[Bug 42288] Review Chan.I.FileTransfer.Metadata

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=42288

--- Comment #3 from Will Thompson <will.thompson at collabora.co.uk> 2011-11-09 08:48:40 PST ---
> ft-client-caps test: change client nameft-metadata
> 
> I've no idea why but if caps/tube-caps.py is run before this test (it
> has the same idea and design as this test was copied from it) then
> this one fails due to the same client name?! Another day...

I think this is probably because the caps end up in the cache so Gabble doesn't
make the disco query that the test expects.

(In reply to comment #2)
> (In reply to comment #1)
> > +  if (field == NULL)
> > +    {
> > +      DEBUG ("ServiceName propery not present in data form; odd...");
> > +      goto out;
> > +    }
> > +
> > +  service_name = g_strdup (field->raw_value_contents[0]);
> > +
> > +out:
> > 
> > Eh, why not
> > 
> >    if (field != NULL)
> >      service_name = ...
> >    else
> >      DEBUG ("ServiceName...");
> > 
> > rather than goto? I don't think the goto makes it clearer.
> 
> OK, done.

No it's not, but in the course of fixing the above complaint you made the goto
justify its own existence. :)

> > Blah, if ServiceName is not going to be stashed in the same data form as the
> > metadata, then it shouldn't be a data form. it should just be
> > 
> >   <service-name xmlns="im:telepathy:file-transfer:service"
> > name='com.example.HiMum'/>
> 
> We discussed this in real life and I thought you conceded that using another
> data form just made things easier?

It means one less XEP to write I guess, but it offends me at some level. :)

> > and… im: is not a URI scheme. http://telepathy.im/...
> 
> Okay. :-)

I hate to be that guy, but I've just noticed that elsewhere we use URLs under
http://telepathy.freedesktop.org/xmpp/ which actually points to documents. So
you could put the proto-XEP there and use it as the capability URI maybe? Sorry
:/

> > wocky_node_add_build (x,
> >   '(', "field",
> >     '@', "var", key,
> >     '(', "value",
> >       '$', val,
> >     ')',
> >   ')', NULL);
> > 
> > or is that less clear?
> 
> Well let's do it and see.

I like it. Particularly getting rid of the casts.

> > I don't really see why the metadata-specific test is the one which explicitly
> > sends no metadata. I guess it's testing edge cases…
> 
> Yes. It tests that no dataform is sent. I didn't want it to send an empty
> dataform in that case.

Ah, okay!

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA Contact for the bug.
You are the assignee for the bug.