Dave,


>>What would be the simplest way to use it from a program?
>>one path to one file for example.
>>(... to generate documentation for generated Ruby files.)
>>
> 
> Well the simplest would be to use #system, I guess.


Unfortunately, AFAICS, that could be platform specific.


> To do it totally programmatically is slightly more complex, as you
> have to set up output formatters and the options object. If this is
> something people want to do, I could look at making it easier.


That would be awesome.

I didn't check the license, but I'd like to redestribute it with a 
current project (if and as soon as it's ready to be released) in a 
simple way. My program is generating libs; after one is finished, it 
could get documented right away using RDoc.

So I'd like to have
1. One directory (-tree) where all files find each other, no install
    required.
2. One object, to which I can pass the same args as to the commandline.
    Something like

      $:.unshift'rdoc/alpha-6'
      require 'rdoc/rdoc'
      r = RDoc.new
      r.document <options> [name...]
      # for example: r.document 'thisfile.rb'

TIA,

Tobi

P.S. An idea: Generate XHTML with REXML/NQXML; then there's less to be 
written to generate (X)HTML, and the generated docs are XML, thus can be 
processed/manipulated/queried with any XML toolkit.

-- 
http://www.pinkjuice.com/