In the past I had a task, run hundreds thousands of regular experssions
on millions text documents, and it should run fast (run time should
be measured by seconds).
Perl regexp alternation may be good option for that, because it put all
regexps on one trie structure and passes them at once, but I needed more
than simple match. I had to know what regexp was matched and exact position
of match, and much more. Trying to optimize matching process, I found that
my options to interact with regexp engine during matching process are very limited.
I decide to…