| 00:00 | <Domenic> | they have document.URL though! not to be confused with window.URL. |
| 00:00 | jgraham | still has to actually do the review to make the plan come together |
| 00:00 | <jgraham> | But I think it's looking good |
| 00:01 | <jsbell> | Domenic: you say "they're all incompatible!". I hear "maybe we can delete it!" |
| 00:01 | <jsbell> | I doubt we're that lucky |
| 00:01 | <Domenic> | jsbell: I wonder if we could. Still for <custom-a> I need access to "the element's base URL" |
| 00:08 | <caitp> | this isn't a custom elements thing because you mentioned IE11, so what is it? o_o |
| 00:20 | <terinjokes> | does IE11 support global.URL? |
| 00:20 | <terinjokes> | s/global/window/ |
| 00:33 | <caitp> | its not URL, really |
| 00:52 | <JonathanNeal> | Let me apologize in advance for not getting it, but is there a browser runnable version of Domenic’s Promise tests? |
| 00:52 | <JonathanNeal> | ^ https://github.com/promises-aplus/promises-tests |
| 02:18 | <JonathanNeal> | Is there a differencee between return undefined; and return; ? |
| 03:19 | <roc> | JonathanNeal: I think we'd like Mozilla Marketplace to work cross-browser/OS. |
| 04:08 | <zcorpan> | JonathanNeal: no |
| 04:09 | <JonathanNeal> | zcorpan: no, there is not a browser runnable version of Domenic’s Promise tests, or no there is no difference between return; and return undefined;, or both? |
| 04:09 | <zcorpan> | the return thing |
| 04:09 | <Domenic> | caitp: it started with some custom elements stuff, then delved into discovering interop hell. |
| 04:10 | <Domenic> | terinjokes: no URL in IE11 |
| 04:10 | <Domenic> | JonathanNeal: the tests are browserifyable |
| 04:10 | <zcorpan> | JonathanNeal: unless you do var undefined or something like that maybe |
| 04:11 | <JonathanNeal> | Domenic: is it obvious or do i have to do something tricky? |
| 04:11 | <Domenic> | JonathanNeal: there is a difference in certain scenarios e.g. `function f(undefined) { return undefined; } f(5);` or `function (undefined){ /* ... 100s of lines later ... */ function g() { return undefined; /* will return 5 */ } /* ... more code here ... */ }(5) |
| 04:11 | <Domenic> | JonathanNeal: if you've used browserify before should be pretty simple |
| 04:49 | <JonathanNeal> | Interesting, IE10’s Object.create method does not assign the correct prototype. |
| 07:25 | <annevk> | Domenic: note that hopefully only document.baseURI is important going forward |
| 07:25 | <annevk> | Domenic: document.body.baseURI will just return that value (if we need to preserve the propery) |
| 07:31 | <annevk> | MikeSmith: I closed it, but I notice now that https://www.w3.org/Bugs/Public/show_bug.cgi?id=27202 is spam |
| 07:31 | <annevk> | MikeSmith: see the URL field |
| 11:46 | <erlehmann> | https://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-November/033914.html does not redirect to http://lists.w3.org/Archives/Public/public-whatwg-archive/2011Nov/0186.html |
| 11:46 | <erlehmann> | and it is not in the wayback machine |
| 11:46 | <erlehmann> | this breaks a link on https://news.ycombinator.com/item?id=3264074 |
| 11:50 | <annevk> | We don't have an easy way of fixing that |
| 11:57 | <erlehmann> | annevk what is it about lists.whatwg.org? |
| 11:57 | <erlehmann> | is it gone? |
| 11:58 | <annevk> | erlehmann: it's not accessible over TLS |
| 12:00 | <erlehmann> | but it was? i see. |
| 12:26 | <Ms2ger> | ... |
| 12:26 | <Ms2ger> | Chrome has a quirk in option.text |
| 13:12 | <Ms2ger> | And IE throws on document.createElement("option").label |
| 13:22 | <annevk> | Film at 11? |
| 13:59 | <Domenic> | annevk: curious in what situations they can be different |
| 13:59 | <Domenic> | annevk: although I guess per spec it seems like "the element's base URL" and "the document's base URL" are the same. |
| 14:00 | <Ms2ger> | xml:base |
| 14:01 | <Domenic> | right -_- |
| 14:12 | <annevk> | Domenic: the plan is to nuke xml:base |
| 14:12 | <annevk> | Domenic: there's various open bugs on browsers |
| 14:12 | <annevk> | Domenic: most browsers don't really respect it... |
| 14:14 | <erlehmann> | what is with normal base element? |
| 14:14 | <annevk> | erlehmann: that sets the document's base URL |
| 14:15 | <annevk> | (and it would be nice if it didn't exist or was not dynamic at least, but...) |
| 14:19 | <annevk> | Domenic: https://www.w3.org/Bugs/Public/show_bug.cgi?id=20976 has my rough plan for replacing the current base URL setup, but I'm waiting for either Chrome or Firefox to remove their remaining xml:base support |
| 14:21 | <Domenic> | annevk: I just wonder if we can fix chrome's baseURI property to return something sensible |
| 14:22 | <Domenic> | wait ... this morning/on my home computer it's being not-retarded... was i hallucinating? will report back when in the office |
| 14:22 | <annevk> | Domenic: make it return the associated document's base URL... |
| 14:23 | <Domenic> | No wait it's retarded in Canary even on this computer... |
| 14:25 | <Domenic> | https://www.dropbox.com/s/hfba5zwbv9ti30y/Screenshot%202014-10-31%2010.25.13.png?dl=0 |
| 14:27 | <annevk> | Domenic: fascinating |
| 14:28 | <annevk> | Domenic: Firefox has about:newtab |
| 14:28 | <Domenic> | Yeah seems Chrome opens up a google.com page that uses AppCache |
| 14:29 | <Domenic> | In yesterday's cursory testing Firefox had sensible baseURI values, Chrome did the weird filename + querystring thing, and IE gave undefined. |
| 14:29 | <Domenic> | This discovery that Chrome 38 gives sensible values too is new |
| 14:30 | <annevk> | Hmm, VMWare is broken in OS X 10.10 |
| 14:30 | <annevk> | If IE returns undefined, can we do away with baseURI? Would be too good to be true I'm sure |
| 14:32 | <Domenic> | OMG |
| 14:32 | <Domenic> | OMG |
| 14:32 | <Domenic> | OMG |
| 14:32 | <Domenic> | Developer tools is fucking with me |
| 14:33 | <Domenic> | document.baseURI |
| 14:33 | <Domenic> | "newtab?espv=2&ie=UTF-8:1" |
| 14:33 | <Domenic> | "hello" + document.baseURI |
| 14:33 | <Domenic> | "hellohttps://www.google.com/_/chrome/newtab?espv=2&ie=UTF-8"; |
| 14:34 | <annevk> | o_O |
| 14:34 | <ondras> | :-) |
| 14:34 | <erlehmann> | lelele |
| 14:35 | <erlehmann> | ondras |
| 14:35 | <erlehmann> | lest we meet again |
| 14:35 | <ondras> | :-) |
| 14:36 | ondras | has been idling on this channel for ages |
| 14:36 | <erlehmann> | never seen you befor |
| 14:36 | <erlehmann> | JonathanNeal how about the table sortable polyfill? |
| 14:37 | <erlehmann> | document.baseURI |
| 14:37 | <erlehmann> | "https://www.google.de/_/chrome/newtab?espv=2&ie=UTF-8"; |
| 14:37 | <erlehmann> | 'hello'+document.baseURI |
| 14:37 | <erlehmann> | "hellohttps://www.google.de/_/chrome/newtab?espv=2&ie=UTF-8"; |
| 14:37 | <JonathanNeal> | erlehmann: what about it? =D |
| 14:37 | <erlehmann> | chromium Version 37.0.2062.120 Debian jessie/sid (281580) |
| 14:37 | <erlehmann> | JonathanNeal how far along is it? |
| 14:38 | <Domenic> | erlehmann: yeah it's a regression from M38 |
| 14:38 | <Domenic> | https://code.google.com/p/chromium/issues/detail?id=429178 |
| 14:38 | <Domenic> | I always wondered if I had a one-character email address (d⊙dm) whether the bug tracker would insert ...s after the character (d...⊙dm). Turns out it does. |
| 14:39 | <JonathanNeal> | Firefox is funny. typeof XMLHttpRequest /* “function" */ XMLHttpRequest instance Function /* true */ XMLHttpRequest.name /* “” */ Function.prototype.toString.call(XMLHttpRequest) /* TypeError */ |
| 14:40 | <JonathanNeal> | erlehmann: paused development a while back, but i’d be happy to work on it today. |
| 14:40 | <Domenic> | JonathanNeal: that *is* weird. |
| 14:41 | <Domenic> | XMLHttpRequest.toString() works |
| 14:41 | <Domenic> | Function.prototype.toString !== XMLHttpRequest.toString apparently |
| 14:42 | <Domenic> | I wonder if that's true for all DOM objects |
| 14:42 | <JonathanNeal> | Maybe. Similarly, Safari considers XMLHttpRequest an object. |
| 14:43 | <annevk> | Domenic: I guess you didn't have any luck with that panel the other day as WebRTC decided to not require TLS |
| 14:43 | <annevk> | Domenic: I'm thinking of putting W3C's infamous Formal Objection to test when the time is there to see if that can be reversed |
| 14:44 | <JonathanNeal> | erlehmann: would you like to help me get back to speed with table sort? |
| 14:44 | <Domenic> | annevk: ummm they were supposed to add a note encouraging browsers to ship with only TLS |
| 14:44 | <annevk> | Domenic: though with the latest STFU to Apple about longdesc I don't have much hope |
| 14:44 | <Domenic> | annevk: I think it's a non-starter to require HTTPS for GUM but not for webkitGUM though |
| 14:44 | <annevk> | Domenic: why can't they make migration happen? |
| 14:44 | <Domenic> | annevk: isn't that up to browsers? |
| 14:45 | <annevk> | Domenic: if we leave security up to browsers, should we leave APIs up to them too? |
| 14:45 | <Domenic> | annevk: IMO they should have a note like https://xhr.spec.whatwg.org/#sync-warning |
| 14:46 | <Domenic> | i will send an email |
| 14:47 | <Domenic> | i looked at too much internet today before breakfast, need to get to the office and eat... |
| 14:52 | <annevk> | Domenic: they can do better than that I think, flag day seems feasible |
| 14:52 | <annevk> | Domenic: but yes, actual deprecation would be good |
| 14:52 | <annevk> | Domenic: the resolution of the bug I tweeted made it sound like they would just add some text about how it's bad but not have an actual plan |
| 14:53 | <JonathanNeal> | When ToLength caps lengths at 2^53-1, why is that? The bit limit in JavaScript? |
| 14:56 | <MikeSmith> | annevk: thanks (about that bug) |
| 14:57 | <JonathanNeal> | ^ http://people.mozilla.org/~jorendorff/es6-draft.html#sec-tolength |
| 14:58 | <boogyman> | JonathanNeal: it would make sense if that were the case |
| 14:59 | <boogyman> | gaps start to appear beyond that definition. |
| 15:03 | <Ms2ger> | JonathanNeal, once you cross that limit, not all integers are representable as IEEE doubles |
| 15:04 | <JonathanNeal> | Thanks. noted. |
| 15:34 | <Domenic> | annevk: "authenticated" is the new "secure" w.r.t. origins? |
| 15:35 | <annevk> | Domenic: it's not stable |
| 15:35 | <annevk> | Domenic: and it's not tied to origins |
| 15:36 | <annevk> | Domenic: https://w3c.github.io/webappsec/specs/mixedcontent/#powerful-features is the current thinking of Mike and I |
| 15:36 | <annevk> | Domenic: mostly Mike |
| 15:43 | <annevk> | https://annevankesteren.nl/2014/10/tls-browser-ui |
| 15:48 | <Domenic> | annevk: enumerateDevices is privacy-sensitive from my understanding, yes? Although just in the fingerprinting sense... |
| 15:49 | <annevk> | Domenic: prolly |
| 15:51 | <Ms2ger> | Hm |
| 15:51 | <Ms2ger> | Is it expected that canvas2d font parsing depends on quirkiness? |
| 16:30 | <Hixie> | erlehmann: lists.whatwg.org was never accessible via TLS, but we have HSTS set up for whatwg.org now so browsers who have visited whatwg.org won't accept to visit lists.whatwg.org except through TLS |
| 16:31 | <Hixie> | erlehmann: if you visit the URL with http:// rather than https:// in a browser that has never seen whatwg.org, you can copy-and-paste the text from that URL into Google and get the w3.org equivalent. |
| 16:31 | <Hixie> | someone should write a cgi script to just do that |
| 16:33 | <erlehmann> | but that means there is a cert for whatwg.org … but not for subdomains? |
| 16:40 | <Hixie> | we have a cert |
| 16:40 | <Hixie> | we just have no way to install it for lists.whatwg.org specifically |
| 16:40 | <Hixie> | because that subdomain is run by dreamhost |
| 16:48 | <Domenic> | Hixie: that clean-browser trick won't work once we make it onto the HSTS preload list |
| 17:14 | <Hixie> | Domenic: yeah, you'll have to use an old browser |
| 17:39 | <annevk> | https://unicorn-wg.github.io/sslv3-diediedie/ heh |
| 18:11 | <annevk> | http://tantek.com/2014/304/b1/url-started-as-udi-conversation-w3c-tpac is relevant to my interests |
| 18:11 | <annevk> | Did not realize URN started at around the same time or was already ongoing |
| 18:26 | <TabAtkins> | JonathanNeal: In particular, between 2^53 and 2^54-1, only even integers can be represented. Between 2^54 and 2^55-1, only integers divisible by 4 can be represented. Etc. |
| 18:29 | <TabAtkins> | JonathanNeal: Re "return undefined;", "undefined" is technically a variable on the global object, and can be overridden. (Unlike true/false/null/etc, which are language keywords and can't be overridden.) Its initial value is the undefined value. |
| 18:31 | <TabAtkins> | So "return;" always returns the undefined value, while "return undefined;" might not. If you ever explicitly want the undefined value, you can use the "void" operator, which is a prefix operator that takes one value, does nothing with it, and returns undefined. |
| 18:31 | <TabAtkins> | Like "void 0" |
| 18:31 | <TabAtkins> | (Which also happens to be shorter to type than "undefined".) |
| 18:33 | <caitp> | to be fair, global.undefined is [[writable]]: false, [[configurable]]: false |
| 18:33 | <caitp> | unless it gets shadowed, which hopefully it doesn't, it doesn't matter |
| 18:35 | <TabAtkins> | Right, but most JS code happens inside of a function, so you shadowing a legit possibility. |
| 18:35 | <caitp> | not if you smack anyone that tries to check in code declaring a variable named `undefined`! |
| 18:42 | <TabAtkins> | Sure, sure. |
| 18:43 | <TabAtkins> | But, for example, jQuery goes out of its way to keep this from being a problem by taking a final argument in its init code that they don't actually pass (just so they get a guaranteed undefined value). |
| 19:05 | <smola> | "file" -> Unfortunate as it is, this is left as an exercise to the reader. When in doubt, return a new globally unique identifier. |
| 19:05 | <smola> | LOLSTANDARD? :p |
| 19:12 | <smola> | annevk: I'm trying to catch up with the URL Standard changes... what's with all the blob stuff? |
| 19:12 | <smola> | having an associated object to a URL seems out of scope |
| 19:14 | <smola> | stuffing File API internals into the URL Standard looks like a call to "include all URL schemes' internals into the spec" |
| 19:14 | <smola> | (e.g. git://, ed2k://, magnet:...) |
| 19:20 | <annevk> | smola: it's something the web platform needs, but you can ignore that |
| 19:20 | <annevk> | smola: i could probably write it down more cleanly |
| 19:22 | <smola> | annevk: it would be nice to have a note, somewhere, noting that "blob" stuff is for people implementing the File API, or for the "web platform", or something like that |
| 19:23 | <smola> | I just want to spare me arguments about why that doesn't need to be implemented in a "non-web context" to comply with the URL parsing standard |
| 19:24 | <MikeSmith> | annevk: sounds like you need to add some text defining different conformance classes |
| 19:24 | <annevk> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=27215 |
| 19:24 | <annevk> | smola: MikeSmith: ^^ |
| 19:24 | <annevk> | patches welcome :p |
| 19:24 | <MikeSmith> | :-) |
| 19:25 | <smola> | annevk: \o/ |
| 20:14 | <Hixie> | if i have an expression with an object literal, as in { foo: 'bar' } |
| 20:14 | <Hixie> | i can add methods to it by making the property value be a function |
| 20:14 | <Hixie> | is there an object literal syntax to add getters and setters? |
| 20:15 | <Domenic> | Wierd, InfoQ basically copied the "model" section of the streams spec + some questions into a news article. http://www.infoq.com/news/2014/10/whatwg-streams |
| 20:15 | <Domenic> | Hixie: yes, `{ get foo() { return 'bar'; } set foo(v) { ... } } |
| 20:17 | <Hixie> | Domenic: does anyone implement that? |
| 20:17 | <Domenic> | Hixie: IE9+ |
| 20:17 | <Domenic> | (and everyone else, obviously) |
| 20:17 | <Hixie> | { set foo(a) { } } in Chrome gets me a syntax error |
| 20:18 | <Hixie> | oh i have to make it an expression |
| 20:18 | <Domenic> | wrap it in parens |
| 20:18 | <Hixie> | i see |
| 20:18 | <Domenic> | yeah |
| 20:18 | <MikeSmith> | InfoQ is hard to take seriously. I wonder who actually reads it, or why |
| 20:19 | <Domenic> | agreed. it seems like a quantity thing? |
| 20:38 | <wanderview> | jsbell: ping |
| 20:46 | <wanderview> | jsbell: wrote a bug instead: https://code.google.com/p/chromium/issues/detail?id=429316 |
| 20:47 | <jsbell> | wanderview: I think the intent is that put() implicitly clones (or eventually tees the stream) |
| 20:48 | <wanderview> | jsbell: I think the decision was not to do implicit clones |
| 20:48 | <wanderview> | let me find the issue |
| 20:50 | <wanderview> | jsbell: it was the result of this (long) issue |
| 20:51 | <wanderview> | seemed to come to the explicit clone conclusion |
| 20:54 | <jsbell> | wanderview: https://github.com/slightlyoff/ServiceWorker/issues/510 |
| 20:55 | <wanderview> | jsbell: that only prevents marking the body used when doing cache.Match(myRequest) |
| 20:55 | <wanderview> | it avoids invoking the Request copy constructor... only create a Request object if a string was passed in |
| 20:56 | <wanderview> | but it doesn't stop the Cache from reading the body in order to store the Request in cache.put() |
| 20:56 | <jsbell> | wanderview: ah, right... |
| 20:56 | <wanderview> | sorry, I have to head out early to take my daughter trick-or-treating |
| 20:57 | <jsbell> | wanderview: kk; I think you should raise a spec issue on this; I don't see anything in the specs that would actually set the used flag, even if conceptually it "should" |
| 20:57 | <wanderview> | k |
| 20:58 | <wanderview> | I just don't see how it could not if we must retrieve the Request in keys()... means the cache had to store the body |
| 21:01 | <wanderview> | jsbell: https://github.com/slightlyoff/ServiceWorker/issues/550 |
| 21:02 | <jsbell> | wanderview: thx; Happy Halloween! |
| 21:02 | <wanderview> | np... Happy Halloween to you too! |
| 22:01 | <annevk> | MikeSmith: InfoQ ends up generating a ton of tweets somehow |
| 22:04 | <MikeSmith> | yeah |
| 22:04 | <MikeSmith> | kind of baffling |
| 22:06 | <Hixie> | anyone know if JS has static properties? |
| 22:06 | <Hixie> | i want my class { } thingy to define a property that appears on the constructor and the instances, readonly |
| 22:10 | <annevk> | Hixie: per IDL constants are non-writable non-configurable properties |
| 22:10 | <Domenic> | Hixie: proper static properties don't exist on the instances, only the class |
| 22:10 | <annevk> | Hixie: not sure if there's sugar though |
| 22:10 | <Domenic> | Hixie: but you can do `class { static foo() { } static get bar() { } } etc |
| 22:12 | <Hixie> | Domenic: k |
| 22:46 | <TabAtkins> | Hixie: Or with today's JS, you just add them as properties on the constructor function, rather than the prototype. |
| 22:46 | <TabAtkins> | Instances wont' get it. |
| 22:47 | <TabAtkins> | You'd have two forms of inheritance (from constructor and from prototype), which we've rejected in general. |
| 22:52 | <Hixie> | yeah |
| 22:53 | <Hixie> | i always get a headache when trying to fiddle with JS prototype stuff |
| 23:02 | <Domenic> | Is https://html.spec.whatwg.org/multipage/rendering.html#phrasing-content-3 missing `:link { cursor: pointer; }` ? |
| 23:04 | <jgraham> | Hixie: With the focus spec, is it possible to have the document node be the focused one? |
| 23:04 | <Hixie> | jgraham: it's possible to have the viewport be focused |
| 23:04 | <Hixie> | jgraham: i don't think the Document node itself can technically be focused itself? but i could be wrong |
| 23:04 | <jgraham> | Hixie: Is that the same? |
| 23:04 | <Hixie> | Domenic: yeah, looks like it |
| 23:05 | <Hixie> | jgraham: not technically |
| 23:05 | <jgraham> | Hixie: So I thought that the focused area was always a node? |
| 23:06 | <Hixie> | jgraham: oh no, definitely not |
| 23:06 | <jgraham> | Or at least always had a DOM anchor |
| 23:06 | <Hixie> | jgraham: there's all kinds of crap that can be focused these days that isn't anode |
| 23:06 | <Hixie> | the viewport uses the document as its dom anchor iirc |
| 23:06 | <jgraham> | OK |
| 23:07 | <Hixie> | jgraham: see https://html.spec.whatwg.org/#focusable-area |
| 23:07 | <Domenic> | Hixie: shall I file a bug or are you fixing already? |
| 23:07 | <Hixie> | Domenic: file a bug |
| 23:07 | <Hixie> | thanks |
| 23:10 | <Domenic> | Wow did not know (remember?) that :link did not match visited links... fun times |
| 23:10 | <Hixie> | :any-link |
| 23:11 | <Domenic> | yepyep. |
| 23:12 | <Domenic> | can't find a straight answer as to whether :any-link is equivalent to `:matches(a[href], area, link)`, hmm |
| 23:13 | <Domenic> | I guess CSS is decoupled from HTML in that respect |
| 23:13 | <Domenic> | CSS just says "represents a link ... for example in [HTML5] ..." |
| 23:14 | <TabAtkins> | Yeah, it's the host language that defines what a link is. |
| 23:16 | <Hixie> | :any-link is equivalent to :matches(:link, :visited) |
| 23:17 | <Domenic> | yeah just curious what :link matches I guess |
| 23:17 | <Domenic> | in terms of elements |
| 23:17 | <Domenic> | gotta trace the references |
| 23:17 | <Hixie> | oh |
| 23:18 | <Hixie> | that's in the spec |
| 23:18 | <Hixie> | html spec |
| 23:18 | <Domenic> | yeah |
| 23:18 | <Hixie> | search for :link |
| 23:18 | <Domenic> | oh it's that easy that sounds nice |
| 23:18 | <Domenic> | hmm no :active in the UA stylesheet either |
| 23:21 | <jgraham> | Hixie: Thanks |
| 23:24 | <Hixie> | jgraham: it's about as far from what browsers implement today as the spec was a few months ago, but the spec has changed drastically in that time, fwiw |
| 23:24 | <Hixie> | jgraham: i think the new spec is way more comprehensive in terms of what browsers actually need to do |
| 23:24 | <Hixie> | jgraham: so if you haven't read that section recently, now's a good time |
| 23:26 | <jgraham> | Hixie: At the moment I'm mainly trying to convince the webdriver people not to define their own concept of focus |
| 23:26 | <jgraham> | This has not been too hard, fortunately |
| 23:26 | <Hixie> | oh man yeah |
| 23:26 | <Hixie> | defining focus is a huge pain |
| 23:27 | <Hixie> | i'm pretty sure what the HTML spec has now is close to complete relative to what the web needs |
| 23:27 | <Hixie> | so they should just use that |
| 23:27 | <wilhelm_> | Other things that should be defined elsewhere and referenced from our spec: visibility, hit testing. |
| 23:29 | Hixie | pretends to have gotten an important phone call and runs out of the room |
| 23:31 | <Domenic> | Is there any way for me to be the first subscriber to an event, but only do my thing if no other subscribers call `e.preventDefault()`? Maybe delaying a turn then checking e.defaultPrevented? |
| 23:32 | <Hixie> | register yourself on the parent element? :-) |
| 23:32 | <Hixie> | what happens if you register yourself during dispatch |
| 23:33 | <Domenic> | oooh interesting |
| 23:33 | <Domenic> | seems like it would work, although maybe browsers do weird things |
| 23:44 | <Domenic> | aww doesn't work |
| 23:44 | <Domenic> | setTimeout(, 0) seems like my best bet |
| 23:46 | <jamesr_> | Hixie: registering yourself during dispatch should do nothing for that event, the propagation chain is computed before firing any handlers |
| 23:46 | <Hixie> | yeah, that seems reasonable |
| 23:46 | <Hixie> | pity |
| 23:49 | <jamesr_> | otherwise it's insanity |
| 23:50 | <Hixie> | totally |
| 23:51 | <Hixie> | would have solved this case though ;-) |
