I've done a bit of testing myself. The figures one gets depend a lot on the kind of testing, but this is a fairly straightforward case when searching for a submatch that doesn't exist in a large string:
Pike: 0.00017900 s Perl: 0.00074000 s Emacs 20: 0.04566600 s Emacs 21: 0.05563044 s TRE: 0.07000000 s GNU Regex: 0.16000000 s
The shortcoming in Pikes Regexp is the lack of features, not the speed.
/ Martin Stjernholm, Roxen IS
Previous text:
2003-09-19 15:30: Subject: Re: Adding PCRE to Pike
where did you see that?
greetings, martin.
/ Brevbäraren
What did you do to the precision in the TRE and GNU tests? Scheduler artefacts?
/ Peter Bortas
Previous text:
2003-09-19 15:48: Subject: Re: Adding PCRE to Pike
I've done a bit of testing myself. The figures one gets depend a lot on the kind of testing, but this is a fairly straightforward case when searching for a submatch that doesn't exist in a large string:
Pike: 0.00017900 s Perl: 0.00074000 s Emacs 20: 0.04566600 s Emacs 21: 0.05563044 s TRE: 0.07000000 s GNU Regex: 0.16000000 s
The shortcoming in Pikes Regexp is the lack of features, not the speed.
/ Martin Stjernholm, Roxen IS
No, it was a result taken from a larger test where times ran away pretty quickly, so there's a bit too low repeat count on those.
/ Martin Stjernholm, Roxen IS
Previous text:
2003-09-19 16:05: Subject: Re: Adding PCRE to Pike
What did you do to the precision in the TRE and GNU tests? Scheduler artefacts?
/ Peter Bortas
pike-devel@lists.lysator.liu.se
-
Martin Stjernholm, Roxen IS ＠ Pike developers forum -
Peter Bortas ＠ Pike developers forum