-dev] A performance question about PCRE library
On Mon, 15 Aug 2011, I wrote:
> I tried studying the combined pattern, assuming it would find a starting
> byte set of < o t and u (which is what studying does), but it didn't.
> This I think is a bug in pcre_study(), which I will investigate (just in
> time for the forthcoming 8.13 release). Looks like it is the initial \b
> in three patterns that is foxing it, because when I remove that, it does
> find the starting bytes.
Yes, it was a simple bug, for which I have committed a patch. If you
study the pattern, the 8.13 release should be somewhat faster. With luck
this release might be out tomorrow.
