| 00:05 | <Philip`> | ment: There aren't attributes in closing tags |
| 00:05 | <ment> | but it's not a parser error either |
| 00:06 | <annevk2> | ment, it is |
| 00:06 | <Philip`> | The spec pretends end tags have attributes so it doesn't have to define a whole new set of states for parsing end tags, if I remember correctly |
| 00:06 | <ment> | DS -> TOS -> CTOS -> TNS -> ... (no error reported along the chain) |
| 00:06 | <Philip`> | but they have no effect other than causing a parse error |
| 00:07 | <ment> | but when does this error reporting occur? |
| 00:07 | <Philip`> | ment: "When an end tag token is emitted with attributes, that is a parse error." |
| 00:07 | <Philip`> | (in the tokenizer intro section) |
| 00:07 | <ment> | ah, it's hidden in the intro |
| 00:07 | <ment> | thanks |
| 00:08 | <Philip`> | Indeed, if by "hidden" you mean "visible and explicitly stated in its own paragraph" ;-) |
| 00:09 | Philip` | does find it easy to lose track of indirect requirements, though |
| 00:23 | <GPHemsley> | Hixie: The most recent one ;) |
| 00:24 | <GPHemsley> | Hixie: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-October/023342.html and its responses |
| 00:43 | <Hixie> | GPHemsley: i remember replying to that, but i can't find the reply... looking still... |
| 00:52 | <Hixie> | weird |
| 00:52 | <Hixie> | i found the e-mail in my sent folder |
| 00:52 | <Hixie> | but it never made it to the list |
| 00:57 | <Hixie> | GPHemsley: resent it |
| 00:58 | <Hixie> | very weird, messages on either side of it made it fine |
| 01:04 | <roc> | Lachy: you might want to read the comments here http://weblogs.mozillazine.org/bz/archives/020267.html |
| 03:10 | <Hixie> | aw, shelley, you're always a member of the whatwg. Everyone's a member of the whatwg! We're an all-inclusive community of people who disagree with each other. :-) |
| 03:11 | <TabAtkins> | So everyone's a member whether they like it or not? ^_^ |
| 03:12 | <othermaciej> | you don't want to join a club that would have you as a member? |
| 03:27 | <Hixie> | TabAtkins: yup :-P |
| 03:28 | <Hixie> | TabAtkins: even people who don't know about the whatwg :-P |
| 03:28 | <TabAtkins> | I will inform them! |
| 03:28 | <Hixie> | excellent, thanks! |
| 03:54 | <GPHemsley> | Hixie: So are you saying that the <cite> element actually has an implied name of <worktitle>? |
| 04:04 | <The_Madcapper> | Could someone please point me to the Javascript channel? :) Is the channel name just "javascript" or is it something special? |
| 04:12 | <GPHemsley> | Also, is the WHATWG so named for the plausible deniability? ("What working group? I don't know what you're talking about.") |
| 04:29 | <Hixie> | GPHemsley: no implied name, but the meaning of the element in html5 is basically "title of work", yes (as per the spec) |
| 04:30 | <GPHemsley> | well, you said that if you had your way it wouldn't be there, and that the name is legacy from HTML4, so figuratively implied, no? |
| 04:30 | <GPHemsley> | (i.e. informatively implied) |
| 04:30 | <Hixie> | not really sure what you mean by an "implied name" |
| 04:31 | <GPHemsley> | implied as in, by "cite", you really mean "worktitle" |
| 04:32 | <Hixie> | the meaning of <cite> in html5 is (currently) "title of work", yes |
| 04:33 | <GPHemsley> | but apparently only in certain contexts? |
| 04:33 | <Hixie> | no? |
| 04:33 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/multipage/text-level-semantics.html#the-cite-element |
| 04:33 | <GPHemsley> | you were giving examples where a title of a work doesn't get <cite>d |
| 04:34 | <Hixie> | i was? |
| 04:34 | <GPHemsley> | I thought |
| 04:34 | <GPHemsley> | hang on |
| 04:34 | <Hixie> | sorry, it's been a while since i wrote that e-mail :-) |
| 04:35 | <Hixie> | ...which still hasn't made it to the archives, wtf |
| 04:35 | <Hixie> | oh i see, it went in at the original date |
| 04:35 | <Hixie> | how confusing |
| 04:35 | <GPHemsley> | oh, maybe you were referring to the old definition |
| 04:35 | <GPHemsley> | oh, it was both |
| 04:35 | <GPHemsley> | here we go |
| 04:35 | <GPHemsley> | : |
| 04:35 | <GPHemsley> | > That leaves usages of 'cite' for both titles of works and authors of |
| 04:35 | <GPHemsley> | > works. Putting aside the issue of styling for a moment, these two pieces |
| 04:35 | <GPHemsley> | > of data both fall under the semantic meaning of 'cite'. |
| 04:35 | <GPHemsley> | No, not at all. Consider: |
| 04:35 | <GPHemsley> | <p>I met Ian today. He was carrying The New Hacker's Dictionary.</p> |
| 04:35 | <GPHemsley> | This contains no citation, but it contains both the name of an author and |
| 04:36 | <GPHemsley> | the title of a work. |
| 04:36 | GPHemsley | just realized that "semantic meaning" is rather redundant |
| 04:37 | <Hixie> | per html5, you can label the title in that paragraph with a <cite>. |
| 04:37 | <GPHemsley> | right |
| 04:37 | <GPHemsley> | and I agree that you should |
| 04:37 | <Hixie> | but it's not a citation |
| 04:38 | <GPHemsley> | true |
| 04:39 | <GPHemsley> | but, and this is perhaps confusing, in the context of my argument, I was (I think) putting aside the issue of "a citation" and just addressing "what can be cited" |
| 04:40 | <GPHemsley> | "sing it as a verb, the |
| 04:40 | <GPHemsley> | definition of 'cite' applies to quotes/quotations, titles, and people, |
| 04:40 | <GPHemsley> | depending on the context." |
| 04:40 | <GPHemsley> | +U |
| 05:22 | <GPHemsley> | Hixie: Is there an updated timeline beyond the FAQ, given today's Last Call? |
| 05:23 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/TIMETABLE is the only timeline so far |
| 05:24 | <GPHemsley> | oh, so, everything's on schedule then? |
| 05:24 | <Hixie> | at least in the WHATWG! |
| 05:24 | <GPHemsley> | heh |
| 05:24 | <Hixie> | The W3C's timetable... not so much |
| 05:25 | <GPHemsley> | I'm leaving the WHATWG as the definitive source for HTML5 ;) |
| 05:25 | <GPHemsley> | is it conceivable that work on HTML6 will begin before REC for HTML5? |
| 05:27 | <GPHemsley> | Also, what's the plan for the test suite(s)? |
| 05:27 | <GPHemsley> | in terms of contributing, etc. |
| 05:43 | <mpilgrim> | http://diveintohtml5.org/semantics.html |
| 05:43 | <mpilgrim> | please stop changing the semantics now |
| 05:44 | <othermaciej> | ah, almost standards mode |
| 05:52 | <zcorpan_> | ok now i've written a bunch of tests for script parsing and the other r|cdata elements |
| 05:54 | <zcorpan_> | http://code.google.com/p/html5lib/source/browse/testdata/tree-construction/tests16.dat |
| 05:54 | <zcorpan_> | hope they're correct |
| 05:57 | <Hixie> | GPHemsley: very conceivable |
| 05:58 | <Hixie> | GPHemsley: no plan really yet... if you want to be in charge of the test suite, it's definitely something that needs doing :-) |
| 05:58 | <GPHemsley> | Hixie: Is there a scratchpad for what might be in HTML6? |
| 05:58 | <GPHemsley> | Hixie: Heh. That seems like quite a task to take on. |
| 05:59 | <erlehmann> | zcorpan_, reading that, all these <!--<script --> cases seem pretty counterintuitive |
| 05:59 | <erlehmann> | :( |
| 05:59 | <Hixie> | GPHemsley: search for "v2" in the source |
| 05:59 | <Hixie> | erlehmann: yeah no kidding |
| 06:00 | <erlehmann> | hmm, if there is a line break between <!-- and <script> or <style>, is it okay then? |
| 06:00 | <GPHemsley> | Hixie: Someone should make a script that scrapes them out. :) |
| 06:01 | <erlehmann> | or doesn't that affect parsing at all? |
| 06:01 | GPHemsley | runs and hides |
| 06:01 | <GPHemsley> | alright, I think it's time for bed |
| 06:02 | <Hixie> | nn |
| 06:02 | <GPHemsley> | g'night/g'morning :P |
| 06:02 | <Hixie> | erlehmann: hm? |
| 06:07 | <zcorpan_> | erlehmann: doesn't affect parsing |
| 06:08 | <erlehmann> | i see. thank you. |
| 06:15 | <zcorpan_> | mpilgrim: there are an infinite number of doctypes that trigger standards mode |
| 06:17 | <othermaciej> | so what's the stylistically preferred way to write the HTML5 doctype in a document? |
| 06:18 | <othermaciej> | I always write <!DOCTYPE html> but maybe all-lowercase would be more elegant |
| 06:18 | <zcorpan_> | i prefer <!doctype html> because i don't like holding down shift |
| 06:19 | <othermaciej> | the WHATWG copy of the HTML5 spec has it as <!DOCTYPE html> |
| 06:20 | <othermaciej> | google's homepage does <!doctype html> |
| 06:24 | <othermaciej> | I <3 Professor Markup |
| 06:26 | <MikeSmith> | zcorpan_: any opinion about http://bugzilla.validator.nu/show_bug.cgi?id=488 (add obsolete attributes to assertions checking) |
| 06:26 | <JoePeck> | othermaciej: probably lowercase, since it might benefit from gzipping html tags that are lowercase |
| 06:27 | <erlehmann> | othermaciej, i bet if you ask hixie he will say that its okay to be inconsistent in examples, to show what is possible ;) |
| 06:27 | <JoePeck> | erlehmann: that is so true =) |
| 06:27 | <othermaciej> | erlehmann: I'm thinking about good style for documents that I write for purposes other than markup pedagogy |
| 06:27 | <MikeSmith> | zcorpan_: I seem to remember having a discussion about that with you, that you reckoned it would be good to explicitly report about obsolete attributes |
| 06:28 | <othermaciej> | I like to use good style if I can |
| 06:28 | <erlehmann> | othree_, all lowercase. uppercase tags are sooooo 1997 |
| 06:30 | <zcorpan_> | MikeSmith: yeah, i think it would be nice |
| 06:36 | <Hixie> | personally i prefer <!DOCTYPE HTML> |
| 06:38 | <Hixie> | MikeSmith: we're going to have to start giving everyone who comments or files on bugs the rights to reopen the bugs, because i've gotten very close to missing bugs because the commentor couldn't reopen it |
| 06:39 | <MikeSmith> | Hixie: OK |
| 06:40 | <Hixie> | ( http://www.reddit.com/r/programming/comments/9yejl/whats_the_biggest_programming_mistake_youve_ever/c0f13p2 ) |
| 06:42 | <MikeSmith> | zcorpan_: OK, I will try to get it added today |
| 06:42 | <zcorpan_> | MikeSmith: cool |
| 06:51 | <othermaciej> | mpilgrim__: IE's DOM s even weirder than what you say - I think it also makes an empty element with a tag name of "/article" |
| 06:57 | <zcorpan_> | mpilgrim__: "To ease migration to and from XHTML, authors may specify an attribute in no namespace with no prefix and with the literal localname "xml:lang" on <html> elements" - should be "HTML elements" |
| 07:33 | <hsivonen> | Spotify sure self-updates often. I wonder what the Web would be like if people got new browsers as often. |
| 07:35 | <erlehmann> | we'll see |
| 07:54 | <zcorpan_> | mpilgrim__: the first article example probably ought to set display:block |
| 07:57 | <MikeSmith> | hsivonen: I think you might consider setting default heapSize for v.nu to 128 |
| 07:59 | <hsivonen> | MikeSmith: ok |
| 08:00 | <MikeSmith> | hsivonen: at least in the case of qa-dev, I can't run v.nu on that machine at all unless I feed it heap=128 |
| 08:00 | <jgraham> | hsivonen: I thought spotify just self-updated to make it harder to hack somehow, not for major functionaility chnages |
| 08:00 | <hsivonen> | jgraham: oh. |
| 08:01 | <hsivonen> | jgraham: recently it has added the ability to download playlists for offline use |
| 08:01 | <hsivonen> | jgraham: and it has added a category for "Purchased" (no idea how to purchase) |
| 08:01 | <hsivonen> | dunno what the update this morning was about |
| 08:02 | <hsivonen> | if it's just for DRM, that sucks |
| 08:02 | <hsivonen> | can't one already "hack" spotify by using the paid-account-only Open Source Linux client? |
| 08:03 | <hsivonen> | or is "hacking" not about DRM of the content but about using the system without ads *and* without paying? |
| 08:03 | <hsivonen> | MikeSmith: fixed |
| 08:04 | <MikeSmith> | thanks |
| 08:05 | <zcorpan_> | mpilgrim__: "and everything else on the page will stay where it is in the document outline." is not true if you're using implied sections too |
| 08:05 | <MikeSmith> | http://qa-dev.w3.org:8888/?doc=http://dev.w3.org/html5/tests/validation/full/invalid/obsolete/longdesc.html |
| 08:05 | <MikeSmith> | zcorpan_, hsivonen : ↑ |
| 08:05 | <MikeSmith> | ideally, we want only one error message for obsolete attributes, right? (not two) |
| 08:06 | <hsivonen> | MikeSmith: yes |
| 08:06 | <hsivonen> | is longdesc obsolete-but-conforming these days? |
| 08:06 | <zcorpan_> | no |
| 08:06 | <MikeSmith> | right, obsolete and not conforming |
| 08:07 | <hsivonen> | but yeah, legacy.rnc should allow it and the obsolete checker should disallow it |
| 08:07 | <MikeSmith> | hsivonen: OK, I'll make the changes to legacy.rnc |
| 08:08 | <hsivonen> | thanks |
| 08:08 | <MikeSmith> | but before I looked back at this today, I hadn't realized how many obsolete-nonconforming attributes there are |
| 08:09 | <MikeSmith> | about obsolete-but-conforming vs. obsolete-and-noconforming, I'm wondering what we can say in the error messages to make it more clear |
| 08:09 | <MikeSmith> | because the error messages for both cases just say, "The foo element|attribute is obsolete. [Use bar instead.]" |
| 08:09 | Hixie | thinks that's ideal :-) |
| 08:10 | zcorpan_ | agrees with Hixie |
| 08:10 | <MikeSmith> | OK |
| 08:11 | <MikeSmith> | anyway, the UI does clearly indicate that first case is a warning and the second and erro |
| 08:11 | <MikeSmith> | *error |
| 08:12 | <zcorpan_> | MikeSmith: maybe you could point to or use text from http://wiki.whatwg.org/wiki/Presentational_elements_and_attributes in some messages? |
| 08:13 | hsivonen | finds it odd that Opera has this new-found interest in @longdesc when they could have implemented @aria-describedby in accordance with the WAI Consensus and skipped @longdesc. |
| 08:13 | <MikeSmith> | zcorpan_: I guess we could experiment with that at least |
| 08:15 | <hsivonen> | grr. A Web site that I wanted to use as a test case is down for maintenance. |
| 08:15 | <hsivonen> | people should keep test cases up |
| 08:16 | <hsivonen> | speaking of house insulation, I should guess a preference between two hotels in California. I know one of them has bad but bearable insulation and I don't know about the winter characteristics of the other. |
| 08:18 | <MikeSmith> | extra blankets |
| 08:19 | <MikeSmith> | hsivonen: btw, I still want to get back to trying to implement the datatypes warning mechanism. I have an idea about what to do in the datatype code itself, but I am realizing that I don't understand the error-handling backend well enough yet. For example, I don't know how to get the backend to actually emit a warning for DatatypeException instead of an error. |
| 08:20 | <hsivonen> | MikeSmith: I suggest adding an IsWarning() method to Html5DatatypeException |
| 08:23 | <hsivonen> | MikeSmith: hmm. the design of MessageEmitterAdapter will make this ugly |
| 08:23 | <hsivonen> | totally ugly |
| 08:23 | MikeSmith | goes to take another look |
| 08:24 | <hsivonen> | you have a change to do things with Html5DatatypeException a few levels of method calls down from the point where it's decided if it's going to be an error or a warning |
| 08:24 | <hsivonen> | s/change/chance/ |
| 08:25 | <MikeSmith> | OK |
| 08:27 | <MikeSmith> | I'm wondering how many cases we might really need this for. So far, I remember it being the messages for grandfathered and redundant language tags that are deprecated |
| 08:27 | <MikeSmith> | maybe there's not enough others to make it worthwhile to add this |
| 08:28 | <MikeSmith> | but end users would at least lose those two if the change isn't made |
| 08:28 | <hsivonen> | MikeSmith: so in the method called message() in MessageEmitterAdaptor, you need to examine the the Expection object and go change the MessageType object if you decide it's going to be downgrade to a Warning |
| 08:28 | <hsivonen> | MikeSmith: now that I've figured out what needs to be done, I could write the code while I'm at it |
| 08:29 | <MikeSmith> | aha |
| 08:29 | <MikeSmith> | well, I can try to in the mean time |
| 08:29 | <MikeSmith> | that last thing you said was the one bit of clue I really needed |
| 08:30 | <hsivonen> | MikeSmith: ok. that works for me, too. |
| 08:30 | <hsivonen> | the check should go as the the first thing in message() even before the if (loggingOk) bit |
| 08:31 | <MikeSmith> | OK, right after line 545 you mean, right? |
| 08:31 | <hsivonen> | yes |
| 08:32 | <MikeSmith> | OK, got it |
| 08:34 | <hsivonen> | MikeSmith: I think the right check is that the message gets downgraded to warning if all the Html5DatatypeExceptions wrapped by the main exception are warnings |
| 08:35 | <hsivonen> | and all there are no non-Html5 DatatypeExceptions there |
| 08:35 | <MikeSmith> | OK |
| 08:36 | <hsivonen> | hmm. are there cases where the schema says "" | w:something-that-emits-a-warning |
| 08:36 | <hsivonen> | what I just said could be wrong in that case |
| 08:36 | <hsivonen> | not sure |
| 08:36 | <hsivonen> | doh. |
| 08:37 | <hsivonen> | I think you need to try what Jing gives you in that case |
| 08:37 | <MikeSmith> | hmm, OK |
| 08:38 | <MikeSmith> | I hadn't noticed any of those before |
| 08:38 | MikeSmith | goes to look now |
| 08:40 | <nessy> | congrats on "Last Call"! |
| 08:40 | <nessy> | I still hope we can get video accessibility into this all before W3C decides to put html5 at last call :-) |
| 08:41 | <Hixie> | we have video accessibility already |
| 08:41 | <MikeSmith> | hsivonen: hmm, I'm confused by 'there cases where the schema says "" | w:something-that-emits-a-warning'.. can you give me an example? |
| 08:42 | <Hixie> | nessy: you mean new features for explicit external subtitles, and the like |
| 08:42 | <hsivonen> | MikeSmith: the lang attribute can take string "" or the language tag datatype |
| 08:43 | <hsivonen> | MikeSmith: so when you have an obsolete language tag, Jing will fail on both string "" and the language tag datatype but the latter says it's only a warning |
| 08:43 | <MikeSmith> | hsivonen: ah, I see now |
| 08:43 | <MikeSmith> | yeah |
| 08:45 | hsivonen | doesn't like debugging stuff by sending printfs to a remote machine and waiting for a couple of hours to get the printout back |
| 08:46 | <othermaciej> | hsivonen: that doesn't sound fun! |
| 08:51 | <hsivonen> | well in the old day, they carries decks of FORTRAN cards to an operator through wind and thick snow to even get a compiler error |
| 08:51 | <hsivonen> | *days |
| 08:53 | <zcorpan_> | mpilgrim__: shouldn't the <time> be inside <header>? |
| 09:01 | <zcorpan_> | mpilgrim__: maybe you should make it clearer that one doesn't have to use h1, and that since software haven't been updated to support the sectioning elements yet, it might be a good idea to continue using lower ranked headings for a while longer |
| 09:23 | hsivonen | tries to figure out a sane way to report line numbers with document.write |
| 09:24 | <hsivonen> | presumably, if the whole doc is document.written, the document.writes from outside the document itself should form a coherent succession of line numbers just like the network stream |
| 09:24 | <gsnedders|work> | You said that yesterday. |
| 09:25 | <hsivonen> | gsnedders|work: still trying |
| 09:25 | <hsivonen> | document.write is trouble in so many different ways |
| 09:26 | <gsnedders|work> | Could you not just report the line on which the document.write call is? |
| 09:26 | <othermaciej> | when I first learned about document.write it totally blew my mind with its craziness |
| 09:26 | <othermaciej> | now it seems totally normal to me |
| 09:26 | <othermaciej> | I have totally forgotten how weird it is |
| 09:26 | <hsivonen> | gsnedders|work: I want the tokenizer and the tree builder not change their behavior |
| 09:26 | <hsivonen> | gsnedders|work: so document.written stuff will increment the line number |
| 09:27 | <gsnedders|work> | I think when I first came across it I was already familiar enough with the web platform to just think, "Oh, that's kinda screwy. Oh well." |
| 09:27 | <hsivonen> | but I need the line number to roll back when the stream reaches a point that was written by an outer script |
| 09:28 | <hsivonen> | hmm. I think I'll make the buffer objects that represent the stream have a start line number that can be -1 |
| 09:28 | <othermaciej> | it's really dismaying that I feel more awake and productive at 2 AM than 2 PM in my local time zone |
| 09:28 | <othermaciej> | good night folks |
| 09:28 | <hsivonen> | nn |
| 09:29 | <hsivonen> | so before document.write inserts to the list of buffers, the current line number well be stored into the the then-first buffer object |
| 09:29 | <hsivonen> | and new buffer objects start out with -1 |
| 09:29 | <hsivonen> | which means don't change the line number |
| 09:30 | <hsivonen> | and otherwise when a buffer with a non--1 line number is reached, the line number of the tokenizer is reset to that number |
| 09:30 | <hsivonen> | I wonder if that makes sense |
| 09:31 | <annevk> | hsivonen, seems you were right after all about the weird pubdate: http://www.w3.org/News/2009.html#entry-8641 |
| 09:32 | gsnedders|work | realizes he hasn't actually looked at the new W3C site yet |
| 09:32 | hsivonen | wonders why there needs to be a "standardized architecture and API for XML and HTML interaction formats implemented in Javascript" |
| 09:32 | <zcorpan_> | gsnedders|work: you know what's wrong with runner.html? |
| 09:32 | <hsivonen> | I guess reading the report would tell me |
| 09:32 | <gsnedders|work> | zcorpan_: A lot |
| 09:32 | <gsnedders|work> | zcorpan_: wrt to query? |
| 09:33 | <zcorpan_> | gsnedders|work: yeah |
| 09:33 | <gsnedders|work> | zcorpan_: The query string makes it run only the TC with that sha1 hash of input, and makes all others pass. don't ask. I just needed a simple way to see if one test pass/failed. |
| 09:33 | <annevk> | hsivonen, sounds like the story from last year's TPAC |
| 09:33 | <gsnedders|work> | zcorpan_: But, please go ahead and make it saner. |
| 09:34 | jgraham | doesn't even understand what an " XML and HTML interaction formats |
| 09:34 | <jgraham> | implemented in Javascript" is |
| 09:34 | <jgraham> | s/formats/format/ |
| 09:35 | <hsivonen> | annevk: do you mean the story Raman and TimBL talked about during the panel or something else? |
| 09:35 | <zcorpan_> | gsnedders|work: i don't understand what sha1 is doing :) |
| 09:35 | <AryehGregor> | jgraham, that makes two of us. |
| 09:37 | <Philip`> | jgraham: It seems to be about writing JS libraries to implement markup features that browser vendors won't implement |
| 09:38 | <Philip`> | like XForms and SMIL |
| 09:38 | <jgraham> | SMIL seems like a poor example since it seems to have some traction |
| 09:39 | <jgraham> | But anyway... yeah, it is unclear why that demands a standard |
| 09:40 | <hsivonen> | jgraham: it doesn't for consumer interop |
| 09:40 | <hsivonen> | jgraham: but it could for authoring tool interop |
| 09:40 | <hsivonen> | jgraham: or interop of certified backplane engineers |
| 09:41 | <annevk> | maybe it was that |
| 09:42 | <annevk> | I remember talk about standardized ways to extend browsers with support for new markup languages through JavaScript |
| 09:42 | <Philip`> | jgraham: You might use lots of different JS libraries implementing features on the same page, and they might conflict and not work nicely together |
| 09:43 | <hsivonen> | annevk: isn't it called XBL2? |
| 09:43 | <Philip`> | (http://code.google.com/p/ubiquity/ seems quite relevant) |
| 09:43 | <hsivonen> | Philip`: why does Ubiquity XForms need anything to be standardized? |
| 09:43 | <hsivonen> | (that isn't already, that is) |
| 09:45 | <annevk> | hsivonen, prolly |
| 09:45 | <Philip`> | hsivonen: Because people make Ubiquity SMIL and Ubiquity RDFa etc, and the extensions might interact poorly unless they follow certain conventions, and the Ubiquity project tries to tie those ones together but third parties might want to use the same mechanism and need a stable standard |
| 09:46 | <Philip`> | I guess |
| 09:46 | <Philip`> | having only briefly glanced at the document |
| 09:46 | <jgraham> | And then a random piece of javascript might come along and mess up the whole thing anyway so you didn't gain much |
| 09:47 | <annevk> | hmm, CSSStyleDeclaration is not really interoperable |
| 09:47 | <annevk> | and also somewhat confusing |
| 09:47 | <Philip`> | jgraham: It's always easy to make a page stop working - what you gain is the ability to make it first start working without so much pain |
| 09:48 | <annevk> | I doubt anyone really sees property priority is separate from property value though it does make some sense I guess |
| 09:49 | <jgraham> | Philip`: But it seems much more likely that a user will be using a js-implemented markup extension and some other random javascript code than that they will be using multiple js-markup extensions |
| 09:49 | <jgraham> | and it seems much easier to simply test the combination of known markup extensions than test markup extensions + all other code |
| 09:51 | <Philip`> | jgraham: Why are users relevant? Authors are in control of what scripted markup extensions are in use on their pages, because they write <script src=ubiquity.js?xforms,smil,rdfa> or whatever |
| 09:51 | <Philip`> | and they can choose to only include scripts which work together nicely |
| 09:52 | <Philip`> | (and so it's useful making more scripts work together nicely) |
| 09:55 | <jgraham> | Philip`: By users I mean "users of the extensions i.e. authors" |
| 09:58 | <Lachy> | is there anyone here who knows anything about american cars? I'm wondering whether a dodge charger would be a reasonable choice for a rental car when we arrive in SF this week |
| 09:59 | <Lachy> | we need it to be roomy enough for 3 of us, and powerful |
| 10:00 | jgraham | notes that powerful sounds like a solution not a problem statement, making it harder to evaluate if it actually fits your needs :) |
| 10:00 | jgraham | also has no idea about cars |
| 10:01 | Philip` | notes that powerful sounds more like a desire than a need |
| 10:02 | <Dashiva> | It sounds like Lachy is turning into a manager |
| 10:02 | <annevk> | he's managing the car part of our trip |
| 10:02 | <jgraham> | Oh except that a surprising number of Swedish people are into American classic cars and, possibly relatedly, |
| 10:03 | <Lachy> | jgraham, the problem is, we're a group of young hooligans who get their kicks from unnecessarily powerful cars. The solution is to get a powerful car :-) |
| 10:03 | <jgraham> | a surprising number of Swedish people can be seen by the side of the road, in the rain, peering at the car engine |
| 10:04 | <jgraham> | Lachy: Oh so the problem statement is something like "should be fun to drive"? |
| 10:04 | <jgraham> | or rather "we want to have fun driving the car" or so? |
| 10:04 | <Dashiva> | Should lead to reckless endangerment of bystanders |
| 10:05 | <annevk> | should be fun to drive, potentially slightly rough terrain |
| 10:05 | <jgraham> | (the difference is that a truck might well be powerful in the sense of having a big engine but be heavy and steer poorly and be no fun at all) |
| 10:05 | <annevk> | should be fun to sit in as well for those without driver's license :) |
| 10:05 | <gsnedders|work> | jgraham: Hey, I got all the way to Stockholm and back in a 70s American muscle car, they aren't all bad |
| 10:10 | <Lachy> | annevk, the dodge charger has a tv in the back for you |
| 10:11 | <annevk> | there we go :) |
| 10:13 | <jgraham> | Just remember kids TV +_speed + rough terrain + hangover === hefty cleaning bill |
| 10:15 | <annevk> | aka jealous :p |
| 10:16 | gsnedders|work | wonders whether jgraham means "kids TV" or "just remember kids," |
| 10:17 | <jgraham> | I don't think I would need the hangover, the (unusually high) speed or the rough terrain to make me sick watching TV in the back of the car |
| 10:17 | <jgraham> | just a few bends in the road |
| 10:17 | <jgraham> | (well I probably wouldn't be sick but I would likely feel sick) |
| 10:18 | <Lachy> | gsnedders|work, of course he meant kids TV. Cartoon Network, nickelodeon and disney channel rule! |
| 10:18 | jgraham | misses rugrats |
| 10:19 | <AryehGregor> | I remember Rugrats. |
| 10:19 | <Lachy> | hah. Was just thinking the same :-) |
| 10:19 | <gsnedders|work> | Man, Rainbow is where it's at. |
| 10:19 | <gsnedders|work> | I mean, they even have a character called Geoffrey! |
| 10:21 | <Lachy> | Rainbow? |
| 10:21 | <Lachy> | never heard of it |
| 10:22 | Lachy | goes back to watching Cabbage Patch Kids and Care Bears |
| 10:22 | <jgraham> | It was a british thing from the 70s or so |
| 10:23 | <gsnedders|work> | As far as I can tell, British children's TV was basically the same in the 90s as it was in the late 70s |
| 10:24 | <jgraham> | So, is there a good, general, way of duplicating a directory tree without performing unnecessary file copies? Something like copy-on-write for files? I guess that is probably a filesystem dependent thing... |
| 10:26 | <Dashiva> | Raid? |
| 10:28 | <jgraham> | Not quite what I want |
| 10:29 | <AryehGregor> | jgraham, the possibility of a reflink() or copyfile() system call was recently being discussed on lkml. As far as I know, no OS currently supports copy-on-write on a per-file basis, for some reason. |
| 10:29 | <AryehGregor> | It would be an awesome feature. |
| 10:32 | <jgraham> | (the problem is that I want to run two instances of a program with different configurations. This is not really easy since it is not possible to change the path to the config file when I launch it so the obvious solution is to make two copies of the whole program) |
| 10:32 | <AryehGregor> | You could use hard links. |
| 10:32 | <AryehGregor> | cp -rl |
| 10:32 | <AryehGregor> | Of course, changes to one side of the directory tree will change the other, if the file is modified in-place. |
| 10:32 | <jgraham> | That could maybe work on linux at least |
| 10:32 | <jgraham> | Dunno about windows |
| 10:32 | <AryehGregor> | But not if it's copied and the copy is rename()d over it, which is usually what's done on Unix. So a poor man's copy-on-write. |
| 10:32 | <AryehGregor> | I think all versions of NT support hard links, or certainly all recent ones. Definitely Vista, anyway, and I think XP also. |
| 10:32 | <AryehGregor> | Of course, you might have to write a script yourself instead of being able to use cp -rl. |
| 10:34 | <AryehGregor> | Or use Cygwin. |
| 10:36 | <Dashiva> | It was added in vista |
| 10:39 | <nessy> | Hixie: subtitles, audio description etc are all regarded as part of "video accessibility" generally - and yes, they would be a new feature to current html5 |
| 10:40 | <nessy> | (sorry for the late reply :) |
| 10:40 | <AryehGregor> | Dashiva, symlink support was added in Vista, but hard link support too? |
| 10:42 | <Dashiva> | Hum, maybe my memory is faulty |
| 10:43 | <Dashiva> | No, it was in 6.0 according to wikipedia |
| 10:43 | <Dashiva> | No |
| 10:43 | <Dashiva> | Ah, there's my error. It was a different way to create them before Vista. |
| 10:43 | <Dashiva> | Carry on :) |
| 11:03 | <Philip`> | jgraham: Use symlinks? |
| 11:03 | <Philip`> | (You can't hardlink directories, if I remember correctly) |
| 11:04 | <Philip`> | (because it could cause cycles, which would make GC hard) |
| 11:13 | <workmad3> | Philip`: correct |
| 11:21 | <gsnedders|work> | jgraham: The Apotek in Stora Torget didn't have any tissues, will try Hemköp later |
| 11:23 | <Philip`> | SSE naming is confusing - after SSE comes SSE2, and after SSE2 comes SSE3 but Linux calls it PNI, and after that is SSSE3 |
| 11:26 | <jgraham> | gsnedders|work: I have bought some from there before |
| 11:27 | <jgraham> | The Apoteket |
| 11:27 | <jgraham> | too many definitives there |
| 11:27 | gsnedders|work | didn't see any on the shelves there |
| 11:28 | <jgraham> | hsivonen: BTW is despotify the open source client you were talking about? |
| 11:28 | jgraham | wonders whether to try that since wine is not so reliable |
| 11:29 | <jgraham> | (so if you have tried it it would be good to know) |
| 11:29 | <jgraham> | (although the curser interface doesn't seem to friendly) |
| 11:29 | <jgraham> | *curses |
| 11:30 | <hsivonen> | jgraham: I don't remember the name |
| 11:41 | <crash\> | Hixie: congratulations for the Last Call draft ;) |
| 12:00 | hsivonen | expects a bikeshed on the URI identifiers used in microdata vocabs |
| 12:09 | <Dashiva> | hsivonen: Good thing microdata is so controversial, it will keep many people out of the shed |
| 12:10 | <jgraham> | Don't you have to be outside the shed to paint it? |
| 12:10 | <jgraham> | At least I never imagine bikeshedding as people decorating the interior of the bikeshed |
| 12:11 | <Dashiva> | The paint color decision meeting is held inside the shed |
| 12:12 | <gsnedders|work> | That must be one large shed |
| 12:12 | <Philip`> | Won't the people inside the shed die from inhaling paint fumes? |
| 12:12 | <Dashiva> | No, because no painting happens until the meeting is done |
| 12:12 | <jgraham> | What if someone were to graffitti the shed while the meeting was happening? |
| 12:13 | <Philip`> | Dashiva: People might paint patches of the shed during the meeting, to demonstrate what it will look like |
| 12:13 | <Dashiva> | Those people should bring gas masks then |
| 12:14 | <Philip`> | I suppose they'll die anyway from radiation from the adjacent nuclear power station that was designed and constructed and switched on during their meeting |
| 12:15 | <jgraham> | More to the point what if the graffitti is actually a bansky containing an ironic depiction of a dysfunctional beuracracy that immediately makes the shed hugely valuable and thus prevents anyone from painting it at all |
| 12:16 | <Dashiva> | Well, when they've invested so much time in deciding the color, they're not going to just stop because of a minor detail like that |
| 12:17 | <hsivonen> | wow. gopher: still works in Firefox |
| 12:17 | <hsivonen> | but not in Safari |
| 12:18 | <gsnedders|work> | Safari never supported it, ever |
| 12:18 | <gsnedders|work> | IE7 removed support for Gopher, IIRC |
| 12:18 | <hsivonen> | can't test Opera. Opera keeps bouncing in the Dock and the bouncing frequency isn't stable |
| 12:19 | <jgraham> | I guess it would be cruel and unfair to draw some analogy between the graffiti artist, the debaters, the WHATWG and the W3C. So I won't do that |
| 12:25 | <hsivonen> | aargh. OS X made fork temporarily unavailable even though I don't even have an insane number of processes |
| 12:27 | <Dashiva> | jgraham: The graffiti artist is lastweek? :) |
| 12:44 | <hsivonen> | The WHATWG HTML5 LC seems to be pretty effective at getting people to reveal their (already known) political alignment on twitter. |
| 13:05 | <TabAtkins> | hsivonen: I need to follow more people on twitter. I've only got a handful of people talking about it. |
| 13:11 | <Philip`> | TabAtkins: Just watch http://search.twitter.com/search?q=html5 |
| 13:11 | <TabAtkins> | Philip`: I prefer http://toriseye.quodis.com/#html5 |
| 13:25 | <Philip`> | "when RDFa in HTML hits the street, the issues will have been resolved" - except that people are already implementing it in HTML and deploying content in HTML, before anyone even started writing a rough draft of how RDFa ought to work in HTML, and there are still lots of unresolved issues that are getting ignored by people who want to use it now |
| 13:25 | <Philip`> | so that doesn't seem great |
| 13:30 | <TabAtkins> | I'm resisting posting by distracting myself with stories about the world's military attacking Heaven. |
| 13:31 | <Philip`> | Seems implausible |
| 13:32 | <Philip`> | The world's military wouldn't just ally like that |
| 13:33 | <TabAtkins> | Well, it happens after Hell tries to invade. |
| 13:34 | <Philip`> | Hell tries to invade the world, or Heaven? |
| 13:34 | <TabAtkins> | The former. |
| 13:34 | <Philip`> | So why attack Heaven? |
| 13:34 | <hsivonen> | I assume a non-obvious definition of hitting the street. |
| 13:34 | <TabAtkins> | Because Hell was allowed to invade by God, who basically said "Fuck off, jerks, Heaven's closed." |
| 13:35 | <Philip`> | Oh, okay |
| 13:37 | <TabAtkins> | Also, I love that the reason the story was implausible was because the world's military wouldn't ally together, not because they were attacking heaven. |
| 13:55 | <Dashiva> | TabAtkins: There's lots of prior art on attacking heaven, none for world unity :P |
| 13:56 | <TabAtkins> | Dashiva: Really? I didn't know there was much about crashing tanks into the pearly gates. |
| 13:57 | <Dashiva> | Well, the specific armaments vary |
| 13:57 | <hsivonen> | hmm. how is the twitter front and search pages different from the user pages in terms of styling mechanics? |
| 13:57 | <hsivonen> | the front page is CSSless for me |
| 13:57 | <Dashiva> | Is it possible to see the front page without logging out? |
| 13:58 | <hsivonen> | dunno |
| 13:58 | <hsivonen> | hmm. looks like they specify different media |
| 13:59 | <hsivonen> | I wonder if I or someone else has broken style sheets with unusual media attributes |
| 13:59 | <Dashiva> | Heh, the w3c twitter RT'd julian's LC comment instead of making a separate statement |
| 14:00 | <TabAtkins> | What's the w3c twitter? |
| 14:00 | <Dashiva> | /w3c |
| 14:00 | <Philip`> | http://twitter.com/w3c amazingly enough |
| 14:00 | <TabAtkins> | I suppose I should have expected that. |
| 14:02 | <Dashiva> | http://twitter.com/shepazu/status/5214106553 |
| 14:02 | <Dashiva> | What is this I don't even |
| 14:08 | <TabAtkins> | Dashiva: I suppose the idea is that he's comparing WHATWG LC to a bait-and-switch? |
| 14:12 | <Dashiva> | I don't really see it, but okay |
| 14:14 | <Philip`> | I thought the idea was cars? |
| 14:15 | <Dashiva> | Elaborate, if you would |
| 14:17 | <brucel> | sense check anyone? what's wrong with <input type=range min=5 max=30 step=5> ? Mr validator tells me "Zero is not a positive floating point number." |
| 14:17 | <TabAtkins> | There is no sense in that. |
| 14:18 | <gsnedders|work> | jgraham: ping |
| 14:18 | <Philip`> | It's true the zero is not a positive floating point number |
| 14:18 | <Philip`> | s/the/that/ |
| 14:18 | <Dashiva> | Probably a bug that snuck in during the recent update |
| 14:18 | <Dashiva> | Or "deployment" as the fancy pants call it |
| 14:18 | <Philip`> | It's not necessarily bad for the validator to say true things |
| 14:18 | <brucel> | it is true that zero isn't a +ve FP number... but I've said step=5, not step=0 |
| 14:19 | <brucel> | tho i like the idea of a random gnomic truism from the validator every time |
| 14:19 | <TabAtkins> | What if the validator just tells you to rethink your life decisions? |
| 14:19 | <jgraham> | gsnedders|work: gnip |
| 14:19 | <Dashiva> | Sounds useful |
| 14:19 | <gsnedders|work> | jgraham: bratell wants you |
| 14:19 | <Dashiva> | Is there a uselessfactlib? |
| 14:21 | <brucel> | there's Daves' Web Of Lies http://www.davesweboflies.com/: lie of the day "If left alone together in a field, two field mice will naturally dig parallel tunnels." |
| 14:22 | <Dashiva> | "Quacking like a duck for five minutes every day will prevent the onset of ague." |
| 14:22 | <Dashiva> | I don't know what ague is, but I don't like the sound of it |
| 14:25 | <Philip`> | Looks like FloatingPointPositive ignores the first digit, when deciding whether something is zero |
| 14:25 | <Philip`> | for (int i = 1; i < literal.length(); i++) { ... looks for digits ... } |
| 14:26 | <TabAtkins> | That's strange. |
| 14:26 | <Philip`> | Looks like FloatingPointPositive is bogus anyway because it doesn't like 'e' in floats |
| 14:27 | Philip` | wonders if he's missing a good reason why i should start at 1 |
| 14:28 | <Dashiva> | Maybe it assumes the presence of + or - |
| 14:28 | <Dashiva> | "Sergei Bubka, the world record pole-vaulter, has recently admitted to filling his shorts with helium for extra lift." |
| 14:28 | <TabAtkins> | That would be my assumption. |
| 14:28 | <zcorpan_> | Hixie: the "editor's draft" link in the SotD is broken in http://www.whatwg.org/specs/web-apps/2009-10-27/multipage/ |
| 14:28 | <Philip`> | Hmm, nope, it's totally wrong |
| 14:29 | <Philip`> | It accepts range=11 and range=x1, and not range=10 or range=1 |
| 14:29 | <Philip`> | Someone less lazy should report it as a validator.nu bug somewhere |
| 14:29 | <Dashiva> | Sounds like a job for you |
| 14:29 | <Philip`> | "less lazy [than me]" |
| 14:29 | <brucel> | i can report it but to whom? |
| 14:30 | <Philip`> | brucel: http://bugzilla.validator.nu/ probably |
| 14:32 | <brucel> | giving a sign for the step <input type=range min=5 max=30 step=+5> suppressed the error message. Tho it's difficult to see how you can get from 5 to 30 in steps of anything other than positive numbers, so the sign seems superfluous. |
| 14:32 | <Philip`> | brucel: That makes your content invalid |
| 14:32 | <Philip`> | (+ is not allowed in valid floating point numbers) |
| 14:32 | <brucel> | (Cue some mathematician telling me that it's perfectly possibe in non-Euclidian Calabai-Yau space) |
| 14:33 | <Philip`> | brucel: It's perfectly possible in modular arithmetic |
| 14:33 | <TabAtkins> | All you need is a finite group. |
| 14:33 | <TabAtkins> | Z/50, frex, would do it. |
| 14:33 | <brucel> | the validator shut its face when I gave it step=+5, though. |
| 14:33 | <Philip`> | brucel: The validator shuts its face when you give it step=x5 too |
| 14:34 | <Philip`> | or step=-5 |
| 14:34 | <brucel> | x5 is perfectly acceptable in Swedenborgian algebraic topology, of course |
| 14:34 | <Philip`> | The shortest thing that's valid and that the validator accepts is step=05, I believe |
| 14:35 | <brucel> | so it's a validator bug and not my brainfart? |
| 14:35 | <Philip`> | (The validator ought to accept "05", "5", "5e0", "5.5e5", etc, and reject "0", "0e5", "1e-100", etc) |
| 14:35 | <hsivonen> | clearly, the datatype for positive integer is bogus |
| 14:35 | <Philip`> | (and it gets almost all of them wrong) |
| 14:35 | <hsivonen> | or positive float |
| 14:35 | <Philip`> | hsivonen: Float, not integer |
| 14:37 | <hsivonen> | well, this is embarrassing |
| 14:38 | Philip` | notes that 1e-100 seems like quite a nasty case, since you really have to parse the string into a float and can't just examine its characters |
| 14:38 | <hsivonen> | the datatype iterates from index 1 not 0 |
| 14:38 | <Philip`> | (because 1e-100 parses into 0, as of yesterday, I believe) |
| 14:40 | <brucel> | Henri, can I formally report the bug to you now please? |
| 14:43 | <hsivonen> | brucel: yes |
| 14:53 | <brucel> | ok, thanks for saving my sanity all |
| 14:55 | Philip` | sees there's already an open validator bug for exponents in floats |
| 14:55 | <gsnedders|work> | Sanity? pff. |
| 14:56 | Philip` | wonders if it's worth filing a new bug for cases like 1. and .1 and 1e-100 |
| 14:59 | <ment> | do you belive there will ever be a conforming implementation of the _whole_ html5 specs? |
| 14:59 | <ment> | or that microsoft will adapt it right? |
| 14:59 | <TabAtkins> | Sure, in 2022. ^_^ (Actually, that's the projected date for *two* of them.) |
| 15:00 | <gsnedders|work> | ment: All W3C specs are required to have two interoperable implementations before going to REC |
| 15:00 | <jgraham> | (some features from the current spec will likely be cut/changed before that happens) |
| 15:00 | <gsnedders|work> | And the intention for HTML 5 is to have two complete interoperable implementations |
| 15:01 | <ment> | so the two will be webkit and gecko? |
| 15:02 | <hsivonen> | ment: the Process make the spec get smaller and smaller until there are two interoperable implementations |
| 15:02 | <jgraham> | gsnedders|work: I think you are actually allowed some freedom in deciding the exit criteria for Rec |
| 15:02 | <hsivonen> | *makes |
| 15:02 | <jgraham> | s/for Rec/from CR/ |
| 15:03 | <hsivonen> | jgraham: sure, the XHTML2 WG can decide that two validators that don't barf on DTDs count as implementations |
| 15:03 | <hsivonen> | jgraham: I expect the bar to be higher in the HTML WG |
| 15:03 | <ment> | hsivonen: huh, what do you mean by that? like you cut off parts of the standard because one of the implementators didn't do it right? |
| 15:03 | <hsivonen> | ment: right |
| 15:04 | <hsivonen> | ment: or more to the point, if there aren't two implementations getting it right, it gets cut |
| 15:04 | <ment> | hsivonen: do you cooperate with microsoft in any way? |
| 15:05 | <Philip`> | Or the spec hangs around in CR until everyone gets fed up and gives up on it |
| 15:05 | <hsivonen> | ment: MS is in the WG |
| 15:06 | <Philip`> | (which isn't necessarily bad, since getting to Rec shouldn't be a goal; the goals should be about having enough interoperability that people don't suffer from lack of it) |
| 15:06 | <hsivonen> | ment: however, it's sufficient for the two impls to be in the set {Gecko, WebKit, Opera} |
| 15:06 | <hsivonen> | ment: IE implementing stuff isn't a requirement for REC |
| 15:07 | <hsivonen> | (there could be a surprise new entrant to the market, too) |
| 15:07 | <ment> | hsivonen: well but from the practical point of view if microsoft doesn't implement html5, that's it for the standard |
| 15:07 | <TabAtkins> | Though, of course, IE *not* implementing stuff is a reason to cut that stuff from the spec, as they're a significant implementor. |
| 15:07 | <hsivonen> | or NetFront becoming conforming or something |
| 15:08 | <hsivonen> | ment: some features of HTML5 (mainly the form stuff) has been explicitly designed so that it can be emulated in IE using JS libs without binary plug-ins |
| 15:09 | <Philip`> | ment: The standard isn't a monolithic thing - it's a set of features, some of which will succeed and some of which will fail, and the standard as a whole won't be a failure just because some features don't get widely implemented |
| 15:09 | <Philip`> | It would be a failure if none got widely implemented, but many already have (including by Microsoft) |
| 15:09 | <zcorpan_> | http://www.peterbuick.com/news/how-to-add-your-own-video-to-firefox-html5 |
| 15:15 | <hsivonen> | brucel: fixed and deployed |
| 15:31 | <gsnedders|work> | Is there any way of knowing what order IRC load events will fire in? |
| 15:31 | <gsnedders|work> | s/events/event handlers/ |
| 15:31 | <Philip`> | IRC has event handlers? |
| 15:31 | gsnedders|work | is getting distracted by people speaking to him |
| 15:31 | <gsnedders|work> | s/IRC// |
| 15:32 | <jgraham> | gsnedders|work: they fire in the order they were registered |
| 15:32 | <TabAtkins> | I need to form a group called ATGNAT - All The Good Names Are Taken. |
| 15:32 | <Philip`> | TabAtkins: That's a good name |
| 15:33 | <TabAtkins> | Philip`: It's like the smallest integer that can't be described in less than twelve words. |
| 15:38 | <jthomas_sb> | Can anyone watch a full video here, with HTML5 embedded video? I don't know if its my Iceweasel that is having issues or the Apache2 server, but all three of those videos stop halfway through and never recover... http://support.sierrabravo.net/files/HowTo--Email_Setups/ The videos worked in the past but don't seem to want to fully play anymore. |
| 15:39 | <hsivonen> | jthomas_sb: WFM in Minefield. |
| 15:40 | <jthomas_sb> | i don't understand |
| 15:41 | <hsivonen> | jthomas_sb: works for me in a build of firefox trunk I made from the mercurial repo |
| 15:41 | <TabAtkins> | works for me in Chrome, too. However, the *sound* goes extra fast and cuts out halfway through. |
| 15:41 | <jthomas_sb> | wfm = works for me, gotcha |
| 15:41 | <hsivonen> | jthomas_sb: is your iceweasel recent enough? |
| 15:41 | <jthomas_sb> | yeah the sound is a seperate issue |
| 15:41 | <jthomas_sb> | perhaps not, but it is 3.5 and it did work before... |
| 15:42 | <jthomas_sb> | if its working for you all i'll go with it, thanks all |
| 15:42 | <hsivonen> | jthomas_sb: if you tweaked mime types, you may have bad stuff in cache |
| 15:44 | <jthomas_sb> | i've not touched the site since i set it up 2mo ago or something, i was wanting to check some of what it is telling the user to do when i noticed the frozen-at-halfway-mark... |
| 15:45 | <brucel> | hsivonen thanks man |
| 15:46 | <hsivonen> | jthomas_sb: does the sound track end half way into the video? |
| 15:46 | <hsivonen> | maybe the old Gecko stops when the sound track ends |
| 15:47 | <hsivonen> | and a newer gecko plays the video track to completion anyway |
| 15:47 | <jthomas_sb> | it seems to, although I've never really worried about it; something about the gtk-recordmedesktop made the audio double-speed even tho I told it not to record audio. That may be the issues exactly, that when the audio ends the video stops.... |
| 15:56 | <brucel> | hsivonen is the source for the validator available for nosing at? |
| 15:59 | <hsivonen> | brucel: yes |
| 15:59 | <hsivonen> | brucel: http://about.validator.nu/#src |
| 16:06 | <brucel> | hsivonen groovy as The Beatles on spacehoppers. Thx. |
| 16:24 | <GPHemsley> | Hixie: Getting back to the test case issue, are there any existing HTML5 test cases? Are we going to modify the existing HTML4 test cases or start from scratch? |
| 16:25 | <Philip`> | GPHemsley: http://wiki.whatwg.org/wiki/Testsuite#Existing_tests |
| 16:25 | <Philip`> | As far as I'm aware, there aren't any HTML4 tests that are particularly useful |
| 16:26 | <GPHemsley> | Philip`: Ah, wasn't aware of that. Hixie said last night that there wasn't a plan yet. |
| 16:26 | <Philip`> | There isn't a plan yet |
| 16:26 | <Philip`> | There's just some people writing some tests |
| 16:27 | <Philip`> | and some rough ideas about what a test suite should look like |
| 16:27 | <GPHemsley> | ah |
| 16:31 | <zcorpan_> | would be nice to be able to measure the test coverage of the spec |
| 16:34 | <Philip`> | It'd be nice if people could point at a paragraph/phrase/sentence/word/etc in the spec and say "there needs to be a test for this" |
| 16:34 | <Philip`> | and then tie test cases to those pointers |
| 16:34 | <zcorpan_> | yes |
| 16:35 | <Philip`> | and then you can work out how many points are not tested at all |
| 16:35 | <Philip`> | (though this wouldn't help with testing interactions between features, or with making sure each point is properly tested) |
| 16:35 | <zcorpan_> | would still be a useful thing to do |
| 16:36 | <Philip`> | The end result could look like http://philip.html5.org/tests/canvas/suite/tests/spec.html :-) |
| 16:47 | <zcorpan_> | "The amount of bugs that are related to contentEditable, text ranges, drag and drop and the Document Object Model in general are staggering." -- http://blog.typlab.com/2009/10/love-web-development-again/ |
| 16:47 | <zcorpan_> | i guess contentEditable and ranges still need a proper spec |
| 16:51 | <annevk2> | like pretty much everything :( |
| 17:02 | <zcorpan_> | i think i'll start to include the string "zcorpan" in my bugs so i can find them and close them later |
| 17:05 | Philip` | was thinking of doing the same |
| 17:05 | <Philip`> | (Well, not exactly the same, I wouldn't put "zcorpan" in them) |
| 17:06 | <jgraham> | If you did then zcorpan would deal with them for you |
| 17:07 | gsnedders | does bug cleanup on html5lib |
| 17:13 | gsnedders | finds a bug in google code |
| 17:15 | <mpilgrim> | zcorpan_: i'm not sure what you mean by "the first article example probably ought to set display:block" |
| 17:15 | <gsnedders> | Hixie: mpilgrim: Could either of you file a bug for me? It's creating a link to http://code.google.com/p/html5lib/issues/list?q=label:0.12-blocker which is (bogusly) failing to find anything. |
| 17:15 | <zcorpan_> | mpilgrim: the part that currently says article { border: 1px solid red } |
| 17:16 | <mpilgrim> | aha |
| 17:16 | <mpilgrim> | thanks |
| 17:17 | <mpilgrim> | ok, next question |
| 17:17 | <mpilgrim> | how should i rephrase "and everything else on the page will stay where it is in the document outline" |
| 17:17 | <mpilgrim> | what i mean is that it won't screw up the header nesting levels |
| 17:17 | <mpilgrim> | because the article is a self-enclosed universe |
| 17:18 | <mpilgrim> | or something |
| 17:18 | <mpilgrim> | i'm drawing a blank on how to explain that |
| 17:19 | <zcorpan_> | but it will screw up the header nesting levels in e.g. <body><h1>a</h1><h2>b</h2>(article here)<h2>c</h2> |
| 17:19 | <mpilgrim> | really? why? |
| 17:20 | <zcorpan_> | because the article becomes a subsection of body and headers after the article become headings of new implied sections that are on the same level as body, iirc |
| 17:21 | mpilgrim | writes a test case |
| 17:21 | gsnedders | doesn't think that's right |
| 17:22 | <Philip`> | gsnedders: http://code.google.com/p/support/issues/list |
| 17:22 | <Philip`> | for filing bugs |
| 17:23 | <mpilgrim> | http://gsnedders.html5.org/outliner/process.py?url=http://diveintomark.org/tmp/headers.html |
| 17:23 | <mpilgrim> | looks like what i would expect |
| 17:23 | <gsnedders> | That's what I'd expect too |
| 17:24 | <gsnedders> | The fact there's an article there should have no effect |
| 17:24 | <jgraham> | It is not impossible that you would expect d as a child of b |
| 17:24 | <zcorpan_> | hmm, then either i was having a different case in mind or i was just wrong |
| 17:24 | <mpilgrim> | ok, well change the question: |
| 17:25 | <mpilgrim> | how do i rephrase the sentence to say that if you do http://diveintomark.org/tmp/headers.html , you get the outline given by http://gsnedders.html5.org/outliner/process.py?url=http://diveintomark.org/tmp/headers.html |
| 17:25 | <mpilgrim> | maybe just say that |
| 17:25 | <zcorpan_> | <h1>a</h1><h2>b</h2><h3>c</h3><article><h1>article</h1></article><h3>d</h3> |
| 17:26 | <zcorpan_> | b and d are the same level |
| 17:26 | <zcorpan_> | without the article, c and d are the same level |
| 17:26 | <zcorpan_> | so inserting the article screws up the level of d |
| 17:27 | <mpilgrim> | interesting |
| 17:31 | <gsnedders> | What's the "lint" filter in html5lib? |
| 17:32 | <zcorpan_> | i think it's like this because otherwise the automatic resizing of h1 wouldn't work without killing performance with a magic selector that depends on the outline algorithm, or something |
| 17:32 | <Rik`> | mpilgrim: on http://diveintohtml5.org/semantics.html, your link to apple documentation is wrong, it should be http://developer.apple.com/safari/library/documentation/AppleApplications/Reference/SafariWebContent/ConfiguringWebApplications/ConfiguringWebApplications.html#//apple_ref/doc/uid/TP40002051-CH3-SW4 |
| 17:33 | <gsnedders> | zcorpan_: In that case the article is a section in and of itself, with a rank of 1, no? |
| 17:33 | <Philip`> | gsnedders: It stops html5lib getting clogged up when you dry markup |
| 17:34 | <zcorpan_> | gsnedders: no it's a subsection of a |
| 17:35 | <mpilgrim> | Rik`: thanks, that'll be fixed in the next update |
| 17:37 | <jgraham> | gsnedders: Well look at the source |
| 17:37 | <jgraham> | It checks some stuff. I don't recall what |
| 17:37 | <jgraham> | Thomas Broyer wrote it |
| 17:38 | <gsnedders> | jgraham: Can you look at #112 and #74? |
| 17:38 | <gsnedders> | jgraham: It touches the content model flags a lot |
| 17:38 | <zcorpan_> | gsnedders: for the html5lib quotes thing, can't we just quote for U+0000..U+0020, U+0060, plus the old opera characters? |
| 17:38 | <jgraham> | gsnedders: Not now, I have to leave (and am out this evening) |
| 17:47 | <gsnedders> | http://code.google.com/p/html5lib/issues/list?q=label:Milestone-Release1.0 |
| 17:48 | <ment> | i thought content model flags are gone |
| 19:30 | <kangax> | Was `StorageItem` interface dropped from the web storage? |
| 20:22 | <krijn> | mpilgrim__: you're missing a </ul> in the fat footers example on http://diveintohtml5.org/semantics.html |