IETF
QUIC
quic@jabber.ietf.org
Tuesday, November 6, 2018< ^ >
Christian has set the subject to: Siren Retry
Room Configuration
Room Occupants

GMT+0
[01:24:26] ekr joins the room
[01:44:37] ekr leaves the room
[01:54:57] ekr joins the room
[01:58:51] ekr leaves the room
[02:01:09] ekr joins the room
[02:22:05] ekr leaves the room
[02:23:18] ekr joins the room
[02:56:49] ekr leaves the room
[03:36:20] ekr joins the room
[03:42:13] ekr leaves the room
[03:53:50] ekr joins the room
[03:54:07] ekr leaves the room
[03:57:53] ekr joins the room
[04:01:47] ekr leaves the room
[04:23:31] ekr joins the room
[05:06:26] ekr leaves the room
[05:08:39] ekr joins the room
[05:22:23] ekr leaves the room
[05:28:35] ekr joins the room
[05:29:14] ekr leaves the room
[06:37:28] 1g1sFs63 joins the room
[06:37:33] Meetecho joins the room
[06:45:09] Yisong Liu joins the room
[06:45:09] Alan Frindell joins the room
[06:45:10] Daniel Havey joins the room
[06:45:13] Nicholas Banks joins the room
[06:45:14] Bernard Aboba joins the room
[06:45:31] Dmitri Tikhonov joins the room
[06:47:34] Praveen Balasubramanian joins the room
[06:48:05] Simon Pietro Romano joins the room
[06:48:15] Juan Pedro Cerezo joins the room
[06:48:43] Klaus Hartke joins the room
[06:49:53] Thomas Peterson joins the room
[06:50:14] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[06:50:22] Daniel Stenberg joins the room
[06:51:28] Victor Vasiliev joins the room
[06:51:45] Thomas Peterson joins the room
[06:51:46] Jonathan Lennox joins the room
[06:52:28] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[06:52:37] Kaoru Maeda joins the room
[06:52:55] Thomas Peterson joins the room
[06:52:57] <Thomas Peterson> I can scribe if need be.
[06:53:04] nat joins the room
[06:53:11] Eric Kinnear joins the room
[06:53:40] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[06:53:55] Bo B joins the room
[06:53:59] Magnus Westerlund joins the room
[06:54:16] <Eric Kinnear> Hi all, I'm Jabber relay today, use "mic: " for comments you'd like relayed at the mic
[06:54:25] Lars Eggert joins the room
[06:54:37] Thomas Peterson joins the room
[06:54:57] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[06:55:12] Magnus Westerlund leaves the room: Stream closed by us: Replaced by new connection (conflict)
[06:55:13] Magnus Westerlund joins the room
[06:55:30] Martin Thomson joins the room
[06:55:33] Thomas Peterson joins the room
[06:55:55] blassey joins the room
[06:56:05] Yisong Liu leaves the room
[06:56:23] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[06:56:29] Mohit Suley joins the room
[06:57:01] g.e.montenegro joins the room
[06:57:09] Thomas Peterson joins the room
[06:57:54] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[06:58:18] frodek joins the room
[06:58:41] Thomas Peterson joins the room
[06:58:58] <Alan Frindell> mic: qpack had 3 implementations
[06:59:23] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[06:59:25] <Alan Frindell> yes
[06:59:27] Luca Niccolini joins the room
[06:59:34] <Alan Frindell> plus two non-interop
[07:00:00] Dragana Damjanovic joins the room
[07:00:10] chi.jiun.su joins the room
[07:00:10] Thomas Peterson joins the room
[07:00:56] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[07:01:43] Thomas Peterson joins the room
[07:02:07] Magnus Westerlund leaves the room: Connection failed: connection closed
[07:02:49] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[07:03:15] Magnus Westerlund joins the room
[07:03:36] Thomas Peterson joins the room
[07:03:45] Thomas Peterson leaves the room
[07:06:20] <Martin Thomson> TIME
[07:06:41] Martin Thomson leaves the room
[07:06:55] Suneesh Babu joins the room
[07:07:07] ekr joins the room
[07:07:21] Ekaterina Volodina joins the room
[07:08:50] Vidhi Goel joins the room
[07:09:05] HAJIME WATANABE joins the room
[07:10:14] Philipp Tiesel joins the room
[07:10:20] <Alan Frindell> (y)
[07:10:51] Thomas Peterson joins the room
[07:11:32] Thomas Peterson leaves the room: Disconnected: Replaced by new connection
[07:13:04] Suneesh Babu leaves the room
[07:16:04] Magnus Westerlund leaves the room: Connection failed: connection closed
[07:16:21] Philipp S. Tiesel joins the room
[07:16:43] Philipp S. Tiesel leaves the room
[07:16:47] Magnus Westerlund joins the room
[07:27:06] ted.h joins the room
[07:27:18] Martin Thomson joins the room
[07:29:11] Magnus Westerlund leaves the room: Connection failed: connection closed
[07:29:24] Magnus Westerlund joins the room
[07:33:15] Thomas Peterson joins the room
[07:37:31] Michele Polese joins the room
[07:37:50] Dan York joins the room
[07:40:00] <Martin Thomson> OK, we are now WAAAAY over time
[07:41:02] Brian Trammell joins the room
[07:41:05] theresa_enghardt joins the room
[07:47:29] janaiyengar joins the room
[07:51:34] Magnus Westerlund leaves the room: Connection failed: connection closed
[07:51:43] <Victor Vasiliev> I am pretty sure the conclusion in TLS was that you always have to do a top-level retry, but I don't believe tlswg ever ended up making concrete recommendations on this topic
[07:51:55] janaiyengar leaves the room
[07:52:06] <ekr> by “top level” —> punt back to the app?
[07:52:12] Magnus Westerlund joins the room
[07:52:16] <Victor Vasiliev> Yeah
[07:52:49] <Victor Vasiliev> There are concerns about ALPN, having exporters in 0-RTT, etc, but I'm not sure we ever wrote them down
[07:53:14] <Martin Thomson> I thought that we had written that down somewhere
[07:53:25] <Martin Thomson> but I guess it slipped through the cracks
[07:53:28] Martin Thomson leaves the room
[07:53:38] <Victor Vasiliev> We might have and I forgot about it
[07:53:44] <ekr> “   A TLS implementation SHOULD NOT automatically resend early data;
   applications are in a better position to decide when retransmission
   is appropriate.  A TLS implementation MUST NOT automatically resend
   early data unless the negotiated connection selects the same ALPN
   protocol.

[07:53:56] <ekr> S 4.2.10
[07:54:14] <Victor Vasiliev> Great!
[07:55:27] <Victor Vasiliev> I think this means the draft should just cite that, or maybe I'm misunderstanding the question?
[07:55:37] <ekr> I tend t think so
[07:59:42] Praveen Balasubramanian leaves the room
[08:00:06] <Dmitri Tikhonov> Yes, skip varints
[08:00:53] ted.h leaves the room
[08:01:11] Magnus Westerlund leaves the room: Connection failed: connection closed
[08:01:42] Magnus Westerlund joins the room
[08:01:52] Martin Thomson joins the room
[08:03:32] ted.h joins the room
[08:05:18] Magnus Westerlund leaves the room: Connection failed: connection closed
[08:05:18] <Alan Frindell> mic: in our implementation, we've decided to default the QPACK table size to 0 if we need to send a response before SETTINGS have arrived.
[08:05:33] Magnus Westerlund joins the room
[08:05:43] <Dmitri Tikhonov> I've done the same today.  +1, Alan
[08:05:57] Mohit Suley leaves the room
[08:07:02] Nicholas Banks leaves the room
[08:07:24] <Dmitri Tikhonov> What's the current holb with streams?  I didn't get it.
[08:07:55] Nicholas Banks joins the room
[08:09:55] <Alan Frindell> It should be rare that SETTINGS is lost
[08:10:36] <Alan Frindell> what was the h2 default?
[08:11:09] <Alan Frindell> mic: in h2 the default was unlimited, right?
[08:11:15] dragana joins the room
[08:11:55] <Eric Kinnear> @Alan, default for what specifically?
[08:12:01] <Alan Frindell> default for header block size
[08:12:26] <Eric Kinnear> Gotcha thanks
[08:12:49] <Eric Kinnear> Still need that at the mic after Mike's comment?
[08:13:02] <Alan Frindell> i guess not
[08:13:49] Magnus Westerlund leaves the room: Connection failed: connection closed
[08:14:06] <Eric Kinnear> 👍
[08:14:47] <Alan Frindell> what is a server supposed to do if it wants to send a response with headers that exceed the client's limit for header block size?  Remove some headers?  500?
[08:14:47] Magnus Westerlund joins the room
[08:17:07] Magnus Westerlund leaves the room: Connection failed: connection closed
[08:17:51] Magnus Westerlund joins the room
[08:18:07] <Martin Thomson> 500 seems fine
[08:18:12] Vidhi Goel leaves the room
[08:18:16] <Martin Thomson> CONNECTION_CLOSE would also work
[08:18:21] <Martin Thomson> but I guess the spec doesn't say
[08:19:04] spencerdawkins joins the room
[08:20:05] <Alan Frindell> send it anyways and let the peer deal with it
[08:20:10] Magnus Westerlund leaves the room: Connection failed: connection closed
[08:21:12] Magnus Westerlund joins the room
[08:22:10] <Alan Frindell> +1 for do something
[08:22:17] <Dmitri Tikhonov> mic: Ian, you'd still have to support peers that frame DATA, so the complexity argument is moot, unless you advocate dropping sized HQ frames.
[08:24:11] <Alan Frindell> lol
[08:25:04] Magnus Westerlund leaves the room: Connection failed: connection closed
[08:25:29] <ted.h> @dmitri Does having this as a different frame type solve your concern there?
[08:25:54] Magnus Westerlund joins the room
[08:25:55] gianluca capitani joins the room
[08:26:28] <Dmitri Tikhonov> ted.h, my only real concern is that the proposed to-the-end-of-stream frame size is so convenient, no one will use sized frames.
[08:27:02] <Victor Vasiliev> Well, you have to use it if you implement trailers, correct?
[08:27:03] gianluca capitani leaves the room
[08:27:23] <Dmitri Tikhonov> Yes.
[08:27:41] <Dmitri Tikhonov> There are many HTTP implementations that don't support trailers.
[08:28:00] <Dmitri Tikhonov> (What was the resolution re: variable-size HQ frames?)
[08:28:25] <Alan Frindell> is it too late to throw out the priority tree and replace with 2 bit priority?
[08:29:00] <Alan Frindell> @dmitri: take it too the list/flesh out concrete proposals
[08:29:14] <ted.h> Mauve, no, puce!
[08:29:22] <ted.h> Maybe a light cerulean?
[08:29:48] <Alan Frindell> And one of the oliver boliver butt!
[08:30:12] <Dmitri Tikhonov> Levity ensues
[08:30:16] ISOC BENIN joins the room
[08:31:31] <Victor Vasiliev> This is amazing!
[08:31:37] ISOC BENIN leaves the room
[08:31:45] <Alan Frindell> @ekr: excellent Sneetches reference
[08:31:48] <Dmitri Tikhonov> Another Dr. Seuss reference!
[08:31:49] ISOC BENIN joins the room
[08:32:54] <Victor Vasiliev> 1.3 actually obsoletes 1.2, but h1 and h2 is... interesting
[08:34:05] Martin Thomson leaves the room
[08:34:25] <Brian Trammell> h(2+1i)
[08:34:30] Martin Thomson joins the room
[08:34:36] <Victor Vasiliev> HTTP/2.5
[08:35:00] <Brian Trammell> (FWIW I think HTTP3 is the right choice, but only if we can keep the logo)
[08:35:21] <Martin Thomson> I would like to point out that we're safe because there is more than one bit involved in this decision
[08:35:21] <Brian Trammell> SPDY/2
[08:35:35] <ekr> I don’t know who 1239…. is, but you are a genius
[08:35:41] <Martin Thomson> that slash in the version number is just weird
[08:37:21] <ekr> Perhaps we could call it HTTP/3 and then use the alpn token hq1
[08:37:35] <ted.h> @ekr  That works for me.
[08:37:54] Brian joins the room
[08:38:47] <Martin Thomson> Every time I see something called "something-1.0" I cringe.  Sure, add "2" to the next version, but own it people.  
[08:39:10] <Martin Thomson> HTTP/Q is fine
[08:39:10] <Brian> wait, wouldn't the right alpn token be h3q1?
[08:39:12] <ekr> I really like 2+1
[08:39:15] <ekr> 2+i
[08:39:20] <Thomas Peterson> HTTP 3.11 for Workgroups?
[08:39:27] <Brian> can you have a + in the version number?
[08:39:39] <ted.h> Is your question Version or ALPN token?
[08:39:41] <ekr> @brian: that’s a complex question. LOL
[08:39:41] nat leaves the room
[08:40:04] <Brian> (apparently i'm 1237... -- the meetecho chat doesn't carry names over for some reason)
[08:40:14] <Martin Thomson> so the answer is that we can call it whatever we like because it won't ever hit any of the protocol elements that carry "legacy" versions, outside of the systems Kazuho mentioned
[08:40:20] <Martin Thomson> and those can just deal, in my view
[08:40:25] <Brian> +1 mt
[08:40:41] <Brian> of course the number of degrees of freedom here is why the mic line is so long
[08:41:06] <ted.h> Wait, is the ALPN token not changing if you run HTTP over some new version of QUIC?
[08:41:15] <Brian> ooof.
[08:41:21] Magnus Westerlund leaves the room: Connection failed: connection closed
[08:41:24] <ted.h> That's what I got from Lars' statement.
[08:41:28] <Brian> doesn't it have to?
[08:41:30] Magnus Westerlund joins the room
[08:41:49] <Martin Thomson> I propose that we call the transport QWIK and the HTTP mapping Boaty McBoatface
[08:41:58] Maxime Piraux joins the room
[08:42:00] <ekr> KWIQ
[08:42:12] <ted.h> It's pretty obvious from the slides what to do:  the name should be "Dave"
[08:42:14] <Martin Thomson> that works too
[08:42:25] Simone Ferlin joins the room
[08:42:25] <Victor Vasiliev> Texty McHypertextface
[08:43:29] <Brian> less facetiously -- the ALPN token carries semantics about the handshake to use for a resumption, no?
[08:43:43] <Brian> so the hXqY token has to encode the handshake at least
[08:43:56] <Brian> HMMMMMMMMMMMMMMMMMM
[08:44:05] Dmitri Tikhonov hum for renaming
[08:44:10] <Victor Vasiliev> hum in support of renaming
[08:45:34] <Martin Thomson> and if you are sick of humming, please hum now
[08:45:50] <ekr> +1 for naming it moonface
[08:46:02] <Brian Trammell> h3q1mf is a nice token
[08:46:09] <Victor Vasiliev> What I hear you saying is "hear you on Thursday at httpbis"
[08:46:15] <Victor Vasiliev> *see you
[08:46:55] ted.h leaves the room
[08:47:42] <Martin Thomson> I love the way that humming in jabber has none of the properties we look for - it is neither semi-anonymous, nor does it actually help answer the question
[08:48:50] <Brian> it's also delayed enough that it is not actually clear which answer is being given
[08:48:57] <Brian> but traditions are nice so
[08:49:06] <Alan Frindell> hmmmm
[08:49:06] <ekr> you say it’s not anonymous, but you could presumably leave jabber and then rejoin with a different name
[08:49:27] <Dmitri Tikhonov> The linkability is trivial
[08:49:31] ted.h joins the room
[08:49:46] <Victor Vasiliev> We might need to define an XMPP extension for this, I heard that's what was in fashion 10 years ago
[08:49:58] <ekr> Not necessarily. I see a lot of people have these numeric meetecho addresses, so I bet theres a way to get a new one
[08:50:18] <ted.h> @ekr  re-register as a new user.
[08:50:23] <Brian> ah! the meetecho numeric addresses are for humming!
[08:50:24] <ekr> So that’s not ideal
[08:50:39] ekr leaves the room
[08:50:40] ted.h leaves the room
[08:50:55] <Martin Thomson> I propose that we instead use the worm: https://en.wikipedia.org/wiki/Worm_%28marketing%29
[08:51:00] <Meetecho> No the numeric address is a fake user that sits there to get the list of attendees and put it on the 2nd screen
[08:51:03] g.e.montenegro leaves the room
[08:51:11] <Meetecho> We'll have to give it a prettier display name :)
[08:51:16] Eric Kinnear leaves the room
[08:51:17] Eric Kinnear joins the room
[08:51:51] spencerdawkins leaves the room: Disconnected: closed
[08:51:52] Eric Kinnear leaves the room: Stream reset by peer
[08:51:57] Alan Frindell leaves the room
[08:52:02] Dan York leaves the room
[08:52:05] Jonathan Lennox leaves the room
[08:52:05] Martin Thomson leaves the room
[08:52:09] Bo B leaves the room
[08:52:11] <Victor Vasiliev> I'm confused by then numbers -- when I look up logs online, I only see the names
[08:52:12] Jonathan Lennox joins the room
[08:52:24] Lars Eggert leaves the room: Disconnected: closed
[08:52:34] Ekaterina Volodina leaves the room
[08:52:34] Klaus Hartke leaves the room
[08:52:34] Brian Trammell leaves the room
[08:52:34] Dragana Damjanovic leaves the room
[08:52:34] Luca Niccolini leaves the room
[08:52:34] Michele Polese leaves the room
[08:52:34] Philipp Tiesel leaves the room
[08:52:34] Kaoru Maeda leaves the room
[08:52:34] Bernard Aboba leaves the room
[08:52:34] Nicholas Banks leaves the room
[08:52:34] Daniel Stenberg leaves the room
[08:52:34] Maxime Piraux leaves the room
[08:52:34] Simon Pietro Romano leaves the room
[08:52:34] Juan Pedro Cerezo leaves the room
[08:52:34] Dmitri Tikhonov leaves the room
[08:52:34] HAJIME WATANABE leaves the room
[08:52:34] ISOC BENIN leaves the room
[08:52:34] Simone Ferlin leaves the room
[08:52:34] Daniel Havey leaves the room
[08:52:34] Victor Vasiliev leaves the room
[08:52:42] chi.jiun.su leaves the room
[08:52:44] <Meetecho> If you mean in general, we do use display names -- maybe your jabber client doesn't display them?
[08:53:06] Meetecho leaves the room
[08:53:11] blassey leaves the room
[08:54:56] Eric Kinnear joins the room
[08:56:15] theresa_enghardt leaves the room
[08:57:54] Brian leaves the room: Disconnected: closed
[09:01:35] Brian joins the room
[09:02:49] frodek leaves the room
[09:04:49] Jonathan Lennox leaves the room
[09:07:58] Jonathan Lennox joins the room
[09:08:17] Magnus Westerlund leaves the room: Connection failed: connection timed out
[09:08:34] Thomas Peterson leaves the room: Disconnected: closed
[09:10:59] dragana leaves the room: Disconnected: closed
[09:11:45] Brian leaves the room: Disconnected: closed
[09:12:45] Jonathan Lennox leaves the room
[09:12:50] Lars Eggert joins the room
[09:13:13] ekr joins the room
[09:13:41] Eric Kinnear leaves the room: Connection failed: connection timed out
[09:15:35] Eric Kinnear joins the room
[09:15:45] Lars Eggert leaves the room
[09:17:11] blassey joins the room
[09:21:03] Martin Thomson joins the room
[09:21:51] Magnus Westerlund joins the room
[09:22:14] chi.jiun.su joins the room
[09:30:13] Magnus Westerlund leaves the room: Connection failed: connection closed
[09:47:58] blassey leaves the room
[10:36:25] spencerdawkins joins the room
[10:36:34] ekr leaves the room
[10:38:46] Martin Thomson leaves the room
[10:40:24] Brian joins the room
[10:44:57] Martin Thomson joins the room
[10:45:06] Martin Thomson leaves the room
[11:04:40] Eric Kinnear leaves the room: Stream reset by peer
[11:14:15] spencerdawkins leaves the room: Disconnected: closed
[11:20:45] ekr joins the room
[11:21:03] ekr leaves the room
[11:32:41] ekr joins the room
[11:33:10] Brian leaves the room
[11:36:35] chi.jiun.su leaves the room
[11:37:37] ekr leaves the room
[11:56:35] chi.jiun.su joins the room
[13:05:49] Brian joins the room
[13:06:15] Brian leaves the room: Disconnected: closed
[13:27:20] chi.jiun.su leaves the room
[16:45:15] Martin Thomson joins the room
[17:08:44] Martin Thomson leaves the room
[22:16:53] ekr joins the room
[22:23:36] ekr leaves the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!