From: Martin Hansen <mail@...>
Date: 2010-04-17T18:27:40+09:00
Subject: Re: Writing a parser

> And you chose ruby? Ruby isn't exactly known for speed. And I can 
> practically guarantee you that the syck parser will be faster for hashes 
> of strings than your parser (as it is currently written and several 
> minor iterations out).
>
> But no, what you described IS yaml, it is just a subset of yaml. You can 
> enforce the subset-ness if you feel like it, but I'm not a big fan of 
> static typing, so I don't recommend it.

Yes, for I want to use Ruby to juggle the records - and then let's see 
about the speed :o). I know from Perl that there is a quite a difference 
between a home rolled parser and YAML-syck (I never saw any subset-ness 
features).

Also, I am learning here. This OO thinking is very exotic ...

> At the very least, you should look at StringScanner and a full rewrite, 
> drop the silly Record class entirely and build up proper hashes of 
> strings. And don't leave those dangling File objects everywhere. Read 
> the whole file or use the block form of file.open.

I will gladly consider a rewrite - I am afraid that I find Robert's 
contribution confusing (am I unfair?). I read about StringScanner in the 
Cookbook IIRC, but I better explore further. Why is the Record class 
silly?


M
-- 
Posted via http://www.ruby-forum.com/.