[15:32:58] Ted joins the room [15:33:52] http://www.ietf.org/proceedings/10mar/agenda/iri.txt is the agenda for today's meeting [15:36:14] Audio stream link should be: http://videolab.uoregon.edu/events/ietf/ietf778.m3u [15:46:14] Exodus joins the room [15:46:26] Hello Ted, [15:46:38] This is Martin Dürst [15:46:42] Hi Martin [15:46:57] I wish I knew how to set nicks in Exodus :-( [15:47:02] You're on the main screen, btw [15:48:55] ok, no problem. I can hear you guys, too [15:48:55] KIM K leaves the room [15:49:06] good morning [15:49:23] Martin is there a lag [15:49:43] I don't know :-) [15:50:10] Julian joins the room [15:50:18] 米谷嘉朗 joins the room [15:50:19] Well, I spoke about the same time I typed, but obviously there could be a lag on both. [15:50:29] 米谷嘉朗 leaves the room [15:50:37] yone joins the room [15:52:10] We'll have someone watching the jabber for comment; please preface them with MIC: [15:52:45] Marc Blanchet joins the room [15:53:01] masinter joins the room [15:53:18] :waves [15:53:19] ylafon joins the room [15:53:22] masinter waves too [15:53:28] Is there anyone in the chat room who does not have access to the audio? [15:53:54] are there any slides? If yes, where? [15:54:03] There are four URIs to load: [15:54:13] http://www.ietf.org/about/note-well.html [15:54:23] http://www.ietf.org/proceedings/10mar/agenda/iri.txt [15:54:31] http://trac.tools.ietf.org/wg/iri/trac/report/1 [15:54:42] https://datatracker.ietf.org/doc/draft-ietf-iri-3987bis/ [15:54:48] No slides per se. [15:55:07] After reviewing note well, we'll bash the agenda and go on from there [15:55:19] The meat of the meeting will be in working through trac on updates. [15:55:51] stpeter joins the room [15:56:26] stpeter is now known as πßå [15:56:54] stpeter joins the room [15:57:09] Peter, do you know how to change nicks in Exodus (or in XMPP)? [15:57:23] typically you can type /nick foo [15:57:33] Exodus is now known as Martin Dürst [15:57:42] ah, thanks! [15:58:06] stpeter is now known as 司馬安平 [15:58:06] http://tools.ietf.org/html/draft-masinter-url-i18n-00 was 1997 [15:58:15] first email post on the topic was 1993 though [15:58:21] Martin Dürst is now known as Martin Dürst, that feels better [15:58:26] sm joins the room [15:59:13] yone is now known as 米谷嘉朗 [16:01:29] With 0 characters, the password would be easy to guess :-) [16:02:20] tonyhansen joins the room [16:02:37] <πßå> meeting has begun [16:02:43] resnick joins the room [16:02:45] πßå is now known as stpeter [16:02:55] Ted leaves the room [16:02:57] "RELAY" [16:03:03] Chair starting the working group meeting [16:03:04] I can relay things to the mic [16:03:04] masinter wonders if we should review charter in agenda [16:03:24] chair presents "Note Well" [16:03:30] ɹəlɒɥʇ əʌɒp joins the room [16:03:42] stpeter has set the subject to: IRI WG | IETF Note Well applies [16:03:46] fujiwara joins the room [16:03:47] ylafon leaves the room [16:03:59] chair: simple agenda. [16:04:12] healthyao2000 joins the room [16:04:35] stpeter says hello to healthyao2000 [16:05:20] healthyao says good morning to stpeter [16:05:27] :) [16:05:29] chair: agenda bashes? [16:05:34] http://tools.ietf.org/wg/iri/charters [16:05:45] larry ask to see the charter [16:06:06] i have a couple of points to emphasize about charter [16:06:25] Thanks, Mark!!!!! [16:06:25] alexey: Marc has agreed to co-chair the wg [16:06:26] Marc Blanchet has agreed to co-chair [16:06:37] :-) [16:07:06] Alexey: looking for another co-chair from the w3c community [16:07:22] Thanks, Peter!!! [16:07:29] Alexey will handover wg AD to St-Peter. [16:08:43] likepeng joins the room [16:09:07] I would prefer to not jabber scribing, so I can listen better to discussions for co-chairing after. my brain is single core. [16:09:39] yu kyung joins the room [16:10:59] we are talking about finding an editor or editors for RFC 4395: "Guidelines and Registration Procedures for New URI Schemes" [16:11:04] need to update RFC4395 [16:11:10] or, specifically, rfc4395bis [16:11:30] Patrick Falstrom would be co-editor. need another. [16:11:36] KIM K joins the room [16:12:00] update rfc4395 to s/URI/IRI/ [16:12:56] w3c may start a bidi wg [16:12:58] (larry) [16:13:13] chair ask Larry to send pointer to this work. [16:13:13] s/Falstrom/Fältström/ :) [16:13:25] There's no WG on bidi in W3C, it's a mailing list, run as part of the W3C Internationlization IG. [16:13:45] ylafon joins the room [16:13:56] thanks Martin [16:14:07] yes, "group", not "working group" [16:14:07] chair: milestone may 2010 for wglc [16:14:21] Larry posted. [16:14:43] The first discussion was probably around 1994~1995 [16:14:54] Julian at the mic [16:14:55] Julian about doc editing mechanics [16:15:12] issue tracker at http://trac.tools.ietf.org/wg/iri/trac/report/1 [16:15:31] W3C bidi mailing list: http://lists.w3.org/Archives/Public/public-i18n-bidi/ [16:15:45] thanks Martin [16:16:09] Julian: in httpbis we don't make changes to the documents unless there is an issue in the tracker [16:16:54] if doc is in svn, then commits logs containing a reference to the ticket will automatically update the ticket [16:16:57] see http://trac.tools.ietf.org/wg/httpbis/trac/report/1 [16:17:05] W3C bidi mailing list currently mainly (or only) working on improving/fine-tuning bidi features in HTML [16:18:40] Anne joins the room [16:18:42] Julian: looking at ticket #29 in httpbis [16:19:47] Julian: showing a diff for ticket #29 that was attached to the ticket. [16:20:08] Joe: how to attach the diff [16:20:21] Julian: do svn diff and attach the diff to the ticket [16:21:30] Joe: add "#bugno" in the svn comment to link with bugno [16:22:06] RELAY: Is there a way to tweak the title of a ticket/issue? [16:22:07] Julian: in httpbis, we recommend to add ticket number in the subject line in mailing list [16:22:28] Julian, the tracker does send a message to the WG mailing list? [16:22:58] Code checkin is no problem for me, all my students use svn! [16:23:05] sm, no, my understanding is that mailing list messages don't get sent automatically [16:23:19] TH: critical thing is "who has authorization to open a ticket"? [16:23:47] TH: co-chairs to start with [16:23:56] MB: I would suggest also document editors [16:24:00] Larry, I saw these notifications in another WG. Tools may know more. [16:24:07] Joe Hildebrand joins the room [16:24:30] Thanks, Julian! [16:24:34] JR: more important is "who can close the ticket?" [16:24:52] JR: in httpbis, only chairs can close design tickets, but editors can close editorial tickets [16:25:05] MB: not sure I agree [16:26:06] MB: I would let editors have more control so that they don't need to wait on the chairs [16:26:09] RELAY: The current -00 draft isn't identical to RFC 3987. Some sections are completely indentical, but some others are "quite in flux" [16:26:44] Martin Dürst: I'll relay that as soon as I can get the mic [16:26:58] Larry is relaying [16:28:07] Larry is requesting confirmation from the group about the differences between 3987 and the current -00 draft [16:28:30] RELAY: I think currently the main work of the chairs should be: Put issues up for discussion [16:28:50] g.e.montenegro joins the room [16:28:51] because the main problem of the WG is that there's no discussion [16:29:10] TH: I would prefer to have discrete tickets instead of one big ticket "differences between 3987 and bis" [16:30:00] ylafon leaves the room [16:30:35] resnick leaves the room [16:30:46] discussion about "document formats", XSLT for HTML deduction, non-ASCII characters in documents? [16:30:58] resnick joins the room [16:30:59] I would be extremely interested! [16:31:18] It would make a lot of sense for this document! [16:31:33] masinter notes that RFCs can have non-normative PDF versions [16:31:57] Martin Dürst: relayed [16:32:02] cnewman joins the room [16:32:06] chair declares concensus on using the tracker. [16:32:09] Martin, I will send more details to the mailing list [16:32:14] for the wg work. [16:32:15] Peter: I can hear everybody very well! [16:32:20] julian: you could write an XSLT that renders the non-ascii glyphs into ASCII art. [16:32:23] Yves Lafon joins the room [16:32:24] Martin Dürst: ok :) [16:32:29] chair will verify concensus to the mailing list. [16:32:35] Larry now on the doc. [16:32:37] Martin Dürst: I hope that I pronounced your name well enough :) [16:33:01] Peter: Well enough. [16:33:40] LM: main change is to make an IRI a protocol element rather than something that is changed into a protocol element [16:34:02] alexey.melnikov joins the room [16:35:14] LM: For some legacy systems, have to take IRI apart and translate pieces separately to URI. [16:35:50] LM: Main special pieces: Authority (part after //), may need punycode instead of %-encoding [16:36:27] LM: handing percent-encoded authority component to DNS would give you a different result than IDNA-encoded A-label [16:36:33] RELAY: use of punycode was already allowed in RFC 3987, the exact conditions may not have been right [16:36:37] TH: two things here... [16:36:48] Martin Dürst: noted [16:39:27] RELAY: Current draft REQUIRES (MUST) to use punycode; I think this is wrong for various reasons. [16:40:28] LM: Nasty problem about form data for forms that didn't use UTF-8 [16:41:04] LM: (1) how domain names are handled (2) query parameters for systems that don't use UTF-8 for encoding form data (3) ... [16:41:20] LM: Forms used the same encoding for submission as the document was in. [16:42:18] RELAY: Special form handling is okay if there is a document enocding known, but we have to define what to do if there is no document encoding context. [16:42:21] this is only for HTML afaict [16:42:26] (last RELAY taken back) [16:42:35] Martin Dürst: ok [16:42:47] Martin Dürst: I will still relay your punycode comments, too [16:42:52] RELAY: Special handling of query part should be limited to http/https scheme. [16:43:02] Martin Dürst: I didn't want to interrupt Larry's train of thought :) [16:43:24] LM: mailto: URIs/IRIs always use UTF-8 in form part. [16:44:50] RELAY: Two different aspects: IRIs are sequences of characters, so not necessarily UTF-8. [16:45:14] UTF-8 is however extremely important for relationship between IRIs and URIs [16:46:34] http://tools.ietf.org/wg/iri/trac/ticket/5 [16:46:41] LM: A is A :) [16:46:56] resnick leaves the room [16:47:04] 司馬安平 leaves the room [16:48:20] RELAY: For Japanese (e.g.), a Japanese IRI on a bus and a Japanese IRI in a computer is about as close as for us with ASCII. So I think this distinction may look somewhat weird. I think we have to look at the actual text. Saying "this IRI is presented on the side of a bus" may work fine, but saying "there is no IRI on the bus" would be really weird! [16:48:36] is the distinction between character and codepoint useful or interesting? [16:49:16] TH at the mic [16:49:27] action item: community review query part [16:49:35] TH: key question is, who has to do the work? [16:52:08] RELAY, answer to TH: Main part of work is done by user in cooperation with basic software (e.g. keyboard drivers/input method editors from OS, or speach-to-text software). Some additional work (transcoding, normalization in some cases) may be necessary. [16:53:16] Please use the MIC! [16:53:52] LM: goal of work is to align with requirements of HTML and XML docs. [16:54:45] LM: XML did not want to refer an internet-draft. [16:55:02] LM references http://www.w3.org/TR/leiri/ [16:55:27] RELAY: Some more issues, for part of changes between RFC 3987 and draft -00. [16:56:47] http://tools.ietf.org/html/draft-ietf-iri-3987bis-00#section-7.1 [16:57:19] TH: they key text seems to be: To convert a string allowed as a LEIRI to an IRI, each character allowed in leiri-ucschar but not in ucschar must be percent-encoded using Section 3.3. [16:57:33] no objections to that text [16:57:35] reviewing 7.1, 7.2 sections [16:57:48] http://tools.ietf.org/html/draft-ietf-iri-3987bis-00#section-7.2 [16:58:06] alexey.melnikov leaves the room [16:58:07] alexey.melnikov joins the room [16:58:25] The XML Core WG has reviewed a (or more than one) previous versions of the LEIRI text, which lead to some tweaks. [16:59:15] RELAY: IRIs go through (plain-text) email as text. [16:59:58] RELAY: in HTML email, I guess it's the HTML browser/display component that the mail UA uses. [17:00:36] CN: The email people participating in the IETF are the email people who dislike HTML in email. [17:01:03] action item: someone to write the tracker ticket about email [17:02:10] Joe Hildebrand: it might be interesting to report back about what IM clients do, too [17:02:43] As late as possible, but not later :-) [17:02:44] TH: if you need to translate IRI to URI, when should it happen? early or late? [17:02:56] LM: push it off as late as possible [17:03:01] Martin Dürst: :) [17:03:24] TH: May need to embedd context with IRI [17:03:31] cnewman leaves the room [17:03:43] محسن joins the room [17:03:46] TH: such as Locale [17:04:24] LM: Can't carry locale context in IRI. Form data should be converted to %-encoding at an earlier stage. [17:04:27] LM: if you convert form data to percent-encoded early on, you can push off everything else until much later [17:04:33] :) [17:04:41] action item: open a ticket to review this issue [17:05:02] Some browsers (in particular Opera) never display non-ASCII in query part, for that reason. [17:05:24] alexey.melnikov leaves the room: Logged out [17:05:30] alexey.melnikov joins the room [17:06:17] Some related stuff about actual implementations: https://bugzilla.mozilla.org/show_bug.cgi?id=552273 [17:06:17] KIM K leaves the room [17:06:39] LM: usually not persistent protocol elements [17:07:44] RELAY: We should advise browsers to display form data for http/https always (for the forseeable future, unless we have UTF-8) %-encoded. [17:07:46] TH: not sure. example: url shortener [17:08:38] #16 may be very easy, or a big rathole [17:08:47] http://tools.ietf.org/wg/iri/trac/ticket/16 [17:09:34] g.e.montenegro leaves the room [17:09:57] RELAY: one big question is what browsers will do, whether they will update to IDNA2008 soon, or keep close to IDNA2003/TR 46 for a longer time. Input from browser people is really important here. [17:10:07] Pete: I have not gone through all the references, but I don't see what else to do [17:11:04] Anne leaves the room [17:11:06] sideshowbarker joins the room [17:11:24] RELAY: It's easy to say that for schemes that use domain names, the domain names SHOULD be Ulabels (or whatever the right term for something clean stuff that doesn't need any mapping anymore). We should definitely say that. [17:11:38] aha, the co-chair-to-be enters the jabber room [17:11:50] Julian: :) [17:12:20] RELAY: The difficult part is how much/what mapping to suggest/recommend so that we are reasonably in accordance with curent and forward-going practice as well as IDNA2008. [17:14:37] LM: this requires review and discussion. why don't we move to another one. [17:14:55] I'm rather sure IDNA2008 doesn't say Ulabel has to be UTF-8. But maybe it doesn't cover things in e.g. Shift_JIS (which are still IRIs). [17:15:29] Ok, so I was right about Ulabel. [17:16:14] Chair: asking Patrick and Pete to look at the IDNA stuff.. [17:17:06] g.e.montenegro joins the room [17:17:33] PR and PF agreed. [17:18:10] http://tools.ietf.org/wg/iri/trac/ticket/15 [17:18:50] resnick joins the room [17:19:00] I'm confused ... IIRC we had many corrections in Hiroshima to the effect that "you can't talk about Unicode code points, you always need to talk about a code point that is encoded with *some* encoding" -- for example, Unicode code points written on a piece of paper, Unicode code points encoded via UTF-32, etc. etc. etc. [17:19:07] LM suggesting separating in multiple documents: core spec, best current practices [17:19:32] RELAY: Lots of text about IRI comparison is copy-paste from RFC 3986 (URI). It's not just a reference because there are minor additions here and there. [17:20:13] RELAY: And we need editors! [17:20:25] Martin Dürst: yes, I made the point about editors [17:20:46] @psa: You can talk about code points, but they don't appear in anything that occurs on the wire. [17:21:59] Peter: You can always *talk* about Unicode code points, but if you want to catch them, you need some kind of 'encoding'. [17:22:07] TH: charter includes BCP but no specific milestones [17:22:18] LM: reading the charter [17:22:29] @resnick: so if an IRI falls in the forest and there's no one to put it on the wire, we can think of it as a series of code points? [17:22:35] The separate document: The rathole document :-(. [17:23:38] The Code Point/encoded set of bytes issue is analogous to the XML InfoSet/<> encoded string from. [17:23:44] The Platonic form of IRIs have codepoints. [17:23:49] yes. [17:24:00] in XML, the Platonic form is the InfoSet. [17:24:08] TH: There's no single exact algorithm, there's a ladder of comparisons.depending on what you are and what you need. [17:24:30] @resnick: got it (it's always easier for me to understand things in philosophical terminology) [17:24:44] issue #18: rewrite security considerations [17:24:54] http://trac.tools.ietf.org/wg/iri/trac/ticket/18 [17:25:03] Sinn und bedeutung [17:25:16] resnick: :) [17:25:22] Pete: Sinn und *B*edeutung [17:25:26] if IRIs are *defined* as their platonic ideal, then we need an ideal form for the domain name portions. [17:25:52] Martin Dürst: natürlich :) [17:25:57] resnick hangs head in shame [17:26:23] is it too late to recharter IDNAbis, to add a doc that does that? or do we need to do it here? [17:26:45] Joe Hildebrand: IDNAbis has been closed :) [17:26:56] sigh. IDNAbisbis. [17:27:58] here a bis, there a bis, everywhere a bis bis [17:28:51] Suggestion: we define a new thing called an D-label, which, when encoded using a standard Unicode Encoding Form (such as UTF-8), is a valid U-label. [17:29:17] (D is for Decoded) [17:30:35] RELAY: To Japanese, Japanes IRIs aren't enormously exacerbatedly more difficult than US-ASCII URIs. [17:31:24] Chris Newman said: auto-fill is keying on domain names, this might be a good thing to mention in the Security Considerations [17:32:36] MB: Guess most spoofing problems are related to domain name part, and domain name spoofing is discussed in the IDNA specs. [17:32:36] cnewman joins the room [17:33:16] Auto-fill of passwords is keyed of domain name. Other auto-fill may not be. [17:35:02] PR: don't close the ticket until we have agreed how to do with idna [17:35:12] Chair: yes. [17:39:47] ticket #13 [17:39:51] http://trac.tools.ietf.org/wg/iri/trac/ticket/13 [17:40:02] LM: mainly editorial [17:42:24] RELAY: On the Net, failure is always an option. [17:43:53] Martin Dürst: I think that was only a joke [17:44:09] Just tried it in Opera. URI is opera:illegal-url-54 [17:44:22] browser page says: [17:44:23] Invalid URL The URL http://xn--ntohunthpnth.xn--ht3nthpua3tcuh.xn--ntohunethunoeth.com contains characters that are not valid in the location they are found. The reason for their presence may be a mistyped URL, but the URL may also be an attempt to trick you into visiting a Web site which you might mistake for a site you trust. [17:45:44] IDNA2003 requires case folding, IDNA2008 assumes somebody else does it. [17:46:08] s/requires/implements/ [17:47:38] RELAY: Why, if something that looks like punycode but can't be converted to Unicode, why don't we just leave that as punycode? [17:49:36] PAF: once again, the question is "where is the protocol line?" [17:50:03] (ietf is very good in _not_ doing layer violation) [17:50:23] TH: Everything like a domain name that's not in the authority section will be treated just as payload, it's somebody else's problem. [17:50:58] TH: For the authority component, we have to say what to do with things that look the same (e.g. NFC,...). [17:51:31] TH: There may be other authorities (than DNS) which may want to use Unicode. [17:52:09] it's all a slough of despond [17:52:32] LM: Hostnames in mailto: URIs/IRIs are not authorities syntactically. [17:53:32] PAF: Use of Unicode in all kinds of parts of IRIs is an indication that we need to be more careful to coordinate things such as casefolding,... in the IETF [17:54:33] PAF: IRI includes IETF protocols, each of them may do differently. newprep showed it. [17:54:46] LM: goal is to have no specific parsing for each scheme [17:54:54] LM: Goal is to have no scheme-independent rules. [17:55:20] MD: *in*dependent? [17:56:06] RELAY: Goal is good, but may not be realistic in all cases. It doesn't make sense for some schemes to use punycode for authority, and so on. Too stong a restriction on new schemes. [17:57:48] Chair asks LM to raise the issue on the mailing list. [17:59:40] LM looks all registered schemes and can not find one would have a problem [17:59:51] TH: If only the list of registered URI schemes were exhaustive [17:59:59] LM partly because most do not allow % in auth section [18:00:23] LM: Until recently, URIs didn't allow %-encoding in authority part. [18:00:52] What issue number now? [18:01:08] 9 [18:01:23] KIM K joins the room [18:01:34] http://trac.tools.ietf.org/wg/iri/trac/ticket/9 [18:02:17] This would make HTML5 a non-formal/informal document :-? [18:02:22] http://tools.ietf.org/wg/iri/trac/ticket/1 [18:02:33] Martin, one could read it that way [18:04:04] Anne joins the room [18:04:14] RELAY: My understanding from face-to-face discussion with Hixie was that we wouldn't need to use exactly the terms they use; we would just need to make sure we had defined the pieces they needed in HTML5, and would be able to make the reference. [18:04:29] cnewman leaves the room [18:05:28] FWIW, it's not just about HTML5; there's also CSS, XMLHttpRequest, etc. [18:05:51] Hello Anne! [18:05:55] (Though for those the URL character encoding is always UTF-8.) [18:06:06] hi Martin [18:06:14] Are you on-site? [18:06:33] no in Utrecht [18:06:38] CSS can be in various encodings with @charset, no? [18:06:40] I was here before, just went to fetch some food [18:06:52] yeah, but it doesn't affect the query parameter [18:07:34] So you mean if I get a CSS document by an URI with a query parameter, the query parameter has to be UTF-8? [18:07:47] it would affect something else in the current IRI spec but hopefully we remove that (something about if the original encoding is not Unicode you need to normalize) [18:08:14] http://lists.w3.org/Archives/Public/public-html/2010Feb/0882.html [18:08:25] Or an IRI in a CSS document with a query parameter is always interpreted via UTF-8, even for the query part, even if the CSS document itself isn't in UTF-8 (or UTF-16,...)? [18:09:06] it means that if you have e.g. an ISO-8859-1 encoded CSS file with a URL that has an e with an accent that would be encoded per UTF-8 [18:09:28] same for XHR [18:09:32] even if the e-accent is in the query part? [18:09:37] yes [18:10:13] In my view, that's good news. That's the right way to go. But we have to make sure the IRI spec gets this right. [18:10:17] it's limited to HTML and location.href [18:10:33] we're scrolling around in http://lists.w3.org/Archives/Public/public-html/2010Feb/0882.html [18:10:35] well, the CSS spec needs to do it right [18:11:01] Ted: It's not Larry who wants, it's HTML5 that needs. [18:11:10] by saying the URL character encoding is always UTF-8 (that should probably be the default in the IRI spec so nobody has to specify that unless it is different) [18:11:22] Ted asks about: Definition: a valid URL is a string that matches the production of "iri-reference" in[draft-ietf-iri-3987bis]. [18:13:26] Marc Blanchet leaves the room [18:13:33] (I think I mentioned this long ago on public-iri including some links...) [18:17:03] Marc Blanchet joins the room [18:17:09] It's am 3:16 here, so the energy here is also completely depleted. [18:17:28] g.e.montenegro leaves the room [18:17:30] Martin Dürst, otsukare [18:18:29] Larry: Preferably, but a bit more background info into the tracker. Just the raw diffs, without diff context, and the very short statements may be very dry for other people to get into. [18:19:39] W3C may also be available to provide some teleconferencing. I don't know if WebEx is more than just teleconfenencing. [18:19:47] resnick leaves the room [18:20:06] KIM K leaves the room [18:20:08] We are usually scarier :-) [18:20:16] stpeter leaves the room: Logged out [18:20:17] sideshowbarker leaves the room [18:20:23] Thanks for everybody relaying! [18:20:24] likepeng leaves the room [18:20:29] sm leaves the room [18:20:33] Good bye! [18:20:44] bye [18:20:59] Martin Dürst leaves the room [18:21:05] Anne leaves the room [18:21:11] Julian leaves the room [18:21:21] 米谷嘉朗 leaves the room [18:21:30] fujiwara leaves the room [18:21:49] Joe Hildebrand leaves the room: Disconnected. [18:23:27] Marc Blanchet leaves the room [18:24:08] alexey.melnikov leaves the room [18:24:45] Yves Lafon leaves the room [18:26:08] masinter leaves the room [18:30:08] healthyao2000 leaves the room [18:31:26] healthyao2000 joins the room [18:36:41] محسن leaves the room [18:37:52] yu kyung leaves the room [18:42:09] healthyao2000 leaves the room [18:49:09] ɹəlɒɥʇ əʌɒp leaves the room [19:53:01] stpeter joins the room [19:55:02] stpeter leaves the room [20:07:00] Joe Hildebrand joins the room [20:07:49] Joe Hildebrand leaves the room [20:26:36] healthyao2000 joins the room [20:27:04] healthyao2000 leaves the room