From: Aristarkh A Zagorodnikov <xm@...3d.ru>
Date: 2001-08-31T16:23:36+09:00
Subject: [ruby-talk:20606] Re: String.subst[!] RCR


> -----Original Message-----
> From: Yukihiro Matsumoto [mailto:matz@ruby-lang.org]
> Sent: Friday, August 31, 2001 11:21 AM
> To: ruby-talk ML
> Subject: [ruby-talk:20605] Re: String.subst[!] RCR
>
>
> Hi,
>
> In message "[ruby-talk:20604] String.subst[!] RCR"
>     on 01/08/31, "Aristarkh A Zagorodnikov" <xm@w3d.ru> writes:
>
> |	I have a proposal ;) Not sure eveyone will need it - maybe
> everyone is
> |satisfied with regexp matching performance, but sometimes I need a lot of
> |string replacements which work fast, and do NOT need complex
> regexp matching
> |rules. I'm pretty sure that regexp matching isn't the fastest
> operation out
> |there (at least I believe it involves some kind of regexp
> compilation). So,
> |I'm proposing the following two methods (to be built in C).
>
> You mean you want "str.sub(pat, new)" without using regular
> expression?
>
> Regular expression search is pretty fast for most of the cases
> (especially for fixed patterns).  It uses BM search inside.
> As a result, str.sub(/foo/, 'bar') is often faster than your
> str.subst('foo','bar').
I'm sure it _is_ faster ... but I believe that's it because of subst is
written in ruby, while sub is in C.

Aristarkh A Zagorodnikov, Lead Programmer, W3D Group
http://www.w3d.ru /// xm@w3d.ru /// ICQ UIN 36987938