| 07:48 | <hsivonen> | hmm. www.w3.org is really slow today |
| 07:50 | <Lachy> | hsivonen, that's probably because they apparently launched the new design for their homepage today, so everyone is looking at it |
| 07:53 | <hsivonen> | the new design looks quite nice, but I suspect they aren't respecting my font size pref |
| 08:07 | hsivonen | wonders if anyone has data on whether it's feasible to get rid of the SVG load event given existing content |
| 08:07 | <hsivonen> | of course there's no SVG-in-text/html existing content |
| 08:07 | <hsivonen> | but it seems bad to zap it only on the text/html side |
| 08:18 | <hsivonen> | reading http://www.xmltoday.org/content/coming-html-5-train-wreck did "the biggest names in XML - from Uche Ogbuji and Micah Dubinko to Liam Quin" ever send their proposal to the HTML WG or the WHATWG? |
| 08:18 | <hsivonen> | the proposal that was on xml-dev didn't fall "on completely deaf ears", since I replied |
| 08:19 | <hsivonen> | (assuming, of course, that I'm part of the HTML5 community) |
| 08:19 | <hsivonen> | and why might it be appropriate to put stuff like "the rich white (primarily American) boys that make up the committee" in the article? |
| 08:21 | <hsivonen> | that is, why would it be appropriate to refer to the wealth, color, nationality, age or gender of the people who make up "the committee"? |
| 08:21 | <Hixie> | primarily american? |
| 08:21 | <Hixie> | are _any_ of us american? :-) |
| 08:22 | <annevk2> | if anything we're biased towards Europe I think |
| 08:24 | <annevk2> | hsivonen, would be great if we could zap the load event imo |
| 08:24 | <annevk2> | hsivonen, I don't think it makes a lot of sense |
| 08:24 | <heycam> | ok i'm going on my few-month break now |
| 08:25 | <heycam> | i'll see you all on irc next year some time! :) |
| 08:25 | <annevk2> | heycam, good luck! |
| 08:25 | <hsivonen> | heycam: see you! |
| 08:25 | <heycam> | erik will be mailing in the svgwg's comments on the svg-in-text/html stuff some time soon |
| 08:25 | <heycam> | annevk2, thanks! |
| 08:25 | <heycam> | later |
| 08:34 | Hixie | comments on the article |
| 08:34 | <Hixie> | i like where he refers to a "recent WHATWG meeting" |
| 08:35 | <Hixie> | are you guys having meetings without telling me? |
| 08:35 | <annevk2> | you don't know about #secrettreehouse? |
| 08:35 | <annevk2> | I guess someone has been masquerading as Hixie then... |
| 08:35 | <Hixie> | hah |
| 08:36 | <hsivonen> | Hixie: I thought the implication that it's OK to refer to wealth, color, nationality, age and gender was so inappropriate that I figured I wouldn't start correcting which parts were factually incorrect and to what degree |
| 08:36 | <Hixie> | oh i didn't correct him |
| 08:37 | <Hixie> | i just said that given where we started (the web today), that if our endpoint is a mere trainwreck, i'll be happy! |
| 08:38 | <hsivonen> | I don't understand why people are still concerned about processing HTML using an XML parser |
| 08:38 | <hsivonen> | if one considers the entire stack, putting an HTML parser in there is a small part |
| 08:38 | <Hixie> | because for years that was put forward as the panacea |
| 08:39 | <Hixie> | and people don't like having to consider changing their world view |
| 08:39 | <annevk2> | I wonder how often he predicted something that's accurate |
| 08:40 | <Lachy> | wtf? Whoever wrote that article seems absolutely clueless about boolean (or, as he calls them, "binary") attributes, and how to use them in XML, and yet he's an XML proponent?! |
| 08:40 | <annevk2> | there was a site to log stuff like this so we can verify in the future if the person was indeed right or just sharing bullshit |
| 08:41 | <Hixie> | hsivonen: just consider the way organised religion (the poster child for sticking to a position beyond it's sell-by date) handled change in the paste, e.g. galileo |
| 08:41 | <Hixie> | in the past |
| 08:41 | <Lachy> | "What has emerged as a result of this "process" is a standard that seems to incorporate just those features that seem to appeal most to the rich white (primarily American) boys that make up the committee" |
| 08:41 | <Hixie> | not the paste |
| 08:41 | <Lachy> | WTF? Who among us in the WHATWG is actually an american? |
| 08:42 | <Hixie> | i'm sure _some_ of us are! |
| 08:42 | <Hixie> | isn't shelley american? |
| 08:42 | <Hixie> | and hey, i consider that quote better than what most people complain about -- usually they say the spec says just what _I_ want |
| 08:42 | <Hixie> | at least this time he suggested you all were to blame too! |
| 08:42 | <Lachy> | yeah, but I can't think of who. At least, none of us who would be considered to be in the "cabal" are American |
| 08:42 | <Lachy> | oh, mjs maybe |
| 08:42 | <annevk2> | no dude |
| 08:42 | <annevk2> | mjs is Polish |
| 08:42 | <Hixie> | he's turkish or something |
| 08:42 | <Hixie> | polish |
| 08:42 | <Hixie> | right |
| 08:43 | <Hixie> | tantek is turkish |
| 08:43 | <annevk2> | Chris_Wilson on the other hand :) |
| 08:44 | <gsnedders> | I mean, there are loads of guys in the USA called Anne |
| 08:44 | <gsnedders> | Obviously we're all American. |
| 08:44 | <gsnedders> | We speak English, so obviously we're American. |
| 08:45 | <Lachy> | oh |
| 08:48 | <annevk2> | I didn't realize before, but I guess in CSSOM I also need to define that the UA is to actually fetch resources linked from Link headers with rel set to stylesheet |
| 08:49 | <Hixie> | yeah |
| 08:49 | <Hixie> | xml-s too |
| 08:52 | <annevk2> | i see |
| 08:52 | <annevk2> | good times |
| 08:53 | <Dashiva> | Oh wow |
| 08:54 | <Hixie> | you could edit xml-s and write an RFC to define Link: rel=stylesheet processing to move it out of CSSOM :-) |
| 08:54 | <annevk2> | i'll keep it simple, thanks :) |
| 08:56 | <Dashiva> | I remember we tried to find an American whatwg teamster before, but nobody was found |
| 08:58 | <Dashiva> | Haha, "The WHATWG was not formed by a standards body. It was a small coalition of AJAX developers" |
| 08:59 | <roc> | no, it's a coalition of browser developers |
| 08:59 | <roc> | no, it's Google |
| 08:59 | <annevk2> | lol, that article sounds like a joke |
| 08:59 | <annevk2> | maybe I should read it |
| 08:59 | <roc> | you really shouldn't |
| 08:59 | <roc> | you especially shouldn't read the comments |
| 09:00 | <annevk2> | you realize you make things more tempting right? anyway, I'll go back to editing CSSOM for now |
| 09:00 | <roc> | this is pretty cool |
| 09:00 | <roc> | http://labs.thecssninja.com/font_dragr/ |
| 09:02 | <Dashiva> | annevk2: Seriously, it's weird. It's like this massive conspiracy theory, but all the posts are old men. |
| 09:02 | <Dashiva> | Well, the pictures next to the posts :) |
| 09:02 | <Hixie> | don't forget, google is supporting html5 because it would have been too expensive for google to support xhtml2 in adsense! |
| 09:02 | <Dashiva> | Actually, that's amusing. That XML is the old man technology now. |
| 09:03 | <Hixie> | (that's still my favourite html5 conspiracy theory) |
| 09:03 | <Dashiva> | Hixie: No, they're supporting HTML5 because with XHTML2 the internet would automatically become semantic and nobody would need Google anymore, they'd just RDF their way to pages! |
| 09:03 | <Hixie> | anyone object to making setting src="" imply a call to load() on <video> and <audio>? |
| 09:03 | <annevk2> | Dashiva, maybe we should make a theory up around that |
| 09:03 | <Hixie> | even if the element isn't in the document or is already playing? |
| 09:03 | <Hixie> | (roc, particularly interested in your opinion on that) |
| 09:03 | <annevk2> | foolip, ^^ |
| 09:03 | <Hixie> | foolip_ too, indeed |
| 09:04 | <hsivonen> | roc: wow. that's cool. Is it all spec-based? |
| 09:04 | <roc> | hsivonen: I haven't inspected it, but I think it's claiming to be |
| 09:05 | <Dashiva> | I wonder how long until the replies appear on the post |
| 09:05 | <roc> | Hixie: it sounds reasonable to me |
| 09:05 | <roc> | Hixie: although I don't keep the resource selection algorithm in my head so I may miss something subtle |
| 09:07 | <Hixie> | k |
| 09:07 | <Hixie> | well i can always change it back |
| 09:07 | <Hixie> | but we got a couple of complaints that setting src="" while it was playing didn't do anything |
| 09:07 | <Hixie> | so this seems like it would at least give them more feedback |
| 09:07 | <gsnedders|work> | Hixie: Get the bounce for what message-id? |
| 09:07 | <Hixie> | (since it would stop the video) |
| 09:07 | <gsnedders|work> | Hixie: I got several emails from you overnight |
| 09:07 | <Hixie> | gsnedders|work: um, i dunno, i deleted it like hours ago |
| 09:08 | <gsnedders|work> | Hixie: I guessed that :P |
| 09:08 | <foolip_> | Hixie: I was just going to reply to that on public-html-comments |
| 09:09 | <foolip_> | Hixie: I think it makes sense, no objections if all the details are sorted out properly |
| 09:09 | <Hixie> | foolip_: k |
| 09:09 | <Hixie> | foolip_: i'm replying to the thread now |
| 09:09 | <Hixie> | foolip_: so hold off and see my reply and let me know if i missed something :-) |
| 09:09 | <Lachy> | "... a small coalition of AJAX developers, most of whom couldn't give a fig for XML" - maybe he forgot that the X in AJAX stands for XML, cause that just doesn't make sense. |
| 09:11 | <hsivonen> | I wonder if people who think we don't care about XML noticed all the Infoset coercion stuff and generally trying hard to make stuff work sanely in both lands |
| 09:11 | <hsivonen> | of course, then RDFa comes along and ruins it |
| 09:13 | <foolip_> | Hixie: also, can we make it explicit what happens if the src attribute is removed (for <img> too if it hasn't been done, implementations vary) |
| 09:13 | <Dashiva> | Lachy: No, it stands for JXSON |
| 09:14 | <hsivonen> | aside: how people view RDFa reveals who belongs to the bits-on-the-wire school of XML and who belongs to the Infoset school of XML. |
| 09:14 | <Dashiva> | Isn't document.all === undefined a red herring, nobody should be using document.all in new code |
| 09:15 | <Hixie> | foolip_: already explicit for <video> (nothing special happens) |
| 09:15 | <Hixie> | foolip_: i'll do it for <img> too |
| 09:15 | <annevk2> | Dashiva, the discussion is about old code |
| 09:15 | <Hixie> | looks like kurt cagle is an xforms user (based on posts to mozilla's xforms newsgroups) |
| 09:15 | <Hixie> | (in 2006) |
| 09:16 | <foolip_> | Hixie: thanks |
| 09:16 | <hsivonen> | Hixie: he is an xml-dev regular |
| 09:16 | <Hixie> | in 2005 i rejected one of his ideas for xbl2 (namely to use xslt instead) |
| 09:17 | <Dashiva> | annevk2: No, === is about people using jslint, which implies the code is being maintained |
| 09:17 | <Hixie> | OHHH! |
| 09:17 | <Hixie> | Kurt Cagle is the reason XBL2 split off from sXBL! |
| 09:17 | <hsivonen> | Hixie: URL? |
| 09:18 | <Hixie> | not really his fault, but he was the catalyst |
| 09:18 | <Hixie> | hold on |
| 09:18 | <Hixie> | http://www.w3.org/mid/41386042.20050901045702⊙wo |
| 09:18 | <jgraham> | Dashiva: It seems like the idea is that people might JSLint old code or have mixed old/new code or something |
| 09:18 | <jgraham> | I think changing === is a bad idea though |
| 09:20 | <roc> | ah, the bad old days |
| 09:20 | <jgraham> | Brendan's proposal 3b + typeof document.all == "undefined" seems like a reasonable solution |
| 09:20 | <hsivonen> | Hixie: oh. that kind of reason. Interesting. |
| 09:20 | <Dashiva> | jgraham: Well, if you update the code you'll also realize the breakage was from your update |
| 09:21 | <jgraham> | Dashiva: depends where you test after the update. |
| 09:21 | <jgraham> | If you only test in IE and Firefox you won't notice that you broke Opera + Webkit iirc |
| 09:22 | <Dashiva> | But isn't this all assuming that FF would align with the spec? |
| 09:22 | <Dashiva> | So you'd notice it in any non-IE browser |
| 09:22 | <jgraham> | I mean at the moment |
| 09:22 | <Hixie> | jgraham: there seems to be some intermittent thing whereby the call to pms for the w3c version (with annotations) doesn't work every now and then, and i have to restart my script, at which point it works fine |
| 09:23 | <jgraham> | Hixie: Hmm. It sounds like something is occasionally timing out or something |
| 09:26 | <hsivonen> | is Web content relying on timeouts being able to call document.write on document.opened documents? |
| 09:27 | hsivonen | wonders if anyone does sync XHR from an XBL constructor |
| 09:29 | <roc> | no |
| 09:29 | <annevk2> | hsivonen, Erik believes we do load pretty much like in HTML though we also dispatch SVGLoad at the moment |
| 09:29 | <annevk2> | hsivonen, so maybe SVG should be changed? |
| 09:30 | <jgraham> | roc: Which question was that a "no" to? |
| 09:30 | <roc> | that wasn't an answer, that was an exclamation |
| 09:30 | <hsivonen> | annevk2: given the data available to me (no data about existing content) I don't object to SVG changing |
| 09:31 | <annevk2> | hsivonen, maybe you should just propose that while you nuke support for the load event? |
| 09:31 | <jgraham> | roc: Ah. Sounds like a use case for an exclaimation mark then ;) |
| 09:31 | <Dashiva> | I know it's the web, but timeout-triggered document.write is mindboggling |
| 09:31 | <annevk2> | hsivonen, I've the feeling the SVG WG is ok nowadays with having more synergy with HTML |
| 09:31 | <jgraham> | timeout triggered document write sounds quite likely |
| 09:32 | <jgraham> | I mean crazy but... |
| 09:32 | <Dashiva> | jgraham: What's the use case? |
| 09:32 | <roc> | annevk2: of course, the WHATWG cabal has seized control of SVG too |
| 09:32 | <roc> | mwahahahaha |
| 09:32 | <Hixie> | we have? |
| 09:32 | <Hixie> | sweet |
| 09:32 | <roc> | yeah |
| 09:32 | <Hixie> | i must have missed that memo |
| 09:32 | <roc> | well, I have |
| 09:32 | <Hixie> | who's the editor! |
| 09:32 | <Hixie> | there are so many changes we can get through now! |
| 09:33 | <Dashiva> | You're the editor |
| 09:33 | <Philip`> | I think we should remove <circle/> |
| 09:33 | <Dashiva> | Or did you mean who the puppet is? |
| 09:33 | <Philip`> | I don't like circles |
| 09:33 | <Philip`> | They're too round |
| 09:33 | <Hixie> | <polygon> vs <polyline> would be my first choice amongst the shapes |
| 09:33 | <hsivonen> | roc: nice |
| 09:34 | <Dashiva> | Replace svg with <embed image.vml> |
| 09:35 | <annevk2> | roc, I'm sitting next to the co-chair at the moment of the SVG WG :) |
| 09:35 | <Dashiva> | chaals? |
| 09:35 | <roc> | Erik |
| 09:36 | <roc> | I presume |
| 09:36 | <annevk2> | yeah |
| 09:37 | <roc> | WHATWG operative sits next to SVG WG co-chair, what more proof is needed? |
| 09:37 | <Dashiva> | Since we do we need proof anyhow? |
| 09:39 | <hsivonen> | what creates a nested event loop other than sync XHR, alert(), confirm(), prompt(), print() and showModalDialog()? |
| 09:40 | <Hixie> | </script> |
| 09:40 | <Hixie> | in certain caises |
| 09:40 | <annevk2> | what do you mean with "nested"? |
| 09:40 | <Hixie> | cases, too |
| 09:40 | <annevk2> | that sounds Gecko-specific |
| 09:40 | <hsivonen> | right, </script>, too |
| 09:41 | <Hixie> | he means running a task while another is running |
| 09:41 | <hsivonen> | annevk2: I thought WebKit and Trident had nested event loops, too |
| 09:41 | <Hixie> | "spin the event loop", in html5 terms, i think |
| 09:41 | <hsivonen> | annevk2: maybe I'm totally wrong about that, though |
| 09:41 | <annevk2> | I suppose they might, but in HTML5 terms everything is blocked, no? |
| 09:42 | <annevk2> | at least for sync XHR the idea is that no events are dispatched etc. |
| 09:43 | <hsivonen> | the case I'm interested in is something dequeueing a JS timeout from an odd place |
| 09:43 | <hsivonen> | when the document has been document.opened() |
| 09:45 | <hsivonen> | After all, if one never document.close()s the document, it should, in principle, be safe to call document.write() from a timeout or interval |
| 09:46 | <annevk2> | in theory a timeout would always dispatch after the syncxhr or after alert() etc. afaik |
| 09:51 | <jgraham> | hsivonen: Is there a fundamental difference between a timeout and some other event here |
| 09:52 | <hsivonen> | jgraham: no fundamental difference between a timeout and event that are dispatched through the event loop |
| 09:52 | <hsivonen> | *events |
| 09:53 | <hsivonen> | do people document.write from arbitrary async event handlers? |
| 09:53 | <jgraham> | It isn't hard to imagine someone doing var doc = document.open(); document.onsomeevent = function() {doc.write("foo")} |
| 09:56 | <hsivonen> | that would be very sad |
| 09:56 | <jgraham> | Why? I mean it doesn't obviously look worse than document.write in general |
| 09:56 | <hsivonen> | it would be feasible to make document.write not tokenize synchronously in crazy cases, if crazy cases must be supported |
| 09:57 | <hsivonen> | jgraham: for the network stream case: I need the speculative parsing thread to know where potential document.write()-caused discontinuities can be |
| 09:58 | <hsivonen> | jgraham: for the document.open case: I don't want to be able to batch notifications (that notify layout that the DOM has changed) |
| 09:59 | <hsivonen> | jgraham: it would be bad to have potential discontinuity points all over the place |
| 10:03 | <hsivonen> | doh. I *do* want to be able to batch notifications |
| 10:04 | hsivonen | sucks at cleaning up negations after editing a sentence. |
| 10:29 | <Hixie> | roc, foolip_: i'm looking at the 'load' thread... any opinion on whether I should just make the progress events use Event instead of ProgressEvent so that there's no 'total' attribute to worry about? |
| 10:31 | <roc> | I'd like that |
| 10:31 | <roc> | we won't be able to implement that immediately though |
| 10:31 | <roc> | and I don't know how much content would break :-( |
| 10:32 | <Hixie> | i've dropped NETWORK_LOADED, which had value 3; should I renumber NETWORK_NO_SOURCE from 4 to 3 or is that not worth it? |
| 10:32 | <roc> | (currently our own controls, and I think other people's controls, use progress 'loaded' and 'total' when they should be using 'buffered', since we don't support 'buffered' yet) |
| 10:32 | <roc> | implementing 'buffered' isn't hard, but there's some stuff we're in the middle of changing and we wanted to implement 'buffered' after that was done |
| 10:32 | <hsivonen> | Hixie: renumbering symbols that have shipped seems scary. those have shipped, right? |
| 10:33 | <Hixie> | i'm fine not renumbering if people think that's wiser |
| 10:33 | <roc> | hsivonen: renumbering symbols seems less scary than removing one |
| 10:33 | <hsivonen> | roc: OK |
| 10:33 | <Hixie> | (i'm not familiar with how widely this is supported) |
| 10:33 | <roc> | I think everyone supports these symbols |
| 10:34 | <roc> | Safari and Gecko will set NETWORK_LOADED in some situations, apparently Chrome never does except for files |
| 10:35 | <Hixie> | should i renumber NETWORK_NO_SOURCE though? |
| 10:35 | <roc> | I think it's probably OK to renumber NETWORK_NO_SOURCE to 3 |
| 10:35 | <Hixie> | k |
| 10:35 | <roc> | and drop NETWORK_LOADED |
| 10:35 | <roc> | we'll see how it goes |
| 10:36 | <roc> | if there's a compat issue we might want to re-add NETWORK_LOADED with value 2 |
| 10:36 | <roc> | I dunno |
| 10:36 | <roc> | hopefully not |
| 10:36 | <Hixie> | i assume you mean value 3 |
| 10:36 | <roc> | no |
| 10:37 | <Hixie> | oh, i see what you mean |
| 10:37 | <roc> | hopefully we can just bludgeon on through |
| 10:39 | <Hixie> | hm, making these not progress events will be more invasive than i expected. let's do that as a second patch in a second. |
| 10:39 | <Hixie> | load event first. |
| 10:39 | <Hixie> | actually if we readd LOADED it should be value 1, i think |
| 10:39 | <Hixie> | (IDLO) |
| 10:39 | <Hixie> | (IDLE, even) |
| 10:44 | <roc> | perhaps so |
| 10:54 | <Hixie> | is zcorpan anywhere nearby? |
| 11:03 | <jgraham> | Hixie: I guess he might be responsive to email |
| 11:14 | <foolip_> | Hixie: I don't have a strong opinion on using progress events or not. They're not particularly useful, though. |
| 11:16 | <foolip_> | not exposing the size in bytes at all will save us an origin-check, I guess |
| 11:17 | <foolip_> | zcorpan is probably en route to the office |
| 11:18 | <Hixie> | i made them simple events |
| 11:18 | <Hixie> | and i replied to zcorpan by e-mail |
| 11:18 | <Hixie> | anyone here have an opinion on whether to remove the character encoding support for scripts in Workers? |
| 11:18 | <Hixie> | and just always use UTF-8? |
| 11:19 | <Lachy> | sure, if it's not too late for browsers to drop support for other encodings in workers, I would |
| 11:21 | <Hixie> | anyone remember why i added support in the first place? |
| 11:26 | <roc> | I thought the reason was that people might want to import existing scripts into a worker |
| 11:27 | <hsivonen> | gsnedders|work: did you already have a demo that shows if document.close() tokenizes? |
| 11:27 | <gsnedders|work> | uh, somewhere |
| 11:31 | <gsnedders|work> | hsivonen: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/272 if that's what you mean |
| 11:31 | <Hixie> | roc: seems likely... is that a good enough reason? |
| 11:31 | <roc> | I don't think so myself |
| 11:32 | <gsnedders|work> | hsivonen: Or for the Gecko bug http://software.hixie.ch/utilities/js/live-dom-viewer/saved/273 |
| 11:37 | <annevk2> | it would make sense to have text/worker or some such for worker scripts then |
| 11:38 | <annevk2> | at least to justify the different encoding rules, it can be ignored in practice I suppose |
| 11:39 | <Hixie> | the mime type is ignored anyway... |
| 11:42 | <hsivonen> | gsnedders|work: thanks |
| 11:47 | <hsivonen> | gsnedders|work: I think that test case doesn't test the right thing |
| 11:47 | <gsnedders|work> | hsivonen: I'm not sure what your meaning |
| 11:47 | <gsnedders|work> | (that you want to test) |
| 11:49 | <hsivonen> | If a document.write() doesn't fully tokenize its argument (due to nested script), I want to see if a subsequent .close() tokenizes what was left in the stream |
| 11:49 | <gsnedders|work> | ah, no, I don't have tests for that |
| 11:49 | <hsivonen> | ok. thanks |
| 11:49 | <hsivonen> | I'll write one |
| 12:05 | <hsivonen> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/276 |
| 12:05 | <hsivonen> | fun results |
| 12:06 | <hsivonen> | I see different output in Firefox trunk with old parser, Firefox trunk with new parser, Safari 4 and Opera 10 |
| 12:06 | hsivonen | starts a VM for IE8 |
| 12:14 | <hsivonen> | grr. I wish IE told me what property or method is missing when I try to invoke something that doesn't exist in IE |
| 12:19 | <hsivonen> | Completely puzzled by IE's behavior |
| 12:19 | <hsivonen> | again |
| 12:20 | <hsivonen> | stringifying an HTMLCollection says it has property length |
| 12:20 | <hsivonen> | but accessing it with .length kills the script |
| 12:23 | <zcorpan_> | Hixie: did you ping me? |
| 12:25 | <hsivonen> | now I'm puzzled by the behavior of IE, Safari and Opera |
| 12:25 | <hsivonen> | they all seem to tokenize to completion on .close() even if there are external scripts in the way |
| 12:27 | <hsivonen> | whoa. whoa. do IE and Safari disable script execution when tokenizing on .close()? |
| 12:28 | <hsivonen> | and Opera |
| 12:28 | <hsivonen> | what makes the src=script scripts not run here: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/278 |
| 12:29 | <hsivonen> | hmm. maybe I'm testing wrong |
| 12:29 | <hsivonen> | this is hard |
| 12:30 | <hsivonen> | woohoo! Safari does what I thought was sane after all |
| 12:33 | <hsivonen> | Opera does something harder |
| 12:33 | <Hixie> | zcorpan_: sent mail replies instead |
| 12:34 | <zcorpan_> | ok |
| 12:35 | <hsivonen> | I've been working on the HTML5 parser for so long that I independently arrive at the same solutions as whoever wrote the parser in WebKit... |
| 12:35 | <hsivonen> | Hixie: I think the document.close() definition is wrong. |
| 12:35 | <hsivonen> | I tentatively suggest changing it to what WebKit does |
| 12:36 | <hsivonen> | which may be what IE does, too |
| 12:36 | <hsivonen> | what opera does is too hard |
| 12:36 | <hsivonen> | and what Gecko does is further away from the other three |
| 12:36 | <hsivonen> | what the spec says is what Gecko does, btw |
| 12:37 | <Hixie> | send mail |
| 12:37 | <annevk2> | Hixie, so elements can now participate in the microdata model? |
| 12:37 | <annevk2> | given the itemValue thingie |
| 12:37 | <Hixie> | hsivonen: (i have no idea what you're talking about :-) ) |
| 12:37 | <Hixie> | annevk2: not any more than before |
| 12:48 | <hsivonen> | Hixie: filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=7917 |
| 12:49 | <Hixie> | thanks |
| 12:51 | <Hixie> | hsivonen: any opinion on the issue of whether we should sniff for charsets inside <?xml?> stuff in text/html? |
| 12:51 | <hsivonen> | Hixie: I'd be happier if I didn't have to write code for that... |
| 12:52 | <hsivonen> | Hixie: what do Trident, WebKit and Presto do? |
| 12:52 | <Hixie> | i meant more an opinion about whether you knew of any bugs because you took that code out |
| 12:52 | <Hixie> | everyone seems to do it according to what i'm told |
| 12:52 | <hsivonen> | Hixie: I don't recall |
| 12:53 | <Hixie> | k. let me know if you stumble across any relevant stuff, i'm likely to look at this again tomorrow. |
| 12:54 | <Hixie> | sweet, i dealt with enough e-mail tonight that my naive script's prediction is back into october |
| 12:57 | <Lachy> | is that the prediciton for Last Call? |
| 12:58 | <Hixie> | yeah |
| 12:58 | <Lachy> | what's the HTML WG's status in relation to publishing a Last Call soon? |
| 12:59 | <Hixie> | i believe the status is "hahaha". |
| 12:59 | <Hixie> | it may best be described by this chart: http://damowmow.com/playground/htmlwg/chart.html |
| 12:59 | <Lachy> | hahaha :-) |
| 13:00 | <Lachy> | what date do you expect to publish a Last Call snapshot on whatwg.org? |
| 13:00 | <Hixie> | sometime this month |
| 13:00 | <Hixie> | i hope! |
| 13:01 | <Lachy> | well, I gathered that much already. I guess that means you don't have a specific date in mind. |
| 13:01 | <annevk2> | on a scale of twenty years a month is quite specific |
| 13:02 | <Lachy> | yeah, but given that it's this month, with only 16 or so days left, I thought it might start getting a bit more specific |
| 13:03 | <hsivonen> | Hixie: if the graph is "Last updated today.", why does it end in mid-September? |
| 13:03 | <Hixie> | it'll happen as soon as all three lines are at zero |
| 13:03 | <Hixie> | (well, the issues line will be at 1, since WebSQL is an XXX issue and I ain't taking that draft to LC this month) |
| 13:04 | <Hixie> | hsivonen: what browser? |
| 13:04 | <hsivonen> | Hixie: Firefox trunk-ish |
| 13:04 | <Hixie> | wfm |
| 13:08 | <Philip`> | Warning: 08 is not a legal ECMA-262 octal constant |
| 13:08 | <Philip`> | Source Code: new Date(2009, 08-1, 26-1), // Aug 26, Chris steps down, Maciej and Paul take over - http://lists.w3.org/Archives/Public/public-html/2009Aug/1293.html |
| 13:08 | <Philip`> | says my Firefox |
| 13:08 | Philip` | doesn't know if that's relevant at all |
| 13:09 | <Lachy> | hsivonen, the graph appears to end in mid-october for me |
| 13:10 | <Philip`> | What month does it start with? |
| 13:10 | <Hixie> | Philip`: that warning should be harmless |
| 13:10 | Philip` | sees it end in mid-September, with FF 3.5.something |
| 13:10 | <hsivonen> | that chart fails even more in Opera and Safari |
| 13:10 | <hsivonen> | yay for interop |
| 13:11 | <annevk2> | we don't have text support yet |
| 13:11 | <hsivonen> | but in safari, too, the last named month is September |
| 13:11 | <hsivonen> | dunno where the chart would end |
| 13:11 | <Hixie> | o_O |
| 13:11 | <hsivonen> | the chart itself isn't drawn although the axes are |
| 13:11 | <Lachy> | Hixie, If I select the "9 months" time scale on the graph, the year says "2010" under the first column. It says 2009 for other time scales. |
| 13:11 | <jgraham> | I see the same as hsivonen |
| 13:12 | <jgraham> | The first marked month is October 2008, the last is September 2009 |
| 13:12 | <Hixie> | if you're not seeing the actual chart, please upgrade to a browser from this year |
| 13:12 | <Hixie> | or month |
| 13:12 | <Hixie> | or week |
| 13:12 | <Hixie> | or however recent it has to be :-) |
| 13:12 | <Philip`> | row.date = new Date(Date.UTC(dateComponents[0], dateComponents[1]-1, dateComponents[2])); |
| 13:12 | <jgraham> | or from the future |
| 13:13 | <Philip`> | Why does that do -1? |
| 13:13 | <Hixie> | Philip`: months start from 0 for some reason |
| 13:13 | <jgraham> | Aren't javascript months number from 0-11 or something? |
| 13:13 | <Lachy> | yes, unfortunately |
| 13:13 | <jgraham> | (whereas days are numbered from 1 and called "dates") |
| 13:15 | <virtuelv_> | fwiw, getDay() returns offset from zero |
| 13:15 | <virtuelv_> | (that's day of week) |
| 13:16 | <virtuelv_> | get*Month() is zero-offset |
| 13:16 | <gsnedders|work> | Who thought JS was consistent anyway? :P |
| 13:16 | <virtuelv_> | also, browsers accept wildly differing input to new Date() |
| 13:17 | <Hixie> | charts look fine to me under firefox trunk and webkit trunk at all zoom levels |
| 13:17 | <Hixie> | so i don't know what y'all are seeing |
| 13:17 | daedb | sees only a single small green box in Opera 10.10 beta |
| 13:17 | <gsnedders|work> | (which causes site compat. bugs more or less whatever you7 do) |
| 13:18 | <Philip`> | The .csv starts with 2007-11-02, but the first label on the chart is October |
| 13:18 | <Philip`> | so it looks like the whole thing is off by one |
| 13:19 | <jgraham> | Please never mention javascript date parsing anywhere where I can hear |
| 13:19 | annevk2 | gets the urge to say something now |
| 13:20 | <gsnedders|work> | jgraham: So never in the Opera bug tracker/ |
| 13:20 | <jgraham> | gsnedders|work: For preference |
| 13:21 | <Hixie> | Philip`: says november for me |
| 13:21 | <Hixie> | maybe it's a timezone issue? |
| 13:21 | <Hixie> | i bet it's a timezone issue |
| 13:21 | <Hixie> | anyway, bed time |
| 13:21 | <Hixie> | nn |
| 13:23 | <Philip`> | Timezone sounds possible |
| 13:35 | <Lachy> | if I set my timezone to Helsinki, same as hsivonen, then it still works fine for me in Firefox trunk. So it can't simply be a timezone issue. |
| 14:55 | <annevk2> | hmm great |
| 14:55 | <annevk2> | I thought I could reverse engineer MediaList from Firefox but it's toally useless |
| 15:12 | <annevk2> | (Opera is slightly better, but not great either, fwiw) |
| 15:12 | <annevk2> | (I don't think this API is used much :) ) |
| 15:53 | <annevk2> | http://dev.w3.org/csswg/cssom/ chapter 4 and 5 are new; I've yet to start on 6 |
| 15:54 | <annevk2> | and with new I mean rewritten |
| 15:54 | <annevk2> | and with new I don't mean entirely done |
| 15:54 | <annevk2> | though it's quite close |
| 15:54 | <annevk2> | (I had to check something in, I was getting afraid of accidentally deleting characters and not noticing because the diff was not usable) |
| 16:07 | <jgraham> | hsivonen: btw it is not clear to me that a table of fallback encodings is the right solution |
| 16:09 | <jgraham> | And I am even less convinced that a table primarilly designed by Firefox localisers is a good idea |
| 16:15 | <adactio> | I just had a thought, reading back through a discussion about the <legend> element ...should the <fieldset> element be a sectioning root? |
| 16:17 | <TabAtkins> | adactio: Hmm, maybe. I know that I sometimes use headings within my <fieldset>s (to label the individual inputs). |
| 16:17 | <adactio> | TabAtkins: and would you want those headings to contribute to the overall outline of the document? |
| 16:17 | <TabAtkins> | No, I wouldn't. |
| 16:18 | <TabAtkins> | I'd like the <legend> to contribute, but not the inner headings. |
| 16:18 | <adactio> | TabAtkins: sounds like <fieldset> could be a good candidate for sectioning root, then. It "feels" kind of like <blockquote> and <td> in that respect. |
| 16:18 | <TabAtkins> | I think you're probably right. |
| 16:19 | <TabAtkins> | You wanna write it up? |
| 16:20 | <adactio> | TabAtkins: I'll fire off an email to the list. |
| 16:20 | <TabAtkins> | kk |
| 16:29 | <TabAtkins> | Man, is the multipage spec broken (still?/)again? |
| 16:30 | <adactio> | Looks that way. |
| 16:30 | <TabAtkins> | ;_; |
| 16:34 | <Philip`> | I blame Hixie again! |
| 16:34 | <TabAtkins> | I'm happy to also blame Hixie. |
| 16:35 | <Philip`> | Hmm, I can't access http://whatwg.org at all |
| 16:35 | <TabAtkins> | Yeah, I was just trying that. |
| 16:35 | <gsnedders|work> | Sounds like Hixie's fault. |
| 16:36 | <Philip`> | If you get a directory listing instead of the multipage spec, that's probably my fault (though I'll blame Hixie) |
| 16:36 | <Philip`> | but it sounds like the whole web server is down |
| 16:36 | <Philip`> | which is nothing to do with me |
| 16:37 | <TabAtkins> | I blame both you and Hixie. And Google. |
| 16:38 | <annevk2> | maybe blame DreamHost? |
| 16:39 | <Philip`> | It still responds to ping |
| 16:39 | <Philip`> | so presumably the server itself is alive, just not Apache |
| 17:09 | <zcorpan_> | hmm, specs on w3.org/tr now have a fancy style |
| 17:09 | <zcorpan_> | might be old news |
| 17:10 | <zcorpan_> | wonder why they have links instead of using media queries for the different views |
| 17:11 | <aroben> | is whatwg.org down? |
| 17:11 | <aroben> | lists.whatwg.org works |
| 17:12 | <aroben> | but I can't get to any specs |
| 17:12 | <aroben> | or the homepage |
| 17:13 | <zcorpan_> | down for me too |
| 18:28 | <a-ja> | Hixie: is http://html5.org/tools/web-apps-tracker broken? |
| 18:31 | <jgraham> | a-ja: Could be; Hixie's server is down. Dunno how much the webapps tracker accesses per request and how much is cahced though |
| 18:32 | <a-ja> | jgraham: tks....will try again later |
| 18:38 | mpt | discovers SWFObject and cries |
| 18:43 | zcorpan_ | wonders why mpt cries |
| 18:44 | <mpt> | Because it means the page doesn't contain an <embed>, and therefore Ubuntu Firefox's plug-in installer doesn't get triggered |
| 18:44 | <zcorpan_> | ah |
| 18:45 | <zcorpan_> | here you go: data:text/html,<embed src=data:application/x-shockwave-flash,> |
| 18:46 | <mpt> | Example #13829 of browsers doing X badly -> Web sites hacking around it -> browsers stymied when they improve X |
| 18:47 | <zcorpan_> | i guess you could file a bug on swfobject |
| 18:50 | <mpt> | so we could |
| 18:59 | <cardona507> | why is whatwg.org down? |
| 19:00 | <jgraham> | cardona507: Why not? |
| 19:00 | <jgraham> | (I think the theoy is that Hixie's apache is down but I guess that might be wrong) |
| 19:00 | <jgraham> | (and he is likely asleep) |
| 19:01 | <cardona507> | ok - thanks jgraham |
| 19:40 | <robertnyman> | This has probably already been discussed, but is there any estimate when http://whatwg.org/ will be up again? |
| 19:45 | <gsnedders|work> | robertnyman: Soon after Hixie gets out of bed. |
| 19:46 | <robertnyman> | Fair enough :-) A little sleep-in is always deserved... |
| 19:50 | <gsnedders|work> | More generally, for people who want to see the spec: http://dev.w3.org/html5/spec/ |
| 20:03 | <robertnyman> | Well, yes, but I have an upcoming article about HTML5 with links to some content only available at whatwg.org - so, I can't publish it till the site is available |
| 21:32 | <robertnyman> | Good to see whatwg.org up again. If interesting, I just published my introduction to HTML5 article at http://robertnyman.com/2009/10/14/an-introduction-to-html5/ |
| 22:00 | <cardona507> | very cool robertnyman |
| 23:30 | <ato> | Lachy: Hope you're alright. |
| 23:30 | <Lachy> | ato, yeah, fine. |
| 23:31 | <Lachy> | the fire was on the other side of the building |
| 23:31 | <ato> | Ah, your apartment faces Waldemar Thranesgate? |
| 23:31 | <Lachy> | yes |
| 23:31 | <ato> | Good to know. |
| 23:32 | <Lachy> | ato, do I know you? |
| 23:32 | <ato> | No. But wilhelm directed me to your Twitter-account. |
| 23:32 | <ato> | (And I live next door.) |
| 23:32 | <Lachy> | ok. |
| 23:34 | <Dashiva> | Darn |
| 23:34 | <Dashiva> | I was hoping lastweek would have an american for us |
| 23:36 | <Lachy> | is MikeSmith American? |
| 23:36 | <MikeSmith> | Lachy: I am a citizen of the world |
| 23:36 | <Lachy> | this has been bugging me all day. There must be at least one American in the cabal. |
| 23:36 | <Dashiva> | MikeSmith is too close to the w3c to count |
| 23:36 | <Dashiva> | And he's basically Japanese anyhow |
| 23:37 | <Rik`> | Lachy: he has a copyright on his name, so he must be american |
| 23:37 | <Dashiva> | Trademark |
| 23:37 | <MikeSmith> | I have a patent on the business process I used to develop the trademark for my name |
| 23:38 | <Rik`> | I think I just sprained my brain |
| 23:39 | <Lachy> | MikeSmith, I assume you also have copyright on the software that implements the process used to develop the trademark for your name too, right? |
| 23:40 | <Philip`> | Lachy: He can't tell you, it's a trade secret |
| 23:41 | <MikeSmith> | the team of lawyers I keep on retainer tells me that EOLAS is infringing on our IP, so I have a pot o' gold waiting for me |
| 23:43 | <MikeSmith> | with that, along with with my Reorg In Box product, I will have enough money to be set for life |
| 23:44 | <MikeSmith> | http://logopoeia.com/softexec/about.html#reorg |