| 01:04 | <Hixie> | heycam: k |
| 01:04 | <Hixie> | heycam: check the recent diff to see if i did the callback stuff right |
| 01:04 | <Hixie> | i have what may be a silly question |
| 01:04 | <Hixie> | how do i update a ubuntu install? |
| 01:05 | <Hixie> | i tried using aptitude but all the sources were either 404 or "ignored" for reasons i could not discern |
| 01:06 | <heycam> | Hixie, this change? http://lists.whatwg.org/pipermail/commit-watchers-whatwg.org/2012/006302.html |
| 01:06 | <Hixie> | yeah |
| 01:14 | <heycam> | Hixie, is [TreatNonCallableAsNull] really needed for setTimeout/setInterval functions? I thought any non-callable object should be converted to a string |
| 01:14 | <heycam> | Hixie, you can also drop the [AllowAny], since the default behaviour for overloading is to select a DOMString overload if it failed to match the others |
| 01:15 | <heycam> | Hixie, if you drop [TreatNonCallableAsNull] there, then I think you can just get rid of ArbitraryCallback and reference http://dev.w3.org/2006/webapi/WebIDL/#common-Function |
| 01:16 | <heycam> | Hixie, "with the callback this value set to the E's currentTarget" =~ s/to the/to/ |
| 01:17 | <heycam> | Hixie, apart from that looks good |
| 01:23 | <Hixie> | can i ask you to file a bug or send mail with those? i'm not able to track things right this second |
| 01:30 | <heycam> | Hixie, sure |
| 01:31 | <Hixie> | thanks dude |
| 01:41 | <MikeSmith> | hmm chrome://tasks/ "This webpage is not available" |
| 07:01 | <annevk> | Hixie: there was a bug for the event handler stuff too |
| 07:03 | <annevk> | closing the bug https://www.w3.org/Bugs/Public/show_bug.cgi?id=15358 |
| 07:04 | <annevk> | looks a lot better now :) |
| 07:39 | <annevk> | Hixie: we could name Document's URL the "document URL" maybe |
| 07:40 | <annevk> | maybe do the same for encoding |
| 07:40 | <annevk> | "document encoding" is somewhat clearer at least |
| 07:40 | <annevk> | which reminds me, I should update specs to use the Encoding spec |
| 08:03 | <zcorpan> | Hixie: shouldn't EventHandler be nullable? |
| 08:09 | <annevk> | zcorpan: that's EventHandlerNonNull |
| 08:17 | <zcorpan> | annevk: oh |
| 08:18 | <zcorpan> | [tm]: why is https://lists.w3.org/Archives/Member/member-webapi-cvs/ Member-only? |
| 08:21 | <smaug____> | odd |
| 08:48 | <jgraham> | Ms2ger: What's with step_func_done? |
| 08:49 | <Ms2ger> | It calls done() after the step |
| 08:49 | <jgraham> | Well, I see that |
| 08:50 | <jgraham> | But why? |
| 09:16 | <annevk> | zcorpan: legacy decision that we have not yet retroactively changed because nobody cared enough last time it was brought up, I think |
| 09:16 | <annevk> | zcorpan: but as gsnedders can attest my memory has failed me before :) |
| 09:19 | <zcorpan> | annevk: ok |
| 09:19 | <zcorpan> | [tm]: is it possible to make it public? :-) |
| 09:29 | <annevk> | updated Notifications and XMLHttpRequest for event handlers |
| 09:29 | <annevk> | not editing anything else I think that has them |
| 09:40 | zcorpan | wonders what https://bitbucket.org/kcomkar/dom-core/ is about |
| 09:45 | <annevk> | spam I think |
| 09:45 | <annevk> | someone should report it to bitbucket |
| 09:46 | <annevk> | I'll do that |
| 09:47 | <annevk> | https://bitbucket.org/site/master/issue/4272/ |
| 10:31 | AryehGregor | wrote a transforms reftest on February 2 and cannot for the life of him figure out what it's meant to test |
| 10:33 | AryehGregor | makes it do something else |
| 10:36 | <[tm]> | zcorpan: make what public? |
| 10:45 | <AryehGregor> | [tm], presumably https://lists.w3.org/Archives/Member/member-webapi-cvs/ |
| 11:00 | <[tm]> | AryehGregor: ah yeah |
| 11:12 | <jgraham> | AryehGregor: I sometimes find that with tests that I wrote today… |
| 11:12 | <AryehGregor> | jgraham, this is why the CSSWG requires <meta name="assert"> that explains what it's testing. |
| 11:12 | <AryehGregor> | Works well enough for reftests. |
| 11:13 | <Ms2ger> | assertDefined(a.initErrorEvent, 'ErrorEvent.initErrorEvent'); |
| 11:13 | <Ms2ger> | zcorpan-- |
| 11:14 | <jarek_> | does SVG1.1 spec define user agent stylings? |
| 11:15 | <jarek_> | I mean default user agent stylings |
| 11:15 | <jgraham> | AryehGregor: I'm not sure that a mandatory comment really solves "what does this code actually do?", which is the common problem |
| 11:15 | <jarek_> | e.g. when you create <rect> without any fill then it's rendered with black fill on Chrome, is this standard behavior or could it have any fill depending on user agent? |
| 11:15 | <AryehGregor> | "What does this code actually do?" is hopefully determinable by reading the code. |
| 11:15 | <jgraham> | I mean <title> is enough to describe what it ought to be testing |
| 11:15 | <AryehGregor> | "Why does it do it?" is the trick. |
| 11:16 | <AryehGregor> | Really? |
| 11:16 | <AryehGregor> | Assert I just wrote: "This tests that rotatex(90deg) scale3d(2, 1, 2) rotatex(-90deg) is the same as scalex(2) scaley(2). Conjugating by the rotation swaps the Y and Z coordinates for the scale." |
| 11:16 | <AryehGregor> | Title: "CSS Test (Transforms): scale3d() and rotatex()". |
| 11:17 | <AryehGregor> | Obviously, a mandatory comment doesn't do much if it's not used right, but at least it prods lazy people like me into writing something meaningful when we otherwise wouldn't. |
| 11:17 | <jgraham> | I doubt most people will write anything like that much for most tests |
| 11:17 | AryehGregor | likes mandatory function-level comments in code for that reason, except that people tend to not bother updating the comments when they update the code |
| 11:17 | <AryehGregor> | Well, I do tend to like to hear the sound of my own . . . typing. |
| 11:17 | <AryehGregor> | Probably my asserts are mostly too long. |
| 11:28 | <roc> | I don't really like mandatory comments |
| 11:28 | <roc> | people who don't want to write good comments will just write bad ones, which are often worse than no comment at all |
| 11:30 | <AryehGregor> | Well, it depends on how much reviewers know or care about good comments. |
| 11:30 | <roc> | I care |
| 11:30 | <AryehGregor> | If it's some corporate policy from on high and the actually programmers don't get it, then yes, it's pointless. |
| 11:30 | <roc> | because I'm a lazy reviewer |
| 11:30 | <AryehGregor> | s/actually/actual/ |
| 11:31 | <Ms2ger> | roc, oh, we know that ;) |
| 11:31 | <roc> | EXPLAIN WHAT THIS CODE DOES SO I DON'T HAVE TO FIGURE IT OUT |
| 11:31 | <jgraham> | AryehGregor: In the case of tests, it helps to assume that there are no reviewers. Since it is basically true |
| 11:31 | <AryehGregor> | jgraham, not in the CSSWG, for better or worse. |
| 11:32 | <jgraham> | Do they review for correctness or just style? |
| 11:32 | <AryehGregor> | I guess there are two types of comments: "summarize what this long function actually does so I don't have to read through the code", and "explain why the heck you're doing this weird-looking thing, even though I understand perfectly well what the immediate effect is". |
| 11:32 | <AryehGregor> | jgraham, apparently it depends on the reviewer! But in principle, both. |
| 11:32 | <Ms2ger> | jgraham, hey, I reviewed the cssom tests from ttwf |
| 11:32 | <AryehGregor> | I definitely got a few test bugs pointed out. |
| 11:32 | <Ms2ger> | Most were wrong |
| 11:32 | <jgraham> | ttwf? |
| 11:32 | <AryehGregor> | "Test the Web Forward" |
| 11:32 | <jgraham> | Oh Test The Web Forward |
| 11:33 | <AryehGregor> | Ms2ger, wrong like how? Also, since when is CSSOM precise enough that you can write useful tests? |
| 11:33 | <jgraham> | AryehGregor: I guess it is nice if they are pointing out real issues |
| 11:33 | <AryehGregor> | (Well, maybe some parts always were, I'm just bitter because basic stuff like "serialization" is almost totally undefined) |
| 11:33 | <jgraham> | But I don't trust a system that is based on people spending unpaid time doing something boring |
| 11:33 | <Ms2ger> | AryehGregor, most of the others are "Specification issue" :) |
| 11:33 | <jgraham> | With no other reward |
| 11:34 | <AryehGregor> | jgraham, the people reviewing my tests are all WebKit people who actually work on implementing transforms. They are paid, although I don't know if anyone is specifically requiring them to spend time on this. |
| 11:34 | <Ms2ger> | As in, using assert_own_property for attributes, while WebIDL puts them on the prototype |
| 11:34 | <AryehGregor> | It's been Dirk Schulze of Adobe and Simon Fraser of Apple. |
| 11:34 | <Ms2ger> | Or writing tests that setting a readonly attribute throws *and* fails silently |
| 11:34 | <jgraham> | AryehGregor: Did they do a big upfront review or just point out bugs they noticed after trying to run the test? |
| 11:35 | <AryehGregor> | jgraham, I have no idea. I'm guessing that running them in an actual UA is part of it, although they'd have to do a special compile, since WebKit still has prefixed transforms last I checked. |
| 11:35 | <AryehGregor> | (maybe they've unprefixed on trunk, I dunno) |
| 11:35 | <AryehGregor> | (Gecko hasn't yet) |
| 11:35 | <AryehGregor> | (but soon!) |
| 11:35 | <Ms2ger> | There's a patch! |
| 11:36 | <AryehGregor> | Yes! It now even doesn't cause test failures, so I can leave it applied in my queue if I want and not get confused by try failures! |
| 11:36 | <AryehGregor> | (at least, I think it doesn't, I haven't tested) |
| 11:36 | <Ms2ger> | It doesn't |
| 11:36 | <AryehGregor> | Yay. |
| 11:36 | <Ms2ger> | But I can't prove it because of the try reset |
| 11:37 | <AryehGregor> | :( |
| 11:42 | <Ms2ger> | zcorpan, alright, done one dir of workers tests, how many do I have left? :) |
| 11:43 | <zcorpan> | Ms2ger: awesome! |
| 11:44 | <zcorpan> | Ms2ger: which dir? |
| 11:44 | <Ms2ger> | webapps/Workers/tests/submissions/Opera/constructors/Worker |
| 11:48 | <zcorpan> | Ms2ger: i can't think of a good way to count directories right now :-P |
| 11:48 | <Ms2ger> | $ ls -R1 | grep html | wc -l |
| 11:48 | <Ms2ger> | 119 |
| 11:48 | <Ms2ger> | :/ |
| 11:49 | <zcorpan> | many dirs have just two tests though |
| 11:50 | <Ms2ger> | Those are the html files :) |
| 11:50 | <zcorpan> | ah |
| 11:50 | <zcorpan> | so you've done 10% |
| 11:55 | <AryehGregor> | find -type d | wc -l |
| 12:14 | <zcorpan> | hmm, wonder why webkit doesn't support display:block on table elements in quirks |
| 12:52 | <zcorpan> | new quirk everyone, http://simon.html5.org/specs/quirks-mode#the-table-cell-height-box-sizing-quirk |
| 13:14 | <AryehGregor> | Yay, quirk! |
| 13:16 | <AryehGregor> | Speaking of which: view-source:https://bug767169.bugzilla.mozilla.org/attachment.cgi?id=635483 |
| 13:16 | <AryehGregor> | / Document must be in quirks mode to trigger the bug. |
| 13:16 | <AryehGregor> | // Document must be in quirks mode to trigger the bug. |
| 13:29 | <annevk> | hmm view-source: |
| 13:32 | <smaug____> | looks like http://krijnhoetmer.nl/irc-logs/ doesn't work |
| 13:32 | <smaug____> | good that we have the backup logs |
| 13:43 | <annevk> | too bad they're crappy |
| 13:43 | <annevk> | I wish rune and I had succeeded in making something happen |
| 13:43 | <annevk> | maybe next time |
| 13:59 | <smaug____> | annevk: what is crappy? |
| 13:59 | <smaug____> | logbot logs? |
| 14:00 | <smaug____> | I could ask glob to change the format or something |
| 14:04 | <annevk> | smaug____: it's hard to get to the latest logs |
| 14:04 | <annevk> | smaug____: have to fill in forms and such |
| 14:04 | <smaug____> | you need to click a button once |
| 14:05 | <smaug____> | but yeah, the date handling in general is odd |
| 14:05 | <smaug____> | latest logs is easy |
| 14:07 | <nw> | 4 |
| 14:07 | <nw> | whoops |
| 14:08 | <annevk> | 17 |
| 14:08 | <nw> | accidental enter =| |
| 14:11 | <Ms2ger> | 42 |
| 14:13 | <annevk> | end thread |
| 14:14 | <annevk> | Ms2ger: smaug____: should we change event dictionaries to have default values? |
| 14:14 | <annevk> | Ms2ger: smaug____: and then define document.createEvent() in terms of the constructor |
| 14:15 | <annevk> | Ms2ger: smaug____: now dictionaries are always present |
| 14:15 | <annevk> | What we should really have I think is IDL for them |
| 14:18 | <annevk> | at least everything but Event itself is really simple |
| 14:22 | smaug____ | can't wait the time when 3G connections are reliable even in trains |
| 14:27 | <smaug____> | it would be nice to see the default values in the event dictionaries |
| 14:28 | <smaug____> | though |
| 14:28 | <smaug____> | those are usually empty string, false, and 0 and null |
| 14:29 | <annevk> | yeah sure, but right now that's all in prose |
| 14:29 | <annevk> | mostly because of document.createEvent |
| 14:29 | <annevk> | but I realized the other day we could define that in terms of constructors |
| 14:29 | <annevk> | should probably ask heycam|away what he thinks |
| 14:40 | <smaug____> | annevk: you'll be on vacation soon? |
| 14:44 | <jgraham> | smaug____: This *is* annevk on vacation |
| 14:45 | <annevk> | indeed :) |
| 14:45 | <smaug____> | ahaa. good. So annevk on vacation may still do some spec editing. |
| 14:46 | <annevk> | just with more intermittent beaches/fun |
| 15:24 | <annevk> | I keep confusing the Glenns when reading email |
| 16:43 | <jgraham> | Has gecko dropped E4X yet? |
| 16:43 | <Ms2ger> | No :( |
| 16:43 | <jgraham> | Cryptic error messages about XML names in ECMAScript are no fun |
| 16:45 | <jgraham> | Also it's kind of a silly idea |
| 16:45 | <jgraham> | H4X seems bad for the same reasons |
| 16:46 | <divya> | whats e4x |
| 16:46 | <divya> | :/ |
| 16:46 | <Ms2ger> | You don't want to know |
| 16:47 | <jgraham> | divya: We operate a strict "you ask, we don't tell" policy on that |
| 16:47 | <divya> | ahahahaha |
| 16:48 | divya | tweeets that. |
| 16:48 | <jgraham> | (s/H4X/E4H/ I guess) |
| 16:49 | <divya> | jgraham: y u no on twitter. |
| 16:50 | <divya> | even Ms2ger is on! |
| 16:50 | <divya> | you got no excuse! |
| 16:50 | <Ms2ger> | Leave Dr. Graham be, or he'll tell you to get off his lawn |
| 16:50 | <jgraham> | Be glad you're not a slug. If you were a slug on my lawn I would cut you in half |
| 16:51 | <divya> | :)) |
| 16:51 | <jgraham> | I think I am on twitter just as much as Ms2ger is |
| 16:52 | <Ms2ger> | "Dude next to me at SFO just said on the phone: "My startup is social and viral." -- What is that? The common flu?!?" |
| 16:53 | <divya> | LOLOLOL |
| 16:53 | <divya> | jgraham: WAITTT where is your twitter account |
| 16:53 | <divya> | Ms2ger: are you paraphrasing someone or quoting yourself :P |
| 16:54 | <Ms2ger> | I'm on the twatters |
| 16:54 | <divya> | https://twitter.com/#!/ms2ger |
| 16:54 | <divya> | totally |
| 16:54 | <divya> | jgraham is not afaik |
| 16:54 | <Ms2ger> | https://twitter.com/pfinette/status/218002531366871041 |
| 16:54 | <divya> | ah |
| 17:02 | <Ms2ger> | http://www.w3.org/Consortium/Recruitment/#techeditor |
| 17:02 | <Ms2ger> | "W3C is seeking a technical editor with Web technologies skills to serve as lead editor for the HTML Working Group." |
| 17:06 | <gsnedders> | jgraham: E4X is forbidden in strict-mode, FWIW |
| 17:07 | <gsnedders> | divya: jgraham's account name follows the corrolation from Ms2ger to Ms2ger, with an equal number of tweets. |
| 17:25 | <jamesr> | hober, yo! |
| 17:33 | <pablof> | it would be nice to have the Archived-At header in whatwg emails, like the w3c has on (most?) of their lists |
| 17:33 | <pablof> | i wonder if that's just a checkbox somewhere in mailmans configs or if it needs actual code to be enabled |
| 17:35 | <Philip`> | That has been proposed before, and I believe it was not found to be trivial to enable |
| 17:37 | <gsnedders> | pablof: Mailman doesn't support it, and it's non-trivial to add. |
| 17:37 | <gsnedders> | Was on the to-do list for Mailman 3, at least |
| 17:38 | <pablof> | gsnedders: hmm, i found this, not sure if it's hard to do, but it doesn't look trivial either. http://packages.python.org/mailman/src/mailman/handlers/docs/rfc-2369.html |
| 17:38 | <pablof> | (and it looks like that's for a 3.0 beta, yeah) |
| 18:07 | <n00lk0w1> | can data-* attr names include underscores in the "*"? |
| 18:07 | <n00lk0w1> | is data-cmtr_name a valid attr name? |
| 18:09 | <miketaylr> | n00lk0w1: sure |
| 18:10 | <n00lk0w1> | miketaylr: thanks :P |
| 18:43 | <jgraham> | http://www.w3.org/Consortium/Recruitment/#techeditor - interesting |
| 20:17 | <hober> | jamesr: yo |
| 20:17 | <jamesr> | hober, about position:sticky - does it do the same in the X and Y? |
| 20:17 | <hober> | jamesr: sorry, was busy obstructing progress in the html wg |
| 20:17 | <jamesr> | ooh that's a fun pasttime |
| 20:18 | <hober> | I suppose I should explicitly </sarcasm> for those reading the logs |
| 20:18 | <hober> | anyway, sticky |
| 20:18 | <jamesr> | you cited news.google.com's sidebar, i was wondering if you noticed that it behaves differently for vertical and horizontal scrolling |
| 20:18 | <jamesr> | it accomplishes the horiz scrolling behavior with script |
| 20:18 | <hober> | yeah, indded |
| 20:18 | <jamesr> | it basically wants to be sticky for vertical scrolling and just move with the page horizontally |
| 20:19 | <hober> | they could do that with foo { position:sticky; top: 10px; } |
| 20:19 | <hober> | that would only stick in the y direction |
| 20:19 | <jamesr> | what dictates that? |
| 20:19 | <jamesr> | since they didn't specify left ? |
| 20:19 | <hober> | yeah |
| 20:20 | <jamesr> | what if they want to alter the static position's left? |
| 20:20 | <hober> | "the static position"? :) |
| 20:20 | <jamesr> | well what if they want to position it and then also have this sticky behavior |
| 20:21 | <jamesr> | abspos container to stick it where they want, then sticky inside that? |
| 20:21 | <hober> | sure, that should work |
| 20:27 | <hober> | wait, why wouldn't you just use margin-left? |
| 20:37 | <TabAtkins_> | Positioning works on the margin box. |
| 20:37 | <TabAtkins_> | Oh, that's what you're saying. I misparsed your sentence. |
| 20:40 | <hober> | yeah, i don't think jamesr needs (or wants) an extraneous container element to achieve the effect he describes; margin-left does it |
| 20:48 | <Hixie> | annevk: or, you know, we could just call it "the document's address"... |
| 20:48 | <anatolbroder> | Hello world! What do you think, will consumer displays (mobile, desktop, tv) exceed 1920 soft pixels in the next 5 years? |
| 21:08 | <anatolbroder> | I mean 1920 CSS pixel. |
| 21:26 | <Hixie> | anyway web video implementors about? |
| 21:39 | <Hixie> | i wonder what to call the attribute that https://www.w3.org/Bugs/Public/show_bug.cgi?id=13359 asks for |
| 21:39 | <Hixie> | TextTrack.doohickey? |
| 22:15 | <zewt> | Hixie: seems like it should just use data-* attributes, at least for now |
| 22:15 | <TabAtkins_> | anatolbroder: That would be a very large display. Remember that the CSS pixel is defined in terms of visual angle, so 1920 pixels is about 40 degrees of vision. |
| 22:15 | <zewt> | at least, until there are use cases for standardizing sub-types |
| 22:16 | <zewt> | TabAtkins_: that's pretty meaningless; no system knows how far the user is from the screen |
| 22:16 | <zewt> | (or some value approaching "none") |
| 22:16 | <TabAtkins_> | zewt: Theoretically, sure. In practice, screens are designed for a particular viewing range. |
| 22:16 | <TabAtkins_> | Phone versus desktop vs TV vs projector. |
| 22:16 | <zewt> | one with a very wide range |
| 22:17 | <TabAtkins_> | Sure, more or less a 2x factor each way, I'd say. |
| 22:17 | <zewt> | projectors are probably the worst case, since projectors don't even know how large a screen they're projecting (beyond some very broad limits) |
| 22:18 | <zewt> | (also a fairly rare case, though) |
| 22:18 | <Philip`> | (Don't they make projectors with automatic range-finders so they can correctly focus themselves?) |
| 22:19 | <zewt> | (if so, I doubt they pass that data on to the source) |
| 22:19 | <zewt> | it would be pretty amusing to have a webpage layout change as you move a projector around, though :) |
| 22:20 | <TabAtkins_> | Hixie: I'm with zewt - as far as I can tell, data-* attributes should be fine. I didn't read the entire bug thread, though, as it was very long. |
| 22:20 | <zewt> | (I only read the OP) |
| 22:21 | <zewt> | (once there's a body of metadata types, and people using them, it would probably make more sense to spec something more concrete) |
| 22:21 | <TabAtkins_> | Yeah, as new @kind values. |
| 22:21 | <zewt> | well, I'd leave that alone, and the spec would just be subcategories of metadata |
| 22:22 | <Hixie> | zewt, TabAtkins_: on what element would you be putting this content attribute? |
| 22:22 | <zewt> | (otherwise, you'd have to spec that unknown @kind means "metadata", for forwards-compatibility, which could cause problems with adding new unrelated @kinds later) |
| 22:22 | <TabAtkins_> | Hixie: The <track>. |
| 22:22 | <zewt> | Hixie: <track> |
| 22:22 | <Hixie> | zewt, TabAtkins_: this is about exposing an in-band text track's doohicket to an IDL object, the only element is the <video> or <audio>, there is no <track>. |
| 22:22 | <Hixie> | doohickey, even |
| 22:22 | <TabAtkins_> | Oh! |
| 22:23 | <TabAtkins_> | How are these doohickey's exposed? Is there a consistent way of encoding and detecting this? |
| 22:23 | <TabAtkins_> | s/'// |
| 22:23 | <Philip`> | zewt: Moving the projector around sounds like something that augmented reality people should be interested in - you could design a room with a web page virtually spread around its walls, and the projector works out where you're pointing and projects the appropriate section of the page to correspond to that wall |
| 22:23 | <zewt> | well, there's a separate discussion about adding metadata headers to VTT; that would, presumably, be exposed with an API on TextTrack |
| 22:23 | <Hixie> | TabAtkins_: it's a mess, but the bug lists some conventions we're going to use |
| 22:24 | <Hixie> | zewt: metadata in vtt is a solution in search of a use case, so far. (might be that the use case is in some of the feedback i haven't read yet) |
| 22:24 | <Philip`> | (It'd fit the usual AR pattern of taking something that you can already do conveniently on a computer, and then adding a layer of awkward UI on top of it so that it's harder to access the content) |
| 22:24 | <zewt> | Hixie: i've listed some concrete use cases (eg. providing info to muxers); this would (perhaps--havn't read into it enough) likely be another use case |
| 22:25 | <zewt> | though, maybe not |
| 22:25 | <TabAtkins_> | Hixie: Hm. In that case, I guess textTrack.metadata |
| 22:25 | <zewt> | since I'd think you'd want this info in <track> so you can access it before the vtt is fetched ... guess I need to read through it (bleh) |
| 22:25 | <zewt> | > Page origin O exposes video/audio from a 3rd party P, which contains/refers to a metadata track sourced by P or sourced by a 4th party M. |
| 22:26 | zewt | squints |
| 22:27 | <zewt> | basically it either fits in <track data-xxx=> for the "available without fetching" optimization, or as VTT file metadata for the "not available without fetching, but can be sourced from a third-party without putting it in <track>" optimization |
| 22:28 | <zewt> | the former is more consistent, though (where you always put that sort of metadata on <track>, even if that means you have to jump some extra hoops for third-party content) |
| 22:33 | <Hixie> | TabAtkins_: that seems a bit overly vague for something that is really just a type identifier for proprietary track dispatch |
| 22:34 | <TabAtkins_> | I'm assuming that more stuff will probably all out of the "get me some data from the stream". |
| 22:34 | <zewt> | parse error |
| 22:34 | <Hixie> | fall |
| 22:35 | <Hixie> | TabAtkins_: yeah, but then we'd want to use "metadata" for that :-) |
| 22:35 | <TabAtkins_> | Not actually sure what I was trying to type there. |
| 22:35 | <Hixie> | maybe media.inBandMetadataDispatchType |
| 22:35 | <Hixie> | or media.inBandMetadataTrackDispatchType |
| 22:36 | <Hixie> | er |
| 22:36 | <Hixie> | track. |
| 22:36 | <zewt> | seems like it'd make more sense to just implement generic VTT headers, and people can stick "Dispatch-Type: parental-advisery", then say track.metadata["Dispatch-Type"] |
| 22:38 | <zewt> | (or "X-Dispatch-Type" or "data-Dispatch-Type" or whatever the equivalent of data-* is in that namespace) |
| 22:40 | <Hixie> | that's like a zillion times more complicated |
| 22:40 | <Hixie> | and wouldn't work :-) |
| 22:40 | <zewt> | why wouldn't it work? |
| 22:40 | <Hixie> | the whole point is to expose specific existing fields that differ by format in a consistent way |
| 22:41 | <currency> | hey |
| 22:41 | <currency> | i need some HTML6, where can i dl that and hav a luk |
| 22:41 | <TabAtkins_> | Try pirate bay, I hear they leaked it early. |
| 22:41 | <zewt> | expose how? isn't this only for JS to know how to interpret the contents of metadata fields, not for UAs exposing it differently? |
| 22:42 | <zewt> | (this bug is a headache to read; comments nested seven deep) |
| 22:42 | <Hixie> | i don't understand the question |
| 22:42 | <zewt> | i don't understand what you mean won't work |
| 22:43 | <Hixie> | this is for exposing specific in-band metadata track data so that a script can know which subscript module to dispatch a track to |
| 22:43 | <Hixie> | there are specific fields in existing in-band formats that they want to use for this |
| 22:43 | <Hixie> | it's not a generic thing |
| 22:43 | <zewt> | not seeing how what I described wouldn't handle that |
| 22:44 | <zewt> | if the metadata fields in a VTT file are JSON with some schema (for example), the script does "if(track.metadata['Type'] == 'my-proprietary-json-format') { handleCueWithMyFormat(cue); }" |
| 22:44 | <currency> | html6 leaked on pirate bay? |
| 22:45 | <Hixie> | this has nothing to do with VTT |
| 22:45 | <zewt> | uh, okay, you could have said that ten minutes ago :) |
| 22:45 | <zewt> | then what does this have to do with HTML? |
| 22:45 | <Hixie> | i don't know how vtt entered the conversation, i keep saying it's about in-band tracks for specific formats :-) |
| 22:46 | <Hixie> | it's about HTML in that it's about the HTML TextTrack API |
| 22:46 | <zewt> | well, the bug is about text tracks, and the only text track format supported by HTML that I know of is VTT :) |
| 22:46 | <Hixie> | HTML is track format agnostic |
| 22:47 | <currency> | did html6 drop yet? |
| 22:47 | <currency> | when is apple going to roll out html6? |
| 22:47 | <zewt> | (of course, you could apply the API just the same, with other formats exposing their type as a metadata field) |
| 22:47 | <zewt> | troll score: 1/10 |
| 22:48 | <Hixie> | afk for a bit |
| 22:48 | <Hixie> | think i'll go with track.inBandMetadataTrackDispatchType for now and hope someone comes up with a better name (or it never gets implemented and we can forget about it) |
| 22:54 | <smaug____> | inBandMetadataTrackDispatchType is almost long enough to be invented by me |
| 22:55 | <hober> | currency: we already shipped html7, duh |
| 22:56 | <currency> | hober : ar eu with apple? |
| 22:56 | <TabAtkins_> | <meme>Unsure if trolling or crazy</meme> |
| 22:57 | <currency> | nice markup |
| 22:57 | <currency> | im not trolling |
| 22:57 | <currency> | i was reading about html6 |
| 22:57 | <currency> | but then i looked at the date it was posted |
| 22:57 | <currency> | and it was april 2nd lol |
| 22:57 | <currency> | of course that was after i joind this channel and already askd about it |