| 00:11 | <MikeSmith> | Lachy: Yeah, I would think the Web interface to the dev stuff could be hosted on the same machines as www.w3.org, just using virtual hosting |
| 00:11 | <MikeSmith> | I would suspect that's already been considered and rejected for some reason, but maybe not |
| 00:11 | <MikeSmith> | I'll ask. |
| 00:12 | <Philip`> | I know exactly nothing about any of this, but I don't see why a simple caching proxy wouldn't solve most of the load problems from people requesting the latest version of the spec |
| 00:27 | <Lachy> | yay, someone contacted me about taking over the about: URI spec. |
| 00:30 | <Hixie> | wtf i broke the websocket spec script |
| 00:30 | <Hixie> | without touching any relevant parts |
| 00:40 | <roc> | Mmmm, RDF-Video. The video can just encode the semantics of what's happening, and the browser can construct a rendering from that. That would get really great compression too. |
| 00:55 | <Hixie> | roc: it'd be interesting to ask pixar what the ratio is from assets to final rendered bitmap, and how many minutes it takes before you do in fact get compression |
| 01:08 | <MikeSmith> | speaking of video, Sun seems to be pretty serious about their OMS Video project |
| 01:08 | <MikeSmith> | at least it appears that they are investing serious money to do patent review (or whatever it's called) for everything in the spec |
| 01:09 | <MikeSmith> | and their goal seems to be to make it at least as good as H.264 |
| 01:10 | <MikeSmith> | the money at stake around licensing costs is really significant |
| 01:10 | <Lachy> | MikeSmith, that's the same goal of Theora and Dirac, isn't it? |
| 01:11 | <MikeSmith> | Lachy: I suppose so |
| 01:12 | <Lachy> | I guess it's good to see more free alternatives, along with Ogg, Matroska, Theora, Dirac and Vorbis. Though having too many competing technologies can be someone counter productive |
| 01:12 | <Lachy> | *somewhat |
| 01:13 | <MikeSmith> | I think the cases where having competing technologies is productive far outnumber the cases where it's not. |
| 01:14 | <MikeSmith> | at least over the long term |
| 01:14 | <Lachy> | eek, they're also doing Project DReaM :-( http://www.openmediacommons.org/ |
| 01:15 | <MikeSmith> | one big goal of any serious codec effort should be to get widspread support for at the hardware level on devices, particularly mobile devices and other constrained devices |
| 01:15 | <Lachy> | I think in the video market, there have been far too many competing technologies, which probably had a lot to do with patents and each vendor wanting to control the market |
| 01:16 | <Lachy> | that's why we still today have few interoperable codecs and container formats supported between the major media players |
| 01:16 | <rubys> | is the root problem vendors wanting control or competition? |
| 01:18 | <Lachy> | the root is probably vendors wanting control, since that creates a lot of competiton from other vendors wanting control |
| 01:18 | <rubys> | +1 |
| 01:18 | <rubys> | ironically, the solution to such problems is often *more* competition instead of less. |
| 01:18 | <Lachy> | although, choice can be good. Different codecs are optimised for different situations |
| 01:19 | <Lachy> | just like we have PNG, GIF and JPG for different types of images, the video container formats and codecs are good for different uses |
| 01:20 | <rubys> | PNG probably wouldn't exist if it weren't for (now expired) patent concerns |
| 01:20 | <Lachy> | like Ogg, I think, is good for streaming, whereas Matroska is good as a general purpose container format |
| 01:20 | <Lachy> | PNG is superior to GIF in many ways, including lossless compression and transparency, so it probably would |
| 01:21 | <rubys> | I'm not so sure. Supperior design is an outcome of what competition produces, but if GIF were "good enough", there might not have been such an effort. |
| 01:22 | <Lachy> | in fact, the usefulness of GIF has shrunk significantly, especially in light if APNG being slowly adopted |
| 01:23 | <rubys> | And yet, GIF will likely be around for as long as RSS 2.0 is. <ducks> :-) |
| 01:23 | <Lachy> | sure, healthy competition is good. The problem is the unhealthy competition stifled by patents |
| 01:23 | <Lachy> | RSS 2.0? Is that crap still around? |
| 01:23 | <rubys> | In abundant quantities. |
| 01:24 | <Lachy> | it's use should die out over the next decade or so as CMSs get upgraded |
| 01:24 | <Lachy> | does WordPress support Atom out of the box yet? |
| 01:25 | <rubys> | yes, but it is hidden and isn't the default |
| 01:26 | <rubys> | your feed is apparently atom produced by wordpress... |
| 01:26 | <rubys> | http://feedvalidator.org/check.cgi?url=http%3A%2F%2Flachy.id.au%2Flog%2Ffeed |
| 01:27 | <Lachy> | yeah, cause I installed the Atom plugin a long time ago |
| 01:28 | <Lachy> | In fact, I only used RSS for a very brief period during my blog's lifetime. It was atom 0.3 originally before finally moving to 1.0 |
| 02:11 | <Lachy> | wow, Giovanni's reasons for wanting the spec modularised are the exact opposite of why I don't. http://lists.w3.org/Archives/Public/www-html/2009Jan/0066.html |
| 02:12 | <Lachy> | "Yes, but it is easier to read only the page containing the feature I want, then to search an heavy and difficult document, with the features I need spreaded all across and concept I may not have any idea what they mean." |
| 02:12 | <Lachy> | whereas I find the monolithic document signifcantly easier and less time consuming than trying to work out which spec, or even which page of the multipage version, I need to look at first |
| 02:16 | <Hixie> | yeah me too |
| 02:16 | <Hixie> | i hate multipage specs |
| 02:16 | <Hixie> | but we have a multipage html5 spec, anyway |
| 02:16 | <Hixie> | and soon apparently we'll have two! |
| 02:16 | <jcranmer> | I prefer multipage |
| 02:16 | <jcranmer> | actually, I prefer PDF |
| 02:16 | <Hixie> | we have that too :-) |
| 02:16 | <Lachy> | the only thing I don't like about the single page spec is the TOC script you have running on it is slow |
| 02:16 | <jcranmer> | but multipage means it's not loading a gazillion-thick page |
| 02:16 | <Hixie> | Lachy: really? |
| 02:17 | <Hixie> | Lachy: that's odd |
| 02:17 | <Hixie> | Lachy: it should be blazing fast |
| 02:17 | <Lachy> | yeah, it takes a few seconds to load in which my browser is frozen |
| 02:17 | <Hixie> | Lachy: you sure you don't mean the issue marker script? |
| 02:17 | <Lachy> | I only started noticing the delay afrter you added the TOC script |
| 02:17 | <Hixie> | there's four scripts in the html5 spec right now |
| 02:17 | <Hixie> | dfn, issues, toc, and updates |
| 02:17 | <Lachy> | but it could be the issue marker script |
| 02:17 | <Hixie> | the only slow one should be the issues one |
| 02:17 | <Hixie> | (er, status, not issues) |
| 02:18 | <Hixie> | reload now and see if it is faster |
| 02:18 | <Lachy> | PDF specs are even worse than multipage spaces |
| 02:18 | <Lachy> | *specs |
| 02:19 | <Lachy> | they're only useful if you have the insane desire to waste a ream of paper |
| 02:20 | <Hixie> | still slow? |
| 02:21 | <Lachy> | woah. the PDF is up to 718 pages now :-) |
| 02:23 | <takkaria> | I hope the HTML 5 spec gets published in paper form at some point |
| 02:23 | <Lachy> | does the TOC script run last? |
| 02:23 | <takkaria> | I want to up my carbon footprint a bit since I don't own a car |
| 02:23 | <Hixie> | Lachy: yes |
| 02:24 | <Lachy> | ok. make it run first. Let the status script go last if it takes the longest |
| 02:24 | <Lachy> | that way I can see if the TOC script is really causing the problem, or if the TOC has loaded and I still get problems |
| 02:25 | <Hixie> | i'm changing the way i run the scripts |
| 02:25 | <Lachy> | I'm also getting a script error: initAnnotations is not defined |
| 02:25 | <Hixie> | yeah i commented out the status script |
| 02:25 | <Hixie> | so if you were still seeing it slow, then it's not that |
| 02:25 | <Lachy> | ok |
| 02:29 | <Hixie> | how is it now? |
| 02:30 | <Lachy> | better |
| 02:30 | <Hixie> | pity |
| 02:30 | <Hixie> | the scripts aren't running at all right now :-) |
| 02:31 | <Lachy> | I noticed |
| 02:32 | <Hixie> | how about now |
| 02:33 | <Lachy> | it's fine |
| 02:34 | <Hixie> | cool |
| 02:34 | <Lachy> | the TOC script is missing still. Was that the culprit? |
| 02:34 | <Hixie> | it is? |
| 02:35 | <Hixie> | wfm |
| 02:35 | <Lachy> | doesn't wfm |
| 02:35 | <Hixie> | error console? |
| 02:36 | <Lachy> | works after I cleared my cache. |
| 02:37 | <Hixie> | weird |
| 02:37 | <Lachy> | that's the second time my cache has caused problems for me today. I wonder what's going on with it |
| 02:37 | <Lachy> | maybe my profile has degraded and I need to create a fresh one |
| 02:38 | <Lachy> | there's still a small delay though, but it's not too bad |
| 02:39 | <Hixie> | ok food time. bbiab. |
| 04:08 | <Lachy> | Hixie, the browsing-context-or-keyword definition doesn't allow "()" as a valid value. The parentheses appear to be used only for grouping, and I think it indicates that target="" is a valid value |
| 04:09 | <Lachy> | but I could be wrong. I find the syntax to be confusing too |
| 04:09 | <Hixie> | that's what i mean |
| 04:09 | <Hixie> | "" is not valid |
| 04:10 | <Lachy> | oh, from your email, it looked like you through target="()" was a valid value |
| 04:10 | <Lachy> | but this just proves that the syntax is confusing :-) |
| 04:10 | <Hixie> | yeah well |
| 04:11 | <Hixie> | no disagreement there |
| 04:12 | <Lachy> | it's fine for people who need such formalisms for developing tools like validators, but definitely not for average authors |
| 04:13 | <Lachy> | woah, I didn't realise that Mike had copied the first paragraph of the Audience section in HTML5 verbatim into his draft till just now |
| 04:13 | <Lachy> | I thought it was something he had written in the scope section before. That's why I wanted to see the CVS revisions so I could check what the previous draft said |
| 04:14 | <Hixie> | ah |
| 05:25 | irc.freenode.net | changes topic to 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Please leave your sense of logic at the door, thanks!' |
| 05:31 | <BenMillard> | now that krijnh is here to log it, IE8 RC1 available: http://blogs.msdn.com/ie/archive/2009/01/26/internet-explorer-8-release-candidate-now-available.aspx |
| 06:30 | <Hixie> | what do you call the function that's the first function in a call stack? that is, the one that started everything? |
| 06:30 | <Hixie> | in particular in JS |
| 06:30 | <Hixie> | would be "the program" in a process stack trace |
| 06:30 | <takkaria> | God |
| 06:30 | <Hixie> | s/stack trace/call stack/ |
| 06:36 | <heycam> | "top execution context" is a phrase in ECMA-232 |
| 06:37 | <heycam> | syntactically, the name of the production is "Program" |
| 06:38 | <heycam> | oops sorry, "top execution context" refers to one most recently put on the stack |
| 06:43 | <heycam> | hmm djokovic retired |
| 07:09 | <Hixie> | heycam: might not be the Program (e.g. from could be a callback) |
| 07:15 | <heycam> | Hixie, top call? |
| 07:15 | <heycam> | i think that's what some JS engines might call it |
| 07:18 | <Hixie> | i just called it the "first script" |
| 07:18 | <Hixie> | can always rename it later |
| 07:31 | Hixie | stares at this case that bz wants fixed: |
| 07:31 | <Hixie> | 1) Setting location.href of an about:blank iframe to a javascript: URI that then sets window.location to a relative URI. |
| 07:33 | Hixie | considers making the answer be "throw an HEADACHE_CAUSED_BY_AUTHOR_ERR" exception |
| 08:20 | <Hixie> | good lord about:blank has a lot of magic |
| 09:34 | <Hixie> | the last few days' worth of checkins are all courtesey of private e-mails from bz containing just links to bugzilla bug comments |
| 09:35 | <Hixie> | such inoccuous-looking e-mails |
| 09:40 | <Hixie> | nn |
| 09:44 | <annevk> | nn |
| 09:59 | <zcorpan> | hmm, doesn't html5 expose event handler attributes on HTMLDocument? |
| 10:02 | <annevk> | Hixie's claim is that e.g. document.onload only works in Opera |
| 10:03 | <zcorpan> | but that's because the load event doesn't bubble through document |
| 10:03 | <zcorpan> | not because document.onload doesn't exist |
| 10:04 | <zcorpan> | document.onclick seems to work everywhere? |
| 10:04 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Cscript%3Edocument.onclick%20%3D%20function(e)%20%7B%20w(e)%20%7D%3C%2Fscript%3E |
| 10:06 | <annevk> | yeah, I don't think that problem is really solved |
| 10:06 | zcorpan | files a bug |
| 10:33 | <annevk> | zcorpan, do you have IE8? |
| 10:33 | <annevk> | zcorpan, any chance you can test if both overflow-x and overflow-y still work or if they only work with a prefix? |
| 10:33 | <annevk> | (if someone else here can do it, great!) |
| 10:36 | <virtuelv> | annevk: when were you going to Oslo again? |
| 10:38 | <annevk> | mañana |
| 10:39 | <annevk> | but late, so I'll see you Thursday |
| 10:43 | <zcorpan> | annevk: they work (in RC1) |
| 10:45 | <zcorpan> | -ms-overflow-x only works in IE8 mode |
| 10:46 | <zcorpan> | unprefixed version works in quirks, IE7 and IE8 mode |
| 10:47 | <zcorpan> | hmm maybe i should test their almost standards mode some time |
| 10:48 | <zcorpan> | window.onerror is null in ie8 |
| 10:49 | <zcorpan> | what's the reason for having onerror be undefined again? |
| 10:52 | <annevk> | thanks |
| 10:52 | annevk | e-mails the CSS WG |
| 10:52 | <annevk> | zcorpan, I don't think there was one in particular |
| 11:00 | <hsivonen> | do Safari, Chrome and Opera have something like chardet? |
| 11:00 | <hsivonen> | if they do have it, do they deliberately not have UI for it? |
| 11:00 | <hsivonen> | I wonder if Firefox could get away with having chardet always on without UI for turning it off |
| 11:02 | <annevk> | we have heuristics for determining the encoding, if that is what you mean |
| 11:02 | <hsivonen> | annevk: that's what I mean |
| 11:03 | <jgraham> | Does anyone have any idea where ubuntu hides the python pstats module these days? |
| 11:03 | <hsivonen> | annevk: so the heuristics are always part of the HTTP, BOM, meta, heuristics chain? |
| 11:04 | <hsivonen> | without UI for changing the chain to HTTP, BOM, meta, windows-1252? |
| 11:05 | jgraham | discovers it is in python-profiler in multiverse despite being part of the stdlib |
| 11:05 | <hsivonen> | jgraham: non-free? |
| 11:06 | <annevk> | I'm not sure I understand encoding UI |
| 11:06 | <hsivonen> | It would be interesting to know if encoding UI is still strictly necessary |
| 11:06 | <jgraham> | annevk: Does anyone? |
| 11:07 | jgraham | guesses many browsers could reduce their encoding UI by 95% |
| 11:07 | <hsivonen> | Safari is pretty reduced, but they opted to have UI for encodings |
| 11:09 | <jgraham> | hsivonen: It seems like pstats has an odd license which probably makes it non-free indeed (but I did not think that Ubuntu were religious about such things) |
| 11:10 | <hsivonen> | jgraham: I think the level of religiosity depends on how generally visible and useful a package is |
| 11:30 | <zcorpan> | http://svn.codetalks.org/repos/trunk/tools/ARIA-DTDs/ |
| 11:30 | jgraham | swears out load as he notices html5lib on the contents list of http://diveintopython3.org/ |
| 11:31 | <jgraham> | I guess that means a Python 3 port would be a good idea |
| 11:33 | heycam | discovers that 'apt-get install git' doesn't install the right thing |
| 11:35 | <krijnh> | "This site is optimized for Lynx just because fuck you." |
| 11:42 | <annevk> | krijnh, the funny thing is that at the top it talks about a color not supported by Lynx at all :p |
| 11:42 | <annevk> | well, actually, the next sentence makes that funny as it implies Mark never looked at anything but Lynx |
| 11:42 | <krijnh> | Yeah, I just search up for 'fuck' and noticed your comment :) |
| 11:43 | <krijnh> | Next time I should search for 'fuck' before I say something :) |
| 11:51 | <Lachy> | wow, about:blank has been submitted to IETF now. http://www.ietf.org/proceedings/staging/draft-holsten-about-uri-scheme-00.txt - I wasn't expecting that to be done so quickly, especially cause Joseph listed me as an editor |
| 11:52 | <Lachy> | looks like he took what I wrote and combined it with what he had and submitted it without even letting me review the changes |
| 11:55 | <Philip`> | Lachy: Does "submitted to the IETF" mean anything more than "uploaded to their web servers as a temporary storage location for early drafts"? |
| 11:57 | <Lachy> | I don't know |
| 11:58 | <Lachy> | I have no experience with the IETF submission process |
| 11:58 | <hsivonen> | Don't submitters need to promise something about rights? |
| 11:58 | <Lachy> | I don't know |
| 12:01 | <xydyx> | hsivonen: what should I do about "Bad MD5 hash for http://archive.apache.org/dist/xerces/j/Xerces-J-bin.2.9.1.zip."; |
| 12:02 | <hsivonen> | xydyx: first you should check if the server gave you a 404 page or somethinga along those lines instead of a zip file |
| 12:03 | <xydyx> | it's a zip, 6952959 bytes |
| 12:04 | <hsivonen> | xydyx: hmm. that's interesting |
| 12:04 | <xydyx> | ETag: "cb4e73-6a17ff-43a2388b9ffc0" |
| 12:05 | <hsivonen> | it has the right length |
| 12:05 | <xydyx> | firefox only downloaded 4923392 bytes |
| 12:07 | <Philip`> | Works fine for me in wget, with MD5 a0e07ede1c3bd5231fe15eae24032b2e |
| 12:07 | <xydyx> | wget got the whole file, I'll try replacing build.py with localhost |
| 12:08 | <Philip`> | You should be able to just copy the downloaded file into build/dependencies, I think |
| 12:08 | <Philip`> | and then build.py won't try downloading it again |
| 12:09 | <Philip`> | Uh, build/../dependencies |
| 12:15 | <xydyx> | that works, thanks |
| 12:19 | <hsivonen> | I also get the right md5 with curl |
| 12:49 | <zcorpan> | http://simon.html5.org/test/html/dom/event-handler-attrs/initial-value.html |
| 13:19 | <zcorpan> | http://www.w3.org/mid/1c8dbcaa0901270455y2b5a1f09w67df621cdce86b1a⊙mgc |
| 13:20 | annevk | was hoping that was over |
| 13:21 | <jgraham> | "It will all be over when you're dead" |
| 13:22 | <annevk> | the problem is, I'm 22 |
| 13:23 | <hsivonen> | I was hoping no one would stir that issue until the group working on ATAG came up with a suggestion on what they want tools to do |
| 13:23 | <jgraham> | Curiously the first verse of the song I was quoting ends "Is it wicked not to care when you've wasted many hours talking endlessly to anyone who's there", which seems curiously apropos |
| 13:27 | <jgraham> | hsivonen: In Gez's list the consumer side of that question is addressed ("with no alt UAs have to guess which results in gibberish") but not the producer side ("in the absence of a human who cares and understands the issue well enough the producer has to guess which often results in authoratitive gibberish") |
| 13:27 | <hsivonen> | Gez's comment form doesn't do Unicode |
| 13:28 | <xydyx> | ./ant warns "You must set JAVA_HOME environment variable to point to the directory where your JDK is installed."; but the build goes on with heaps of errors then fails |
| 13:28 | <zcorpan> | hsivonen: have you let him know? |
| 13:29 | <hsivonen> | zcorpan: only by leaving comments with broken characters |
| 13:29 | <hsivonen> | xydyx: you need to set JAVA_HOME |
| 13:30 | <xydyx> | I know, but I have to scroll up a lot to see the message |
| 13:30 | <hsivonen> | xydyx: yeah, the build script sucks in many ways |
| 13:30 | <annevk> | Lachy, it seems better to not make about: depend on HTML5 as I do not think it is needed |
| 13:30 | Philip` | wonders to what extent the mathematical syntax used in theoretical computer science is determined by the availability of nice-looking symbols in LaTeX |
| 13:33 | <Lachy> | annevk, if it's going to define about:blank returns a text/html document, shouldn't it reference something? |
| 13:33 | <annevk> | there's a text/html RFC |
| 13:33 | <annevk> | which HTML5 will update in due course |
| 13:34 | <Lachy> | ok. Reply to my mail about that on the list to let Joseph know |
| 13:34 | <annevk> | sure, it just seems that without dependency on unstable drafts getting an RFC out there is considerably easier |
| 13:34 | <annevk> | it can always be updated later once everything is in place |
| 13:35 | hallvors | notes that about:blank must be a document in quirks mode with no whitespace in body except for a BR tag.. or something like that :-p |
| 13:36 | <zcorpan> | hallvors: why quirks mode? |
| 13:37 | <hallvors> | we've got bugs when editors set designMode on and apply CSS, apparently quirksmode is/was expected .. ;-] |
| 13:37 | <hallvors> | I think |
| 13:37 | <xydyx> | hsivonen: checking for "JAVA_HOME" not in os.environ in buildAll or buildJing should work |
| 13:37 | <hallvors> | but that's from memory so I may be plain wrong. Perhaps it was the other way around, even. |
| 13:38 | <annevk> | you're right, though zcorpan discovered another potential fix |
| 13:38 | <annevk> | but quirks mode seems safer |
| 13:40 | <zcorpan> | hallvors: i only recall one bug (which didn't actually use css but the result was "broken" because of p's default margins) |
| 13:41 | <zcorpan> | how would they apply css to about:blank? |
| 13:42 | <zcorpan> | oh by inserting elements with script |
| 13:42 | <hsivonen> | xydyx: thanks. recorded http://bugzilla.validator.nu/show_bug.cgi?id=439 |
| 13:43 | <zcorpan> | annevk: we might have to do change default block anyway to make google people happy |
| 13:43 | <hallvors> | zcorpan: probably the P margin issue I was thinking of.. |
| 13:45 | Philip` | wonders why Firefox requires 20% of his CPU just to display the "an upgrade is available" box |
| 13:45 | <annevk> | zcorpan, yeah, though quirks mode stays safer :) |
| 13:45 | <zcorpan> | annevk: standards mode is nicer :) |
| 13:46 | <jgraham> | zcorpan: WWhen did we ever get to do something nice in HTML? |
| 13:46 | <annevk> | zcorpan, you mean no quirks mode :p |
| 13:46 | <zcorpan> | jgraham: true... otoh, interop is nice and if we change we have interop |
| 13:50 | <Philip`> | annevk: I think you mean "slightly fewer quirks mode" |
| 13:51 | <annevk> | Philip`, not found |
| 13:51 | <Philip`> | If you removed all the quirks, HTML wouldn't even be HTML any more, it'd be an entirely different language |
| 15:31 | <annevk> | https://bugzilla.mozilla.org/show_bug.cgi?id=472529 (Web Sockets) |
| 15:36 | gsnedders | hasn't got into Imperial |
| 15:37 | zcorpan | should read the whole thread before replying :( |
| 15:40 | annevk | reminds himself to not reply to trivial questions because everyone else does it too |
| 16:25 | jgraham | is sure that making hundreds of megabytes of LaTeX a dependency for python-matplotlib is overkill |
| 16:26 | <Philip`> | You should use Gentoo, then you could set USE=-latex and it wouldn't need those dependencies |
| 16:33 | <dfjoerg> | jgraham: can you fix setup_base.py? package_dir should be 'src', not 'src/' |
| 16:34 | <dfjoerg> | jgraham: the latter fails on windows |
| 16:41 | <jgraham> | dfjoerg: Yes, but not right now. I guess I can do it in an hour or so |
| 16:41 | jgraham | thought that got fixed once before |
| 16:42 | <dfjoerg> | no need to hurry, just wanted to see it fixed for the next release |
| 17:07 | zcorpan | hates dom events |
| 17:26 | gsnedders | sighs |
| 17:26 | <gsnedders> | YouTube's new RSS feeds are invalid |
| 17:27 | <rubys1> | url? |
| 17:27 | <gsnedders> | http://www.youtube.com/rssls |
| 17:28 | <gsnedders> | <author> containing username |
| 17:30 | <rubys1> | http://feedvalidator.org/check.cgi?url=http%3A%2F%2Fgdata.youtube.com%2Ffeeds%2Fbase%2Fvideos%2F-%2Fmonkey%3Fclient%3Dytapi-youtube-browse%26v%3D2 |
| 17:31 | <rubys1> | drop alt=rss |
| 17:54 | <Philip`> | LVM scares me |
| 17:54 | <Philip`> | The problem is that I only want to use it once every six months or so, so I never really learn how it works |
| 17:55 | <Philip`> | but then occasionally I want to expand a disk onto a RAID volume and don't want to accidentally destroy all my data |
| 18:32 | Philip` | discovers a 32GB partition, and has no idea where it came from or who made it |
| 18:32 | <Philip`> | (and it doesn't seem to be mounted anywhere) |
| 18:33 | <dfjoerg> | secret pr0n collection of your gf? |
| 18:34 | <Philip`> | No, it's my work computer :-p |
| 18:34 | <Philip`> | and I don't think the sysadmins would store that kind of thing on random users' machines |
| 19:49 | <Philip`> | Hmph, silly dmoz.org having 11667 URLs from newadvent.org |
| 19:50 | <Philip`> | vs 13010 from wikipedia.org |
| 21:15 | <alyoshka> | for some reason I had trouble getting to this channel, so I asked a question on W3C's html-wg irc channel, but this channel is more active usually, so I guess I can ask here and the message on html-wg can be ignored |
| 21:15 | <alyoshka> | What is the <figure><legend> supposed to look like by default? does it look like the <fieldset><legend> or does it have the caption(legend) at the bottom like the classic photography style? |
| 21:16 | <Hixie> | we don't know yet |
| 21:16 | <Hixie> | :-) |
| 21:17 | <alyoshka> | uhh...that leaves me with a problem on how to fix it in older browsers.lol |
| 21:18 | <alyoshka> | well, is there any special behavior that current graphical browsers don't have for it? Last time I checked, there wasn't, but maybe I need to look again |
| 21:19 | <alyoshka> | I always imagined it looking like an inline-block element with the caption under the image |
| 21:43 | <Hixie> | alyoshka: it should be block-level, but yeah, other than that that's basically it |
| 21:43 | <Hixie> | alyoshka: probably legend above if the legend is at the top, and below otherwise |
| 21:44 | <Hixie> | you could center the resource and the legend, too |
| 21:44 | Philip` | 's web page collection is far too old :-( |
| 21:44 | Philip` | tries downloading another hundred thousand |
| 21:45 | <Hixie> | so probably, like: figure, figure > legend { text-align: center; margin: 0 auto; display: block; } |
| 21:46 | <alyoshka> | OK, makes sense |
| 21:48 | <alyoshka> | for default settings. and I figure the legend would be more of plain text rather than the way <fieldset><legend> works |
| 21:48 | <Hixie> | yeah, no default borders or anything |
| 21:51 | <alyoshka> | what about <details><legend>? no borders, but just open/close action for all contents, right? |
| 21:53 | <alyoshka> | <details><legend>You don't...</legend> want to know the details</details> would render "You don't..." without a style by default except maybe looking like a hyperlink? |
| 21:58 | <Hixie> | are you familiar with the Mac OS X "disclosure triangle" widget? |
| 21:59 | <Hixie> | that's what I had in mind when speccing <details> |
| 22:06 | <alyoshka> | um, I think I know what you're talking about, Gnome has the same thing too (if I understood what you're talking about) |
| 22:06 | <alyoshka> | I'm primarily a KDE user tho |
| 22:07 | <ap> | Hixie: hmm, is noupdate event no longer dispatched in the main resource was loaded from cache? is this intentional? |
| 22:08 | <Hixie> | ap: not sure, send e-mail -- i'm looking at this stuff this afternoon |
| 22:08 | <ap> | k |
| 22:33 | <Hixie> | so does anyone know what IE8's "clickjacking protection" consists of? |
| 22:35 | <alyoshka> | apparently it's a tag a developer puts in the document that prevents IE8 from ever framing the page |
| 22:36 | <Hixie> | i was hoping for someone who knew what the "tag" was :-) |
| 22:38 | <heycam> | https://bugzilla.mozilla.org/show_bug.cgi?id=475530 claims it's actually the "X-I-Do-Not-Want-To-Be-Framed-Across-Domains" header |
| 22:39 | <heycam> | should be simple to test, i guess |
| 22:40 | <Hixie> | that doesn't solve clickjacking, unless you consider "not having a car" to be a solution to "it's possible for someone to break into my car" |
| 22:40 | <Hixie> | if it is what they did, though, i'm amused -- that was originally a Google proposal on the WHATWG list |
| 22:41 | <heycam> | and presumably a strawman header name |
| 22:42 | Philip` | 's new download storage format is sufficiently less dumb that his old format that it now takes 15 seconds to scan through 130K pages, rather than ten minutes |
| 22:43 | <alyoshka> | such a funny name for the header |
| 22:43 | <Philip`> | ...although I appear to have a page which sends the Validator.nu parser into either an infinite or a very slow loop |
| 22:43 | <Hixie> | http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Jan/0132.html "Most RDFa parsers are able to handle broken XHTML" ... |
| 22:43 | <alyoshka> | I hate having pages that needed to have something pasted in them every time |
| 22:44 | <alyoshka> | one reason I hated the XHTML doctypes |
| 22:44 | <alyoshka> | it's very easy to get a typo when typing that into the page, so I would usually paste it |
| 22:45 | <Philip`> | Hixie: Who needs parsing specs when you can just pass everything through Tidy or whatever other tool you happen to have lying around? |
| 22:45 | <alyoshka> | but with the header it's more important not to make a typo, since it actually does something, unlike the doctype |
| 22:56 | <annevk> | broken XHTML, is that like invalid XHTML? |
| 22:56 | <annevk> | please don't answer, I'm being needlessly pedantic |
| 22:58 | Philip` | likes how his Java stuff now uses ~390% CPU, rather than being constantly blocked on IO |
| 22:58 | <Philip`> | (but I don't like how it makes it hard for me to find this maybe-infinite loop :-( ) |
| 22:58 | <Hixie> | can't you figure out which page is doing it? |
| 23:01 | <Philip`> | I'm trying to, but it's taking me far too much effort |
| 23:02 | <Philip`> | My current attempt is to set each thread's name to the URL it's currently processing, and then I can use ctrl-\ to print a stack trace of all the threads once it's got to the end and has one stuck |
| 23:03 | <Hixie> | just run them all on one thread and output the url before each one |
| 23:03 | <Philip`> | That would probably be sensible, but it'd take much longer to run |
| 23:03 | <Philip`> | Found it now, so that's okay |
| 23:12 | <annevk> | Lachy, what is the normalization issue with Selectors? |
| 23:12 | <annevk> | I suppose I can imagine some myself |
| 23:12 | annevk | hopes there is no normalization whatsoever |
| 23:24 | <Philip`> | Hmm, looks like the page's body randomly stopped and then was followed by a quarter of a megabyte of null bytes |
| 23:24 | <Philip`> | (a quarter of a megabyte being the limit I imposed on the response body lengths) |
| 23:25 | <Philip`> | and that happened to be in the middle of a table just after a </tr> |
| 23:25 | <Philip`> | so it wants to foster-parent a quarter of a million characters |
| 23:25 | <Philip`> | and presumably there's some non-linear behaviour in there |
| 23:26 | <annevk> | interesting testcase |
| 23:26 | <Philip`> | The actual page itself looks fine, it might have just been a weird artifact of my downloader |
| 23:28 | <Philip`> | hsivonen: It'd be nice if foster-parenting didn't take seemingly non-linear amounts of time in SAXTreeBuilder.insertCharactersBefore |
| 23:30 | <Hixie> | All U+0000 NULL characters in the input must be replaced by U+FFFD REPLACEMENT CHARACTERs. Any occurrences of such characters is a parse error. |
| 23:31 | <Philip`> | hsivonen: Test case: http://philip.html5.org/misc/table-fostering.html shouldn't take multiple seconds to parse with HtmlParser(XmlViolationPolicy.ALLOW) |
| 23:32 | <takkaria> | 23:35 < Hixie> All U+0000 NULL characters in the input must be replaced by U+FFFD REPLACEMENT CHARACTERs. Any occurrences of such characters is a parse error. |
| 23:32 | <takkaria> | oops, sorry |
| 23:36 | <Philip`> | Urgh, looks like dozens of pages are stuff with nulls :-( |
| 23:48 | annevk | grmbls in the general direction of HTTP |
| 23:51 | Philip` | learns that thread contention hurts performance |
| 23:51 | annevk | leaves it at that for tonight |
| 23:51 | <annevk> | nn |
| 23:51 | <Philip`> | It took over two minutes to analyse these pages with 64 threads, but 75 seconds with 5 threads |
| 23:51 | <Philip`> | (on a quad-core machine) |
| 23:52 | <Philip`> | (64 threads is useful when you're downloading from the web, though) |
| 23:58 | <Philip`> | http://pypi.python.org/pypi/rdfadict |
| 23:58 | <Philip`> | html5lib seems to be getting around a bit |