A couple of weeks ago, while playing around with some of the shootout benchmarks, I came up with a version of the regex-dna benchmark:

http://shootout.alioth.debian.org/gp4/benchmark.php?test=regexdna&lang=dlang&id=0

In order to get better results, I had to improvise an awful workaround which can
be seen in the first foreach.
Unfortunately, RegExp.replace shows incredibly low performance when paired with
long streams of data (5MB). Another workaround, in the fashion of the one inside
the first foreach loop, would be possible, but not efficient.
I wonder if there's work already in progress, in Phobos or Ares, to solve this.
Otherwise, although I cannot make any promises, I'd try to get to the bottom of
this.


-- Sebastián.

Sebastián E. Peyrott wrote:
>
> I wonder if there's work already in progress, in Phobos or Ares, to solve this.
> Otherwise, although I cannot make any promises, I'd try to get to the bottom of
> this.

At the moment, I'm waiting to see what Don and Eric come up with for regex parsing so far as Ares is concerned.


Sean