Re: [PATCH] RFC2449 "CAPA" support

On Friday 26 August 2005 06:47 am, Solar Designer wrote:
> On Thu, Aug 25, 2005 at 06:40:06PM -0700, N.Cat wrote:
> > Since some clients want capabilities to be described (e.g. KMail
> > complains that it can't determine UIDL support),
>
> I was unaware that such clients existed.  Although adding CAPA support
> to popa3d has been on TODO for years, I was not planning on doing that
> until popa3d would have to say anything beyond RFC 1939's definitions. (*)
>

I think it's a good idea to advertise the capabilities for commands defined as 
optional in RFC 1939. This prevents clients from making incorrect assumptions 
if they don't resort to extensive probing with CAPA unavailable.

> > I implemented the "CAPA" command as per RFC 2449.
> > The capabilities set also includes "pass" (it's identical to NGPopper's).
>
> RFC 2449 doesn't define a capability named PASS.  Unless there's a
> client which depends on it, I would rather not declare it.

Right. That should be unnecessary since the USER tag is supposed to indicate 
support for the PASS command.

> (*) Well, frankly, I was about to add CAPA support into popa3d in 1999,
> but then put it on the back burner.  There are also some problems with
> RFC 2449 (especially with PIPELINING) that noone has bothered to comment
> on (let alone fix them in a subsequent revision of the RFC):
>
> 	http://www.imc.org/ietf-pop3ext/mail-archive/msg00083.html