CZECH IT OUT!

+      DEBUG ("don't know %u's presence; assuming text chat caps.", handle);

I don't like telling you assume text caps there as it's a decision of im_factory. So if we change there, likely that debug will become wrong.

It is not unlikely that at some point Call cap will be added for offline contacts, for voicemail. AFAIK Skype does it already.

But that's just my 2cents, I don't know much that gabble code.

(In reply to comment #2)
> +      DEBUG ("don't know %u's presence; assuming text chat caps.", handle);
> 
> I don't like telling you assume text caps there as it's a decision of
> im_factory. So if we change there, likely that debug will become wrong.

Well, right, I agree -- if the IM factory behaviour changes, this will break, which is why I added the big NOTE in that function. I also think the test is strict enough to only allow the text caps, isn't it?

> It is not unlikely that at some point Call cap will be added for offline
> contacts, for voicemail. AFAIK Skype does it already.

Sure. At that point of course we'll have to change this behaviour, but it's not so bad now, is it? Perhaps we could add a GabbleCapsChannelManagerClass->get_assumed_caps vfunc so we can iterate through the channel managers and get text caps from im-factory and Call caps from call-factory, but that can be done later, no?

I think the implementation is functionally fine. But here's a bunch of trivia.

+  /* IM factory */
+  GabbleImFactory *im_factory;

This comment does not add anything except unnecessary scrolling.

+"""
+Test for fd.o#32874
+"""

I'm offline on a train. What was that bug?

+    # new ContactCapabilities
+    ccaps_map = conn.ContactCapabilities.GetContactCapabilities([bob_handle])
+    assertEquals(1, len(ccaps_map))
+
+    assertEquals(1, len(ccaps_map[bob_handle]))

assertLength

+    assertSameSets({cs.CHANNEL_TYPE: cs.CHANNEL_TYPE_TEXT,
+                    cs.TARGET_HANDLE_TYPE: cs.HT_CONTACT}, fixed)

assertEquals

+    print all_caps

No.

I'm upset that there's no convenience function to expect a roster query and send a reply, but I won't hold it against you.

(In reply to comment #4)
> +  /* IM factory */
> +  GabbleImFactory *im_factory;
> 
> This comment does not add anything except unnecessary scrolling.

I've added a comment.

> +"""
> +Test for fd.o#32874
> +"""
> 
> I'm offline on a train. What was that bug?

I've added yet another comment.

> assertLength
...
> assertEquals
...
> +    print all_caps
> 
> No.

All three fixed.

hooray!

Awesome.