| 00:00 | <Hixie_> | queued tasks don't fire because tasks get delayed, but that doesn't help with dispatchEvent()... |
| 00:01 | <Hixie_> | annevk? |
| 07:17 | <MrTango> | hi, i will add new DC-Tags like DC.format and DC.type which are widely used in CMS Plone, can anyone create me an account http://wiki.whatwg.org with the username MrTango please? |
| 07:22 | <annevk> | heycam|away: ah yeah, it's just a return value so sequence could work I suppose |
| 07:22 | <annevk> | heycam|away: but yeah, I think we should make it more obvious how this works |
| 07:22 | <annevk> | heycam|away: because the current situation is too confusing |
| 07:24 | <annevk> | Hixie_: not that I know |
| 07:25 | <annevk> | Oh you filed a bug |
| 07:25 | <annevk> | MrTango: I can help you out I guess |
| 07:26 | <annevk> | MrTango: I need an email for that though |
| 07:27 | <MrTango> | annevk: md⊙dd |
| 07:27 | <annevk> | "A randomly generated password for MrTango has been sent to md⊙dd" |
| 07:28 | <MrTango> | annevk: thank you |
| 08:09 | <zcorpan> | jgraham: should this review be dropped? https://critic.hoppipolla.co.uk/showcomment?chain=393 |
| 08:12 | <Ms2ger> | jgraham, how much of a review do you want for the html5lib update? |
| 08:13 | <zcorpan> | TabAtkins: "This style sheet could be used by an implementation as part of its default styling of HTML4, XHTML1, XHTML1.1, XHTML Basic, and other XHTML Family documents." http://tabatkins.github.io/specs/css-color/Overview.html#sample |
| 08:14 | <zcorpan> | 2003 called and wants its XHTML Family members back |
| 08:16 | <zcorpan> | TabAtkins: also the stylesheet doesn't match what's in HTML's rendering section. maybe drop that section altogether? |
| 08:20 | <zcorpan> | TabAtkins: wonder if we should spec what the system colors map to |
| 08:25 | <wilhelm_> | I have just encountered RDF in the Real World for the first time. I expected it to be bad, but, oh, $deity, please get me away from here. |
| 08:38 | <hsivonen> | wilhelm_: What was the context? |
| 08:42 | <Ms2ger> | People here might be interested in http://mozilla.pettay.fi/workerconsole/ too |
| 08:44 | <wilhelm_> | hsivonen: My customer (which is the Norwegian equivalent of Encyclopedica Britannica) is importing a corpus from a different publisher. Our stuff is HTML in an SQL database. Theirs involves RDF, SPARQL and everything else you can think of from that house of horrors. |
| 08:44 | <hsivonen> | wilhelm_: nice |
| 08:53 | <SimonSapin> | zcorpan, TabAtkins: spec’ing reasonable defaults for system colors: yes please |
| 08:54 | <zcorpan> | SimonSapin: has anyone documented what they map to on different systems? |
| 08:55 | <SimonSapin> | zcorpan: I think that "current browsers" implement it for real: ask the system for colors from the user’s theme |
| 08:55 | <SimonSapin> | but I don’t really want to bother, in Servo |
| 08:56 | <zcorpan> | SimonSapin: yes. my question stands :-) |
| 08:56 | <SimonSapin> | oh, what they map |
| 08:56 | <SimonSapin> | I was thinking of a "name -> color" map |
| 08:57 | <jgraham> | zcorpan: I don't know. I think Aryeh should decide if there is anything he wants to salvage |
| 08:57 | <jgraham> | Ms2ger: Not too much I think |
| 08:57 | <jgraham> | I wouldn't review the generated files (which is why they are in their own commit) |
| 08:58 | <annevk> | zcorpan: should be pretty easy to generate this map |
| 08:58 | <jgraham> | Feel free to review the update code, but it's pretty ugly and probably not worth spending time on small issues |
| 08:59 | <zcorpan> | annevk: yeah, my problem is just that i don't have different systems. but thinking about it there are browser screenshot services to solve that |
| 09:00 | <Ms2ger> | jgraham, heh, namespaces |
| 09:01 | <Ms2ger> | jgraham, how about you fix the trailing whitespace and I claim I've reviewed it all? |
| 09:02 | <annevk> | zcorpan: I can get you Windows (7 I think) and Mac OS X |
| 09:02 | <zcorpan> | annevk: i have those |
| 09:02 | <annevk> | zcorpan: can also crowdsource if you can figure out some way to communicate back in 140 chars |
| 09:04 | <zcorpan> | a color is 24 bits... there are 28 system colors... |
| 09:04 | <hsivonen> | annevk: Crowdsourcing hasn't worked for me in the case of trying to find a twitter follower who runs the Greek version of Windows or the Welsh version of Windows. :-( |
| 09:06 | <annevk> | zcorpan: you have 140 Unicode scalar values even, seems like it ought to be doable |
| 09:06 | <zcorpan> | annevk: 96 ascii chars seems like it could encode the data |
| 09:07 | <annevk> | hsivonen: hmm yeah, localization is harder |
| 09:07 | <zcorpan> | not printable ascii though |
| 09:09 | <SimonSapin> | The thing with system colors is that there is not just one set of them per OS, they also change when the users change the desktop "theme" |
| 09:10 | <zcorpan> | SimonSapin: yeah, but if we're going to spec a fixed set of colors, the default themes seems like what we should care about |
| 09:10 | <SimonSapin> | zcorpan: or just pick one default set, regardless of OS |
| 09:11 | <jgraham> | Ms2ger: Sounds like it could work to me :) |
| 09:11 | <SimonSapin> | Tab’s draft suggest white for background-ish system colors, black for the rest |
| 09:11 | <SimonSapin> | all that matters is keeping things readable in legacy content |
| 09:12 | <zcorpan> | SimonSapin: yes. i'm trying to figure out what the set should be. just black/white seems unnecessarily boring |
| 09:12 | <SimonSapin> | fair enough |
| 09:14 | <zcorpan> | turns out grepping for system colors in webdevdata is a good way to fill up the hard drive (not from matches but from memory usage) |
| 09:15 | <SimonSapin> | gecko’s implementation is in mozilla-central/widget/*/*LookAndFeel* |
| 09:21 | <jgraham> | hsivonen: I wouldn't be surprised if the total number of users of the welsh version of windows is in the thousands |
| 09:21 | <jgraham> | So you owuld need a big crowd |
| 09:23 | <jgraham> | (it seems that there are about half a million Welsh speakers in total, so it could be higher. But very few of those don't also speak English and I imagine there isn't much Welsh-localised software so it is probably easier to have English everywhere than a mix) |
| 09:34 | <zcorpan> | SimonSapin: annevk: http://www.browserstack.com/screenshots/67610195dac7cc4cb12206dd2589f08b3d0a7c02 |
| 09:38 | <annevk> | zcorpan: now all you need is wget plus OCR |
| 09:39 | <jgraham> | Ms2ger: Updated |
| 09:50 | <zcorpan> | looks like iOS and Android have the same colors |
| 09:50 | <zcorpan> | that by itself seems like a reasonable candidate |
| 10:19 | <matjas> | annevk, zcorpan: as for crowdsourcing, maybe you could use Browserscope? (see <iframe> here: http://mathiasbynens.be/demo/xhr-responsetype) |
| 10:19 | <matjas> | browserscope requires the results to be normalized to numbers, but that’s not a problem in this case :) |
| 10:20 | <zcorpan> | matjas: interesting |
| 10:20 | <matjas> | just create an HTML page like that, tweet the link, and watch the results flow in |
| 10:20 | <zcorpan> | matjas: i think i have enough data already, but feel free to gather more if you like |
| 10:21 | <annevk> | if we map locale to a number... |
| 10:21 | <annevk> | and encoding too I suppose |
| 10:21 | <zcorpan> | matjas: also, does it differentiate OSes or just browser versions? |
| 10:22 | <matjas> | just browser versions afaik |
| 10:23 | <matjas> | a Browserscope clone that accounts for OS diffs would be neat |
| 10:29 | <zcorpan> | sent http://lists.w3.org/Archives/Public/www-style/2013Aug/0667.html |
| 10:30 | <annevk> | oh, so that wouldn't even work for what I had in mind |
| 11:09 | <annevk> | aah |
| 11:09 | <annevk> | "ZIP – a base version of this data compression and archive file format is in the public domain, but newer versions have some patented features[13][14][15]" |
| 11:12 | <MikeSmith> | that doesn't sound good |
| 11:15 | <annevk> | It seems that's related to encryption, which we won't have. |
| 11:16 | <annevk> | In fact, from reading it seems the zip format we care about is in the public domain so we can do whatever we want. |
| 11:16 | <annevk> | Not sure I'm keen on defining it though. |
| 11:19 | <jgraham> | annevk: Probably should have chosen a different career then :) |
| 11:21 | <MikeSmith> | well Tracking Protection WG needs a chair |
| 11:21 | <MikeSmith> | that'd be a nice career change |
| 11:22 | <MikeSmith> | I can't find an html5lib test for the plain case of <table><input></table> |
| 11:22 | <MikeSmith> | that is, without the "hidden" attribute |
| 11:23 | <annevk> | MikeSmith: you mean type=hidden? |
| 11:23 | <Ms2ger> | MikeSmith, add one! |
| 11:23 | <MikeSmith> | ah yeah that I mean |
| 11:23 | <annevk> | MikeSmith: also, don't wanna get demoted |
| 11:23 | <MikeSmith> | hah |
| 11:24 | <MikeSmith> | that job is just a stepping stone man |
| 11:24 | <MikeSmith> | nobody takes that job for the job itself |
| 11:24 | <MikeSmith> | is the track it provides to other jobs |
| 11:25 | <Ms2ger> | The tracking, if you will |
| 11:25 | <MikeSmith> | yeah, like that |
| 11:25 | <MikeSmith> | you take the job and then who knows your might be moved up to CSS WG chair next |
| 11:26 | <MikeSmith> | and then Web Foundation CEO maybe |
| 11:32 | <annevk> | Does "utf-8 bomless decode" make sense? Or should it be "utf-8 BOM-less decode"? |
| 11:32 | <annevk> | I guess the latter |
| 11:33 | <SimonSapin> | annevk: the latter looks better |
| 11:33 | <annevk> | Not to me :) |
| 11:33 | <annevk> | But we haven't really introduced bom as a noun so I guess... |
| 11:34 | <annevk> | On the other hand, you might as well just use the utf-8 encoder directly in that case, that's shorter :) |
| 11:35 | <SimonSapin> | annevk: and while you’re at it, it would be nice to avoid similar names like decode vs decoder for different concepts :) |
| 11:35 | <annevk> | Not constructive enough for me to care, sorry |
| 11:39 | <jgraham> | "the bom" -> sounds like a noun to me |
| 11:45 | <SimonSapin> | annevk: uhg. ok. Let’s try more constructive. The "encode" concept does nothing but call an encoding’s encoder, I’m not convinced it’s worth keeping at all. "decode" looks for a BOM, then calls an encoding’decoder. I (and I believe other people) have been confused before by the similarity in the name. Perhaps this could be fixed by renaming "decode" to "BOM-decode" or something more specific, or |
| 11:45 | <SimonSapin> | even split the bom detection from the actual decoding. |
| 12:42 | <annevk> | SimonSapin: this is that split |
| 12:42 | <annevk> | SimonSapin: the idea of this section was to have an API-layer between specs and the encoder/decoders |
| 12:43 | <annevk> | SimonSapin: however, it hasn't completely materialized yet what the details of that layer should be |
| 12:43 | <SimonSapin> | annevk: I mean having something like "detect the encoding" that return an encoding |
| 12:43 | <jgraham> | Any opinions on https://critic.hoppipolla.co.uk/showcomment?chain=424? |
| 12:44 | <annevk> | SimonSapin: seems like that would be horrible misunderstood and would also get the precedent order wrong |
| 12:44 | <SimonSapin> | yeah, I agree |
| 12:52 | <annevk> | jgraham: seems okay given that testing the tokenizer separately is really an implementation-specific thing |
| 13:11 | <annevk> | SimonSapin: if you have time at some point we could go through thinking about how to restructure the Encoding Standard |
| 13:15 | <SimonSapin> | annevk: later today? |
| 13:15 | <annevk> | SimonSapin: sure |
| 13:16 | <jgraham> | annevk: Yeah, but in this case the tests would be "treebuilder" tests, just using the html5lib tokenizer tests as a source of "interesting" trees |
| 13:16 | <jgraham> | Or interesting inputs, rather |
| 13:21 | <annevk> | jgraham: in that case... |
| 13:21 | <annevk> | jgraham: can't you convert them to the other format then? |
| 13:21 | <annevk> | jgraham: and remove dupes |
| 13:22 | <jgraham> | I did at some point. But it rather relies on html5lib as a reference implementation. |
| 13:23 | <jgraham> | Seems like the tokenizertotree code still exists, so maybe I should just do it |
| 13:25 | <annevk> | I relied on url.js when generating URL tests... |
| 13:36 | <annevk> | Domenic_: will you post consensus check or should I? |
| 13:36 | <Domenic_> | annevk: I got it |
| 13:37 | <annevk> | Domenic_: I'd kinda prefer a cc to public-script-coord and www-dom |
| 13:37 | <annevk> | Domenic_: to catch any implementers |
| 13:37 | <Domenic_> | annevk: ok, sure. if you'd rather write it up that's fine too |
| 13:37 | <annevk> | Domenic_: nah, you should take credit, was just offering in case you were busy |
| 13:40 | <annevk> | Domenic_: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cscript%3Ew%28PromiseResolver%29%3C%2Fscript%3E ... |
| 13:41 | <annevk> | Domenic_: but yeah, that can go away |
| 13:43 | <Domenic_> | annevk: not sure what i should be seeing there; PromiseResolver is not defined (but I'm only running FF 23) |
| 13:45 | <annevk> | Domenic_: ah, gotta run non-beta/non-stable I think |
| 13:52 | <annevk> | Domenic_: not sure if I told you, but I also put this proposal on the TC39 agenda; kinda crappy they operate that way, but so be it; hoping it's more or less settled before than though |
| 13:52 | <Domenic_> | annevk: yeah i saw that. I think I'll actually be able to make that meeting since it's in Boston. |
| 13:53 | <Domenic_> | (I'm in NYC) |
| 13:53 | <annevk> | Domenic_: oh that'd be nice |
| 13:53 | <annevk> | Domenic_: you gonna be at Edge? |
| 13:53 | <Domenic_> | annevk: Don't think so? What's Edge? |
| 13:53 | <annevk> | Domenic_: conf in NY |
| 13:53 | <annevk> | Domenic_: Monday after TC39 |
| 13:53 | <annevk> | Domenic_: http://edgeconf.com/ |
| 13:54 | <annevk> | Domenic_: 100 USD, seems they might have a couple of seats left |
| 13:54 | <Domenic_> | annevk: hmm seems fun, might be able to beg another day off work for it. |
| 13:58 | <annevk> | Domenic_: thanks for that email |
| 13:59 | <Domenic_> | annevk: np |
| 14:48 | <Domenic_> | lol mark way to make things way more confusing with your second paragraph haha |
| 14:51 | <annevk> | "Note that this turns the common use case on its head." o_O |
| 14:51 | <annevk> | Did we email this at the wrong time or something? |
| 14:55 | <annevk> | I'm the 1%: https://twitter.com/dontcallmeDOM/status/373444615199682560 |
| 16:05 | <SimonSapin> | annevk: ping |
| 17:34 | <Ms2ger> | r? https://critic.hoppipolla.co.uk/r/294 |
| 18:10 | <rniwa> | annevk: yt? |
| 18:11 | <annevk> | rniwa: for a bit |
| 18:11 | <rniwa> | annevk: do you know if xhr's response supposed to keep the same document or not |
| 18:11 | <rniwa> | annevk: when the type is a document? |
| 18:11 | <annevk> | rniwa: yeah is |
| 18:11 | <rniwa> | annevk: okay |
| 18:11 | <rniwa> | annevk: which part of the spec says that? |
| 18:11 | <annevk> | rniwa: http://xhr.spec.whatwg.org/#document-response-entity-body |
| 18:12 | <rniwa> | annevk: it doesn't say to use the same document... |
| 18:12 | <annevk> | rniwa: it says to return the response entity body, and only if that has no value are you allowed to create a new one |
| 18:12 | <annevk> | document response entity body* |
| 18:13 | <rniwa> | annevk: ah, now i see it |
| 18:13 | <rniwa> | annevk: thanks |
| 18:13 | <annevk> | I should more properly reset it though :/ |
| 18:14 | <annevk> | I'll file a bug on doing that |
| 18:15 | <annevk> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=23102 |
| 18:19 | <TabAtkins> | Do Event names have to be unique, or just unique to the interfaces that can see them? |
| 18:20 | TabAtkins | is wondering whether Bikeshed needs to consider <dfn event> to be a type that requires a for='' attribute. |
| 18:22 | <Ms2ger> | TabAtkins, event types are not constrained |
| 18:24 | <TabAtkins> | Ms2ger: What does that mean in the context of my question? |
| 18:24 | <Ms2ger> | I'm not sure |
| 18:24 | <TabAtkins> | Heh, kk. |
| 18:24 | <Ms2ger> | I don't understand your question |
| 18:24 | <Ms2ger> | What are Event names? |
| 18:24 | <TabAtkins> | Yeah, thought so. Let me rephrase. |
| 18:25 | <TabAtkins> | Bikeshed/Shepherd parse specs for definitions, and categorize them across types. The hope is that, most of the time, just specifying what you want to link to, and what type it is, should be enough to uniquely identify the right link target. |
| 18:26 | <TabAtkins> | Some types are roughly a global namespace, so that's good. Others aren't, like property values, so their definitions additionally get a for='' attribute specifying what they're for (the thing they're for is hopefully global). |
| 18:26 | <TabAtkins> | When I'm wrapping an event name in a <dfn>, do I need a for='' or equivalent? |
| 18:26 | <Ms2ger> | What is an event name? |
| 18:27 | <Ms2ger> | And why do you have a dfn for it? |
| 18:27 | <Ms2ger> | If you want to fire an event, you say "fire an event named 'x'" |
| 18:27 | <TabAtkins> | Oh wait, duh, I just answered my own question. Yes, I need a for='', to point to the relevant event interface. Names are global *when constrained to a relevant interface*. |
| 18:27 | <TabAtkins> | Right, but you still want a place where there's a canonical definition of what a given event is *for*. |
| 18:29 | <Ms2ger> | jgraham, https://github.com/w3c/web-platform-tests/pull/248 doesn't seem to have a critic review? |
| 18:30 | <annevk> | TabAtkins: load is used in incompatible ways, often with the same interface |
| 18:30 | <annevk> | TabAtkins: same for readystatechange |
| 18:30 | <TabAtkins> | annevk: Oh, hrm. |
| 18:30 | <TabAtkins> | Annoying, then. |
| 18:30 | <annevk> | TabAtkins: events are not really defined as such though |
| 18:31 | <annevk> | TabAtkins: events follow from an algorithm that dispatches them |
| 18:31 | <annevk> | TabAtkins: the original DOM series about Events distracted people by making them appear different |
| 18:32 | <TabAtkins> | Perhaps they're unique when constrained to a given event target? |
| 18:32 | <TabAtkins> | The issue is, Anne, that I want it to be easy to tell the difference between, say, "readystatechange" when fired for different types of algorithms, so you can link to that particular usage. |
| 18:34 | <TabAtkins> | At least, I *think* that's a reasonable thing to want to do. Is it? |
| 18:35 | <TabAtkins> | On a related note, should I kill CSSFontFaceLoadEvent, or is that a reasonable thing to do? What's the current thinking on how to do events? |
| 18:35 | <annevk> | Yeah, I think event names are generally unique per target. |
| 18:37 | <annevk> | I think events are still okay, we just haven't needed them much for a lot of new stuff... With regards to fonts. It seems people want explicit control over fonts in JavaScript. |
| 18:37 | <annevk> | With fonts as first-class primitives. |
| 18:37 | <TabAtkins> | Yes, I'm reworking the spec right now, but starting with the load events. |
| 18:37 | <TabAtkins> | I need to come up with something that'll work in Workers, though. |
| 18:39 | <TabAtkins> | Which probably means coming up with a first-class Font object, yeah. |
| 18:42 | <annevk> | I recommend starting with the primitive. And adding the event for "all fonts are here" later. No need to solve all the use cases at once. |
| 18:47 | <annevk> | zewt: if you could add why you think APPNOTE.TXT is not an option on http://wiki.whatwg.org/wiki/Zip that'd be nice |
| 18:47 | <annevk> | zewt: would be good to have that documented somewhere |
| 18:47 | <annevk> | zewt: especially given how many other standards bodies didn't seem to care |
| 18:50 | <Ms2ger> | jgraham, r? https://critic.hoppipolla.co.uk/r/216 |
| 18:56 | <zewt> | annevk: i hope we care |
| 18:57 | <zewt> | i didn't even know I had an account on this wiki, but apparently Firefox remembered |
| 18:57 | <annevk> | zewt: we'll see, I'm not too interested in working on it |
| 18:58 | <zewt> | i don't blame you for that, heh |
| 19:11 | <jgraham> | Ms2ger: Don't know what happened there |
| 19:13 | <jgraham> | Ms2ger: I will tey to look at the review, but internet access might be spotty |
| 19:14 | <Ms2ger> | jgraham, thanks |
| 21:33 | <miketaylr> | back |
| 21:56 | <TabAtkins> | annevk: I already have strong use-cases for "tell me when X font is loaded", "tell me when all fonts are loaded", "load this font", and "tell me if X font is already loaded". |
| 22:01 | <gsnedders> | TabAtkins: Speak to the pdf.js guys, they have a fair few requirements. |
| 22:01 | <gsnedders> | And ugly hacks. |
| 22:01 | <TabAtkins> | Heh, yeah. |
| 22:01 | <TabAtkins> | I'm guessing similar reqs to the Google Docs people, as they're doing similar types of things. |
| 22:01 | <TabAtkins> | Though I guess the PDF.js people are loading up fonts from the binaries. |
| 22:01 | <gsnedders> | Yup. |
| 22:03 | <TabAtkins> | that would be interesting. Constructor takes a CSSFontFaceRule, or a string-url, or a TypedArray, maybe. |
| 22:03 | <tantek> | HTML5 slide frameworks could also use the "tell me when X font is loaded" event |
| 22:03 | <TabAtkins> | Well, it'll be a factor function because it needs to be async. |
| 22:04 | <tantek> | since modern HTML5 slide framework templates typically use custom/nicer webfonts that are dynamically loaded these days |
| 22:04 | <TabAtkins> | tantek: Right, that one's easy to sell. (Though it's a promise, not an event.) |
| 22:04 | <gsnedders> | TabAtkins: Currently they go to a data URI. Obv. would be more efficient to from TypedArray. |
| 22:04 | <tantek> | TabAtkins - I submit to the bikeshedding/reframing :) |
| 22:04 | <gsnedders> | s/Currently/Back when I looked at this, when I cared about Presto support in pdf.js,/ |
| 22:05 | <TabAtkins> | Font.create() is the preferred naming for factories, right? |
| 22:05 | <gsnedders> | Why a factory, why not new Font? |
| 22:06 | tantek | goes back to the standards factory. |
| 22:06 | <TabAtkins> | gsnedders: Need async. |
| 22:06 | <TabAtkins> | Though, hm, I'm going to be vending not-yet-completed Font obects anyway. |
| 22:07 | <TabAtkins> | So might as well vend one immediately from a constructor, and rely on the .ready() promise. |
| 23:48 | <gsnedders> | Where do U+D800 and the like become U+FFFD now? |