| 02:03 | <Hixie> | Philip`: added 404 |
| 02:03 | <Philip`> | http://msdn.microsoft.com/en-us/library/cc288325(VS.85).aspx - IE=EmulateIE7 looks new |
| 02:04 | <Philip`> | Hixie: Thanks, though I guess that means I need to update my script instead of going to sleep :-p |
| 02:04 | <Hixie> | heh |
| 02:04 | <Hixie> | no rush |
| 02:04 | <Hixie> | you can trigger an update now anyway |
| 02:04 | <Hixie> | just run your script |
| 02:04 | <Hixie> | and it'll all work |
| 02:21 | <Hixie> | ok wow. data-* is awesome. |
| 02:21 | <Hixie> | it's been the answer to almost every single question people have asked me about in the last week here at google. |
| 02:22 | <othermaciej> | it's pretty cool, indeed |
| 02:22 | <othermaciej> | on my list of top 5 HTML5 things that I wish were already in WebKit |
| 02:23 | <Hixie> | it is already in webkit :-) |
| 02:23 | <Hixie> | well, not the .dataSet thing |
| 02:23 | <Hixie> | or .dataset |
| 02:23 | <Hixie> | or whatever |
| 02:23 | <Hixie> | but that's not a big deal |
| 02:23 | <Hixie> | people can use getAttribute() for now |
| 02:23 | <Hixie> | .dataset is just extra sugary goodness |
| 02:23 | <othermaciej> | well ok the dataset part is the part I wish were there |
| 02:23 | <othermaciej> | but sure |
| 02:23 | <othermaciej> | (the other sugary API I really want is classList) |
| 02:25 | <Hixie> | oooh yeah that's nice too |
| 02:26 | <Philip`> | Hixie: I guess none of those people asked a question like "how can I do X, without relying on scripting"? |
| 02:27 | <othermaciej> | that doesn; sound like a very googley consideration |
| 02:27 | <othermaciej> | *doesn't |
| 02:28 | <Hixie> | indeed! |
| 02:28 | <Hixie> | most of the question are "so i have this script that needs to get information X, how do i include it in the markup?" |
| 02:29 | <othermaciej> | the web is an application platform |
| 02:29 | <Hixie> | now it is |
| 02:29 | <othermaciej> | I admire the holdouts who would like to still think of it as a document-oriented hypertext system for their tenacity |
| 02:29 | <othermaciej> | but the race is over |
| 02:29 | <othermaciej> | and it's about time they concede |
| 02:31 | <Philip`> | data-* is like RDF, minus all the bits about types and validation and URIs and namespaces and theoretical basis |
| 02:33 | <Philip`> | It seems to be applicable to any problem of the form "I want to do stuff with data" |
| 02:35 | <Hixie> | not all of them, but many, yes |
| 02:35 | <Hixie> | gotta go, bbl |
| 03:07 | <Philip`> | Hixie: "you can trigger an update now anyway [...] just run your script [...] and it'll all work" - is it meant to be really slow (like more than ten minutes)? |
| 03:38 | <MikeSmith> | Philip`: btw, did you update your multipage-generator script yet? |
| 03:39 | <MikeSmith> | I'd like to use it for this WD publication if possible |
| 05:36 | hsivonen | mumbles something about using Jython and the V.nu parser to get a faster spec gen and running it as a Web service in a memory-resident servlet container to avoid JVM startup delay |
| 06:54 | <Hixie> | wow, go MikeSmith |
| 06:54 | <Hixie> | finally someone taking the authority to clean up the trolls in public-html |
| 06:54 | <Hixie> | MikeSmith: he did, and it's running already |
| 06:55 | <Hixie> | MikeSmith: he gave a url for where to see the files generated by it |
| 06:56 | <Hixie> | it's in the logs somewhere :-) |
| 06:56 | <Hixie> | Philip`: it should work in seconds |
| 07:14 | <hsivonen> | Hixie: do you have plans wrt. complex ruby? |
| 07:26 | <Hixie> | not really |
| 07:26 | <Hixie> | i'm sure it's very important to some users |
| 07:26 | <Hixie> | just like, say, explicit poetry markup |
| 07:27 | <hsivonen> | ok. I take it that implementing ruby parsing still carries an above-normal risk |
| 07:27 | <hsivonen> | but not too high a risk I guess |
| 07:28 | <Hixie> | i have no plans to change it, though of course that might change when i look more carefully at the feedback we got |
| 07:29 | <hsivonen> | I zapped UTF-32 from Validator.nu |
| 07:30 | <Hixie> | cool |
| 07:30 | <hsivonen> | unsupported encoding count as document errors, although for theoretical purity, they probably should count as non-document-errors |
| 07:31 | <hsivonen> | It's amazing how much trouble encoding proliferation causes implementation-wise even when the decoders are available off the shelf |
| 07:32 | <hsivonen> | (the UTF-32 code path had some issues even though the decoder was an off-the-shelf library decoder) |
| 07:45 | <heycam> | Hixie, you have a couple of references to [DOM3VIEWS], but i think these should be [DOM2VIEWS] |
| 07:46 | <heycam> | dom 3 views seems to be pretty different |
| 07:46 | <heycam> | it claims to "[build] on the Document Object Model Views Level 2", but that doesn't really seem to be the case |
| 07:56 | <Hixie> | if you care a lot about this, e-mail me |
| 07:56 | <Hixie> | in general, i plan to go through all the references several years from now |
| 07:56 | <Hixie> | just before LC |
| 07:56 | <Hixie> | well, several months |
| 07:56 | <Hixie> | maybe not years |
| 07:56 | <heycam> | ok, i don't care that much, was just a small thing i noticed |
| 07:57 | <Hixie> | cool |
| 07:57 | <Hixie> | thanks for mentioning it |
| 07:57 | <heycam> | but a quick :%s/DOM3VIEWS/DOM2VIEWS/g should fix it ;) |
| 07:58 | heycam | home |
| 08:19 | <MikeSmith> | Hixie: about Philip` script, thanks |
| 08:19 | <MikeSmith> | will go back through logs when I got time |
| 08:22 | <hsivonen> | hmm. contrived mislabeled bogus UTF-16 test cases indeed cause ungraceful errors... |
| 08:23 | <hsivonen> | I don't know how to fix with hindering authors from locating stray \0 bytes in 8 bit data |
| 08:27 | <hsivonen> | s/with/without/ |
| 09:04 | <annevk> | http://weblogs.mozillazine.org/roc/archives/2008/06/applying_svg_ef.html |
| 09:04 | <annevk> | http://lists.w3.org/Archives/Public/www-style/2008Jun/0003.html |
| 09:19 | <annevk> | hsivonen, "<event-source> in <head> is XHTML-only" |
| 09:19 | <annevk> | that's not true |
| 09:20 | <annevk> | it's just that new elements haven't been integrated in the parser yet |
| 09:20 | <annevk> | (not sure about <command>, but the idea for <event-source> is to make it work in <head>) |
| 09:20 | <hsivonen> | annevk: ah. |
| 09:21 | <hsivonen> | it would be good to figure out the parsing of the new elements sooner than later |
| 09:21 | <hsivonen> | especially if browsers start implementing the elements as bolt-ons to existing parsers |
| 09:22 | <annevk> | yeah, and if browsers implement HTML5 parsing they might as well get the parsing right for the new elements |
| 09:22 | <annevk> | shouldn't be a big deal if the elements don't do anything yet |
| 09:22 | <hsivonen> | it would be good to make the new sectioning elements report themselves to AT ASAP |
| 09:40 | <hsivonen> | oh. great. according to the IANA, image/svg+xml still does not exist |
| 09:44 | <hsivonen> | lots and lots of other +xml types in the registry these days |
| 09:46 | <annevk> | I'm not sure if there's any benefit to image/svg+xml or any XML media type intended for Web browsers |
| 09:47 | <annevk> | It seems that using text/xml, application/xml, and */*+xml and dispatching on namespaces works just fine |
| 10:18 | <annevk> | Ah, the TAG is informed about data-* |
| 10:19 | <othermaciej> | are they hopping mad? |
| 10:19 | <annevk> | not yet, but they might |
| 10:20 | <annevk> | It seems that Julian forgot to mention dataset though which would not really work well with arbitrary namespaces |
| 10:38 | hsivonen | rereads http://www.flightlab.com/~joe/sgml/sanity.txt |
| 10:52 | <annevk> | I actually thought that xmlns:x="" in XML 1.1 meant to assign the no namespace to the x prefix. Apparently this is not exactly true. |
| 10:53 | <annevk> | Hmm, Opera now treats <test xmlns:x=""/> as a fatal error in both XML 1.0 and XML 1.1 |
| 10:54 | <annevk> | Oh well, with XML 1.0 (5ed) XML 1.1 becomes obsolete and we can fix this somehow. |
| 10:55 | <hsivonen> | XML 1.0 5th ed. is a bad idea too |
| 10:55 | <hsivonen> | According to John Cowan, it's about justice |
| 10:56 | <MikeSmith> | I was going to say that the TAG already was informed about data-* if they had read the "areas that are introducing major architectural features" message at http://lists.w3.org/Archives/Public/www-tag/2008May/0087.html |
| 10:56 | <annevk> | It's a first step in the right direction |
| 10:56 | <MikeSmith> | ..but I see now that that data-* attributes and dataset were not mentioned at all in that |
| 10:56 | <hsivonen> | but when nations smaller than China have the choice of minting XML tags in their own language, it is still in their best interest to mint them in ASCII-based ENglish |
| 10:56 | <annevk> | MikeSmith, I suppose it's debatable whether it's architectural |
| 10:57 | <MikeSmith> | yeah |
| 10:57 | <MikeSmith> | anyway, I think there were other previous messages to the TAG list that mentioned data-* |
| 10:57 | <hsivonen> | so XML 1.0 5th ed. will be a lot of grief in order to give Cambodians, etc. the same freedom of choice Europeans have not to use native-language tags |
| 11:00 | <annevk> | It also makes increemental improvements to XML 1.0 acceptable, which seems like a huge win |
| 11:01 | <hsivonen> | annevk: I think going through the pain n many increments is not a good idea in the context of Draconian error handling |
| 11:03 | <hsivonen> | I'm sure native English speakers who advocate justice in XML element names are well-meaning, but effectively they are still suggesting that foreigners lock their data formats into national languages and not come compete in the global software market |
| 11:04 | <annevk> | I don't think that's the major win with XML 1.0 5ed |
| 11:04 | <hsivonen> | annevk: that's the stated win |
| 11:04 | <annevk> | (That it matches XML5 is nice though.) |
| 11:05 | <annevk> | hsivonen, if that's what it takes to get improvements to XML going, fine with me :) |
| 11:05 | <hsivonen> | annevk: umm. 5th ed. still makes arbitrary restrictions on Name characters |
| 11:05 | <hsivonen> | they change it but still not all the way |
| 11:05 | <annevk> | ok |
| 11:11 | <annevk> | anyway, we'll see what happens with that in due course |
| 11:12 | <hsivonen> | I had a lengthy politically-incorrect blog post drafted about 5th ed, but I got distracted |
| 11:15 | <hsivonen> | are HTML 4.01 MultiLengths actually supported on table columns in browsers: http://www.w3.org/TR/REC-html40/types.html#h-6.6 |
| 11:23 | annevk | finds more Mozilla <canvas> extensions in https://bugzilla.mozilla.org/show_bug.cgi?id=430906 |
| 11:25 | <hsivonen> | does it count as a distributed extension? |
| 11:27 | <annevk> | what is that? |
| 11:28 | <hsivonen> | an instance of exercising distributed extensibility |
| 11:28 | <annevk> | I guess not, it seems to be a new DOM attribute called mozOpaque |
| 11:29 | <annevk> | I don't think distributed extensibility works for the DOM |
| 11:29 | <hsivonen> | but moz is effectively a namespace prefix |
| 11:29 | <annevk> | effectively it's a short namespace |
| 11:50 | <annevk> | http://lists.w3.org/Archives/Public/www-archive/2008Jun/ :o |
| 12:05 | <hsivonen> | zcorpan: I backported area shape datatypes and others to HTML 4.01 / XHTML 1.0. thanks |
| 12:13 | takkaria | sighs at RB |
| 12:14 | <takkaria> | the reason I stop replying to him is just because it's not worth my time to try and clarify things to, not because I'm hostile towards him personally |
| 12:16 | <annevk> | I'm not sure why he's so hung up on his example. It seems pretty clear that any change increases implementation complexity. But then he believes it has zero implementation cost so maybe that's why... |
| 12:37 | <zcorpan> | hsivonen: cool |
| 12:52 | <Philip`> | Whoops, I should have woken up four hours ago |
| 12:52 | <Philip`> | MikeSmith: http://canvex.lazyilluminati.com/wa1/multipage-w3c/Overview.html is the new spec-splitter output |
| 12:52 | <MikeSmith> | Philip`: good morning and thanks :) |
| 12:53 | <Philip`> | MikeSmith: It's far from morning now :-( |
| 12:54 | <MikeSmith> | when you get up, that's morning |
| 12:55 | <MikeSmith> | in Japan, people who work in bars and clubs and other food-service jobs always say "Ohayo" or "Ohayo gozaimasu" (good morning) to greet each other, regardless of what time of day (or night) it actually is. |
| 13:03 | <Philip`> | Hmm, the multipage spec heading seem okay in IE6 (though with bad layout) and fine in IE8, so it's just IE7 that fails to draw the headings |
| 13:04 | <Philip`> | (or, rather, draws them behind the green box, I guess) |
| 13:14 | <jgraham> | I wonder if I should try to point out to Rob Burns that taking up everyone else's time is irritating and antisocial. I think I shouldn't because I don't think I would succeed but something really has to be done. |
| 13:16 | <MikeSmith> | Philip`: If anybody can cook up a CSS workaround for the IE7 issue, we can put up a modified CSS stylesheet and point the multipage version at that. |
| 13:16 | <annevk> | or we put a link in there to opera.com/download |
| 13:16 | <Philip`> | jgraham: If you don't succeed, you would just be taking even more of everyone's time, so that doesn't sound entirely productive |
| 13:16 | <MikeSmith> | annevk: :) |
| 13:17 | Philip` | wonders if he should bother going in to work today |
| 13:17 | <Philip`> | since it'll be about 2:30pm by the time I get there |
| 13:18 | <annevk> | Philip`, by that agument you might as well do nothing as you never know whether something will succeed |
| 13:18 | <jgraham> | Philip`: Obviously I wouldn't CC anything other than www-archive so it would only be a waste of my own time |
| 13:18 | <MikeSmith> | Philip`: work at the office til 10:30pm and you've got a full day in there |
| 13:20 | <Philip`> | jgraham: You count as one of "everyone", so that would still be wasting the (collective) everyone's time |
| 13:20 | <Philip`> | and people like me can't resist reading everything on www-archive anyway :-p |
| 13:23 | <jgraham> | Yeah it's only worthwhile if P(success)*(time saved if successful) > (1-P(success))*(time wasted if unsuccessful) |
| 13:25 | <annevk> | It might be that with RB P(success) is always 0 though. |
| 13:25 | <Philip`> | MikeSmith: It would be kind of cold and dark walking back home then, and also I'd starve :-( |
| 13:25 | <annevk> | Personally I sometimes draft responses and then delete them when I realize what I'm doing... |
| 13:25 | <jgraham> | P(success) is certianly close to zero |
| 13:26 | <jgraham> | but the time being wasted is really high |
| 13:36 | <annevk> | http://www.fightfortheopenweb.com/2008/04/15/why-html5-is-not-enough/ |
| 13:39 | <annevk> | (it seems to be a site written by a guy who has taken it on himself to find a way to fight SilverLight/Flash) |
| 13:41 | <jgraham> | So his point is basically that we need WF2 and XBL2? |
| 13:42 | <Philip`> | It's a shame that nothing like WF2 and XBL2 exists - someone should invent them |
| 13:43 | <Dashiva> | jgraham: Looks like he wants joined client and server code as well |
| 13:44 | hsivonen | disagrees about not assuming CSS and JS |
| 13:46 | <Dashiva> | Assuming JS in general is okay, but there are some sites that abuse it enough that I disable it :) |
| 14:01 | <hsivonen> | hmm. comments on the blog aren't working |
| 14:03 | <MikeSmith> | hsivonen: recent mysql upgrade on the server? or PHP maybe? |
| 14:03 | <hsivonen> | dunno |
| 14:03 | <hsivonen> | I meant on the blog annevk referred to above |
| 14:03 | <MikeSmith> | oh |
| 14:04 | <annevk> | there's http://groups.google.com/group/fight-for-an-open-web if you really want to make your comment reach him I guess |
| 14:21 | <BorekB> | Hi all, I've got some questions regarding opening links into new tabs. Is this the right place to discuss it or should I post to the mailing list? |
| 14:22 | <annevk> | If you're not sure about your comment you can discuss it here first |
| 14:23 | <annevk> | (Though you can also post it directly to the mailing list, it's really up to you :) ) |
| 14:23 | <BorekB> | Thanks, I'll try it here :) |
| 14:24 | <BorekB> | I am aware that there is no way to inform the browser that it should open the link in a new tab. However, I think this would be useful in some situations. I just wanted to know if this is something that is being considered or discussed |
| 14:26 | <hsivonen> | BorekB: tabs vs. windows is considered a matter of browser UI |
| 14:26 | <annevk> | It has been raised and the general consensus is that target=_blank covers that use case |
| 14:26 | <hsivonen> | BorekB: so whether to open new "windows" as tabs is a browser pref |
| 14:28 | <BorekB> | From a user perspective, opening links in a new window or in a new tab is very different. If there is only one way to express this in HTML/JS (target="_blank" or window.open), there is an obvious mismatch. Let me demonstrate on a specific use case.. |
| 14:31 | <zcorpan> | 2541 lowsrc |
| 14:32 | <BorekB> | In web applications where you work with many external links, it is quite common to open links that you want to read later to new tabs. Google Reader would be an example of such an application (quickly skimming through the new posts and opening the ones of your interest to new tabs). There is no way to implement it by the GReader team at the moment. |
| 14:32 | <zcorpan> | lowsrc is more common than 2337 xmlns:o |
| 14:33 | <zcorpan> | interesting |
| 14:33 | <BorekB> | (I mean you can always middle-click the link in Firefox but if you use keyboard shortcut 'v', the links are opening in new windows by default which kind of defeats the point of having tabbed browser.) |
| 14:33 | <zcorpan> | oh wait, this was total occurances |
| 14:34 | <zcorpan> | pages: 2223 xmlns:o, 681 lowsrc |
| 14:34 | <BorekB> | The workaround is to go to Firefox settings and force it to open all "_blank" links to the tabs instead of windows. |
| 14:35 | <BorekB> | With this setting in place, however, some other sites where the new window would be more appropriate that a new tab (small help window for instance) will be force to use new tabs as well. |
| 14:37 | <BorekB> | I don't like this as a user - the web application should be able to express its intention to either open the link in a new window or in a new tab. (There can still be an option to override this intention on a browser level, this is nothing against it.) |
| 14:39 | <BorekB> | (Sorry for spelling errors, I hope it's clear what my concern is.) |
| 14:40 | <BorekB> | I am obviously not alone - just try googling "_tab" |
| 14:40 | <BorekB> | What do you think? |
| 14:41 | <annevk> | mpt proposed the opposite a while ago: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2008-April/014564.html |
| 14:42 | <BorekB> | I see it more like an addition than opposite |
| 14:43 | <BorekB> | Furthermore, target="_tab" would be much easier to implement for browser vendors - all necessary infrastructure (tabs!) is already there |
| 14:43 | <annevk> | Well, with his proposal in place it might be feasible to do target=_blank in a new tab |
| 14:45 | <Philip`> | Opera already opens target=_blank in a new tab by default |
| 14:45 | <annevk> | true :) |
| 14:46 | <annevk> | Anyway, I don't really feel strongly about this. I guess I don't feel a compelling need for either |
| 14:46 | <BorekB> | Philip, my concern isn't about what is or what isn't default. It's more about inability to distinguish between the two obviously different behaviours in HTML/JS |
| 14:46 | <Philip`> | Of course it also does "Cycle [tabs] in recently used order" by default, so being default UI doesn't mean it's not a horribly awful choice :-p |
| 14:47 | <annevk> | Philip`, aren't the newer builds better with that by default? |
| 14:47 | <BorekB> | Philip: I agree that Opera got the default settings much better than the other browsers out there |
| 14:48 | <Philip`> | annevk: They still default to "recently used order", which is impossible to use since the whole list gets reshuffled every time I switch tab |
| 14:49 | <BorekB> | OK thanks for initial discussion, I think I will send this to the mailing list so that we can have broader discussion |
| 14:49 | <Philip`> | Opera also does window.open() in a tab, including special not-full-sized window-like tabs that are drawn on top of the background tab, for when you specify width and height |
| 14:50 | <Philip`> | so it doesn't have as much of a distinction between windows and tabs as other browsers do |
| 14:50 | <BorekB> | Now we just need a solution for 98% of browsers out there :) |
| 14:51 | <Philip`> | BorekB: 98% are IE6 and IE7 and FF2 etc, so any new feature is definitely not going to work in them :-) |
| 14:53 | <Philip`> | The most common browser doesn't even have tabs |
| 14:53 | mpt | reads backlog and discovers that annevk read his mind |
| 14:56 | <mpt> | With target="_reference" available, browsers could open target="_blank" in a new tab |
| 14:58 | <mpt> | That might take longer than introducing target="_tab" |
| 14:58 | <BorekB> | mpt: exactly |
| 14:59 | <BorekB> | target="_tab" would be very quick fix for most browsers |
| 14:59 | <mpt> | But by the time *either* of those happen, whether something opens in a new window or a new tab might not be under the control of the browser any more. |
| 15:00 | <mpt> | In which case target="_tab" would be a dead distinction, functionally identical to target="_blank". |
| 15:00 | <BorekB> | how do you mean that? not under the control of the browser? |
| 15:01 | <hsivonen> | oh well. I didn't resist replying to elharo. |
| 15:01 | hsivonen | hopes the subthread doesn't grow long |
| 15:04 | <mpt> | BorekB, handled by the window manager or operating environment instead |
| 15:05 | <mpt> | BorekB, an increasing amount of Web browsing is being done on small-screen devices that sometimes have tabs, but if they do, make no distinction between tabs and windows |
| 15:05 | <BorekB> | mpt: Why would OS prevent opening new windows? |
| 15:07 | <mpt> | I don't understand that question |
| 15:08 | <BorekB> | ok i get what you mean. your comment is valid for devices but probably not so appropriate for desktop operating systems |
| 15:15 | <mpt> | Not appropriate right now, certainly. |
| 15:42 | <Philip`> | Hixie: "it should work in seconds" - it's been twelve hours and it still doesn't seem to have updated :-( |
| 15:43 | <Philip`> | (e.g. http://www.whatwg.org/specs/web-apps/current-work/multipage/fragment-links.js is not the version now in http://canvex.lazyilluminati.com/misc/multipage/multipage.tar.bz2) |
| 15:43 | <Philip`> | (and I did call the do-pubrules-update thing after updating that file) |
| 15:50 | <Philip`> | MikeSmith: If you want to use the multipage version of the W3C version of the spec, let me know if there's anything that ought to be changed (like if the header bit is missing stuff or has too much stuff, or if the <title>s ought to be distinct between pages, or whatever) since it ought to be fairly quick to fix |
| 15:51 | <MikeSmith> | Philip`: OK, after I have had a chance to look at it more carefully |
| 15:53 | <Philip`> | (The main differences are that the header thing in the source file has changed, and the spec-splitter outputs an HTML4 doctype and <div> instead of <nav> and Overview.html instead of index.html and no link-fixup script and that's all) |
| 15:53 | <Philip`> | (*differences compared to the WHATWG version) |
| 15:54 | <MikeSmith> | OK |
| 15:57 | <Philip`> | (Oh, and the W3C one has index.html instead of section-index.html, because the WHATWG version would have a conflict between index.html and $sectionname.html when $sectionname = 'index' hence it gets renamed to section-index.html) |
| 15:57 | <Philip`> | (But I think that's all :-) ) |
| 16:13 | <zcorpan> | hmm. setting innerHTML in XML |
| 16:13 | <zcorpan> | step 2 |
| 16:15 | <zcorpan> | does "declaring all the namespace prefixes that are in scope on that element in the DOM" mean xmlns:foo attributes or elements with prefix, or both? |
| 16:21 | <zcorpan> | document.createElementNS('http://www.w3.org/1999/xhtml','foo:div').innerHTML = '<foo:img/>'; |
| 16:22 | <zcorpan> | time to send email |
| 16:23 | Philip` | wonders how that div changed into an img |
| 16:23 | <Dashiva> | A wizard did it |
| 16:24 | <Dashiva> | zcorpan: Definitely vague |
| 17:09 | <MikeSmith> | about the following in the current HTML5 draft: |
| 17:09 | <MikeSmith> | http://www.w3.org/html/wg/html5/#html-namespace |
| 17:10 | <MikeSmith> | do current browsers actually already "place elements in HTML in the http://www.w3.org/1999/xhtml namespace, at least for the purposes of the DOM and CSS" ? |
| 17:11 | Philip` | likes that he can go to http://www.whatwg.org/specs/web-apps/current-work/multipage/#html-namespace and get sent to the right page |
| 17:11 | <MikeSmith> | or is that something that they will need to be changed to do in order to be conformant? |
| 17:12 | <Philip`> | MikeSmith: WebKit does; others don't |
| 17:12 | <Philip`> | (http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cbody%20onload%3Dw(document.body.namespaceURI)%3E ) |
| 17:13 | <myakura> | I thought Firefox does too (for CSS) |
| 17:15 | <Philip`> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cstyle%3E%0A%40namespace%20a%20%22http%3A%2F%2Fwww.w3.org%2F1999%2Fxhtml%22%3B%0A%40namespace%20b%20%22http%3A%2F%2Fwww.w3.org%2F1999%2Fxhtml1%22%3B%0Aa%7Cdiv%20%7B%20background%3A%20green%20%7D%0Ab%7Cdiv%20%7B%20background%3A%20red%20%7D%0A%3C%2Fstyle%3E%0A%3Cbody%20onload%3Dw(document.body.namespaceURI)%3E%0A%3Cdiv%3ETest%3C%2Fdiv%3E |
| 17:16 | <Philip`> | myakura: Ah, okay - it looks like Firefox and Opera both do for CSS |
| 17:17 | <MikeSmith> | OK |
| 17:18 | <Philip`> | (and IE is 'special') |
| 17:20 | <MikeSmith> | yeah, that at least I knew already |
| 17:30 | Philip` | wonders if MikeSmith is having much fun in www-archive |
| 17:31 | <MikeSmith> | Philip`: I think you can guess what the answer to that question is :) |
| 17:37 | <smedero> | oh cool, people are putting words in my mouth. |
| 17:37 | <smedero> | hooray. |
| 17:38 | <smedero> | apparently I gave an authoritative view on something instead of simply citing what the chairs have said time and time again. |
| 17:38 | <smedero> | sigh, this kind of stuff bores the hell out of me. |
| 17:41 | <takkaria> | Philip`: you have any stats on the <q> tag? |
| 17:42 | <Philip`> | takkaria: Any particular stats? |
| 17:42 | <Philip`> | I see it on 59 pages out of 130K |
| 17:42 | <takkaria> | whether people provide their own quotes inside, outside, or not at all |
| 17:46 | <Philip`> | Looks like sometimes inside, sometimes outside, mostly not at all |
| 17:46 | Philip` | will upload the stuff once his grep finishes |
| 17:47 | <takkaria> | thanks |
| 17:47 | <Philip`> | (I've no idea how many of these pages define their own styling for <q>) |
| 17:49 | <Philip`> | (<q> is the 125th most common element, barely half as common as the <align="center"> element) |
| 17:49 | <takkaria> | <align="center"> sounds fun |
| 17:51 | <Philip`> | That seems to be used on a wide variety of pages, not just lots of instances of the same site |
| 17:55 | <Philip`> | takkaria: http://philip.html5.org/data/q-tags.txt |
| 17:56 | <Hixie> | Philip`: do you get a 500 or anything when calling the do-multipage-update script? |
| 17:56 | <Philip`> | Hixie: do-multipage-update? |
| 17:57 | <Hixie> | isn't that the page i said to use to ping when there's an update? |
| 17:57 | <Philip`> | If that exists, you never told me about it :-p |
| 17:57 | <Hixie> | oh |
| 17:57 | <Hixie> | what uri are you calling |
| 17:57 | <Philip`> | 11:54 < Hixie> can you make it ping http://www.whatwg.org/specs/web-apps/current-work/do-pubrules-update when it's done? |
| 17:57 | <Hixie> | oops |
| 17:58 | <Hixie> | i gave you the wrong one! |
| 17:58 | <Hixie> | s/pubrules/multipage/ |
| 17:58 | <Hixie> | sorry! |
| 17:58 | <Hixie> | pubrules is the one that the other host uses once the spec has been run through the pubrules checker |
| 17:58 | <Philip`> | Aha - I'll try that :-) |
| 17:58 | <Philip`> | Success |
| 17:59 | <Hixie> | woo |
| 17:59 | <Philip`> | http://www.whatwg.org/specs/web-apps/current-work/multipage/section-introduction.html - hurray |
| 17:59 | <Hixie> | cool |
| 17:59 | <Hixie> | sorry about that |
| 18:02 | Philip` | wonders how you make bash scripts abort when one of the commands failed |
| 18:02 | <gsnedders> | Philip`: With far too many if statements |
| 18:02 | <gsnedders> | (like, one per command) |
| 18:03 | <Philip`> | I don't even know how to do if statements in bash :-p |
| 18:04 | <gsnedders> | I don't know off the top of my head :P |
| 18:04 | <gsnedders> | And I'm finding a lack of xrefs |
| 18:07 | <gsnedders> | "The x element" has no xref |
| 18:08 | <gsnedders> | Philip`: etree.tostring(Element, encoding=unicode, method='text') includes Element.tail! |
| 18:09 | <Philip`> | e=deepcopy(Element);e.tail='';etree.tostring(e, ...) :-) |
| 18:09 | <Philip`> | That does sound kind of irritating |
| 18:10 | <gsnedders> | It breaks a heckuva lot of cross-refs too :( |
| 18:14 | <gsnedders> | Philip`: Just stripping len(Element.tail) chars if Element.tail is probably quicker |
| 18:15 | <gsnedders> | It isn't, actually |
| 18:15 | <gsnedders> | Yes, it is. |
| 18:16 | <gsnedders> | 3x slower |
| 18:17 | <Philip`> | gsnedders: Or use tostring(..., with_tail=False) |
| 18:17 | <gsnedders> | Philip`: Hah! |
| 18:18 | <Philip`> | There's .xpath("string()") too |
| 18:19 | <gsnedders> | Philip`: That's really expensive |
| 18:19 | <Philip`> | Ah |
| 18:21 | <gsnedders> | Philip`: The difference between with_tail and stripping len(Element.tail) is only a few thousandths |
| 18:21 | <Philip`> | But one of those isn't a nasty hack :-) |
| 18:22 | <gsnedders> | Philip`: Exactly :) |
| 18:22 | <gsnedders> | (This is a few thousandths on the entire HTML 5 spec) |
| 18:23 | <Philip`> | (A few thousandths of what?) |
| 18:23 | <gsnedders> | A second. |
| 18:23 | <Philip`> | Ah |
| 18:24 | <gsnedders> | And the random uncertainty seems to be around ±2ms |
| 18:24 | <gsnedders> | which makes a difference of around 5ms between the means even less meaningful |
| 18:25 | gsnedders | stares at the WHATWG version of the HTML 5 spec |
| 18:25 | <gsnedders> | yay! more undefined behaviour in the real spec-gen! |
| 18:27 | <gsnedders> | "Cross-references are made by marking the defining instance of a term with <dfn>." |
| 18:27 | <gsnedders> | Not <hx>. |
| 18:30 | <gsnedders> | No, it is only <dfn>. Some other bug. |
| 18:38 | <gsnedders> | Oh, it's because textContent() keeps the whitespace |
| 18:54 | <Hixie> | MikeSmith: yt? |
| 18:57 | <jgraham> | Philip`: Can't you use && to exit if a command fails in bash? |
| 18:57 | jgraham | knows very little bash scripting |
| 18:58 | <Philip`> | jgraham: I'm not confident in how that interacts with things like output redirection, and also I'd prefer not to make my entire script be a single line :-) |
| 18:59 | <MikeSmith> | Hixie: here now |
| 18:59 | <Philip`> | Currently I'm sticking with the approach that probably nothing is going to fail too badly or too frequently so there's no need for detecting errors |
| 18:59 | <jgraham> | Hmm. I'm not sure why it would require the whole script to be a single line |
| 18:59 | <jgraham> | biab |
| 19:11 | <Hixie> | sweet |
| 19:11 | <Hixie> | rob told mike that he was out of line |
| 19:11 | <Hixie> | that's hilarious |
| 19:12 | <Hixie> | it's sad that the chairs are making you do the leg work here |
| 19:12 | <Hixie> | if this was going on in a community i had authority over he'd have been banned for a week already |
| 19:28 | <Hixie> | The aria-* syntax creates a multitude of attributes. It is as if one decided to go for style-font="*" instead of style="font:*". |
| 19:28 | Hixie | wonders if anyone told this guy about SVG |
| 20:05 | <Dashiva> | Hixie: I wonder if he would be any happier with aria="a:b;c:d;" :) |
| 20:10 | <Hixie> | lord |
| 20:12 | <deltab> | Philip`: run the shell with -e, use use set -e in a script |
| 20:12 | <deltab> | ^or use |
| 20:13 | <deltab> | and use foo || true; if you need to ignore the result of foo |
| 20:24 | gsnedders | guesses he should make his xref have whitespace normalisation |
| 20:57 | <annevk> | yup |
| 20:58 | <annevk> | it's typical for <dfn> to contain a newline and several consecutive spaces |
| 21:02 | <gsnedders> | Anyone feeling like being useful, so I can do my homework? |
| 21:03 | <gsnedders> | Make http://hg.gsnedders.com/spec-gen/file/0b4dc9d2fc3e/runtests.py run the tests/basic |
| 23:15 | gsnedders | wishes Mozilla supported display: run-in |
| 23:15 | <gsnedders> | I mean, even IE8 does! :P |
| 23:15 | <gsnedders> | (There again, that doesn't have a ten year old layout engine in standards mode) |
| 23:27 | <roc> | you're the first person I've ever heard ask for it |
| 23:27 | <Dashiva> | gsnedders: inline-block only took 10 years, cut them some slack :) |
| 23:28 | <gsnedders> | roc: Heh. I use it on the print layout of my blog, though as Philip` will undoubtedly point out if he sees this, nobody prints blogs. |
| 23:28 | <gsnedders> | (Prince is actually the only thing that gets the print version of my blog right) |
| 23:29 | <gsnedders> | Dashiva: the run-in bug will be 10 in Dec :P |
| 23:30 | <Philip`> | gsnedders: I suppose some people might print some blogs, when they're running out of toilet paper |
| 23:30 | gsnedders | sighs |
| 23:34 | <Philip`> | It's really quite convenient how computers have a button that causes paper to be extruded at you |