From: Alan Chen <alan@...>
Date: 2002-10-22T01:24:32+09:00
Subject: Re: URI bug or bad URI?

On Mon, Oct 21, 2002 at 08:05:46PM +0900, Brian Candler wrote:
> On Sun, Oct 20, 2002 at 08:24:07PM +0900, akira yamada wrote:
> > 
> > >>>>> In [ruby-talk : No.53493] 
> > >>>>>	Alan Chen <alan@digikata.com> wrote:
> > > mailto = URI.new( 'mailto:user@domain.com?subject=test%20(@test=10)
> > >
> > > Are '@' symbols allowed in uri's in this context or do they need to be
> > > escaped?
> > 
> > '@' is OK.  But '=' cannot be put in hvalue.  I think that 'subject:
> > test (@test=10)' should be written as 'subject=test%20(@test%3D10)' in
> > URI.
> 
> RFC2368: ``Note that all URL reserved characters in "to" must be encoded:
>   in particular, parentheses, commas, and the percent sign ("%"), which
>   commonly occur in the "mailbox" syntax.
> 
>   "hname" and "hvalue" are encodings of an RFC 822 header name and
>   value, respectively. As with "to", all URL reserved characters must
>   be encoded.''
> 
> But that conflicts with RFC1738 which explicitly allows parentheses and
> commas as unreserved characters which can always be used unencoded in a URL.
> Grr.
> 
> Anyway, if you want to err on the safe side, you might choose
> 
> subject=test%20%28@test%3D10%29

I really wan't choosing this uri encoding, it was something that my
application encountered "in the wild" of the internet. I guess I can
write some sort of pre-filter for the URI or add some code to extend
URI.parse.  Maybe I'll modify it for URI.parse(uristring, parsetype)
where parsetype could be :rfc2368, :rfc1738, :rfc2368_lax or similiar.


-- 
Alan Chen
Digikata LLC
http://digikata.com