| 00:17 | <MikeSmith> | heycam: so I couldn't figure out specifically what might have caused that problem |
| 00:17 | <MikeSmith> | other than the file system just running out of space completely |
| 00:17 | <heycam> | MikeSmith, ok np. I'l l let you know if it comes up again. |
| 00:20 | <MikeSmith> | I think we just don't have enough space on that machine. There isn't a separate mount for /tmp on it. It's on the same mount as the main filesystem, which is only 4.3G and 93% full. So there less than 300MB free on it right now. |
| 00:20 | <MikeSmith> | anyway I'll talk to the systems team about it |
| 00:20 | <MikeSmith> | heycam: I don't mind getting pinged on the weekend, btw. But if you need to ping me and I don't seem to be around on IRC, feel free to e-mail at mike⊙wo |
| 00:21 | <MikeSmith> | If you do I'll get a notification on my mobile right away, and I can actually ssh into that machine from there |
| 00:22 | <heycam> | MikeSmith, righto, thanks |
| 00:53 | <bk> | hi |
| 00:55 | <bk_1> | is anyone here |
| 00:55 | <bk_1> | hmm |
| 01:01 | <|bk|> | anyone knows a good way to start with xhtml |
| 01:01 | <|bk|> | not sure where else to ask |
| 01:09 | <zewt> | "don't" |
| 01:11 | <GPHemsley> | <audio>/<video> should be able to contain <embed> as a fallback, right? |
| 01:34 | <GPHemsley> | I'm told that fallback content is only used by browsers that don't actually support <audio>/<video> |
| 01:34 | <GPHemsley> | But if the browser doesn't support the media type of the actual file, shouldn't it be able to treat it as it would <embed>? |
| 01:57 | <zewt> | GPHemsley: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2010-February/025028.html |
| 01:58 | <zewt> | nonsense, in my opinion (codec compatibility will *never* become a non-issue, so long as new codecs are being deployed that not everyone has) |
| 02:00 | <GPHemsley> | zewt: I didn't find the argument particularly convincing, particularly from a user experience point of view |
| 02:18 | <Hixie> | zewt: new video codecs come out even more rarely than new image formats |
| 02:18 | <Hixie> | (which have changed like, once, in the history of the web) |
| 02:59 | <zewt> | Hixie: what? they come out far, far more frequently (i'm watching stuff on my htpc right now that I had to buy an htpc to watch, because it's so new none of my set-top stuff can play it) |
| 03:00 | <zewt> | (namely, 10-bit h264, which both requires a lot of CPU and is not decode-compatible with regular h264) |
| 03:02 | <zewt> | (or whatever lower profile of h264 is widely used; not familiar with the particulars) |
| 03:23 | <Hixie> | zewt: browser-supported codecs? |
| 03:23 | <Hixie> | zewt: sure, new image formats and video codecs come out all the time, but we're talking specifically about browser-supported ones |
| 03:24 | <Hixie> | (and so far we don't even have one that's uniformly supported) |
| 06:09 | <MikeSmith> | anybody familiar with geolocation-api implementations, I'm wondering if it's really possible in practice for a Web app to get old position data for a user from the browser position-object cache |
| 06:11 | <MikeSmith> | I thought I saw something before that suggested browser implementations only allow a Web app to access position information which the user has already authorized that app to access |
| 06:12 | <MikeSmith> | I also found http://www.w3.org/2011/09/07-geolocation-minutes.html#item06 |
| 06:13 | <MikeSmith> | where the Geolocation WG discussed this |
| 06:13 | <MikeSmith> | "The wg consensus is that no API change is required to address this, but that UAs should perhaps consider protecting against sites receiving old position data shared with other sites" |
| 06:13 | <MikeSmith> | but I don't understand why they decided the spec should not have some explicit requirement with regard to this |
| 08:03 | <hsivonen> | MikeSmith: looks like validator.nu defaults to HTML5 + SVG 1.1 + MathML 3.0 + RDFa Lite 1.1 for text/html with <!DOCTYPE html> but to XHTML5 + SVG 1.1 + MathML 3.0 for application/xhtml+xml with XHTML root namespace |
| 08:03 | <hsivonen> | MikeSmith: that can't be intentional, can it? |
| 08:03 | <hsivonen> | looks like there's a bug in the preset file |
| 08:07 | <hsivonen> | looks like the CML validator has moved and no longer builds on Validator.nu |
| 08:35 | <MikeSmith> | hsivonen: yeah that bug in the presets is almost certainly my fault |
| 08:35 | <MikeSmith> | sorry |
| 08:41 | <hsivonen> | MikeSmith: I can fix it later today |
| 08:41 | <MikeSmith> | ok |
| 08:41 | <hsivonen> | MikeSmith: I'm almost done with the statistics feature |
| 08:42 | <MikeSmith> | excellent |
| 08:42 | <hsivonen> | I'll put it behind a JVM system property |
| 08:42 | <MikeSmith> | once that lands I can enable it in the W3C backend as well, if you think that would be useful |
| 08:43 | <hsivonen> | MikeSmith: ok. right now, the schema stats don't support the W3C list of presets |
| 08:43 | <hsivonen> | but that could be arranged |
| 08:43 | <MikeSmith> | yeah I could tweak it locally afterward for now if it came to that |
| 08:44 | <MikeSmith> | I mean locally on those server instances |
| 08:44 | <MikeSmith> | I think we currently have 4 hosts behind validator.w3.org/nu |
| 09:27 | <matjas> | http://whatwg.org/html/webappapis.html#dom-windowbase64-btoa why was btoa specced to not support code points above U+00FF in the input string? was that how UAs had implemented it? |
| 09:31 | <gsnedders> | Yes. |
| 09:31 | <matjas> | gsnedders: thanks |
| 09:31 | <gsnedders> | Not sure if all did, but certainly most. |
| 09:32 | <AryehGregor> | All, IIRC. |
| 09:32 | AryehGregor | wrote the spec for that |
| 09:32 | <matjas> | Too bad. I’m used to stuff above U+FFFF breaking, but not > U+00FF. |
| 09:33 | <matjas> | what’s the reason UAs had implemented it that way? is there a good one? |
| 09:33 | <AryehGregor> | matjas, basically, btoa() treats its input as a sequence of bytes, not a string. |
| 09:33 | <AryehGregor> | The base64 algorithm acts on bytes, after all, not characters. |
| 09:33 | <AryehGregor> | How would you have said it should work? |
| 09:33 | <AryehGregor> | What would you do with codepoints above U+FF? |
| 09:34 | <matjas> | rebuild http://software.hixie.ch/utilities/cgi/data/data in HTML + JS, for example |
| 09:34 | <matjas> | it’s possible only if you don’t use `btoa` but instead write your own method |
| 09:35 | <matjas> | dev expectation (at least for me, without checking the spec) was that you can just feed any string to btoa() and get the base64-encoded result back, but that’s not true |
| 09:36 | <AryehGregor> | matjas, I mean, how would you modify the btoa() algorithm so that it does something sensible with chars above U+00FF? What should it do with them? |
| 09:36 | <AryehGregor> | I'm not sure there's anything to do with them that makes sense. |
| 09:37 | <gsnedders> | Encode as UTF-8, like Encode() does. |
| 09:37 | <matjas> | encode strings per UTF-8, then base64 encode that? |
| 09:37 | <matjas> | but maybe that’s just me |
| 09:38 | <AryehGregor> | I guess that might have been plausible if it was done from the beginning, but it's not backward-compatible now. |
| 09:38 | <AryehGregor> | At the time the algorithm was designed, UTF-8 by default would have been a lot weirder than now, I think. |
| 09:38 | <matjas> | so instead of e.g. btoa('foo♥bar') you have to use btoa(unescape(encodeURIComponent('foo♥bar'))) |
| 09:39 | <matjas> | AryehGregor: Yeah. I was just wondering why browsers never implemented it that way, as it seems to make much more sense IMHO. |
| 09:39 | <AryehGregor> | matjas, does encodeURIComponent actually always use UTF-8? |
| 09:40 | <matjas> | AryehGregor: it does as per ES 5.1; not sure about implementations though |
| 09:40 | <matjas> | http://ecma-international.org/ecma-262/5.1/#sec-15.1.3.4 |
| 09:40 | <gsnedders> | All impls do. |
| 09:52 | <matjas> | function base64encode(string) { return btoa(unescape(encodeURIComponent(string))); } function base64decode(string) { return decodeURIComponent(escape(atob(string))); } |
| 09:52 | <matjas> | there we go. |
| 10:10 | <jgraham> | Why do you need [un]escape? |
| 10:12 | <annevk> | what's the difference between encodeURI and encodeURIComponent? |
| 10:13 | <annevk> | ah |
| 10:13 | <annevk> | encodeURI does not escape # |
| 10:14 | <annevk> | oh and whatever is uriReserved |
| 10:49 | <annevk> | welcome to the WHATCG odinho |
| 10:50 | <hsivonen> | MikeSmith: http://validator.nu/stats.html http://html5.validator.nu/stats.html |
| 10:51 | <hsivonen> | in the latter case, the schema is always automatically chosen but is never reported as such |
| 10:51 | <[tm]> | hsivonen: ok |
| 10:52 | <hsivonen> | there's no server-side persintence |
| 10:53 | <hsivonen> | the stats could be persisted by running wget on stats.html via cron or something |
| 10:53 | <hsivonen> | from my point of view, the time window between validator redeployments should be enough to get interesting numbers |
| 10:54 | <hsivonen> | already two things surprise me: validations per second < 1 |
| 10:54 | <hsivonen> | and the Web service APIs seem more often used than the Web UI |
| 10:58 | <hsivonen> | [tm]: OK if I fix the bug in w3c-presets.txt, too, even though the bug is latent there (has no effect, AFAICT)? |
| 10:58 | <hsivonen> | (the bug about the default schema for XHTML, that is) |
| 11:04 | <matjas> | jgraham: [un]escape is not really needed here (it is only needed if you want to do proper UTF-8 {en,de}coding), but using it results in much smaller base64-encoded strings, and makes it match the output of other base64-encoders |
| 11:07 | <jgraham> | matjas: Right, that makes sense |
| 12:31 | <[tm]> | hsivonen: ok by me |
| 12:32 | <[tm]> | interesting that the Web-services API is used more than the Web UI |
| 12:39 | <hsivonen> | when going to TPAC, is it a good or a bad idea to take a TGV from CDG to Lyon? |
| 12:41 | <AryehGregor> | <http://html5.org/specs/dom-parsing.html>; is loading slowly. |
| 12:42 | <AryehGregor> | Oh, it seems whatwg.org isn't loading. |
| 12:42 | <AryehGregor> | That's why. |
| 12:43 | <hsivonen> | [tm]: I pushed a fix for w3c-presets.txt |
| 12:44 | <AryehGregor> | This doesn't seem to say that any escaping happens? http://html5.org/specs/dom-parsing.html#concept-serialize-xml |
| 12:45 | <AryehGregor> | (for text nodes) |
| 12:45 | <AryehGregor> | That's surely a bug. |
| 12:49 | AryehGregor | files |
| 13:00 | <MikeSmith> | hsivonen: thanks |
| 13:05 | <MikeSmith> | hsivonen: the fact that 75% of the requests are from the Web-services client makes me wonder if many of those are coming from the same small set of clients |
| 13:06 | <MikeSmith> | in other words, if there are is a small set of users or third-party services that account for most of the requests |
| 13:07 | <MikeSmith> | with users of the "normal" Web-based direct use case being the other 25% |
| 13:13 | <hsivonen> | MikeSmith: it would be interesting further research to try to figure out how varied the IP addresses for the Web service requests are |
| 13:15 | <MikeSmith> | hsivonen: yeah |
| 13:19 | <MikeSmith> | hsivonen: incidentally the W3C systems team has deployed a mechanism that when it detects excessive traffic from a particular IP, shows a page that says, |
| 13:19 | <MikeSmith> | Excessive traffic pattern blocked |
| 13:19 | <MikeSmith> | Due to excessive traffic we are temporarily blocking your IP address, please rate limit your use of W3C's Validator. |
| 13:19 | <MikeSmith> | Your IP address will clear automatically. |
| 13:19 | <hsivonen> | MikeSmith: is this in a load balancer independent of the validator code? |
| 13:19 | <MikeSmith> | yes |
| 13:20 | <MikeSmith> | I don't know the details of the rate-blocking part of that, though |
| 13:20 | <MikeSmith> | nor do I know how long they plan to set the blocking period in production |
| 13:20 | <MikeSmith> | nor how many requests per minute you have to make before it will block you and show that message |
| 13:21 | <MikeSmith> | previously it did not show that message |
| 13:21 | <MikeSmith> | it would just return a 403 |
| 13:23 | <hsivonen> | It's annoying that SNCF won't sell train tickets more than three months in advance and the end of TPAC is more than three months away. |
| 14:08 | <hsivonen> | also very annoying that SNCF sells out some trains that are almost 3 months away |
| 14:09 | <hsivonen> | so I booked the Paris Lyon train now and will book Lyon Paris once available |
| 14:09 | <hsivonen> | at this rate, the Paris-Lyon tickets would be gone by the time the Lyon-Paris tickets go on sale |
| 14:09 | <hsivonen> | crazy |
| 14:26 | <icaaq> | Hi, if I have a sentence that someone has to say to make an accomplishment, would it be appropriate to use the q-element then? |
| 14:29 | <annevk> | <q> is annoying, better use "" |
| 14:30 | <icaaq> | why is it annoying? |
| 14:31 | <karlcow> | icaaq: It is not annoying. |
| 14:32 | <annevk> | icaaq: because it sometimes and sometimes not renders quotes |
| 14:33 | <annevk> | I guess most browsers do render quotes nowadays, but not always those you want |
| 14:33 | <annevk> | and since you never want to style it you might as well use "" |
| 14:34 | <icaaq> | annevk: ok, but that I dont care about :) |
| 14:35 | <icaaq> | I'm more interested in the right semantics |
| 14:35 | <karlcow> | or 「」(lang:ja) or « quote » (lang:fr) depending on the language. |
| 14:35 | <karlcow> | The question was about semantics. So icaaq is asking if it is appropriate to markup the expression of someone with q. |
| 14:36 | <icaaq> | karlcow: a future quote, so to speak, that not has been said yet |
| 14:36 | <karlcow> | icaaq: http://dev.w3.org/html5/spec-author-view/the-q-element.html#the-q-element |
| 14:36 | <jgraham> | karlcow: The author is pretty much the only person who can tell what the right quote character is, so that seems like a disadvantage |
| 14:37 | <karlcow> | icaaq: is more about contexts than content |
| 14:37 | <karlcow> | does it come from elsewhere with regards to the text you are writing. |
| 14:39 | <karlcow> | Basically, you are saying: I'm putting this (text) in here (<q></q>) and it's coming from there (cite="<uri>") |
| 14:40 | <icaaq> | Yes, but it has not been said yet ;) |
| 14:40 | <karlcow> | icaaq: it's a provenance context, not a type of discourse. :) |
| 14:41 | <karlcow> | basically it is not about narrative |
| 14:43 | <icaaq> | it's this pull request I'm wondering about. https://github.com/mozilla/bedrock/pull/296 |
| 14:44 | <karlcow> | he is right, no need of q there |
| 14:45 | <hsivonen> | [tm]: whoa! somebody uses the legacy XHTML+SVG+MathML schema! No one validates SVG, though. http://validator.nu/stats.html |
| 15:04 | <david_carlisle> | hsivonen: some of them might have been me (but you can discount those:-) |
| 15:04 | <icaaq> | karlcow: ok, thanks! |
| 17:07 | <dglazkov> | good morning, Whatwg! |
| 20:10 | <tantek> | wow - in an example of the validator actually *not helping* looks like a lot of folks running wordpress were very confused by rel="category tag" being marked as invalid |
| 20:11 | <tantek> | e.g. http://wordpress.org/support/topic/wordpress-abuses-rel-tag |
| 20:15 | <tantek> | (only discovered today when someone created a placeholder page for rel-category_tag due to someone adding "category tag" as a *single* value to the rel-registry - since corrected) |
| 20:16 | <TabAtkins> | It's a category *and* a tag, right? Not a "category tag"? |
| 20:17 | <Hixie> | hah, that thread is awesome |
| 20:17 | <Hixie> | i love it when people say "it's not finalised, so it can change at any time" |
| 20:18 | <Hixie> | as if everything in HTML4 is still in HTML today unchanged, and as if HTML4 not changing itself was a good thing :-) |
| 20:28 | <tantek> | for more: https://www.google.com/search?q=rel-category+tag |
| 20:35 | <tantek> | TabAtkins, exactly. And as a result I had to write some text here: http://microformats.org/wiki/rel-category_tag |
| 21:43 | <Hixie> | bah, html sucks. |
| 21:43 | Hixie | finds himself wanting a non-rectangular cell in a table |
| 21:45 | <annevk> | sounds like a CSS problem |
| 21:45 | <edwardbc> | tables suck |
| 21:45 | <TabAtkins> | No, it's HTML. |
| 21:45 | <TabAtkins> | He means one that spans rows and columns without spanning the full cross of them. |
| 21:46 | <TabAtkins> | Not a cell that is, by itself, non-rectangular a la CSS Exclusions. |
| 21:46 | <annevk> | still a CSS problem |
| 21:46 | <TabAtkins> | I... don't understand. It's content. |
| 21:46 | <annevk> | HTML does not dictate how to render a cell |
| 21:46 | <TabAtkins> | It's no more a CSS problem than @rowspan is. |
| 21:47 | <annevk> | I guess I misunderstood what Hixie meant by non-rectangular |
| 21:47 | <TabAtkins> | This is just a more complicated form of combining @rowspan and @colspan. |
| 21:47 | <annevk> | sleepy time |
| 21:47 | <TabAtkins> | That's why I tried to explain it. ^_^ |
| 22:03 | <Hixie> | wow, i accidentally started a whole conversation of confusion there |
| 22:03 | <Hixie> | sorry about that |
| 22:06 | <Hixie> | hober: you editors had your meeting yet? |
| 23:07 | <UltraParadigm> | Will my webpage look right in windows xp ie if I use <header> <footer> instead of <div> |
| 23:13 | <zewt> | if you care about how your page looks in a browser, you're going to need to test in that browser :) |
| 23:15 | <UltraParadigm> | Yeah that's the best way but i think it's a pain in the butt get old version of IE. |
| 23:16 | <gsnedders> | UltraParadigm: Do you mean IE6/7/8? |
| 23:17 | <UltraParadigm> | I don't even use windows anymore except at work. We still have some xp at work u guess I can use to test. |
| 23:18 | <zewt> | if you aren't going to test in it yourself, you may as well give up, heh |
| 23:19 | <UltraParadigm> | yeah, alot of the world still uses xp unfortunatly, probably a good 30% at my work |
| 23:19 | <UltraParadigm> | Yeah gs |
| 23:20 | <UltraParadigm> | Gsnedders: ms wont let xp users upgrade to a html5 version if ie. |
| 23:21 | <gsnedders> | UltraParadigm: No, I meant, which IE versions do you care about? |
| 23:24 | <UltraParadigm> | 7 and 8. |
| 23:25 | <gsnedders> | Okay, IE8 is fine. IE7 you can fool into working by creating such elements through scripts before the actual elements in the page. |
| 23:26 | <UltraParadigm> | Hopefully everyone has upgraded past 6 by now. Although I know at my last job (cable vision tech support) they still had it. Too bad for them |
| 23:26 | <gsnedders> | <script>document.createElement("header");</script><header>foo</header> parses differently to just <header>foo</header> |
| 23:26 | <UltraParadigm> | Gsnedders: neat! |
| 23:53 | <Hixie> | aim:setstatus?message=wtf |
| 23:53 | <Hixie> | who designed this protocol |
| 23:53 | <abarth> | for real? |
| 23:54 | <Hixie> | http://webcache.googleusercontent.com/search?q=cache:http://dev.estage.aol.com/aimclient/OpenAIM182/technotes/AIMURL.html |
| 23:55 | <abarth> | AddBuddy? |
| 23:55 | <abarth> | :) |
| 23:55 | <Hixie> | also, jabber uses "jabber:" schemes for their namespace |
| 23:55 | <abarth> | is this actually used for something? |
| 23:55 | <Hixie> | abarth: apparently it's what the aim client supports |
| 23:56 | <Hixie> | i wonder how many people have yelled at the jabber folk for not buying into the distributed extensibility nonsense |
| 23:58 | <abarth> | I think we've got aim blacklisted in the external protocol handler\ |
| 23:58 | <Hixie> | seems wise |
| 23:58 | <gavinc> | Hixie: errr, they did buy into it |