Looks nice to me. In connection with the PREG_LENGTH_CAPTURE option floated in a previous post, this would easily let the wikimedia/remex-html package parse HTML in a streaming fashion; it would fill up a buffer array and then do an incremental parse, stopping as soon as a (hard) partial match was found, then move the prefix returned to the start of the buffer, wait for the buffer to fill more, then restart. --scott
On Wed, Mar 20, 2019 at 5:32 AM Nikita Popov <nikita....@gmail.com> wrote: > Hi internals, > > PCRE has some very nice partial matching functionality described at > https://www.pcre.org/current/doc/html/pcre2partial.html. This is useful > for > streaming processing, as it allows you to distinguish between "there's > definitely no match here" and "this could match starting from position N, > but we need more data to find out". > > Here is a PR to expose this functionality from PHP: > https://github.com/php/php-src/pull/3969 The PR has a basic description of > the API. > > What do you think? > > Nikita > -- (http://cscott.net)