[Psi-Devel] Feature request: Handling of im: URIs

textshell-I1QKlO at neutronstar.dyndns.org textshell-I1QKlO at neutronstar.dyndns.org
Fri Nov 9 14:58:28 PST 2007


On Fri, Nov 09, 2007 at 11:04:29AM -0700, Peter Saint-Andre wrote:
> textshell-I1QKlO at neutronstar.dyndns.org wrote:
> > On Fri, Nov 09, 2007 at 04:17:05PM +0200, Norman Rasmussen wrote:
> >> On Nov 9, 2007 3:59 PM, Kevin Smith <kevin at kismith.co.uk> wrote:
> >>> I may have misunderstood what you wanted. I thought you meant handling
> >>> of xmpp: links within Psi, whereas now I guess you meant using Psi as
> >>> the default handler for system xmpp: links. The latter is Textshell's
> >>> baby in the IPC stuff. If you're interested in the former, it
> >>> shouldn't be too hard.
> >> Mmm, personally if getting xmpp links to work correctly with-in Psi is
> >> easy, then I think we should go for that.
> > 
> > That's mainly the missing part of the current code. And it's not quite that
> > easy because we need to do it in an secure and unobstrusiv way.
> > And i didn't see any xmpp / im uris send via xhtml-im or normal chat
> > messages yet, so i think it's not really higher priority than gettig the
> > ipc and this work.
> 
> In general I think that xmpp: URIs will be found on web pages. So for
> instance you might find links like this:
> 
> <a href='xmpp:jdev at conference.jabber.org?join'>jdev chatroom</a>
> 
> It would be good for the browser to pass that URI along to Psi and for
> Psi to do the right thing (in this case, prompt the user to join the
> chatroom).
> 
> I don't think that IM users will send these URIs very often. We have
> better ways to handle such use cases natively over Jabber (chatroom
> invitations, roster item exchanges, and the like).
> 

right, that's what i meant. for me, xmpp uri handling is interesting mostly
with ipc and hooked up in the OS/browser uri handling settings.

btw, is there any recommendations for handling xmpp uri:s in a secure way?
there's a lot potential for information disclosure and privacy violations
if implemented naively... 

 - Martin


More information about the Psi-Devel mailing list