XMPP: Composing events and <id/> tags

[Mark Doliner]

On Tue, 24 Apr 2007 14:56:18 -0400, Etan Reisner wrote
> On Tue, Apr 24, 2007 at 01:56:04AM -0500, Mark Doliner wrote:
> > Nope!  jm->id specifies the id attribute on the message element, which we
> > already set.  This spec is talking about the id element that is a child of the
> > <x xmlns='jabber:x:event'> element.
> 
> The spec looks to be talking about the id attribute on the message element
> actually:
>     (http://www.xmpp.org/extensions/xep-0022.html#sect-id2251213)
>     "Note that the <message/> element requesting event notification
>     contains an 'id' attribute. While these attributes are optional 
> in the    Jabber protocol, messages that contain event notification 
> requests MUST    contain an 'id' attribute so that raised events may 
> be matched up with    their original requests."

What?  The spec talks about the 'id' in both places.  The except you pasted is
from the section ABOVE the one that linked to by you and Evan.  I don't think
Evan was asking about the 'id' attribte on the message stanza, I think he was
asking about the 'id' tag within the <x xmlns='jabber:x:event'> tag.

> > We're definitely in violation of the spec here.  I don't think adding an empty
> > <id> tag would help much, we'd still be in violation of the spec.  Really I
> > think we should only be sending XEP-0022 <composing> messages if we're
> > replying to a message.  We should not send them on the initial message to
> > someone.  And the id tag should contain the value of the id attribute from the
> > message tag that you received from the other person.
> 
> If we want to be requesting XEP-0022 events then we should be 
> setting ids on our messages in the first place, if we don't want to 
> request that information then we don't need to do that.

As I said, we're already setting ids on our messages.

> The id element under the jabber:x:event element is something else, 
> that we should be sending when sending <composing/> events and if we 
> aren't doing that now that's a problem and should be fixed.

Right.

> > But it's kinda minor.  It would be easier to just remove support for XEP-0022,
> > since it's superseded by XEP-0085 anyway.
> 
> I haven't followed any Standards-JIG type stuff for a while but last 
> I was paying attention I thought this was a hotly debated issue 
> since while they overlap they don't both hit all of the same points. 
> Is that no longer the case?

I wasn't aware of that.  If that's true then fixing it is definitely a better
idea than removing it.

> > -Mark
> > 
> > > [1] http://www.xmpp.org/extensions/xep-0022.html#sect-id2251213
> 
>     -Etan