| 00:10 | <dglazkov> | wicked |
| 00:12 | <dglazkov> | I met annevk, fantasai, hakon, jens, and arron in one day. |
| 00:19 | <Hixie> | hm, looks like something is checking in changes under me |
| 00:19 | <Hixie> | it's causing my w3c-html upload script to fail |
| 00:23 | <JonathanNeal> | hey all |
| 01:26 | <othermaciej> | Hixie: did you figure out the problem? |
| 01:27 | <othermaciej> | I see a bunch of commits from mike marked [updated by splitter] |
| 01:32 | <Hixie> | MikeSmith: your temporary workaround for the bug in anolis is affecting unrelated parts of the spec |
| 01:32 | <Hixie> | it's also affecting my checking in, but i've worked around that by just forcing the checkin each time |
| 01:32 | <Hixie> | in my script |
| 01:32 | <MikeSmith> | Hixie: OK, I'll turn it off |
| 01:33 | <Hixie> | well it's not really causing me any problems at this point |
| 01:34 | <Hixie> | just thought you should know :-) |
| 01:34 | <MikeSmith> | OK |
| 01:35 | <MikeSmith> | so I guess I'll leave it on for now, and try to see if I can refine it so that it doesn't affect any other part of the spec |
| 01:36 | <MikeSmith> | maybe I should instead have my makefile for the multipage build just drop the whole named-character-references table for now |
| 01:38 | <Hixie> | really we just need anolis fixed :-) |
| 01:40 | <MikeSmith> | strangely enough, I running the multipage build script with the workaround removed, but I'm not getting the errors I was getting before |
| 01:48 | <myakura> | Hixie: per 2.2.2 Extensibility it follows vendor-specific attributes start with an underscore, but the examples shown above all start with a dash... |
| 02:31 | <Hixie> | man, lots of feature requests in this batch of bugs |
| 03:46 | <MikeSmith> | Hixie: I tweaked my script to just have the multipage build remove the whole named-character-references table |
| 03:46 | <MikeSmith> | and also fixed it so that I think you won't need to force your commit |
| 03:47 | <MikeSmith> | you probably will need to force it the next time you commit |
| 03:47 | <MikeSmith> | but I think you won't need to after that |
| 04:29 | <Hixie> | MikeSmith: cool |
| 05:11 | <MikeSmith> | Hixie: copy-paste error in you r4941 commit |
| 05:11 | <MikeSmith> | http://html5.org/tools/web-apps-tracker?from=4940&to=4941 |
| 05:11 | <MikeSmith> | + <dt><dfn title="attr-body-margintop"><code>marginheight</code></dfn> on <code>body</code> elements</dt> |
| 05:11 | <Hixie> | i'll file a bug thanks |
| 05:12 | <MikeSmith> | hai |
| 05:15 | <rjett> | Is this the correct place to ask a question about an issue I've run in to when tinkering with Web Sockets? |
| 05:26 | <Hixie> | rjett: sure |
| 05:27 | <rjett> | I've been having trouble getting my client to connect to my server from my University, but it works everywhere else |
| 05:27 | <rjett> | Shouldn't Web Sockets get passed any firewall issues? |
| 05:42 | <Hixie> | rjett: which port? |
| 05:44 | <rjett> | for the Web Socket? I'm using 12345... |
| 05:44 | <Hixie> | can you telnet to that port from your university network? |
| 05:44 | <rjett> | I couldn't ping my server from the University, but I could access web pages from the server |
| 05:47 | <rjett> | I did not try to telnet to that port |
| 05:49 | <Hixie> | sounds like you have a proxy that's blocking outgoing connections |
| 05:49 | <Hixie> | i recommend either looking to see if you can use a SOCKS proxy, assuming your browser supports that, or using port 443 and TLS-encrypting the websocket connection |
| 05:50 | <Hixie> | shepazutoo: yt? |
| 05:53 | <rjett> | Thanks a lot. I really appreciate your help. :) |
| 05:53 | <Hixie> | np |
| 05:53 | <Hixie> | i expect this to be a common problem |
| 05:54 | <rjett> | From what I had read I was under the impression that Web Sockets would magically handle these kinds of issues |
| 05:54 | <Hixie> | sadly not |
| 06:25 | <wirepair> | quick question, that validator site is a neat idea, but aren't you worried about it getting abused? |
| 06:25 | <wirepair> | the 'address' option in particular |
| 06:27 | <Hixie> | validator site? |
| 06:27 | <wirepair> | http://html5.validator.nu |
| 06:27 | <wirepair> | i was under the impression some one here ran it |
| 06:27 | <Hixie> | abused how? |
| 06:27 | <wirepair> | as in i can make it scan any site i want |
| 06:28 | <Hixie> | how is that different than, say, validator.w3.org? |
| 06:28 | <wirepair> | it's not |
| 06:28 | <wirepair> | i could use that too ;> |
| 06:30 | <wirepair> | but if you're not worried about people using it as an open proxy and attacking other sites, then don't worry about it ;> |
| 06:36 | <Hixie> | i don't really see how they could attack other sites |
| 06:41 | <wirepair> | well, i could easily script up a tool that would point it at a site |
| 06:41 | <wirepair> | then look for common vulnerabilities, such as sql injection |
| 06:42 | <wirepair> | and since i can see the source of the responses |
| 06:42 | <wirepair> | i can just parse out the response |
| 06:43 | <wirepair> | and all of the attacks would look like they are coming from the validator, not from the attacker |
| 06:52 | <Hixie> | i expect henri has set some rate-limiting on the validator |
| 06:55 | <wirepair> | yeah that or some form of a captcha would be good |
| 06:56 | <wirepair> | but yeah i was just curious if anyone thought about it before, so i brought it up. |
| 06:59 | <Hixie> | hsivonen is the guy who runs it, so he's the one to ask |
| 07:03 | <wirepair> | will do, thanks |
| 07:55 | <boblet> | Hixie: you there? |
| 07:55 | <Hixie> | yeah |
| 07:56 | <Hixie> | notwithstanding the weird <wbr> interaction with nowrap, is it the same as ZWSP? |
| 07:56 | <Hixie> | anyone know? |
| 07:56 | <boblet> | re: Ruby, I noticed the w3 ruby spec uses annotating WWW with World Wide Web as an example. I’m wondering how using ruby like that relates to the abbr element |
| 07:56 | <Hixie> | relates how? |
| 07:56 | <boblet> | http://www.w3.org/TR/ruby/ |
| 07:57 | <Hixie> | i mean, they're similar... how do you mean? |
| 07:57 | <Hixie> | ruby renders the annotation by default and abbr doesn't, is the main difference |
| 07:57 | <boblet> | well, if you’re adding ruby to an abbreviation, then I’m guessing that you could also wrap the ruby word in abbr |
| 07:58 | <boblet> | but the abbr title text would duplicate the rt text, so maybe not necessary |
| 08:00 | <boblet> | abbr and ruby seem pretty similar, so I guess I’m wondering if there’s any difference apart from the way they display the annotation |
| 08:01 | <JonathanNeal> | hey all |
| 08:01 | <JonathanNeal> | hi Hixie |
| 08:02 | <boblet> | finally since complex ruby didn’t make it in, would using a title on abbr or span be a way to add the second term |
| 08:02 | <boblet> | hey JonathanNeal |
| 08:02 | <Hixie> | boblet: well there's lots of differences, e.g. they use different element names, and i wouldn't recommend using <abbr> for annotating bopomofo readings... dunno what to tell you |
| 08:02 | <Hixie> | hey JonathanNeal |
| 08:03 | <JonathanNeal> | hey boblet |
| 08:03 | <Hixie> | boblet: what is the complex ruby example you want to do? |
| 08:03 | <JonathanNeal> | For a while I thought no one could read my messages. Yay, I exist. |
| 08:03 | <boblet> | Hixie: ok, I’ll mock up some code and ping you for feedback later |
| 08:04 | <Hixie> | the amount of work required to add a new element to the spec is inane |
| 08:04 | <boblet> | Hixie: writing an article on it, so referring to the W3 Ruby doc example of Japanese with furigana and romaji |
| 08:04 | <Hixie> | i really need to work on this |
| 08:04 | <Hixie> | boblet: ah |
| 08:04 | <JonathanNeal> | What element are you thinking of adding? |
| 08:04 | <boblet> | JonathanNeal: it’s the nsfw element |
| 08:05 | <boblet> | hixie is trying to sneak it in while no one is looking ;-) |
| 08:05 | <Hixie> | trying to turn <wbr> from an obsolete elements into a valid element |
| 08:05 | <Hixie> | it's proving quitethe tedious task |
| 08:06 | <Hixie> | k i think i have it. |
| 08:06 | Hixie | regens |
| 08:06 | <JonathanNeal> | Shouldn't that be left to a style? |
| 08:06 | <JonathanNeal> | How is <wbr> meaningful as information? |
| 08:07 | <Hixie> | you can't really do it in style |
| 08:08 | <boblet> | Hixie: btw, what was the reason complex ruby was dropped? too complex for not enough use cases? |
| 08:09 | <Hixie> | it wasn't dropped |
| 08:09 | <Hixie> | it was just never added |
| 08:09 | <boblet> | or just deferred til later? |
| 08:09 | <Hixie> | it was not added because IE didn't do it |
| 08:09 | <Hixie> | HTML5's ruby is a reverse engineering of IE's, I didn't pay much attention to the ruby spec |
| 08:09 | <boblet> | dropped meaning it was specced in the ruby spec |
| 08:09 | <boblet> | aah |
| 08:10 | <boblet> | any thoughts of adding it later? say post-HTML5? |
| 08:10 | <Hixie> | are there real use cases for it? |
| 08:11 | <Hixie> | and can those use cases not be handled by simply nesting two rubys in each other? |
| 08:11 | <Hixie> | (i don't know the answers to these questions) |
| 08:11 | <othermaciej> | Hixie: I did check what at least some sites use <nobr> for |
| 08:12 | <othermaciej> | Hixie: the only one I recall offhand is YouTube, which uses it for the username of the person uploading a video in some screens, presumably because it is atomic and must not be broken across lines, even if it has a space |
| 08:12 | <Hixie> | JonathanNeal: the main use case is the one shown by the example in the spec: http://www.whatwg.org/specs/web-apps/current-work/#the-wbr-element |
| 08:12 | <boblet> | Hixie: I’ve got some introductory Japanese textbooks that do the kanji with kana and English. will mock up some nested ones and see how I go |
| 08:12 | <Hixie> | othermaciej: sounds like a stylistic concern. If my screen is small, I'd want it to wrap. |
| 08:12 | <boblet> | prolly not enough use cases to justify complexity |
| 08:12 | <JonathanNeal> | Hixie, checking |
| 08:13 | <othermaciej> | Hixie: yeah, I will look around and see if I find any real uses that seem more semantic |
| 08:13 | <Hixie> | boblet: yeah it's a hard call, when there are valid use cases but they are rare... it might just be something to do in a future version, or it might be something we never bother with because it's not that common, or it might be something that there are acceptable workarounds for (like nesting rubys). I'd love to see mockups or photos of real use cases though. |
| 08:14 | <JonathanNeal> | Weeeeird |
| 08:14 | <JonathanNeal> | Niche |
| 08:14 | <Hixie> | it is indeed niche, much like <br> |
| 08:15 | <Hixie> | but since it's already implemented and widely included in tutorials already, it's almost free to support |
| 08:15 | <Hixie> | so since there's a use case, why not |
| 08:15 | <othermaciej> | Hixie: do you recall what's partial about the acceptance on <http://www.w3.org/Bugs/Public/show_bug.cgi?id=9350>; (the <wbr> bug)? |
| 08:15 | <othermaciej> | Hixie: wondering if I should check anything before I mark it CLOSED |
| 08:15 | <Hixie> | othermaciej: just the parenthetical, i think |
| 08:15 | <boblet> | Hixie: k. also thinking that for non-bopomofo use ruby is almost like a phrase version of small with pre-set styling (“side comments”) |
| 08:16 | <Hixie> | boblet: yeah. it's mostly intended for the CJK typographical ruby thing, the other uses are somewhat esoteric. |
| 08:16 | <Hixie> | othermaciej: i.e. partial acceptance of the argument, not of the change. |
| 08:17 | <boblet> | I was somewhat surprised by the WWW example |
| 08:17 | <boblet> | (in W3 Ruby) |
| 08:17 | <Hixie> | yeah me too to be honest |
| 08:17 | <Hixie> | you'll notice the html5 spec doesn't show any such examples |
| 08:17 | <Hixie> | :-) |
| 08:17 | <othermaciej> | Hixie: I would consider that fully accepted, but fine by me either way |
| 08:17 | <boblet> | heh |
| 08:18 | <boblet> | that’s what made me wonder if such examples were legit |
| 08:18 | <Hixie> | othermaciej: to be perfectly honest i marked it partially accepted because i have two buttons, one that says "accepted" and prefills the rationale as concuring with the reporter, and one that says "partially accepted" and leaves the rationale blank, and so since i wanted to write a more detailed rationale, i clicked the second one to save me the selecting the rationale and replacing it :-P |
| 08:19 | <othermaciej> | Hixie: lol |
| 08:19 | <Hixie> | boblet: yeah, i dunno. might be worth asking richard ishida or other people more expert in that area what they think. |
| 08:20 | <boblet> | will do |
| 08:35 | <JonathanNeal> | Oh man ... PeeWee Herman makes me smile. |
| 08:45 | <JonathanNeal> | MikeSmith has quit |
| 08:45 | <JonathanNeal> | There goes HTML5. |
| 08:45 | <othermaciej> | umm, just from IRC |
| 08:48 | <JonathanNeal> | There goes IRC. |
| 08:48 | <kennyluck> | He is now sleeping :) |
| 08:49 | <JonathanNeal> | There goes being awake. |
| 08:49 | <Bolkonskij> | good morning |
| 08:49 | <JonathanNeal> | There goes it being night. |
| 08:49 | <Bolkonskij> | Heh |
| 08:50 | <JonathanNeal> | Next thing you know <wbr> will be back in the spec. |
| 08:51 | <Hixie> | ok time for bed |
| 08:51 | <Hixie> | nn |
| 08:52 | <JonathanNeal> | Night Hixie! |
| 12:45 | <gsnedders> | Hixie: Well, really, we just need Python to be fixed : |
| 15:51 | <annevk> | why is http://dev.w3.org/csswg/css3-2d-transforms/#transform-property not comma-separated? |
| 16:05 | <JonathanNeal> | mornin' |
| 17:03 | <JonathanNeal> | Hi all! |
| 17:15 | <JonathanNeal> | Anyone here versed in zen html coding? |
| 17:15 | <JonathanNeal> | Where you write HTML like this: a#skip-to-main-content[href=#main-content] |
| 17:15 | <boblet> | woah, crazy. ruby-position:inline mentioned in #5 Properties Index, but not in #4.1 Ruby Positioning. Wonder how inline differs from right? http://www.w3.org/TR/css3-ruby/ |
| 17:16 | <boblet> | JonathanNeal: I saw a screencast once |
| 17:17 | <JonathanNeal> | I just don't get how you move out of a selector. |
| 17:18 | <JonathanNeal> | Say I want to make <div><span></span></div><div><p></p></div>a |
| 17:20 | <boblet> | JonathanNeal: I’m pretty sure they did multiple blocks in the screencast. I’d guess whitespace (extra returns?) |
| 17:21 | <JonathanNeal> | (parens) |
| 17:21 | <tabatkins> | What do you mean by "move out of a selector", JonathanNeal? |
| 17:21 | <JonathanNeal> | "JonathanNeal, This is #html, not #wackyhtmltemplateengines." |
| 17:25 | <JonathanNeal> | tabatkins, I meant I didn't know how to write css that didn't constantly move down the chain |
| 17:25 | <JonathanNeal> | But I realized you can control the scope with parens |
| 17:25 | <boblet> | nice |
| 17:26 | <tabatkins> | JonathanNeal: Um? You can do what with the who now? |
| 17:26 | <tabatkins> | Selectors move down the chain [or forward, with sibling selectors]. That's how they're made. |
| 17:26 | <tabatkins> | Are you trying to abuse selectors to do something crazy? |
| 17:31 | <JonathanNeal> | Yes, and I just did. |
| 17:32 | <JonathanNeal> | You've never heard of zen coding? |
| 17:32 | <JonathanNeal> | html>(head>meta[charset=UTF-8]+title)+body>#main-document>(a#skip-to-main-content[href=#main-content])+(header#banner[role=banner]>(hgroup#heading>(h1.site-title[href=@site-url@]>span)+(h2.page-title[href=@page-url@]>span))+(nav#navigation>h1+ul>li*3))+(div#content>(nav#breadcrumbs>(h1>span)+p)+div#main-content)+footer#content-info[role=content-info]>p |
| 17:32 | <JonathanNeal> | That's a complete html5 page with heading, navigation, breadcrumb, content, footer |
| 17:35 | <tabatkins> | That looks like a combination of disgusting, unreadable, and ridiculous. |
| 17:35 | <AryehGregor> | Parentheses don't work like that in CSS. That looks like a syntax error to me. |
| 17:35 | <AryehGregor> | If it's meant to be an actual selector. |
| 17:38 | <annevk> | o_O |
| 17:38 | <lazni1> | it works backward from a selector: serialize a tree matching that ext-CSS |
| 17:38 | <jgraham> | I assume it is just supposed to be some way of generating a document structure |
| 17:39 | <jgraham> | That happens to reuse selector synatx, mostly |
| 17:39 | boblet | is surprised y’all are not up with what the kids are playing with these days |
| 17:39 | <boblet> | it’s just generating html verbiage from quasi-CSS, to cut down on typing |
| 17:39 | <annevk> | boblet, like engineer barbie? |
| 17:40 | <AryehGregor> | Oh, I see. |
| 17:40 | <AryehGregor> | That seems kind of pointless. It's not that much shorter. |
| 17:40 | <AryehGregor> | Although it's true that the redundancy of closing tags in SGML-ish languages is annoying. |
| 17:41 | <boblet> | annevk: Google is my friend http://www.ni.com/news/releases/april0601.htm ;-) |
| 17:41 | <jgraham> | Reusing selector syntas to make things more readable is an... odd approach |
| 17:42 | <JonathanNeal> | annevk, exactly like engineer barbie |
| 17:42 | <jgraham> | You could of course use sexprs and cut out all the end tags |
| 17:42 | <tabatkins> | I've certainly seen that stuff before. It's just horrifying when extended beyond toy examples. |
| 17:42 | <boblet> | TextMate does most of it for me, so the only appeal of zen coding is prolly making tables |
| 17:42 | <annevk> | boblet, it's not a joke anymore I believe, but it doesn't really matter :) |
| 17:43 | <boblet> | annevk: There’s a computer engineer Barbie, but I can’t find the one with the wrench & grease on her face |
| 17:44 | <JonathanNeal> | that would be a strange computer engineer. |
| 17:44 | <jgraham> | Yeah, who keeps a wrench on their face? |
| 17:46 | <annevk> | that is not how it works? |
| 17:46 | <JonathanNeal> | No idea. |
| 17:47 | <AryehGregor> | I use screwdrivers sometimes, but not sure where you'd use a wrench in computing. |
| 17:47 | <jgraham> | In other news I think I have a fix for the spec generation bug |
| 17:47 | <boblet> | Computer Engineer Barbie; “I ran the measurements on my leg-to-torso ratio against the 95 percentile, and the results are … well…” |
| 17:48 | <jgraham> | Nah, computer engineers son't know statistics |
| 17:48 | <jgraham> | *don't |
| 17:49 | <JonathanNeal> | Computer Engineer Barbie: "I will be on vacation tomorrow." |
| 18:01 | <annevk> | is anyone able to dig up that email on default styles for form controls on the WHATWG list? |
| 18:01 | <annevk> | someone from Google wrote it I believe |
| 18:01 | <annevk> | (could've been a different list maybe, but I doubt it) |
| 18:01 | <JonathanNeal> | What do we want? Self Closing DIVS and SPANS! When do we want them? NOW! |
| 18:02 | <annevk> | I'm trying to write a few words about HTML5 forms and it'd be nice to include a pointer to that research |
| 18:03 | <JonathanNeal> | html5 forms seems to be the big thing lately, please link me to that when you're done annevk |
| 18:03 | <JonathanNeal> | I'd love to learn what it is you folks are doing to them. |
| 18:05 | <jgraham> | OK, just updated pms.net |
| 18:06 | <annevk> | e.g. stuff like http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-April/019478.html |
| 18:07 | <seutje> | cool, specs have to list every browsers behaviour? I figured it was meant to set a standard and hope browsers will addapt |
| 18:08 | <AryehGregor> | seutje, some specs try that, but they aren't particularly useful to anyone. |
| 18:09 | <AryehGregor> | HTML5 doesn't try to list every browser's behavior, but it tries to match their behavior when it's consistent. |
| 18:09 | <AryehGregor> | Or, at any rate, it doesn't require things that it doesn't think implementers will refuse to do. |
| 18:09 | <jgraham> | It tries to specify a single behaviour that, due to legacy, msut be almost the same as existing deployed behaviour from market leaders |
| 18:13 | <AryehGregor> | At least in cases where sites rely on the behavior. |
| 18:13 | <seutje> | I see |
| 18:14 | <JonathanNeal> | It would be cool if Simon's html5-elements page could load up individual pages for those elements, instead of having to #get-there |
| 18:14 | <seutje> | u have a funny way of putting it though |
| 18:14 | <seutje> | "it doesn't think.." |
| 18:14 | <jgraham> | Which is more cases than you can imagine |
| 18:14 | <seutje> | the spec became self-aware! |
| 18:15 | <jgraham> | (well it is hopefully as many cases as we can collectively imagine or there will still be cases with required but undefined behaviour) |
| 18:15 | <jgraham> | (so what I really mean is that it is more cases then most people naively imagine) |
| 18:15 | <jgraham> | *than |
| 18:17 | gsnedders | thought jgraham was going away |
| 18:18 | <gsnedders> | Yet he fixed the bug |
| 18:18 | <jgraham> | gsnedders: Maybe I am using my telekinetic powers to bug fix by remotly typing on someone else's keyboard. Maybe it is your keyboard. Have you been watching for unexpected movements? |
| 18:19 | <gsnedders> | Well, no. |
| 18:19 | <jgraham> | Tell you what, I fixed the bug so you can fix the tests :) |
| 18:22 | <gsnedders> | However, I need to go and cook now ;P |
| 18:23 | <JonathanNeal> | gsnedders, master chef |
| 18:26 | <seutje> | hmm, that reminds me, shouldn't forget to eat |
| 18:31 | <dglazkov> | Hixie: is this the latest? http://dev.w3.org/html5/spec/forms.html#the-select-element |
| 18:32 | <annevk> | should be |
| 18:32 | <dglazkov> | Hixie: I want to add a corner case for handling default-selected item in select element |
| 18:32 | <dglazkov> | annevk: http://trac.webkit.org/export/57003/trunk/LayoutTests/fast/forms/select-change-popup-to-listbox.html |
| 18:33 | <dglazkov> | it's a bit weird, but all browsers seem to agree, so probably should be spec'd |
| 18:33 | <dglazkov> | for size == 1, first item is selected by default |
| 18:33 | <dglazkov> | if no other items are selected |
| 18:34 | <dglazkov> | but then if you change the size, the default selection is preserved. |
| 18:34 | <dglazkov> | annevk: sorry, I did't try on Opera! |
| 18:34 | <dglazkov> | :) |
| 18:34 | <annevk> | no worries |
| 18:35 | <dglazkov> | I wonder how mobile browsers deal with this? |
| 18:36 | <annevk> | http://dev.w3.org/html5/spec/the-xhtml-syntax.html#the-select-element-0 |
| 18:36 | <annevk> | defines UI |
| 18:36 | <annevk> | i.e. what to render for various values of size |
| 18:38 | <annevk> | I think it is defined also that the first is selected |
| 18:38 | <annevk> | "If the multiple attribute is absent, whenever there are no option elements in the select element's list of options that have their selectedness set to true, the user agent must set the selectedness of the first option element in the list of options in tree order that is not disabled, if any, to true." |
| 18:38 | <annevk> | (that's in section 4.10 again) |
| 18:41 | <dglazkov> | annevk: great! do you think it's worth outlining the case of preserving the selectedness when the size changes? |
| 18:41 | <dglazkov> | (that's the one I am interested about) |
| 18:42 | <dglazkov> | hang on.. reading this one more time :) |
| 18:42 | <annevk> | I suppose, but that would have to be added to the rendering section; you could add a comment via http://whatwg.org/html#the-select-element-0 |
| 18:43 | <annevk> | (if you fill the thing in at the bottom that will file a bug for you) |
| 18:44 | <annevk> | (not sure how many other such liveness things there are though and whether they all need to be pointed out; in general nothing changes unless stated otherwise) |
| 18:52 | <JonathanNeal> | So what's so bad about allowing xml style self-closing on all elements? |
| 18:52 | <JonathanNeal> | Or on divs and spans? |
| 18:53 | <Philip`> | Legacy compatibility |
| 18:54 | <AryehGregor> | JonathanNeal, no existing browser parses them that way in text/html. |
| 18:54 | <AryehGregor> | And none can change, because there are pages that use the self-closing syntax but actually rely on the fact that it's not self-closing. |
| 18:55 | <tabatkins> | I've never been a fan of the term "rely" in that context, because it implies they did it on purpose, when really they just did it, then hacked around it until it sort of worked. |
| 18:56 | <Philip`> | Why do you think reliance is purposeful? |
| 18:57 | <Philip`> | I rely on lots of things that I've never even considered |
| 18:57 | <dglazkov> | annevk: actually I think the spec is perfectly fine now that I read it |
| 18:57 | <dglazkov> | (read it vs. skim through it) |
| 18:57 | <Philip`> | and I can't think of a better word than "rely" for them |
| 19:00 | <tabatkins> | Hmm, not sure what a better term would be. |
| 19:07 | <annevk> | dglazkov, cool |
| 19:17 | <JonathanNeal> | AryehGregor, so you don't do it so it's more compatible with the old html version? |
| 19:18 | <AryehGregor> | JonathanNeal, there is no "old HTML version". HTML5 is the one and only specification to fully specify the details of the text/html format that has been used for years. |
| 19:18 | <JonathanNeal> | ok |
| 19:18 | <AryehGregor> | All these implementation requirements (or variants there of) were a de facto standard for a long time, now they're becoming de jure. |
| 19:19 | <AryehGregor> | Previous HTML specs just didn't say anything about error handling by implementations. |
| 19:20 | <AryehGregor> | So <div /> was non-conforming in HTML prior to XHTML, and browsers consistently ignored the "/" (which was allowed by the HTML specs of the time). |
| 19:21 | <JonathanNeal> | So it's just a property of hypertext to not have self-closing elements? |
| 19:21 | <JonathanNeal> | or bi-closing elements? |
| 19:21 | <AryehGregor> | The closing "/" was not valid in any version of HTML prior to XHTML 1.0. |
| 19:22 | <AryehGregor> | It's an XML feature, not a text/html feature. |
| 19:22 | <AryehGregor> | It would raise validation errors if you used it, which no one did, because it did nothing anyway. |
| 19:23 | <AryehGregor> | XHTML allowed the self-closing feature, but browsers didn't implement it in text/html for compatibility reasons. |
| 19:23 | <AryehGregor> | And still don't, and presumably never will. |
| 19:24 | <Philip`> | Actually it wouldn't raise validation errors, because SGML-based HTML4 parsers (e.g. validator.w3.org) consider <br/> to be correct and equivalent to <br>> or whatever it is |
| 19:25 | <JonathanNeal> | Right, and there is meta and link too |
| 19:25 | <JonathanNeal> | and img |
| 19:25 | <Philip`> | It's never meant "self-closing" in non-X HTML |
| 19:25 | <AryehGregor> | Oh, really? Interesting. |
| 19:26 | <Philip`> | SHORTTAG and NET are relevant keywords |
| 19:27 | <Philip`> | http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2004-August/001940.html |
| 19:27 | <Philip`> | (The W3C validator does warn about these things now, though) |
| 19:30 | <JonathanNeal> | So <br> <img> are bad? |
| 19:32 | <Philip`> | No |
| 19:32 | <Philip`> | Why would they be? |
| 19:32 | <JonathanNeal> | Ah, so it's because they're NOT self-closing, right? |
| 19:32 | <Philip`> | I have no idea what you're talking about :-) |
| 19:33 | <JonathanNeal> | <img> is okay because it's not an element that opens and closes, it just is. |
| 19:33 | <Philip`> | You've always been able to write "<br>" and "<img>" in HTML, because that's how the syntax is defined |
| 19:33 | <tabatkins> | JonathanNeal: In old-style SGML, the / was a way of closing a tag without explicitly using the end tag. |
| 19:33 | <JonathanNeal> | So ... <br /> is also valid. |
| 19:33 | <JonathanNeal> | So what would be bad about <div /> then? |
| 19:33 | <tabatkins> | So you could, frex, do "<div>stuff goes here/" and be equivalent to "<div>stuff goes here</div>" |
| 19:33 | <Hixie> | not quite |
| 19:34 | <Hixie> | <div/stuff goes here/ |
| 19:34 | <Hixie> | or <br/ (since there's no end tag, no end / either) |
| 19:34 | <tabatkins> | Ah yeah, right. |
| 19:34 | <JonathanNeal> | No > ? |
| 19:34 | <tabatkins> | Nope. |
| 19:34 | <Hixie> | the first / takes places of the / and the next / takes place of the entire end tag |
| 19:34 | <Hixie> | er |
| 19:34 | <Hixie> | let me retype that |
| 19:34 | <tabatkins> | Which means that, if browsers were conforming SGML parsers, you'd see a linebreak followed by a > character in your page. |
| 19:35 | <Hixie> | the first / takes place of the > and the next / takes place of the entire end tag |
| 19:35 | <JonathanNeal> | Wow! That's not implemented that way in any browser, is it? |
| 19:35 | <AryehGregor> | No, browsers don't actually implement SGML. |
| 19:36 | <Hixie> | there's a reason html5 gives up on the sgml fiction :-) |
| 19:36 | <JonathanNeal> | So, you can do <div /> because it wouldn't be properly conforming to something no browsers actually implement? |
| 19:36 | <JonathanNeal> | *can't |
| 19:36 | <tabatkins> | No, you can't do <div /> because it looks exactly like "<div>" to all browsers. |
| 19:36 | <JonathanNeal> | Well, in HTML4. |
| 19:36 | <tabatkins> | [That is, it looks like an open tag.] |
| 19:36 | <Hixie> | here, if you want to blow your mind: http://damowmow.com/playground/mental.html http://damowmow.com/playground/not-html-yet-valid.html |
| 19:37 | <Hixie> | both of those validate in an html4 validator |
| 19:37 | <tabatkins> | Html nothing. I'm talking about browsers here, not languages. All existing browsers, current and legacy, treat it like that. |
| 19:37 | <tabatkins> | [Unless they are specifically in an XHTML mode, of course.[ |
| 19:38 | <tabatkins> | So allowing <div /> as a short form of <div></div> wouldn't be backwards compatible in any relevant browser straight away, let alone all the pages that accidentally use <div /> and would be screwed up if it actually acted like <div></div>. |
| 19:39 | annevk | wonders when the SGML interest will die off |
| 19:40 | <tabatkins> | Dammit, JonathanNeal. I didn't realize you'd dropped, and typed several lines at you. >_< |
| 19:40 | <JonathanNeal> | tabatkins, not only did i drop |
| 19:40 | <JonathanNeal> | my whole computer restarted |
| 19:40 | <tabatkins> | That's fun. |
| 19:40 | <JonathanNeal> | hixie's links really blew my mind |
| 19:40 | <JonathanNeal> | and lappy's. |
| 19:40 | <tabatkins> | Anyway, check the logs. I don't want to retype those. |
| 19:40 | <JonathanNeal> | I was like "This is a cell" and my computer was like "this fuuuuuuuuuuuuuuu" |
| 19:40 | <JonathanNeal> | Where do I find the logs? |
| 19:40 | <Philip`> | tabatkins: There's this fancy new concept known as "copy and paste" which you could use |
| 19:40 | <tabatkins> | Check the topic. |
| 19:41 | <tabatkins> | Philip`: Too fancy for me. |
| 19:41 | <JonathanNeal> | checking |
| 19:42 | <JonathanNeal> | tabatkins, aren't nav and video not backwards compatible? |
| 19:42 | <JonathanNeal> | I believe <section> causes fuglies in ff2 and ie6-7-8 |
| 19:42 | <tabatkins> | They have fallback. |
| 19:43 | <JonathanNeal> | they don't have fallback ... not in those browsers. What do you mean by "fallback"? |
| 19:43 | <tabatkins> | ff2 is luckily ignorable. |
| 19:43 | <tabatkins> | ie6-8 has the hack, which is nice. |
| 19:43 | <JonathanNeal> | ie6-7-8? |
| 19:44 | <JonathanNeal> | so it's okay because there's a hack? :-| |
| 19:44 | <tabatkins> | The hack makes it usable *now*, so... yes? Things are a lot less appealing when there's no way to use them until all browsers update. |
| 19:45 | <tabatkins> | That doesn't mean we'll avoid *all* things without fallback, but they need a better reason to exist. |
| 19:45 | <tabatkins> | Self-closing syntax has no real reason to exist besides "It's what XML does." |
| 19:45 | <JonathanNeal> | ie supports it. |
| 19:45 | <Philip`> | No it doesn't |
| 19:45 | <Philip`> | except sometimes |
| 19:46 | <Hixie> | ie's parsing behaviour is seriously messed up |
| 19:46 | <Hixie> | and very internally inconsistent |
| 19:46 | <Philip`> | like in unknown elements that are recognised due to being declared with xmlns:foo or with the createElement hack, or due to being any normal unknown element in IE9 |
| 19:46 | <Philip`> | (i.e. <section/> closes the element in IE9) |
| 19:47 | <Philip`> | (but <div/> doesn't) |
| 19:49 | <JonathanNeal> | Ah, I see. |
| 19:50 | <JonathanNeal> | I tried This is <span style="background: #0F0; padding: 10px;" /> self closing. and it failed in all browsers. |
| 19:51 | <JonathanNeal> | So it can't be implemented because it would break on all browsers. You're lucky you got any new elements in :) |
| 19:51 | <JonathanNeal> | Considering how ff used to handle these things. |
| 19:51 | <tabatkins> | If FF still handles unknown elements as badly as it used to, I suspect the new elements would have had a much more uphill battle. |
| 19:52 | <JonathanNeal> | aye |
| 20:00 | tabatkins | goes to learn himself a python for great justice. |
| 20:10 | <mr_daniel> | is it really possible that Firefox still has no support for websockets? Just downloaded the newest version 3.6.3, but the included websocket chat in Jetty fails to run |
| 20:12 | <AryehGregor> | mr_daniel, https://bugzilla.mozilla.org/show_bug.cgi?id=472529 |
| 20:13 | <AryehGregor> | Work ongoing. |
| 20:15 | <Hixie> | mr_daniel: websockets isn't really ready yet, the wg is still discussing the protocol |
| 20:20 | <annevk> | Philip`, apparently their plan was to conform to HTML5 |
| 20:20 | <annevk> | Philip`, maybe you should file bugs? |
| 20:25 | <MikeSmith> | mr_daniel: if you are comfortable building from sources, you can download the patch from that bug and apply it and build and it should work |
| 20:26 | <MikeSmith> | Hixie: it seems ready enough for experimenting with at least |
| 20:26 | <MikeSmith> | and for playing Quake2 |
| 20:44 | <Hixie> | MikeSmith: we're making non-backwards-compatible changes to the protocol, so everything will break |
| 20:46 | <MikeSmith> | Hixie: yeah, I realize |
| 20:46 | <MikeSmith> | but it's understandable that people are anxious to try it out in the mean time in spite of any instability of the protocol spec |
| 20:47 | <Hixie> | sure, i've done it myself :-) |
| 20:47 | <Hixie> | in my spare time i'm writing a web-based mud using websockets |
| 20:48 | <annevk> | mud? |
| 20:48 | <annevk> | multi-user d...? |
| 20:48 | <paul_irish> | daemon |
| 20:50 | <MikeSmith> | Hixie: you should make a mud where the character is a spec editor who needs to try to create spec for Web-platform features and get them standardized and implemented |
| 20:50 | <Hixie> | heh |
| 20:50 | <Hixie> | annevk: multi-user dungeon |
| 20:51 | <paul_irish> | oops. :x http://en.wikipedia.org/wiki/MUD |
| 20:51 | MikeSmith | thought it was "domain" |
| 20:53 | MikeSmith | reads annevk blog and finds out about https://wiki.mozilla.org/User:Mounir.lamouri/HTML5_Forms |
| 20:53 | <MikeSmith> | would be interesting to know how soon active work might get started on all the items in the red rows |
| 20:54 | <MikeSmith> | oh, I see that some of them do have bugs associated with them already |
| 20:56 | <MikeSmith> | I wonder if anybody knows/remembers how long it took Opera engineering to finish their implementation of Web Forms 2 at the time |
| 20:56 | <MikeSmith> | or how many people-hours it took |
| 21:03 | MikeSmith | wonders if recent TC39 f2f spent any time discussing https://cvs.khronos.org/svn/repos/registry/trunk/public/webgl/doc/spec/TypedArray-spec.html or other related proposals |
| 21:03 | <annevk> | one person on and off for half a year or so? |
| 21:03 | annevk | doesn't remember the details |
| 21:03 | <MikeSmith> | annevk: ok |
| 21:04 | <MikeSmith> | it seems like Webkit devs have been working on it very actively for more than 6 months now |
| 21:04 | <annevk> | btw, http://en.wikipedia.org/wiki/MUD |
| 21:05 | <annevk> | it could be one person for a year on and off |
| 21:06 | <annevk> | but not more than a year and not more than one person developing; I created tests together with someone else |
| 21:06 | <MikeSmith> | ok |
| 21:06 | <annevk> | but then our UI is not great... |
| 21:06 | <annevk> | but then WebKit does not have UI yet I believe... |
| 21:06 | <annevk> | oh well |
| 21:07 | <MikeSmith> | I guess it's hard to decide what's the appropriate UI for some of these controls |
| 21:07 | <MikeSmith> | btw, W3C Bugzilla instance was upgraded today |
| 21:08 | <tabatkins> | A decent rule, of course, is "Do whatever the most popular equivalent jQuery plugin does." |
| 21:08 | <annevk> | well, you need a UI engineer, not someone who's typically deeply involved in DOM/parsing/etc. code |
| 21:08 | <MikeSmith> | right, exactly |
| 21:10 | <estellevw> | i've been doing some testing of the various form elements in recent browser releases. webkit seems to understand from a DOM perspective new input types, but does nothing with them |
| 21:10 | <othermaciej> | WebKit's UI for many of the controls is not so great |
| 21:10 | <estellevw> | in terms of UI |
| 21:11 | <annevk> | estellevw, yeah, that is somewhat troublesome |
| 21:12 | <annevk> | estellevw, but as long as nobody ships final releases I guess it could work |
| 21:13 | <estellevw> | no, i have been testing the actual releases, not the nightly builds... so it is what has been released |
| 21:13 | <MikeSmith> | annevk: or even not just a UI engineer, but a UX expert too |
| 21:14 | <annevk> | just a little insight might have done wonders, but we were short on time too |
| 21:14 | <annevk> | and with no competition and not much real world usage we didn't move it beyond experimental |
| 21:15 | <annevk> | estellevw, i.e. in Safari 4 and Chrome 4? |
| 21:15 | <estellevw> | for example, chrome supports color, date and email, but doesn't validate |
| 21:15 | <MikeSmith> | yeah, but all told, it's still a pretty good implementation |
| 21:15 | <estellevw> | yeah, exactly |
| 21:15 | <annevk> | ugh :/ |
| 21:15 | <annevk> | MikeSmith, yeah, it's pretty cool |
| 21:15 | <MikeSmith> | Hixie: W3C bugzilla is down |
| 21:16 | <estellevw> | i've been compiling results here: http://www.standardista.com/html5/html5-web-forms (not finished ... a work in progress) |
| 21:16 | <MikeSmith> | Hixie: systems team is working on it |
| 21:16 | <miketaylr> | estellevw: this might be helpful for you: http://www.miketaylr.com/code/input-type-attr.html |
| 21:16 | <miketaylr> | also, http://www.miketaylr.com/code/html5-forms-ui-support.html |
| 21:16 | <estellevw> | MIke- i came across your site last week, and found it extremely useful |
| 21:17 | <miketaylr> | ahh, k |
| 21:17 | <estellevw> | and then started visually comparing |
| 21:17 | <miketaylr> | yeah i'm just doing feature testing |
| 21:17 | <estellevw> | so, assumed DOM support from your excellent resource, and then sandboxed each |
| 21:18 | <estellevw> | nice to virtually meet you, by the way |
| 21:18 | <miketaylr> | :) |
| 21:18 | <miketaylr> | same |
| 21:19 | <MikeSmith> | are all 192.* IP addresses reserved (or whatever it's called), or just 192.168.* ones? |
| 21:22 | <annevk> | oh hey, someone is picking up the ball I sort of dropped: http://mail.apps.ietf.org/ietf/charsets/msg01877.html |
| 21:24 | <annevk> | MikeSmith, do you know what happened to http://lists.w3.org/Archives/Public/ietf-charsets/ ? |
| 21:25 | <annevk> | it scores highest on Google for ietf-charsets but has dropped the ball seemingly after 2004 |
| 21:25 | <MikeSmith> | annevk: wfm |
| 21:25 | <MikeSmith> | oh, I see |
| 21:25 | <MikeSmith> | no idea |
| 21:25 | <MikeSmith> | it's an IETF list, just hosted at W3C |
| 21:26 | <MikeSmith> | maybe Richard or Martin Duerst would know |
| 21:26 | <annevk> | the actual list is hosted by IANA |
| 21:26 | <annevk> | well, it uses an IANA email address |
| 21:27 | <MikeSmith> | other IANA lists are hosted somewhere else |
| 21:27 | <MikeSmith> | like the language-subtags list |
| 21:27 | MikeSmith | tries to find |
| 21:27 | <MikeSmith> | http://www.alvestrand.no/mailman/listinfo/ietf-languages |
| 21:28 | <MikeSmith> | but that seems to be the same list as ietf-languages⊙io |
| 21:29 | <annevk> | W3C lists are nice |
| 21:29 | <Philip`> | MikeSmith: http://en.wikipedia.org/wiki/Private_network#Private_IPv4_address_spaces |
| 21:29 | <MikeSmith> | Philip`: ah, thanks |
| 21:30 | <annevk> | WHATWG mailing list archive links have been broken at least once (and not fixed) and now several different entry points are used |
| 21:30 | <annevk> | it's a bit of a mess |
| 21:30 | <MikeSmith> | so it is only the 192.168.* range |
| 21:31 | <Philip`> | MikeSmith: Yep |
| 21:31 | <MikeSmith> | annevk: ietf-charsets⊙io |
| 21:31 | <MikeSmith> | but that appears to lead nowhere |
| 21:33 | <annevk> | MikeSmith, the actual archive is now at http://mail.apps.ietf.org/ietf/charsets/ |
| 21:33 | <annevk> | MikeSmith, just wondering why the W3C one was not kept updated |
| 21:34 | <MikeSmith> | annevk: dunno, I'll ask RIchard |
| 21:41 | <jgraham> | Hixie: Did you try regenning the spec recently? Did the non-BMP entity problems go away? |
| 21:42 | <MikeSmith> | jgraham: fwiw, I can try it now |
| 21:43 | <Dashiva> | "Developers should know better than to use a GET request for teleportation." |
| 21:44 | <annevk> | for that td:last-child:hover effect a transform might look better |
| 21:45 | <annevk> | though might also work less well come to think of it |
| 21:45 | <annevk> | unless we get some hit testing control |
| 21:45 | <MikeSmith> | annevk: so I'm told the w3c ietf-charsets list was shut down at the beginning of 2004 due disuse and to being overwhelmed with spam |
| 21:46 | <annevk> | kk, too bad |
| 21:46 | <annevk> | with the ietf mailing list archive you get everything on a single page |
| 21:46 | <jgraham> | MikeSmith: OK |
| 21:46 | <annevk> | it 1) takes ages to load and 2) doesn't browse very well |
| 21:47 | jgraham | is futzing about with python on the server at the moment as well so it is not impossible that something will accidentially break |
| 21:47 | <jgraham> | (it shouldn't but...) |
| 21:48 | <gsnedders> | (I wouldn't trust jgraham.) |
| 21:49 | <MikeSmith> | annevk: if the list is no longer spam-inundated, I can ask for the mirroring to be turned back on |
| 21:49 | <annevk> | nah it's okay, thanks |
| 21:49 | <JonathanNeal> | hey again |
| 21:50 | <jgraham> | Dashiva: where is that from? |
| 21:50 | <Dashiva> | http://code.google.com/p/chromium/issues/detail?id=31482 |
| 21:50 | <annevk> | going to bed, hopefully I can sleep ten or so hours and defeat da jetlag |
| 21:50 | <MikeSmith> | annevk: oyasumi |
| 21:50 | annevk | only slept like two hours on the plane... |
| 21:51 | <annevk> | MikeSmith, ta! |
| 21:56 | <MikeSmith> | jgraham: this only affects the serializer behavior, right? |
| 21:56 | <gsnedders> | MikeSmith: Yes |
| 21:56 | <MikeSmith> | ok |
| 21:59 | <Hixie> | jgraham: trying now |
| 22:00 | <jgraham> | MikeSmith: You will need to wait a bit I think |
| 22:00 | <MikeSmith> | jgraham: hai |
| 22:00 | <jgraham> | The python messing changed some paths, I am just fixing now |
| 22:00 | <jgraham> | (well actually I am taking the oppertunity to upgrade some of the installed versions of things at the same time) |
| 22:05 | <JonathanNeal> | I wish there was a way to tell the w3 validator to ignore certain things, like X-UA-Compatible |
| 22:05 | <JonathanNeal> | Just because it makes me sad when I see the red tape. |
| 22:05 | <Hixie> | jgraham: getting 500s |
| 22:10 | gsnedders | blames jgraham, looks innocent himself |
| 22:13 | <jgraham> | MikeSmith, Hixie: Should be back up now |
| 22:14 | <jgraham> | JonathanNeal: That is sort of the point |
| 22:14 | <MikeSmith> | jgraham: hg pull says no updates found |
| 22:14 | <jgraham> | MikeSmith: Pull of what? |
| 22:14 | <JonathanNeal> | jgraham, but poor IE! |
| 22:14 | <MikeSmith> | um, from html5lib repo |
| 22:15 | <MikeSmith> | I thought that's where you were making the changes |
| 22:16 | <gsnedders> | He did a lot earlier |
| 22:16 | <jgraham> | MikeSmith: I made a change there earlier (rev. 4113ad9d98) but I was just futzing with the server |
| 22:16 | <jgraham> | now |
| 22:16 | <jgraham> | I needed to upgrade python and that changed the default python path |
| 22:16 | <MikeSmith> | aha |
| 22:17 | <MikeSmith> | I was confused |
| 22:17 | <gsnedders> | MikeSmith: like normal |
| 22:17 | <MikeSmith> | heh |
| 22:17 | <MikeSmith> | I had forgotten that this was a change needed to gsnedders adonais tool |
| 22:18 | <jgraham> | (the need to upgrade python was entirely unrelated to this issue) |
| 22:19 | <MikeSmith> | I guess I will need to wait til Hixie checks in again |
| 22:20 | <MikeSmith> | because I'm just working with his post-adonais generated output |
| 22:20 | <jgraham> | OK, well I am going to sleep in a moment so if it is broken it will have to stay broken for a while |
| 22:23 | <JonathanNeal> | http://achecker.ca/checker/index.php <html lang="en"> fails ... is "en" not valid ISO 639 specification for the HTML lang attribute? |
| 22:24 | <Hixie> | regenning again |
| 22:26 | <Hixie> | hrm i overloaded pms |
| 22:26 | <Hixie> | let me add more delays and try again |
| 22:29 | <Hixie> | hm, wait, it's not my fault |
| 22:29 | <Hixie> | jgraham: i'm getting xml syntax errors |
| 22:29 | <Hixie> | jgraham: when doing the specs that have the annotations |
| 22:30 | <Hixie> | jgraham: http://www.whatwg.org/specs/web-apps/current-work/.w3c-html-core/Overview.html |
| 22:30 | <jgraham> | Oh. That sounds bad |
| 22:32 | <jgraham> | Hixie: Would it be a problem to turn off the annotations until I have a chance to fix? |
| 22:34 | <Hixie> | no |
| 22:34 | <Hixie> | not for me, at least |
| 22:34 | <Hixie> | can't speak for everyone else... there's always someone who'll fine a problem with something |
| 22:35 | <Hixie> | what parameter do i need to remove to stop it annotating? |
| 22:35 | <Hixie> | just annotation= ? |
| 22:35 | <Hixie> | or annotate_w3c_issues=? |
| 22:35 | <Hixie> | both? |
| 22:36 | <Hixie> | i prefixed both with xxx-, we'll see if that helps |
| 22:38 | <Hixie> | MikeSmith: checking in |
| 22:39 | <MikeSmith> | OK |
| 22:39 | <MikeSmith> | http://lists.w3.org/Archives/Public/www-validator/2010Apr/0011.html is entertaining |
| 22:39 | <MikeSmith> | Jukka Korpela |
| 22:39 | <MikeSmith> | '"HTML 5" is codename for a sketch of a draft of a working paper. "Validating" against something as vague as that is pointlss.' |
| 22:39 | <jgraham> | Hixie: It might well work with annotations now |
| 22:40 | <MikeSmith> | dude is such a sweetheart, you just gotta love him |
| 22:40 | <Hixie> | jgraham: k trying |
| 22:40 | <Hixie> | MikeSmith: i guess he hasn't looked at the spec recently |
| 22:41 | <Hixie> | MikeSmith: or he draws WAY bigger sketches than i do |
| 22:41 | <MikeSmith> | heh |
| 22:41 | <tabatkins> | I can't imagine the draft of a working paper that HTML5 is a sketch of, let alone the working paper itself. |
| 22:41 | <Hixie> | i mean, there are lots of ways to dismiss html5, but calling it a "sketch" is somewhat silly |
| 22:41 | <MikeSmith> | he's an expert, so he's not obligated to actually read specs |
| 22:42 | Hixie | lies to his script to get it to actually regen the specs even though he didn't change anything |
| 22:42 | <tabatkins> | Hixie: irssi actually isn't that bad. I'll continue to use that on my linux boxes I guess. |
| 22:43 | <Hixie> | dude i've been telling you that for ages :-P |
| 22:43 | <tabatkins> | If "for ages" you mean "once, last week", then yes. |
| 22:44 | <Hixie> | jgraham: looks like it worked |
| 22:44 | <Hixie> | tabatkins: i live in the now, anything in the past happened ages ago :-P |
| 22:44 | <Hixie> | MikeSmith: ok, looks like the astral plane stuff is all fixed |
| 22:46 | MikeSmith | runs splitter on W3C copy |
| 22:47 | <MikeSmith> | yippee |
| 22:47 | <MikeSmith> | http://dev.w3.org/html5/spec/named-character-references.html |
| 22:47 | <MikeSmith> | back to normal |
| 22:47 | <Hixie> | now if only browsers supported absolute positioning correctly in columns |
| 22:47 | <Hixie> | so the hover effect worked in more than the first column |
| 22:48 | <JonathanNeal> | It's kinda weird how the letter disappears |
| 22:48 | <JonathanNeal> | when you hover |
| 22:49 | <JonathanNeal> | Oh ... I see ... it's just not working right in Chrome. |
| 22:49 | <MikeSmith> | people should just ignore those other columns |
| 22:49 | <tabatkins> | Wow, that's... quite a table. |
| 22:49 | <MikeSmith> | http://www.w3.org/Bugs/Public/ appears to back in action too |
| 22:50 | <tabatkins> | JonathanNeal: ?_? |
| 22:51 | <JonathanNeal> | tabatkins, what are you ?_? ing? |
| 22:51 | <tabatkins> | Your nick change. |
| 22:51 | <paul_irish> | tabatkins: http://paulirish.com/i/ec20.png |
| 22:51 | <tabatkins> | Ah, I see. |
| 22:52 | <JonathanNeal> | And http://www.imdb.com/title/tt0087182/quotes#qt0328198 |
| 22:52 | <JonathanNeal> | And it was Paul ... there was no resisting. |
| 22:53 | MikeSmith | wonders what a "mouse locking feature" is and why WebGL-based Quake2 implementation would highlight the need for one |
| 22:54 | <Philip`> | MikeSmith: You should be able to spin around by moving your mouse rightwards a lot |
| 22:54 | <Philip`> | for a longer distance than the width of your screen |
| 22:55 | <Philip`> | without the mouse leaving the web page window and stopping controlling the game |
| 22:55 | <MikeSmith> | I see |
| 22:55 | <MikeSmith> | isn't this what joysticks are for? |
| 22:56 | <AryehGregor> | Because joysticks are much better supported by the web platform, right? |
| 22:56 | <MikeSmith> | heh |
| 22:56 | <tabatkins> | Do you own a joystick that works with your computer, MikeSmith? |
| 22:56 | <AryehGregor> | (and no, no one uses joysticks for FPSes on the PC, only for flight sims or something) |
| 22:57 | <MikeSmith> | tabatkins: I don't own any joystick at all |
| 22:57 | <MikeSmith> | unless you meant the question figuratively |
| 22:57 | <Philip`> | I own two dead joysticks |
| 22:57 | <Philip`> | I think the modern term you want is "joypads" or "gamepads" or whatever |
| 22:58 | <Philip`> | which people are more likely to own |
| 22:58 | <Philip`> | but they don't exactly work on the web either |
| 22:58 | <divya> | A question, In CSS3 the display property refers to block/inline elements http://www.w3.org/TR/css3-layout/#declaring-templates-the-display-property but HTML5 does not define block/inline elements. Is there something I am missing? |
| 22:58 | <Philip`> | and they're rubbish compared to mouse+keyboard for FPSes anyway |
| 22:59 | <tabatkins> | divya: In terms of CSS, a block/inline element is one which has display:block/inline set on it in the default UA stylesheet. |
| 22:59 | <AryehGregor> | divya, look in the rendering section. |
| 22:59 | <AryehGregor> | http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#rendering |
| 23:00 | <divya> | tabatkins: so does this mean display no longer alters the "content model" of an element? |
| 23:00 | <Hixie> | a block/inline element in CSS is one which has display:block/inline as its computed value, not necessarily in the UA style sheet |
| 23:00 | <Hixie> | divya: it never did |
| 23:01 | <tabatkins> | Hixie is more correct than I was. I was referring to what elements were by default. |
| 23:02 | <Hixie> | technically even its _used_ style, not computed style, though i think with 'display' they're always the same |
| 23:02 | <Hixie> | (or does 'position' not affect the computed style?) |
| 23:03 | <Hixie> | an element with display:inline;position:absolute is block-level |
| 23:03 | <Hixie> | ok yeah for such an element the computed value is 'block' |
| 23:03 | <Hixie> | so yes, computed value |
| 23:05 | <divya> | Hixie: I was under the illusion display can alter the content model because of this sentence in the HTML4.01 spec http://www.w3.org/TR/html401/struct/global.html#h-7.5.3 "Style sheets provide the means to specify the rendering of arbitrary elements, including whether an element is rendered as block or inline. " |
| 23:05 | <tabatkins> | Rendered as a block or inline, yes. That doesn't change what kind of elements are legal to put inside of it, though. |
| 23:06 | <Hixie> | divya: yeah, that confusion is why i stopped using "block" and "inline" in html5 |
| 23:06 | <Hixie> | it confused a lot of people |
| 23:06 | <Hixie> | you're not alone, by a long shot :-) |
| 23:07 | <divya> | Hixie: thanks, that is a comfort :) |
| 23:07 | <Hixie> | :-) |
| 23:08 | <divya> | so in a sense display:inline/block are legacy CSS syntaxes with no real-world relationship with the elements? |
| 23:09 | <tabatkins> | I'm not sure what you mean by "legacy" or "real-world" there. CSS has no interaction with what's legal or illegal in HTML. |
| 23:09 | <Hixie> | that's not how i would put it |
| 23:09 | <Hixie> | i'd just say display:inline/block conrtol the rendering |
| 23:09 | <Hixie> | and that's all :-) |
| 23:09 | <divya> | ok that makes sense :) |
| 23:10 | <divya> | (I think it is hard for me to not get confused between css inline/block and html4 inline/block) |
| 23:10 | <tabatkins> | Easiest is to probably just ignore the HTML4 definitions entirely. ^_^ |
| 23:11 | <divya> | tabatkins: :) |
| 23:11 | <AryehGregor> | I bet that's why HTML5 doesn't use those terms to begin with. |
| 23:12 | <divya> | I wonder why not many people brought this up. :/ |
| 23:12 | <divya> | (when they were writing their "Top Ten CSS Tricks" posts) |
| 23:13 | <AryehGregor> | Why would they bring it up? |
| 23:14 | <divya> | AryehGregor: It is confusing and deserves clarity. More than cool tricks that can be accomplished with display: inline/block |
| 23:14 | <estellevw> | i found the distinction of block, inline and flow very helpful to understanding the semantics of elements |
| 23:14 | Hixie | has no idea what to do about character references that don't end with ; |
| 23:14 | <AryehGregor> | Hixie, are they used too much to just be dropped? |
| 23:15 | <Hixie> | yes |
| 23:15 | <Hixie> | there are bugs that ask for various changes to the way they're parsed |
| 23:15 | <Hixie> | 9351, 9207 |
| 23:15 | <Hixie> | or reported as errors |
| 23:15 | <Hixie> | 9352 |
| 23:16 | <Hixie> | we could make unescaped & not be an error but leave charrefs that don't have a ; be an error, i guess |
| 23:16 | <Hixie> | so "?a=b&c=d" would be fine |
| 23:16 | <GarethAdams|Home> | how do you tell the difference? |
| 23:17 | <Hixie> | but "?apple=banana©=ditto" would be an error (since the © gets parsed in an unexpected way) |
| 23:17 | <GarethAdams|Home> | oh, so from the fixed list of known entities |
| 23:17 | <Hixie> | or we could make ©= not be parsed as a char ref in attributes |
| 23:17 | <Hixie> | just like ©pasta already isn't |
| 23:18 | <Hixie> | but that makes attributes even more different than content |
| 23:18 | <tabatkins> | Are we sure that char references without a trailing ; appear in attributes, especially @href and similar, often enough to cause a problem? |
| 23:18 | <Hixie> | tabatkins: cause problems in which direction? |
| 23:18 | <tabatkins> | In the direction of, if we make them not parse as entities, links will break. |
| 23:19 | <Hixie> | tabatkins: i expect there are plenty of pages that say "?foo=bar©=ditto" and are broken, and i expect there are plenty of pages that say "this file © 2010 by me" and expect it to be an entity. |
| 23:19 | <tabatkins> | Sure. |
| 23:19 | <Hixie> | the latter is even valid in ghtml4 |
| 23:19 | <Hixie> | s/g// |
| 23:20 | <Hixie> | so yes, things will break if we change anything, and things are broken if we don't change anything |
| 23:20 | <tabatkins> | I think I'm fine with attributes being different from content here. |
| 23:20 | <Hixie> | i'm only talking about attributes here |
| 23:21 | <Hixie> | they're already different than content, so that ship has sailed -- teh question is whether we can keep that to a minimum, so that people moving stuff in and out of attributes don't get screwed |
| 23:21 | <Hixie> | i'd really like to make any valid content in one have the same meaning in the other |
| 23:21 | <Hixie> | invalid content is less of an issue |
| 23:22 | <tabatkins> | I'd like links to be easy to author, and not have to worry about the names of my query paramaters. |
| 23:22 | <AryehGregor> | It's too bad we can't just make © equivalent to &copy in content as well. |
| 23:22 | <AryehGregor> | But the extra difference between content and attributes seems like the lesser evil here. |
| 23:22 | <AryehGregor> | It's not like it should come up very often in practice. |
| 23:23 | <AryehGregor> | You could make "©" and such raise a warning or error if used in content. |
| 23:23 | <tabatkins> | It would be interesting to try and see how much outcry there is (try making entities require a ;, and allow unescaped & otherwise). |
| 23:23 | <AryehGregor> | Yeah, I wonder if implementers would do that. How many pages are we talking about? It wouldn't be huge breakage in any case, if it's largely in copyright boilerplate then most users probably won't even notice. |
| 23:24 | <AryehGregor> | (even if they view an affected page) |
| 23:24 | AryehGregor | wonders why tabatkins has decapitalized himself |
| 23:24 | <tabatkins> | Indeed. |
| 23:24 | <Hixie> | we're talking billions of pages here |
| 23:24 | <tabatkins> | Eh, accidental. irssi is taking my account name from linux, and I'm not bothering to correct it. |
| 23:24 | <Hixie> | AryehGregor: it would be a disaster in non-english locales |
| 23:25 | <AryehGregor> | Oh well. |
| 23:25 | <Hixie> | e.g. è is one of these entities |
| 23:27 | <Hixie> | suggests the simplest option to improve compatibility on the web is to make = make these entities not get replaced |
| 23:27 | <Hixie> | er, http://www.w3.org/Bugs/Public/show_bug.cgi?id=9207#c5 suggests |
| 23:27 | <tabatkins> | = and _, apparently. |
| 23:27 | <AryehGregor> | That sounds like a clever solution. |
| 23:27 | <Hixie> | _ was not quite as clear |
| 23:29 | <Hixie> | ok let's add = and then reexamine the conformance problem. |
| 23:31 | <Hixie> | i suppose i could make &foo= be not a parse error in attributes, but otherwise leave these rules alone |
| 23:31 | <Hixie> | or i could just make all &foo that don't match an entity be not parse errors |
| 23:32 | <Hixie> | the latter seems simpler, though it would lead to authors never bothering to escape their &s, which could get them in trouble if they accidentally walk into &s |
| 23:38 | <Hixie> | this would make &foo; (where "foo" isn't anything known) into _not_ an error... |
| 23:42 | <dglazkov> | Hixie: I demand a recount |
| 23:42 | <dglazkov> | :) |
| 23:43 | <dglazkov> | can you point me to the place where the spec says that if the size of select is >1, multiple is not set, and no options have selectedness set to true, we shouldn't mark any item as selected? |
| 23:45 | <Hixie> | no, because the spec doesn't say that :-) |
| 23:45 | <Hixie> | "If the multiple attribute is absent, whenever there are no option elements in the select element's list of options that have their selectedness set to true, the user agent must set the selectedness of the first option element in the list of options in tree order that is not disabled, if any, to true." |
| 23:45 | <dglazkov> | but all browser that I know of do it. |
| 23:45 | <Hixie> | do what? |
| 23:45 | <dglazkov> | not select if size > 1 |
| 23:46 | <dglazkov> | annevk told me to do the comment thingy |
| 23:46 | <dglazkov> | should I go ahead and do it? |
| 23:47 | <Hixie> | how weird |
| 23:48 | <Hixie> | why would size="" have any effect on this? |
| 23:48 | <Hixie> | please do file a bug |
| 23:48 | <Hixie> | it's an easy fix |
| 23:48 | <Hixie> | i wonder why browsers act like this |
| 23:50 | <othermaciej> | Hixie, dglazkov: I suspect the browser rule is "if it renders as a pop-up menu select the first item, if it renders as a listbox, don't select anything" |
| 23:50 | <othermaciej> | because listboxes have a reasonable "nothing selected" state |
| 23:50 | <dglazkov> | othermaciej: yes, something like that |
| 23:50 | <othermaciej> | that would be why size= has an effect |
| 23:50 | <dglazkov> | but if you change the size from X to 1 and then back to X (X>1), you'll have a selected element. |
| 23:51 | <othermaciej> | sure, because size changes don't cause the selected element to be deselected |
| 23:51 | <dglazkov> | but they do cause the element to be selected :) |
| 23:51 | <othermaciej> | and browsers don't keep track of whether something got selected by default |
| 23:51 | <dglazkov> | yep |
| 23:51 | dglazkov | is writing a layout test for that |
| 23:51 | <othermaciej> | yes, if you enter the "must have a selection" state (no multiple or size > 1), then an item gets default-selected |
| 23:52 | <othermaciej> | but when you leave that state, nothing gets deselected |
| 23:52 | <othermaciej> | seems logical to me, at least to the extent one can expect logic of the Web platform |
| 23:54 | <Hixie> | i guess it makes sense given visual UAs using different widgets and how single-select list boxes can have an (normally unreachable) no-selection state |
| 23:54 | <Hixie> | doesn't make much sense from an abstract point of view though |
