From: Sean Chittenden <sean@...>
Date: 2002-05-10T08:24:31+09:00
Subject: Re: cgi params api

> > cgi.rb has undocumented method "CGI#param".
> > 
> > cgi.param["name"]    # == cgi.params["name"].join("\0")
> > 
> > 
> > 
> > now    : cgi["name"] == cgi.params["name"]
> > future?: cgi["name"] == cgi.param["name"]
> > 
> > if cgi["name"] == cgi.param["name"] then
> > 
> > cgi["name"].to_a    # ["value1", "value2", "value3",...]
> > 
> > how about?
> 
> This looks better than my suggestion.  I like that you make the single
> param with the join.  to_a is also a better name than multiple.

Before these cgi param threads get too far down whatever path they're
going, let me drop in a few words.

*) cgi.rb's behavior of having multiple args per key _is_ correct
*) a common line that I use is:

   username = CGI::escapeHTML(cgi['username'].to_s) if cgi.has_key?('username')

*) I'm working on parsing apache requests into an Apache::Table
 instance for mod_ruby.

   reqs = r.get_params()  # Just GET parameters
   reqs = r.post_params() # Just POST parameters
   reqs = r.params()      # Both GET and POST parameters

  reqs will be an instance of Apache::Table.  Apache::Table allows for
  multiple values per key, which is correct for HTTP.  If you access
  reqs['key'], then you'll always get a non-array object.
  reqs.get_array['key'] will return you an array like cgi.rb does
  currently.

*) As for breaking out some of it's other, more useful features, IMHO
 such as escapeHTML and escapeURI, I'm hoping those get placed into an
 Apache::Util module.

-sc

-- 
Sean Chittenden