[{"author": "Will Law", "text": "<p>We can't hear anything in Shenzhen, if you are talking.</p>", "time": "2026-03-17T03:32:32.000Z"}, {"author": "Will Law", "text": "<p>yes</p>", "time": "2026-03-17T03:33:02.000Z"}, {"author": "Tommy Pauly", "text": "<p>Now we can</p>", "time": "2026-03-17T03:33:03.000Z"}, {"author": "Lucas Pardue", "text": "<p>I can hear you remote too</p>", "time": "2026-03-17T03:33:16.000Z"}, {"author": "Mike English", "text": "<p><span aria-label=\"clap\" class=\"emoji emoji-1f44f\" role=\"img\" title=\"clap\">:clap:</span></p>", "time": "2026-03-17T03:40:40.000Z"}, {"author": "Jonathan Lennox", "text": "<p>Tempted to suggest 418</p>", "time": "2026-03-17T03:47:11.000Z"}, {"author": "Alan Frindell", "text": "<p><a href=\"https://http.cat/405\">https://http.cat/405</a></p>", "time": "2026-03-17T03:47:25.000Z"}, {"author": "Alan Frindell", "text": "<p><a href=\"https://http.cat/406\">https://http.cat/406</a></p>", "time": "2026-03-17T03:47:40.000Z"}, {"author": "Ted Hardie", "text": "<p>426 seems closer, though it would be a weird view of \"Upgrade\".</p>", "time": "2026-03-17T03:51:11.000Z"}, {"author": "Jonathan Lennox", "text": "<p>Is there a converse issue of what you should do if someone tries to do a GET on an extended-connect-only resource?</p>", "time": "2026-03-17T03:51:34.000Z"}, {"author": "Mike English", "text": "<p>@Ted I was also looking at 426, but it is awkward</p>", "time": "2026-03-17T03:52:42.000Z"}, {"author": "Junxiao Shi", "text": "<p>415 Unsupported Media Type<br>\nWebTransport can be seen as a \"media type\".</p>", "time": "2026-03-17T03:53:29.000Z"}, {"author": "Mike English", "text": "<blockquote>\n<p>The server MUST send an Upgrade header field in a 426 response to indicate the required protocol(s)</p>\n</blockquote>", "time": "2026-03-17T03:53:29.000Z"}, {"author": "Lucas Pardue", "text": "<p>new problem type?</p>", "time": "2026-03-17T03:54:31.000Z"}, {"author": "Lucas Pardue", "text": "<p><code>:protocol not supported for resource</code></p>", "time": "2026-03-17T03:54:57.000Z"}, {"author": "Benjamin Schwartz", "text": "<p>It's not ambiguous whether the server supports WebTransport.  That's already clear from the SETTINGS</p>", "time": "2026-03-17T03:56:48.000Z"}, {"author": "Benjamin Schwartz", "text": "<p>The point is that the status code is basically useless because these errors are all unrecoverable</p>", "time": "2026-03-17T03:57:49.000Z"}, {"author": "Kazuho Oku", "text": "<p>405 + Proxy-Status?</p>", "time": "2026-03-17T03:58:11.000Z"}, {"author": "Bron Gondwana", "text": "<p>I put my hand up to say \u201cnew status code\u201d too</p>", "time": "2026-03-17T03:58:18.000Z"}, {"author": "Martin Thomson", "text": "<p>I just think that 405 works perfectly fine for this: it's a situation where the CONNECT isn't viable.</p>", "time": "2026-03-17T03:58:45.000Z"}, {"author": "Harald Alvestrand", "text": "<p>does it matter what existing implementations (without WT but with connect) will return if you try to talk WT to them?</p>", "time": "2026-03-17T03:58:53.000Z"}, {"author": "Ted Hardie", "text": "<p>@benjamin  that seems like 400 is okay,if we don't expect the client to be able to do anything sensible to recover.</p>", "time": "2026-03-17T03:58:54.000Z"}, {"author": "Kazuho Oku", "text": "<p>IMO, if \"what to do next\" is going to be specific to WebTransport, then using a header field to indicate that \"what\" makes more sense</p>", "time": "2026-03-17T03:59:07.000Z"}, {"author": "Jonathan Lennox", "text": "<p>Yeah, this problem seems to be \"human [or LLM] needs to fix something in the code or configuraiton\", not something where a client can try something else.</p>", "time": "2026-03-17T03:59:15.000Z"}, {"author": "Benjamin Schwartz", "text": "<p>Yes, I think 400 or whatever is fine.  This situation can only occur in the event of a serious misconfiguration, and is unrecoverable regardless of status code.</p>", "time": "2026-03-17T03:59:28.000Z"}, {"author": "Martin Thomson", "text": "<p>It's not like we really care about the distinction between \"failed because CONNECT isn't supported\" and \"failed because CONNECT with webtransport isn't supported\" isn't helpful.</p>", "time": "2026-03-17T03:59:34.000Z"}, {"author": "Alan Frindell", "text": "<p>are there two cases?  resource is not a wt resource and wt protocol negotiation failed because no mutual option was found</p>", "time": "2026-03-17T03:59:35.000Z"}, {"author": "Benjamin Schwartz", "text": "<p>(To be clear: I fully expect different implementations to produce different error codes, and that is fine because there is nothing useful to do with the error code anyway.)</p>", "time": "2026-03-17T04:00:38.000Z"}, {"author": "Martin Thomson", "text": "<p>It might help to have a clear articulation of what a client might do when it receives this error response.  Trying again on a different SVCB path seems to be the only thing.  In which case, a 400 would work.</p>", "time": "2026-03-17T04:00:55.000Z"}, {"author": "Lucas Pardue", "text": "<p>400 would be fine by me</p>", "time": "2026-03-17T04:01:31.000Z"}, {"author": "Alan Frindell", "text": "<p>it's useful for interop.  I think I was failing wt-avail protocol negotiation with a js application in chrome and the failure code was...not helpful</p>", "time": "2026-03-17T04:01:31.000Z"}, {"author": "Christian Huitema", "text": "<p>The privacy issue that I think of is kinda related to David Schinazi unprompted authentication. Field a server that appears innocuous to unauthenticated users, but expose data if authenticated. In that case, unauthenticated user trying to use the service will get \"no such thing here\".</p>", "time": "2026-03-17T04:01:47.000Z"}, {"author": "Benjamin Schwartz", "text": "<p>IMHO all SVCB endpoints are required to be semantically interchangeable unless marked with a mandatory SvcParam, so there should be no need to retry/falback.</p>", "time": "2026-03-17T04:02:07.000Z"}, {"author": "Victor Vasiliev", "text": "<p>mic question: are there H/2 implementations interoperable?</p>", "time": "2026-03-17T04:02:40.000Z"}, {"author": "Benjamin Schwartz", "text": "<p>@Alan: That can be handled by including a useful error message in the response body, regardless of status code</p>", "time": "2026-03-17T04:02:52.000Z"}, {"author": "Martin Thomson", "text": "<p>A server that requires application protocol negotiation MAY reject the session if the <code>WT-Available-Protocols</code> header field is absent or if it is malformed and therefore ignored. --&gt; if nothing is acceptable, could you respond with <code>WT-Protocol: error</code>.</p>", "time": "2026-03-17T04:02:54.000Z"}, {"author": "Alan Frindell", "text": "<p>ERR_QUIC_PROTOCOL_ERROR: please call Lucas</p>", "time": "2026-03-17T04:04:02.000Z"}, {"author": "Lucas Pardue", "text": "<p>i need a :hardcry: emoji here</p>", "time": "2026-03-17T04:04:46.000Z"}, {"author": "Alan Frindell", "text": "<p>Sounds good thanks!</p>", "time": "2026-03-17T04:07:09.000Z"}, {"author": "Dmitri Tikhonov", "text": "<p>+1 To Marten</p>", "time": "2026-03-17T04:07:14.000Z"}, {"author": "Lucas Pardue", "text": "<p>so Alan, Dmitri and Christian can onboard by end of week?</p>", "time": "2026-03-17T04:08:01.000Z"}, {"author": "Marten Seemann", "text": "<p><a href=\"https://github.com/quic-interop/quic-interop-runner/blob/master/webtransport.md\">https://github.com/quic-interop/quic-interop-runner/blob/master/webtransport.md</a></p>", "time": "2026-03-17T04:08:24.000Z"}, {"author": "Alan Frindell", "text": "<p>@lucas: I asked our dev who's leading the WT charge to do so, yes</p>", "time": "2026-03-17T04:08:46.000Z"}, {"author": "Martin Thomson", "text": "<p>here's to no meeting in IETF 126</p>", "time": "2026-03-17T04:08:49.000Z"}, {"author": "Dmitri Tikhonov", "text": "<p>I'll shoot for end of next week maybe</p>", "time": "2026-03-17T04:08:51.000Z"}, {"author": "Christian Huitema", "text": "<p>I will certainly try, but it depends how much I need to implement.</p>", "time": "2026-03-17T04:09:58.000Z"}, {"author": "Junxiao Shi", "text": "<p>H2 in interop-runner, is it too much to ask for?</p>", "time": "2026-03-17T04:11:03.000Z"}, {"author": "Mike English", "text": "<p>I am excited to see an interop runner for WebTransport</p>", "time": "2026-03-17T04:11:03.000Z"}, {"author": "Alan Frindell", "text": "<p>+1 for h2, but I suppose I need to onboard for h3 before I ask Marten for more work</p>", "time": "2026-03-17T04:11:34.000Z"}, {"author": "Marten Seemann", "text": "<p>I did implement reliable resets</p>", "time": "2026-03-17T04:12:36.000Z"}, {"author": "Christian Huitema", "text": "<p>I did indeed implement Reliable Reset.</p>", "time": "2026-03-17T04:12:50.000Z"}, {"author": "Dmitri Tikhonov", "text": "<p>So did I</p>", "time": "2026-03-17T04:13:10.000Z"}, {"author": "Lucas Pardue", "text": "<p>cool, thanks for the information!</p>", "time": "2026-03-17T04:13:39.000Z"}, {"author": "Martin Duke", "text": "<p>We implemented RR for Quic, victor can confirm if it\u2019s wired through wt</p>", "time": "2026-03-17T04:13:48.000Z"}]