Logs for jdev@conference.jabber.org
[00:04:24] * waqas joined the chat.
[00:08:20] * Tobias left the chat.
[00:08:24] * Tobias joined the chat.
[00:10:30] * Maranda left the chat.
[00:11:01] * Flow left the chat.
[00:12:57] * Maranda joined the chat.
[00:14:58] * Maranda left the chat.
[00:15:36] * KevWalke left the chat.
[00:16:54] * Maranda joined the chat.
[00:18:03] * Maranda left the chat.
[00:25:39] * Maranda joined the chat.
[00:26:14] * tato left the chat.
[00:38:09] * Maranda left the chat.
[00:38:10] * Maranda joined the chat.
[00:51:50] * Maranda left the chat.
[00:52:26] * Maranda joined the chat.
[00:53:21] * westsibe joined the chat.
[01:00:25] * westsibe left the chat.
[01:17:43] * Maranda left the chat.
[01:17:44] * Maranda joined the chat.
[01:37:44] * Maranda left the chat.
[01:41:06] * Irdis left the chat.
[02:31:19] * Neustradamus joined the chat.
[02:46:06] * Neustradamus left the chat.
[03:52:00] * bear joined the chat.
[03:54:36] * jcea left the chat.
[03:57:25] * Tobias joined the chat.
[04:00:25] * Tobias left the chat.
[04:23:36] * naw left the chat.
[04:36:51] * MattJ left the chat.
[04:50:52] * bear left the chat.
[04:56:08] * bear joined the chat.
[05:07:22] * MattJ joined the chat.
[05:27:03] * MattJ left the chat.
[05:49:59] * Kev left the chat.
[06:11:40] * bear left the chat.
[06:22:38] * ermine joined the chat.
[06:26:12] * Lance left the chat.
[06:42:40] * ermine left the chat.
[06:42:42] * ermine joined the chat.
[06:55:21] * waqas left the chat.
[06:57:31] * waqas joined the chat.
[07:02:51] * Neustradamus joined the chat.
[07:05:39] * waqas joined the chat.
[07:14:55] * Neustradamus joined the chat.
[07:15:33] * Tobias joined the chat.
[07:53:14] * edwinm joined the chat.
[07:58:22] * 0xAFFE joined the chat.
[08:01:05] * Kev joined the chat.
[08:14:04] * westsibe joined the chat.
[08:15:05] * westsibe left the chat.
[08:23:18] * kevin joined the chat.
[09:03:20] * edwinm left the chat.
[09:03:20] * edwinm joined the chat.
[09:13:49] * KevWalke joined the chat.
[09:16:55] * Zash joined the chat.
[09:19:07] * Tobias left the chat.
[09:23:25] * edwinm left the chat.
[09:23:28] * edwinm joined the chat.
[09:26:39] * Tobias joined the chat.
[09:57:47] * Link Mauve joined the chat.
[10:04:57] * xnyhps joined the chat.
[10:21:33] * bear joined the chat.
[10:33:16] * kevin left the chat.
[10:36:40] * Tobias left the chat.
[10:36:44] * Tobias joined the chat.
[10:51:43] * Maranda joined the chat.
[10:58:03] <> Will there be anything else done tomorrow beside gathering statistics from the test :)?
[11:09:02] * bear left the chat.
[11:53:16] * Tobias left the chat.
[12:37:59] * Tobias joined the chat.
[12:54:27] * scippio joined the chat.
[13:02:10] * deryni joined the chat.
[13:21:39] * xnyhps left the chat.
[13:21:46] * xnyhps joined the chat.
[14:11:17] * naw joined the chat.
[14:12:04] * waqas joined the chat.
[14:38:33] * jcea joined the chat.
[14:53:41] * Zash left the chat.
[14:53:44] * Zash joined the chat.
[15:01:22] * Maranda left the chat.
[15:04:10] * Maranda joined the chat.
[15:09:03] * Zash left the chat.
[15:10:23] * Zash joined the chat.
[15:11:27] * MattJ joined the chat.
[15:18:45] * jcea left the chat.
[15:24:20] * jcea joined the chat.
[15:27:48] * Flow joined the chat.
[15:47:44] * waqas left the chat.
[15:48:04] * waqas joined the chat.
[15:48:59] * waqas left the chat.
[16:01:40] * Maranda left the chat.
[16:26:26] * Maranda joined the chat.
[16:33:29] <> /me grins.
[16:35:28] <> /me thinks that from what he observed, implementing some stuff for the manifesto tomorrow "fireworks are to be expected".
[16:36:35] * JBSregath joined the chat.
[16:37:38] * waqas joined the chat.
[16:38:59] * JBSregath left the chat.
[16:39:07] <> If servers are just doing what the manifesto says (not requiring strong auth, falling back to dialback), then the only thing
that's added is that they'll not connect to servers not offering starttls. Other than (at the moment) Google-hosted domains,
I think that's a relatively small proportion of servers.
[16:51:47] <> Kev, falling back to dialback means they should perform starttls prior to that, and apparently that broken interoperability
with a good chunk of ejabberd servers and all Openfire ones here.
[16:52:59] <> Oh, you mean starttls on the dialback stream itself? That seems largely pointless, at first glance.
[16:53:57] <> Kev, not all servers use a separate streams to perform dialback (even because without bidi there's little point into)
[16:54:08] <> Kev, not all servers use a separate stream to perform dialback (even because without bidi there's little point into)
[16:54:23] <> There's no such thing as a dialback stream in prosody. And unless Maranda added it, in Metronome.
[16:54:55] <> Prosody does starttls if offered and supported.
[16:55:18] <> Well, there is a dialbackstream in the sense of a stream on which dialback happens.
[16:55:24] <> It doesn't have to be uniquely a dialback stream.
[16:55:41] <> If it is uniquely a dialback stream, there's not much gained by unauthenticated TLS.
[16:55:50] <> Actually, that's not true, because you can do dwd and things.
[16:56:46] <> Prosody will reuse or start a normal s2s stream for dialback.
[16:57:16] <> but dwd has interop problems out there IIRC
[16:57:39] <> Maranda, what incompatibilites are there with ejabberd? At least newer versions should behave ok enough.
[16:57:47] <> Kev, I think what broken a lot of servers is that I removed the dialback stream namespace declaration on the stream header
if encryption for s2s streams is required.
[16:57:50] <> Openfire............
[16:57:52] <> Zash ^
[16:58:19] <> They just throw random invalid-namespace errors which is hilarious ^^
[16:58:25] <> Fun
[16:59:21] <> Maranda: Is that right?
[17:00:00] <> Kev, xep-220 says that stream namespace declaration is susperseeded isn't it?
[17:00:17] <> /me is sure he read that line correctly
[17:00:59] <> "Note: All XML elements qualified by the Server Dialback namespace MUST be prefixed with the namespace prefix for the 'jabber:server:dialback'
namespace as advertised on the stream header originally sent by the entity sending the element."?
[17:01:39] <> That's done.
[17:01:40] <> In fact, later on it explicitly says that you have to use stream namespace declarations for dialback, even though other protocols
don't.
[17:01:47] <> ...
[17:01:49] <> "Note: Although in general advertising protocol support by means of an XML namespace declaration has been superseded by the
use of stream features as originally defined in RFC 3920 and updated in RFC 6120, the server dialback protocol predates the
existence of stream features and therefore the namespace declaration method is still used in this instance."
[17:02:40] <> /me underlines "in this istance"
[17:03:05] <> Yes.
[17:03:12] <> So other protocols don't do it, but dialback does.
[17:03:17] <> I think you've misread this.
[17:03:50] <> The line I just pasted is immediately subsequent to "Support for the traditional server dialback protocol (originally specified
in RFC 3920) is indicated by inclusion of the dialback namespace declaration in the stream header."
[17:03:59] * aman joined the chat.
[17:04:29] <> I'm inclined to say that not advertising dialback in the stream header is broken.
[17:06:20] <> Honestly that's hardly interpretable that way, "this instance" more goes for "we show it here for documenting reasons."
[17:07:09] <> <<the namespace declaration method MUST still be used in this instance.>> would be correct likewise if it still has to be
included.
[17:08:22] <> I don't really see how you can read "Support for the traditional server dialback protocol (originally specified in RFC 3920)
is indicated by inclusion of the dialback namespace declaration in the stream header." and come to the conclusion that support
isn't signalled with a namespace declaration in the stream header.
[17:09:06] <> or <<the namespace declaration method MUST still be used in for the purpose of this document>>
[17:10:06] <> I don't know what you're saying with that.
[17:10:36] <> It's late Friday afternoon, I might be going insane.
[17:10:46] <> "MUST still be used in this instance" would certainly be an improved formulation of that sentence though I'm not sure I agree
that the current one is confusing.
[17:10:52] <> MattJ / waqas / other server devs. Am I going mad here?
[17:11:59] <> Kev: I'm half-awake, but you seem sane
[17:12:16] <> Ta.
[17:14:53] <> Kev, because there isn't any mandating statement in phrase and frankly the RFC document formatting provides that kind of terminology
which would avoid misinterpretation, and maybe my english is not so good, but for sure <<the namespace declaration method
MUST still be used for the purposes of this document>> or <<MUST still be used in this instance>> isn't subject to interpretations,
the text there is.
[17:15:01] <> waqas, jet lag? :)
[17:15:05] <> Kev, because there isn't any mandating statement in that phrase and frankly the RFC document formatting provides that kind
of terminology which would avoid misinterpretation, and maybe my english is not so good, but for sure <<the namespace declaration
method MUST still be used for the purposes of this document>> or <<MUST still be used in this instance>> isn't subject to
interpretations, the text there is.
[17:15:44] <> Maranda: I'm not arguing that a MUST wouldn't be appropriate, but the text that's there is not ambiguous.
[17:16:00] <> Sorry to insist but it is
[17:17:18] <> Yeah, I don't think you're going to persuade me here.
[17:21:05] <> ..and even so my implementation is correct in which I do appropriately namespace all db elements, perhaps I would think most
(6120 compliant) servers should have a sane enough XML parsing to not throw (wrongly) random invalid-namespace errors but
that's another topic.
[17:22:22] <> Kev, ditto but maybe the fact that there're so many erratic implementations out there is because technical specs docs aren't
as clear as you think ;)
[17:22:50] <> I doubt anyone has implemented off XEP-220, TBH.
[17:23:04] <> Most dialback implementations come from 3920 days, when it was much less clear.
[17:26:36] * 0xAFFE left the chat.
[18:03:15] * Tobias left the chat.
[18:14:41] <> Kev, I re-read it carefully (including footers...) and you're right, I'm convinced now. I think I got mislead by the "superseeded"
and "updated" in the notes.
[18:15:15] <> <<Removed stream feature for advertising mere protocol support, using it only for advertising support for enhanced error handling.>>,
*goes to remove code*
[18:24:34] * Darlan joined the chat.
[18:35:50] <> Great :)
[18:41:05] <> And thus peace was restored to the realm
[18:41:20] <> :D
[18:43:23] <> waqas, as long as removing the dialback element doesn't break anything...
[18:43:27] * Tobias joined the chat.
[18:43:31] <> (from features that is)
[18:43:43] <> The dialback element should still be in feature,s shouldn't it?
[18:43:49] <> No
[18:43:49] * aman left the chat.
[18:44:01] <> You put it in features to show you support the extended errors, and the stream header to show you support the old version.
[18:44:08] <> The doc says otherwise since Metronome doesn't support graceful error handling
[18:44:29] <> OK, if you don't support the newer errors then yes.
[18:44:35] <> Kev, I managed to imprint my brain right now ;)
[18:49:34] * Flow left the chat.
[18:49:46] * Flow joined the chat.
[19:04:39] * scippio left the chat.
[19:06:06] * KevWalke left the chat.
[19:06:06] * jcea left the chat.
[19:06:06] * KevWalke joined the chat.
[19:24:36] * bear joined the chat.
[19:43:37] * kevin joined the chat.
[19:43:38] * bear joined the chat.
[19:54:14] * bear left the chat.
[20:03:43] * Tobias left the chat.
[20:07:46] * Tobias joined the chat.
[20:21:45] * Maranda left the chat.
[20:29:09] * Maranda joined the chat.
[20:35:52] * kevin left the chat.
[20:41:56] * naw left the chat.
[20:54:15] * dezant joined the chat.
[21:07:28] * Tobias left the chat.
[21:08:54] * Tobias joined the chat.
[21:15:18] * dezant left the chat.
[21:15:50] * dezant joined the chat.
[21:27:16] * dezant left the chat.
[21:27:19] * dezant joined the chat.
[21:39:51] * Lance joined the chat.
[21:59:42] * jabberjocke joined the chat.
[22:01:58] * Lance left the chat.
[22:03:12] * Lance joined the chat.
[22:05:16] * Lance joined the chat.
[22:32:26] * Lance left the chat.
[22:32:27] * Lance joined the chat.
[22:35:20] * bear left the chat.
[22:36:02] * bear joined the chat.
[22:42:27] * bear left the chat.
[22:42:35] * bear joined the chat.
[23:01:54] <> Kev, out of curiousity M-Link does mandate dialback if it doesn't trust the receiving entity certificate?
[23:05:48] * jabberjocke left the chat.
[23:07:54] * jabberjocke joined the chat.
[23:21:52] * Lance left the chat.
[23:27:05] * Lance left the chat.
[23:35:23] * Valérian joined the chat.
[23:35:37] * Valérian left the chat.
[23:37:06] * Lance joined the chat.