| 01:03 | <heycam> | annevk, there's a difference between including a final "z" and not. if you don't have it, it's an open path, and you won't get a stroke on that final edge |
| 01:03 | <heycam> | Hixie, mm I think you may have asked for something like this before. or it was "do nothing" in the method. |
| 01:05 | <heycam> | Hixie, it feels like it's just on the other sound of my "too hacky" boundary. :) what mehod do you want it for? |
| 01:05 | <heycam> | annevk, (oh I see rcabanier already answered you) |
| 02:22 | <annevk> | heycam|away: yeah, but I'm only filling so it appears to be fine |
| 02:22 | <annevk> | heycam|away: |
| 02:22 | <annevk> | heycam|away: there's a few bug reports, but I guess you'll get to them at some point |
| 02:27 | <annevk> | matjas: let me know when you create whatwg/javascript, I'll set up the rest then |
| 03:34 | <Hixie> | heycam|away: getContext()'s first argument, which is either '2d' or 'webgl' |
| 03:34 | <Hixie> | heycam|away: what i really want isn't so much that it return null, but that the enum get coerced to null or something i can detect |
| 03:35 | <Hixie> | heycam|away: but there's not really a good reason to do it that way rather than prose other than being clear in the IDL about the intentions |
| 03:52 | <Hixie> | how come, in Gecko, does a.document instanceof Document return true, where a is some other Window? |
| 03:53 | <Hixie> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1905 |
| 03:53 | <Hixie> | (also, anyone with IE around? what does IE put in the log for that test?) |
| 04:51 | <Hixie> | annevk: what does "standards must define scheme data either literally as or as a subset of one or more URL units that do not start with "?"." mean? |
| 07:45 | <annevk> | Hixie: trying to say that if you define e.g. the javascript URL scheme you need to limit it to the allowed syntax of scheme data |
| 07:48 | <annevk> | Hixie: so either you define it the same way (literally), as one or more URL units that does not start with "?", or you define it otherwise, e.g. as a sequence of ASCII digits |
| 07:50 | <annevk> | Hixie: my plan is to give some examples by defining the javascript and about URL schemes in the specification |
| 08:07 | <annevk> | "X-XSS-Protection" |
| 08:07 | <annevk> | when will the X- stop? |
| 08:07 | <annevk> | apparently http://tools.ietf.org/html/rfc6648 was not read |
| 08:56 | <hsivonen> | too bad mark pilgrim doesn't blog anymore. I'like to see his take on polyglot |
| 08:58 | <annevk> | GPHemsley: created http://resources.whatwg.org/logo-mime.svg for you |
| 08:58 | <annevk> | hsivonen: I would love him to blog about a lot of things |
| 08:58 | <annevk> | anything really |
| 08:59 | <annevk> | created http://resources.whatwg.org/logo-dom.svg for the DOM, okay Ms2ger? |
| 08:59 | annevk | is not entirely sure about that one |
| 08:59 | <Ms2ger> | annevk, what's it supposed to be? :) |
| 09:00 | <annevk> | Ms2ger: a tree |
| 09:01 | <Ms2ger> | Ah |
| 09:01 | <Ms2ger> | wfm |
| 09:01 | <annevk> | cool, I hope I don't have to explain that all the time, but I guess it's all open source so we can just ask for patches :) |
| 09:06 | <SimonSapin> | annevk: I can see the binary tree once you tell me it’s there … but not really obvious otherwise |
| 09:07 | <annevk> | DOM API is non-intuitive too :p |
| 09:08 | <SimonSapin> | is it in r.w.o? |
| 09:09 | <annevk> | SimonSapin: what do you mean? |
| 09:09 | <MikeSmith> | annevk: suggest U+1F4A9 for the DOM logo |
| 09:09 | <Ms2ger> | MikeSmith, I *knew* it! |
| 09:09 | <SimonSapin> | annevk: I understood you already had a logo for DOM API |
| 09:10 | <MikeSmith> | Ms2ger: quit reading my mind. It makes it hard for me to concentrate |
| 09:10 | <annevk> | SimonSapin: not before the one I just created, but that one is hosted on resources.whatwg.org now |
| 09:10 | <annevk> | MikeSmith: I like how I know what that is by the number |
| 09:10 | Ms2ger | wonders about that girl MikeSmith met last week |
| 09:11 | <annevk> | MikeSmith: but my SVG path skills are not good enough for that |
| 09:11 | <SimonSapin> | oh, DOM API is the same as DOM |
| 09:12 | <MikeSmith> | Ms2ger: :-) |
| 09:13 | <annevk> | SimonSapin: ah, I see what tripped you now :) |
| 09:15 | <SimonSapin> | so "DOM API" is the long form to disambiguate from "DOM Parsing", just like "strictement supérieur à" :p |
| 09:22 | <annevk> | well with DOM API I meant to refer to the API part of the DOM, which is generally considered non-intuitive and crappy |
| 10:05 | <annevk> | I wonder at what point Apache will stop emitting stuff like <!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN"> |
| 10:05 | <annevk> | once Roy gets veto power over HTML? |
| 10:06 | <Stevef> | othermaceij: am I right in thinking that a patch to add <main> to webkit would be pretty much the same as the patch for the <nav> element? https://bugs.webkit.org/attachment.cgi?id=35098&action=prettypatch |
| 10:08 | <Stevef> | othermaciej: (sorry mispelt nick first time around) am I right in thinking that a patch to add <main> to webkit would be pretty much the same as the patch for the <nav> element? https://bugs.webkit.org/attachment.cgi?id=35098&action=prettypatch |
| 10:15 | <annevk> | yeah |
| 10:16 | <SimonSapin> | so Google never dropped h264, did they? |
| 10:23 | <Stevef> | annevk: whas that a yeah to my question or some other? |
| 10:32 | <annevk> | yeah |
| 10:32 | <annevk> | SimonSapin: nope |
| 10:32 | <annevk> | hehe http://t.co/qApp3RWc |
| 10:34 | <Stevef> | thanks |
| 10:36 | <annevk> | Stevef: ;) |
| 10:49 | <Philip`> | annevk: The logo looks to me like a Darwinian who has suffered a terrible accident |
| 10:49 | <Philip`> | (http://i.neoseeker.com/n/8/multiwinia2_thumb.jpg) |
| 10:50 | <annevk> | yay, Philip` is back |
| 11:31 | <asmodai> | annevk: aside from the spec, any other SVG documents that are a must-read? |
| 11:55 | <annevk42> | http://code.google.com/p/html5lib/issues/detail?id=216 seems like Google is porting html5lib to Dart |
| 12:01 | <smaug____> | crazy stuff |
| 12:02 | <smaug____> | I already imagined Dart kind of died |
| 12:53 | <annevk42> | http://homepages.cwi.nl/~steven/vandf/2012-11-teneuro.html |
| 13:20 | <gsnedders> | annevk42: https://github.com/dart-lang/html5lib |
| 13:52 | <gsnedders> | Anyone happen to know off-hand what Philip` used to generate all those tests for the tokenizer (from OCaml?)? |
| 13:54 | <Philip`> | I used some OCaml code that tried every possible interesting input to the tokeniser (and then pruned inputs that took the tokeniser into states it had already been in) |
| 13:54 | <gsnedders> | So no library to generate it from the model? |
| 13:54 | <gsnedders> | Okay. |
| 13:55 | <Philip`> | http://canvex.lazyilluminati.com/svn/tokeniser/test_gen.ml |
| 13:56 | <Philip`> | No libraries |
| 13:57 | gsnedders | was planning on prescan tests this evening |
| 14:25 | <annevk42> | didn't know http://www.w3.org/TR/media-frags/ was a rec |
| 16:05 | <GPHemsley> | DreamHost suspended my account for lack of payment without even notifying me. |
| 16:06 | <GPHemsley> | And they have this message on the web panel: "get paid up to prevent permanent data loss (we only have your data for as long as we have backups of it, usually a week or two from day of suspension)" |
| 16:06 | <GPHemsley> | So it's not even clear that my website will be as I left it when payment is made. |
| 16:06 | <GPHemsley> | This is supremely frustrating. |
| 18:10 | <GPHemsley> | ah, well, at least they were quick to restore |
| 18:19 | <heycam> | annevk, I like the logos. :) SVG spec bug reports very welcome too. |
| 18:23 | <heycam> | Hixie, filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=19936 for the enum thing |
| 18:28 | <Hixie> | GPHemsley: what has the "has content type" flag? is it a system-global flag? |
| 18:29 | <Hixie> | annevk42: i don't think "literally" means what you think it means, then |
| 18:29 | <GPHemsley> | annevk: Here's a case for you: a file:/// page with a // href - the domain disappears |
| 18:29 | <GPHemsley> | Hixie: If I understand what you're asking, yes. |
| 18:30 | <Hixie> | heycam: thanks |
| 18:30 | <GPHemsley> | Hixie: It is set in the "supplied media type detection algorithm" and checked again in the "media type sniffing algorithm" |
| 18:30 | <Hixie> | GPHemsley: so if two pages are both going to read a resource's mime type at the same time, they interfere? |
| 18:31 | <GPHemsley> | Hixie: Oh, it's not that global. This entire spec assumes each resource analysis is self-contained. |
| 18:31 | <Hixie> | heycam: yeah, that'd be neat. Purely editorial, in effect, but would be nice to make the IDL clearer that the argument expects a particular set of values, not just arbitrary strings. |
| 18:32 | <Hixie> | GPHemsley: "assumes" gets us into a lot of trouble :-P |
| 18:32 | <GPHemsley> | Hixie: Yeah... hadn't considered the possibility |
| 18:32 | <Hixie> | i recommend just making it a return value |
| 18:32 | <Hixie> | return a pair (type, had-content-type) |
| 18:32 | <heycam> | Hixie, I think not using null would make me feel better about it, because that then means you could only use the enum if it's nullable. |
| 18:32 | <Hixie> | or some such |
| 18:32 | <heycam> | (like how I used "") |
| 18:33 | <Hixie> | heycam: yeah "" would be fine. It only needs to be distinguishable from valid values. |
| 18:33 | <heycam> | ok |
| 18:33 | <GPHemsley> | Hixie: Would you mind filing a bug for the issue? |
| 18:33 | <Hixie> | heycam: i like the explicit "this is the default" syntax, fwiw |
| 18:33 | <Hixie> | GPHemsley: sure |
| 18:33 | <heycam> | cool |
| 18:34 | <Hixie> | man every time i file a bug the whatwg product has more components |
| 18:34 | <GPHemsley> | annevk: Oh, that's also the reason the dfn script doesn't work locally. Interesting. |
| 18:35 | <GPHemsley> | Hixie: That's good, isn't it? |
| 18:37 | <GPHemsley> | annevk: BTW, the WHATWG logo has an opaque white center; the mimesniff logo is transparent. |
| 18:39 | <Hixie> | GPHemsley: neither good or bad, just amusing |
| 18:39 | <GPHemsley> | annevk: Oh, also, thanks. :) |
| 18:41 | <Hixie> | GPHemsley: so i'm confused by the "Configurable flag" thing |
| 18:42 | <Hixie> | GPHemsley: are you saying that browsers can have any of these arbitrarily enabled or disabled? |
| 18:42 | <Hixie> | GPHemsley: that seems like it'd mean that almost any behaviour is conforming... |
| 18:42 | <GPHemsley> | Hixie: Yes, either based on their own decisions, or by leaving the decision in the hands of the user. |
| 18:42 | <Hixie> | why? |
| 18:42 | <GPHemsley> | Hixie: Well, one of the overarching themes of this spec was that UAs SHOULD NOT sniff. |
| 18:43 | <GPHemsley> | Hixie: That idea was pre-existing when I came on the scene. |
| 18:43 | <GPHemsley> | The only behavior that is required is for unknown types |
| 18:43 | <Hixie> | i'm fine with a browser never doing sniffing, but if it's going to sniff, it should do it consistently for everything |
| 18:43 | <Hixie> | having many different conforming configurations defeats the point of having a spec |
| 18:44 | <GPHemsley> | Hixie: Well, without it, I think there will be a lot of pushback in certain situations. |
| 18:44 | <Hixie> | from whom? implementors? |
| 18:44 | <GPHemsley> | Hixie: For example, I don't think Mozilla wants to sniff content tagged as "text/plain". |
| 18:44 | <GPHemsley> | (They don't do it now.) |
| 18:44 | <Hixie> | in what situation? |
| 18:44 | <Hixie> | they certainly do for <img> |
| 18:45 | <GPHemsley> | Step 4 of "media type sniffing algorithm" |
| 18:45 | <GPHemsley> | Doing it for <img> is a different case. |
| 18:45 | <Hixie> | they definitely do for "text/plain; charset=ISO-8859-1" |
| 18:45 | <GPHemsley> | Sniffing for <img> is defined in the "rules for sniffing images specifically", which are invoked by HTML directly. |
| 18:46 | <GPHemsley> | Hixie: Do they? Well, then I suppose I'll have to follow up on that. |
| 18:46 | <GPHemsley> | Hixie: But they don't do it for text/plain—which I discovered by accident. |
| 18:46 | <Hixie> | the goal here should be to define the exact rules required to be compatible with the web |
| 18:46 | <Hixie> | if we don't have to do it for text/plain, then nobody does |
| 18:46 | <Hixie> | if we do, then everybody does |
| 18:46 | <Hixie> | there might be some cases where people disagree about how important it is to do it, but then it's your job to decide who's right :-) |
| 18:47 | <GPHemsley> | Very well. |
| 18:47 | <gsnedders> | And to convince everyone else you are. :) |
| 18:47 | <Hixie> | imho you just want the algorithms to have a single bail-out point at the top, along the lines of "if you want to not be interoperable with anyone but want to be strictly correct per the IETF, then use the specified type, otherwise, follow the rest of this algorithm to the letter" |
| 18:47 | <GPHemsley> | Like I said, most of these concepts were written by abarth. I don't necessarily fully understand the motivation behind them (yet). |
| 18:48 | <GPHemsley> | But I see your point. |
| 18:48 | <GPHemsley> | I suppose he did have something to that effect in the old draft, so maybe this was my addition. |
| 18:48 | <GPHemsley> | I'll have to go back and check. |
| 18:49 | <GPHemsley> | Hixie: But there is the situation where the user might want to override the UA. |
| 18:49 | <GPHemsley> | One way or the other. |
| 18:49 | <Hixie> | there is? |
| 18:49 | Hixie | is not aware of any UA that supports that |
| 18:52 | <GPHemsley> | Hixie: Well, they may not currently support it... but it's a relatively common requested feature, AFAICT. |
| 18:53 | <GPHemsley> | (though perhaps not in all cases) |
| 18:54 | <gsnedders> | Has anyone got any intention of supporting it? |
| 18:54 | <GPHemsley> | No idea |
| 18:54 | <Hixie> | GPHemsley: that's actually setting an override type, though, right? not just disabling specific sniffers. |
| 18:54 | <GPHemsley> | hmm... perhaps |
| 18:55 | <Hixie> | GPHemsley: i would put that closer to where the type gets processed, not in the sniffing spec. e.g. in the navigation algorithm, or something. |
| 18:55 | <GPHemsley> | Hixie: Alright, if you wanna take it on |
| 18:55 | <GPHemsley> | It would simplify things, I think. |
| 18:59 | <annevk42> | Hixie: maybe not :) |
| 19:01 | <annevk42> | GPHemsley: the file: thing you encouter is a Gecko-ism |
| 19:02 | <annevk42> | GPHemsley: I made the WHATWG logo no longer white I think |
| 19:02 | <GPHemsley> | annevk42: What about the icon? Does that maybe not use SVG? |
| 19:03 | <GPHemsley> | (or it could just be a cache thing) |
| 19:03 | <GPHemsley> | annevk42: Do you suppose the URL spec will encourage them to fix that Gecko-ism? |
| 19:04 | <annevk42> | GPHemsley: well yes |
| 19:04 | <annevk42> | GPHemsley: btw, I wrote it as utf-8 because that's the canonical case these days |
| 19:04 | <annevk42> | GPHemsley: and you want file-bug.js from resources.whatwg.org, not from dvcs |
| 20:31 | gsnedders | wonders if there's any bug report on non-IE browsers on removing __lookupGetter__, etc. |
| 22:32 | <GPHemsley> | annevk: What are the PNGs for? |
| 22:32 | <annevk> | |
| 22:33 | <annevk> | and whatever other crappy piece of software does not accept SVG |
| 22:34 | <GPHemsley> | ah |
| 22:37 | <annevk> | didn't know people viewed specs locally, then of course scheme-relative URLs are going to fail |
| 22:52 | <annevk> | maybe we should make xn--zn7c.spec.whatwg.org redirect to encoding.spec.whatwg.org for fun |
| 22:55 | <SimonSapin> | annevk: is that the replacement character? |
| 22:55 | <annevk> | yes |
| 22:56 | <SimonSapin> | do browsers support it? (depending on the IDNA version…) |
| 22:56 | <annevk> | (which is not valid in an IRI, and probably not a valid IDN code point either, but fuck the police) |
| 22:56 | <annevk> | SimonSapin: it would highly surprise me if that lookup failed |
| 22:57 | <annevk> | SimonSapin: but I need to do more testing still |
| 23:00 | <annevk> | SimonSapin: Safari and Opera both do a lookup for that if you type in http://�.spec.whatwg.org even |
| 23:03 | <SimonSapin> | annevk: firefox gives an error page, chromium does a search of http:// .spec.whatwg.org/ (with a space) |
| 23:04 | <annevk> | SimonSapin: the error page Firefox gives suggests a DNS lookup took place |
| 23:04 | <SimonSapin> | Epiphany (WebKit-GTK) gives an error saying http://%20.spec.whatwg.org/ |
| 23:05 | <annevk> | oh sweet, works in Safari's address bar, but not through <a> |
| 23:06 | <annevk> | browsers are silly |
| 23:06 | <annevk> | Film at 11; bedtime here |
| 23:06 | <SimonSapin> | wireshark shows no sign of a DNS request from firefox |
| 23:06 | <annevk> | SimonSapin: hmm, "Firefox can't find the server at �.spec.whatwg.org." is kinda misleading then |
| 23:07 | <SimonSapin> | it is |
| 23:08 | <annevk> | SimonSapin: Firefox prolly uses system DNS so if it was already cached there you might not see another DNS request, fwiw |
| 23:10 | <SimonSapin> | yeah, I tried ��.spec.whatwg.org and a�.spec.whatwg.org |
| 23:11 | <annevk> | interesting |
| 23:15 | <annevk> | nn |
