From: Charles O Nutter <headius@...>
Date: 2006-07-25T18:11:29+09:00
Subject: Re: rdoc grows to large size.

------=_Part_167653_22682038.1153818685561
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

I've studied the rdoc code a bit (only so much as to maintain my sanity) and
wondered aloud: is this all hand-written, hand-optimized code? It did not
appear to be generated, and I have seen hand-rolled patches trickle in over
time. Is there perhaps a better way to parse all that content and generate
those docs that would not require a hand-made implementation?

Of course there's a bit of an alterior motive for me: JRuby's interpretation
is still several times slower than C Ruby's, so having rdoc be inefficient
in C Ruby is magnified under JRuby. If there's any way to improve this
implementation across the board, I think it would help everyone.

On 7/25/06, Eric Hodel <drbrain@segment7.net> wrote:
>
> On Jul 24, 2006, at 6:09 AM, Hugh Sasse wrote:
>
> > While working on that .document patch I noticed that rdoc grew to
> > about 124MB.
>
> $ rdoc --help
>
> RDoc V1.0.1 - 20041108
>
> Usage:
>
>    rdoc [options]  [names...]
>
> Files are parsed, and the information they contain
> collected, before any output is produced. This allows cross
> references between all files to be resolved. [...]
>
> It is possible that the structure created is inefficient.  The code
> that builds the structure is a forbidding forest of mystery, I rarely
> dare venture there.
>
> --
> Eric Hodel - drbrain@segment7.net - http://blog.segment7.net
> This implementation is HODEL-HASH-9600 compliant
>
> http://trackmap.robotcoop.com
>
>
>
>


-- 
Charles Oliver Nutter @ headius.blogspot.com
JRuby Developer @ www.jruby.org
Application Architect @ www.ventera.com

------=_Part_167653_22682038.1153818685561--