| 01:21 | <kochi> | morning, whatwg |
| 01:23 | <jamesr> | evening, kochi |
| 03:46 | <MikeSmith> | kochi: hey |
| 03:47 | <MikeSmith> | kochi: I think we should aim to publish an updated WD of the IME API on April 2nd |
| 04:13 | <kochi> | MikeSmith: hi |
| 04:14 | <kochi> | MikeSmith: agreed, what steps do I need to make it an updated WD? |
| 04:16 | <kochi> | (... and wondering why Apr. 2 specifically? to avoid April fool that the spec isn't a joke? :) |
| 04:17 | <kochi> | maybe people consider any published spec that has date on April 1st is a joke spec (like joke RFCs) |
| 04:23 | <MikeSmith> | kochi: we publish only on Tuesdays and Thursdays each week |
| 04:24 | <kochi> | MikeSmith: i see. |
| 04:24 | <MikeSmith> | kochi: I think you don't need to do much to prepare the draft. I'll take a look at it next week and if there are any problems, I'll let you know. Otherwise, if there are no problems, I'll just go ahead and stage it up for publication. |
| 04:25 | <kochi> | MikeSmith: Thanks!! |
| 09:25 | <zcorpan> | turns out "HTML 5.0" is causing implementors trouble. who knew? http://www.w3.org/mid/514AC4B1.9070507⊙oc |
| 10:25 | <Ms3ger> | jgraham: can you get me a critic review for https://github.com/w3c/web-platform-tests/pull/18 ? |
| 10:42 | <jgraham> | http://critic.hoppipolla.co.uk/r/12 |
| 10:43 | <jgraham> | Ms3ger: ^ |
| 10:43 | <jgraham> | In case you didn't already find it |
| 10:43 | <Ms3ger> | Ta |
| 10:55 | <jgraham> | Ms3ger: Isn't the prototype part of those tests covered by idlharness? |
| 10:55 | <Ms3ger> | Mm |
| 10:55 | <Ms3ger> | Only if we're creating all those elements |
| 10:57 | <Ms3ger> | And if we actually have an idlharness test for HTML |
| 11:00 | Ms3ger | wonders if http://critic.hoppipolla.co.uk/showbatch?review=12 is supposed to show something |
| 11:02 | <Ms3ger> | Also wondering why gmail shows a photo of jgraham next to this critic email |
| 11:02 | <jgraham> | That's terrifying |
| 11:03 | <jgraham> | Ms3ger: If you are asking whether it imports the existing comments, it doesn't |
| 11:03 | <jgraham> | If that is a problem you'll have to use github for that review. |
| 11:04 | <Ms3ger> | No, the question was why I was redirected to an empty page after submitting the review |
| 11:06 | <jgraham> | Oh, I don't know, I have never been redirected to that page |
| 11:06 | <jgraham> | You should end up on http://critic.hoppipolla.co.uk/showcommit?first=52ee9be5&last=f2476047&review=12 |
| 11:06 | <jgraham> | Or something, depending on which commits you were reviewing |
| 11:13 | <Ms3ger> | I see |
| 11:13 | Ms3ger | wanders off to lunch |
| 12:11 | <jgraham> | Anyone know what happened to the Mozilla wiki page about writing stable tests? |
| 12:14 | <gsnedders> | jgraham: https://developer.mozilla.org/en-US/docs/Mozilla/QA/Avoiding_intermittent_oranges |
| 12:18 | <jgraham> | And yet a search for "avoiding intermittent orange site:mozilla.org mdn" didn't find it |
| 12:18 | <jgraham> | Thanks |
| 12:22 | <gsnedders> | "avoid orange mozilla" did :) |
| 13:14 | <jgraham> | Maybe Google just hates me then |
| 13:19 | <karlcow> | jgraham: you are not clicking often enough on ads :p |
| 16:25 | <aklein> | hsivonen: yt, by any chance? |
| 16:26 | <aklein> | hsivonen: (interested in your thoughts on namespace-sensitive comparisons on the stack of open elements) |
| 16:31 | <jgraham> | aklein: Projecting hsivonen for a moment, I expect he would argue that compat is better than theoretical purity at this point |
| 16:50 | <aklein> | jgraham: we already don't have compat. what I'm trying to avoid is crashiness |
| 16:51 | <aklein> | jgraham: but I understand what you mean, and I agree that for any of these cases it doesn't particularly matter what we choose to do since the markup makes no sense (e.g., <svg><table><foreignContent><table>) |
| 16:52 | <aklein> | it just so happens that comparing namespaces as well as localNames seems like a nice way to explain how these cases get resolved |
| 16:52 | <jgraham> | Well, personally I think comparing namespaces makes more sense |
| 16:53 | <jgraham> | And I am happy to change html5lib |
| 16:53 | <jgraham> | But "it doesn't particularly matter waht we do because the markup doesn't make sense" isn't a great argument since authors don't seem to care if the markup makes sense |
| 16:54 | <aklein> | jgraham: fair enough. that's basically what Hixie's response to these questions has been though |
| 16:58 | <aklein> | I'll try to put together this change today or tomorrow; I'm curious how many html5lib tests it'll affect |
| 17:01 | <MikeSmith> | aklein: you writing a parser? |
| 17:24 | <annevk> | MikeSmith: the one in WebKit ;) |
| 17:25 | <MikeSmith> | oh |
| 17:27 | <annevk> | From mutation observer fame and now <template> |
| 17:28 | <MikeSmith> | I'd thought just abarth and Eric Seidel were working on the WebKit parser |
| 17:28 | <MikeSmith> | annevk: aklein is implementing the <template> parsing? |
| 17:28 | <annevk> | afaik |
| 17:29 | <MikeSmith> | ok |
| 17:29 | <annevk> | <template> seems to be implemented by other people than those who wrote the parser architecture, at least in Gecko/WebKit |
| 17:29 | <TabAtkins> | Correct. |
| 17:29 | <MikeSmith> | who implemented it in Gecko? |
| 17:29 | <annevk> | And it seems to be specified in such a way too, it being a separate specification (for now) |
| 17:29 | <MikeSmith> | yeah |
| 17:30 | <annevk> | MikeSmith: I think wchen (not here) is/will be doing it, but not a 100% sure |
| 17:31 | <MikeSmith> | hey TabAtkins I'm in America's heartland, visiting my girlfriend in Indiana. She tells me to quit saying Howdy to people here. She says they don't say Howdy here. I feel more comfortable in Texas. Should have convinced here to go to graduate school at UT. |
| 17:31 | <MikeSmith> | annevk: ah yeah I vaguely remember |
| 17:34 | <TabAtkins> | MikeSmith: Hah, awesome. Never been to Indiana, but I still say howdy here in Cali sometimes. |
| 17:34 | <MikeSmith> | heh yeah I bet |
| 17:35 | <TabAtkins> | I'm rather glad that y'all has spread over the country pretty well in the past decade or so. |
| 17:35 | <MikeSmith> | yeah. Next we spread "all ya'll" |
| 17:36 | <MikeSmith> | speaking of y'all, before you hear the bad news from somebody else: I added longdesc support to the W3C validator |
| 17:36 | <MikeSmith> | holding my nose while doing it |
| 17:36 | <TabAtkins> | Nuuuuuuuu! |
| 17:36 | <paul_irish> | omg |
| 17:36 | <MikeSmith> | under a flag |
| 17:36 | <MikeSmith> | so it's not in validator.nu |
| 17:36 | <MikeSmith> | figured it was better to get it over with |
| 17:37 | <MikeSmith> | wonderful thing to see politics and aggressive lobbying triumphing over sound technical arguments |
| 17:37 | <MikeSmith> | anyway, tired of fighting |
| 17:37 | <nimbu> | trolls need to be fed MikeSmith |
| 17:37 | <nimbu> | thanks for fighting. |
| 17:37 | <MikeSmith> | heh |
| 17:58 | <Ms2ger> | jgraham, do you know what function is being called at https://github.com/w3c/testharness.js/blob/master/testharness.js#L1485 ? |
| 18:02 | <Ms2ger> | "I know there's a past history of sketchy PR games around testing. Nevertheless, I would like to start afresh and trust everyone to behave properly." |
| 18:04 | <annevk> | nimbu: I am curious now whether I'm the youngest, but I can't find a list |
| 18:05 | <MikeSmith> | Ms2ger: where's that quote from? |
| 18:05 | <Ms2ger> | MikeSmith, tobie, in the thread titled... |
| 18:05 | <Ms2ger> | Review of tests upstreamed by implementors |
| 18:06 | <MikeSmith> | ok |
| 18:06 | <Ms2ger> | On test-infra |
| 18:06 | <annevk> | Yeah, all the bitching about longdesc will sure make it used more and turn it into a huge success |
| 18:07 | <TabAtkins> | Haha, ConceptionConnection (from today's SMBC) is a real thing already: http://zirak.github.com/ConceptionConnection/ |
| 18:07 | <TabAtkins> | Sounds like my parents were probably aroused to unprotected sex by the withdrawal of Israeli troops from Lebanon. |
| 18:08 | <MikeSmith> | wtf |
| 18:08 | <MikeSmith> | oh, related to that link |
| 18:08 | <TabAtkins> | Hahahahahaha |
| 18:09 | <annevk> | "The first Nintendo home video game console in United States is released by the Nintendo Entertainment System." |
| 18:09 | <annevk> | I can relate to that |
| 18:09 | <TabAtkins> | That's an excellent reason! |
| 18:09 | <annevk> | though I think my parents can't |
| 18:09 | <annevk> | it's great that it handles negative days |
| 18:09 | <annevk> | whoa |
| 18:09 | <annevk> | also |
| 18:09 | <annevk> | "The Free Software Foundation is founded in Massachusetts, USA." |
| 18:09 | <annevk> | read over that somehow |
| 18:10 | <annevk> | but seems relevant too |
| 18:10 | <TabAtkins> | Damn bro, you got an auspicious conception date. |
| 18:10 | <annevk> | and "The first flight of Space Shuttle Atlantis." which seems relevant to architecture astronauting |
| 18:10 | <annevk> | TabAtkins: this is kinda weird |
| 18:10 | <annevk> | but so is Nostradamus |
| 18:11 | <Hixie> | annevk: i had to make an edit to the fetch algorithm |
| 18:11 | <Hixie> | minor edit |
| 18:11 | <Hixie> | just adding a paragraph about how to handle appcache loads |
| 18:11 | <TabAtkins> | annevk: Context: http://www.smbc-comics.com/ |
| 18:11 | <MikeSmith> | mine mentions the Lockheed SR-71 Blackbird |
| 18:12 | <Hixie> | because i realised that the navigate algorithm totally failed to actually call fetch if it used an appcache, which resulted in all the things that relied on fetch failing :-) |
| 18:12 | <MikeSmith> | annevk: btw glad that Web Notfications support finally landed in Gecko but that notification widget is fugly |
| 18:12 | <Hixie> | annevk: might be worth putting the other appcache things directly into fetch |
| 18:12 | <annevk> | TabAtkins: hehehe |
| 18:12 | <annevk> | Hixie: yes |
| 18:13 | <annevk> | Hixie: that's my plan |
| 18:13 | <annevk> | well, long term plan |
| 18:13 | <annevk> | Hixie: so I have http://fetch.spec.whatwg.org/ written down |
| 18:13 | <annevk> | but it misses a bunch of detail stuff |
| 18:13 | <annevk> | s/stuff/still/ |
| 18:13 | <Hixie> | k |
| 18:13 | <annevk> | "Legacy" is mapping HTML5 to the new stuff... |
| 18:14 | <annevk> | I should have time next week to flush it out more |
| 18:14 | <Hixie> | roger |
| 18:14 | <annevk> | MikeSmith: euhm yeah, there's a bug on native theming |
| 18:14 | <MikeSmith> | ok |
| 18:16 | <annevk> | Hixie: so btw wycats__ brought up a number of excellent points on restructuring bits of HTML so we can make subclassing of <input> and such possible in due course |
| 18:16 | <Hixie> | lordy |
| 18:16 | <Hixie> | yeah |
| 18:16 | <Hixie> | eventually we should probably do that |
| 18:16 | <Hixie> | let's fix the bugs first... |
| 18:17 | <annevk> | Hixie: e.g. currently it's not clear how new elements could hook into <label>; <form> submission algorithm should call into the control and get back the value from it rather than defining the value for each control, etc. |
| 18:17 | <Hixie> | i've been doing too much refactoring and so on in the past few months and not enough actual bug fixing and feature speccing |
| 18:17 | <Hixie> | need to make more forward progress before i spend more time making things nicer |
| 18:17 | <Hixie> | (nicer for spec authors, i mean) |
| 18:17 | <annevk> | I think wycats__' plan is to create some kind of list and email this |
| 18:18 | <Hixie> | k |
| 18:18 | <Hixie> | this would be for web components v2? |
| 18:18 | <annevk> | I also think there's some material effect on implementors as we want implementors to implement elements the same way developers implement custom elements, eventually... |
| 18:18 | <annevk> | I'm not sure what's in components v1 vs v2 to be honest |
| 18:19 | <annevk> | according to dglazkov it's going to solve everything :-) |
| 18:19 | <Hixie> | well i doubt we're doing <input> extensions in v1 |
| 18:19 | <Hixie> | so we probably have many months if not years before this becomes an actual issue |
| 18:29 | <annevk> | On the level of X extends HTMLInputElement this may be true, but WebKit is e.g. starting to use ShadowRoot here already and allows others to add additional ShadowRoot elements. Of course that's not quite subclassing completely yet. |
| 18:29 | <annevk> | Anyway, we'll see. Wanted to give you a heads up mostly. |
| 18:29 | <Hixie> | thanks |
| 18:29 | <Hixie> | we'll need implementation experience before we can really do it anyway |
| 18:29 | <TabAtkins> | We're not subclassing yet largely because the DOM element stuff is super-sucky. |
| 18:40 | Hixie | resists the temptation to call the algorithm that strips and collapses whitespace the "space vacuum" |
| 18:41 | <WeirdAl> | call it a white hole? |
| 18:44 | <annevk> | heh |
| 18:46 | <Hixie> | feedback in a bug: "The postMessage design outlined in the W3C document edited by Ian Hickson is |
| 18:46 | <Hixie> | not good! The design of the cross document messaging by Ian Hickson (Google, |
| 18:46 | <Hixie> | Inc.) is very bad. Even the last version is not good either." |
| 18:46 | <Hixie> | :-( |
| 18:46 | <TabAtkins> | Maybe try to do better? |
| 18:47 | <Hixie> | heh, i missed the end of that bug report in my last read-through |
| 18:47 | <Hixie> | "If you can do this before march 1st, 2013, I will give you $10." |
| 18:47 | <Hixie> | d'oh! |
| 18:49 | <annevk> | such are the riches of being a spec writer |
| 18:49 | <jamesr> | takes more than $10 to have Hixie redesign postMessage, apparently |
| 18:52 | <Hixie> | probably would take less, you just have to make sure i know about it! :-D |
| 18:55 | <tobie> | Ms2ger, MikeSmith: context is to trust tests upstreamed from implementors to have been properly reviewed and to allow review by colleagues. |
| 18:56 | <Ms2ger> | tobie, fwiw, I would not assume tests checked into Mozilla to have been reviewed |
| 18:58 | <Hixie> | oh i misread it after all |
| 18:58 | <Hixie> | the money was for implementors, not me |
| 18:58 | <annevk> | aah, money always goes to them |
| 18:58 | <annevk> | $10000 for dhyatt |
| 18:58 | <Hixie> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=20790 |
| 18:59 | <Hixie> | if anyone wants to proof-read my attempt at rewrapping that comment, please be my guest |
| 19:00 | <annevk> | seems okay |
| 19:00 | <annevk> | sounds like he wants sharing of data between origins, which localStorage offers via events iirc, but we don't really have an actual API for |
| 19:01 | <Hixie> | zcorpan: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=15993 |
| 19:01 | <annevk> | well, I guess SharedWorker offers it to some extent |
| 19:02 | <carlos_antonio> | oh hixie is here, fantastic post on drm :) |
| 19:03 | <Hixie> | zcorpan: thanks |
| 19:03 | <Hixie> | er |
| 19:03 | <Hixie> | carlos_antonio: thanks |
| 19:03 | <TabAtkins> | annevk, MikeSmith: Using a different ConceptionConnection <http://nahueljose.com.ar/conception-connection/>;, I found an even more likely event to cause my conception - the founding of Hezbellah. |
| 19:03 | <TabAtkins> | s/bell/boll/ |
| 19:04 | <carlos_antonio> | yw, thank you for your work |
| 19:04 | <zewt> | Hixie: heh i remember that guy |
| 19:04 | <zewt> | at first i thought "i wonder if this guy knows his english is so bad he sounds like a prick", then i realized he just is a prick |
| 19:04 | <zcorpan> | Hixie: do you want me to reply to comment 14? |
| 19:05 | <jgraham> | Ms2ger: https://github.com/w3c/testharness.js/blob/master/testharness.js#L2060 I assume |
| 19:05 | <Ms2ger> | jgraham, ... oh good |
| 19:06 | Ms2ger | looked but didn't find |
| 19:06 | <Hixie> | zcorpan: if you're proposing something that doesn't match IE, and if there's a way to convince them, yes please |
| 19:07 | <zcorpan> | Hixie: yeah, sure, just wanted to make sure i'd focus on the thing you pinged me for :-) |
| 19:08 | <Hixie> | zcorpan: :-) |
| 19:08 | <annevk> | Hixie: btw, http://f.cl.ly/items/2L0a0B0R130G3u0k0b1y/Layering.pdf is the background to wycats__' points |
| 19:09 | <annevk> | Hixie: at some point I think it touches on the HTML form stuff |
| 19:13 | <Hixie> | annevk: i'm already convinced that we need a way to extend <input> |
| 19:15 | <WeirdAl> | input already has over 20 types defined for it... how many more do we need? :s |
| 19:17 | <annevk> | I find it kinda humorous that we are now on the verge of having distributed extensibility. |
| 19:18 | <annevk> | (Of course, this is a whole other ball game than "just" allowing "XML Namespaces".) |
| 19:20 | <odinho> | WeirdAl: Many more, and that's the problem ;] |
| 19:21 | <MikeSmith> | TabAtkins: auspicious |
| 19:22 | <Hixie> | annevk: distributed extensibility with semantic fallback is fine, it's the syntax and lack of fallback that's the main problem with xmlns |
| 19:22 | <Hixie> | problems, i guess |
| 19:23 | <MikeSmith> | annevk: how are we on the verge of distributed extensibility? Web Components? |
| 19:23 | <jwalden> | annevk: if you got first-flight-of-Atlantis, we must be really close in birthdays, although I didn't see any of the other ones you noted for me |
| 19:24 | <annevk> | MikeSmith: yes |
| 20:16 | <zcorpan> | Hixie: it seems making the quirk not apply to <hr> would make it a bit closer to IE's behavior |
| 20:22 | <zcorpan> | Hixie: and making the margin not collapse when a margin was specified in author css... |
| 20:23 | <annevk> | notes from the TAG F2F: http://annevankesteren.nl/2013/03/tag-f2f |
| 20:25 | <Ms2ger> | s/developers have not/developers do not have/ |
| 20:26 | <Ms2ger> | And [citation needed] on "IDL not encouraging the kind of APIs that are close to JavaScript" |
| 20:29 | <annevk> | heh |
| 20:31 | <Ms2ger> | Well, I guess you can cite a lot of people on that |
| 20:31 | <Ms2ger> | So [weasel words] instead? |
| 20:31 | <annevk> | I think we have to address the mismatch whether it's true or not |
| 20:32 | <annevk> | So it doesn't really matter if our goal is to work with TC39 |
| 20:32 | <Ms2ger> | Well, the question is whether we need technical changes or just convince TC39 that it's not as bad as they want to think :) |
| 20:33 | <tobie> | s/TC39/the entire web developer community/ |
| 20:33 | <annevk> | I think having wording changes might at least help. The wording not being aligned with JavaScript does not help. |
| 20:33 | <annevk> | In the IDL syntax that is. |
| 20:33 | <tobie> | :) |
| 20:34 | <Ms2ger> | I don't think "the entire web developer community" knows the difference between JS and DOM |
| 20:34 | <tobie> | Ms2ger: for a good reason. there is none. |
| 20:34 | <wycats__> | Hixie: glad to hear that you're already convinced :) |
| 20:34 | <wycats__> | tobie: O_O |
| 20:34 | <tobie> | Both are jQuery/ |
| 20:35 | <tobie> | :D |
| 20:35 | <TabAtkins> | Haha. |
| 20:36 | <zcorpan> | tobie: but html5 has better perf! |
| 20:37 | <Ms2ger> | You guys want me to pull out all my hair, do you? :( |
| 20:37 | <annevk> | Ms2ger: reportedly you have a lot of it |
| 20:37 | <Ms2ger> | Has mounir been talking? |
| 20:38 | <annevk> | Ms2ger: In any event, the larger issue is addressing a) communication between TC39 and the rest of the standards world and b) whatever comes out of that hopefully better mutual understanding |
| 20:39 | Ms2ger | approves of better mutual understanding |
| 20:39 | <tobie> | alright. bye folks. |
| 20:41 | <annevk> | I should go too |
| 20:42 | <annevk> | I'll return tomorrow morning (Europe time) as zombie-annevk |
| 20:42 | <zcorpan> | aklein: i wouldn't be surprised if it would affect few or none html5lib tests. it'd be good to have new tests that target this specifically |
| 20:43 | <Ms2ger> | See you |
| 21:01 | <zcorpan> | "Using objects without accounting for the possibility of their death" https://developer.mozilla.org/en-US/docs/Mozilla/QA/Avoiding_intermittent_oranges sounds like a bug, the timer probably shouldn't be collected if its callback can be invoked |
| 21:01 | <Ms2ger> | Yeah |
| 21:01 | <jgraham> | zcorpan: Isn't that point XPCOm-specific? |
| 21:01 | <Ms2ger> | Our XPCOM timers have some issues |
| 21:02 | <Ms2ger> | Then again, XPCOM have some issues |
| 21:02 | <Ms2ger> | *has |
| 21:04 | <gsnedders> | "some"? Doesn't that rather imply there's a part without issue? |
| 21:06 | <zcorpan> | "Using magical delay numbers is never necessary." - i find myself using timeouts when i need to test that something *doesn't* happen |
| 21:07 | <Ms2ger> | gsnedders, thank you |
| 21:07 | <Ms2ger> | zcorpan, well, how do you know it won't happen *after* that delay? :) |
| 21:08 | <zcorpan> | Ms2ger: i don't |
| 21:09 | <zcorpan> | but i still think such tests can be useful, at least i've found bugs with such tests |
| 21:09 | <Ms2ger> | Fair |
| 21:10 | <jgraham> | Yeah, timeouts are needed for sure |
| 21:11 | <zcorpan> | e.g. opera would sometimes fire several 'close' events with websockets, which is why some of our websocket tests have a timeout after the close event to make sure another close isn't fired immediately afterwards |
| 21:11 | <jgraham> | I think that point would be better as "avoid unneeded timeouts" |
| 21:12 | <zcorpan> | though i'm probably guilty of using timeouts when i shouldn't have, too :-) |
| 21:12 | <jgraham> | Yeah, when writing or reviewing tests it's always worth asking "could this timeout be avoided" |
| 21:12 | <jgraham> | +? |
| 21:15 | <zcorpan> | speaking of tests, i'm currently testing structured clone with web workers and i'm creating a new worker for each subtest because it was easy to make the worker know which test to expect. but this is pretty inefficient since there are many tests by now |
| 21:17 | <zcorpan> | if i want to only start one worker, i'd probably have to come up with a protocol where i first send a message with the test ID, then a message with the test data. and same in the other direction since i test structured clone both ways |
| 21:18 | <jgraham> | I would prefer inefficient and simple over efficient and complex |
| 21:18 | <jgraham> | For tests |
| 21:18 | <jgraham> | Is the performance actually a problem? |
| 21:19 | <Ms2ger> | What he said |
| 21:19 | <zcorpan> | i guess it's not really a problem |
| 21:19 | <zcorpan> | the worker should be cached since it's only the hash that is different |
| 21:20 | <zcorpan> | and this test is complex already :-| |