From: Kristof Bastiaensen <kristof@...>
Date: 2004-07-31T02:16:46+09:00
Subject: Re: [ANN] ri-emacs

On Fri, 30 Jul 2004 17:49:35 +0100, Alex Fenton wrote:

> Kristof Bastiaensen wrote:
> 
>> It seems that the ruby-script is returning an error.  I have yet to find
>> a way to give a meaningfull message when the script dies.
>  >
>> Could you run it on the command-line, and see if it gives an error?
> 
> Yep, it's failing because it can't load 'rdoc/ri/display'; the class
> RiDisplay seems to be defined in $RUBY/bin/ri in my installed version. My
> RDoc seems to be 1.0pr1 - what version were you working on?
> 

ri --version returns V1.0pr1 (on Debian linux).  That ruby
script is the weak point, since it is closely tied to the 
ri-implementation, but this is the only way I could have
fast completion (or writing my own version of ri).
Anyway it is strange that RiDisplay is 
included in bin/ri.  My version of ri contains only:

ri = RiDriver.new
ri.process_args

The obvious solution would be to include the RiDisplay class
in the package, and load it if necessary.  I still wonder
why it is there, though.  (Perhaps the authors of ri didn't
expect the implementation to be used?)

Regards,
KB