Regex DNA benchmark

[Jascha Wetzel]

bearophile wrote:
> On the Shootout site there are two D versions, one of them uses std.regexp, the other PCRE. Even if the PCRE-based test is slower than the Python/Perl versions (but they use the trick of splitting the RE, so they can't be fully compared), it shows D will need a much much faster std.regexp (10 seconds with PCRE, versus 117 seconds with std.regexp).

i have implemented a regex engine that is supposed to be categorically 
faster than the one in Phobos (it's part of the apaged package). i will 
do some polishing soon, in order to make it useful as a library and do 
some benchmarking. i'm pretty confident that it can improve on that 
shootout-benchmark, not only because it can compile regexps to D code.