| 04:36 | <hemanth> | meow |
| 04:37 | hemanth | is watching Domenic's https://github.com/domenic/Array.prototype.contains awesome progress! [ Take a bow ] |
| 04:52 | <caitp> | meow |
| 04:54 | <caitp> | it would be a nice feature to have, just like all of the es5 features nobody ever uses because they need to support ie4 |
| 04:55 | <hemanth> | caitp, hope there will be one day, where all the browser are in sync... |
| 04:56 | <caitp> | such a day will never come, because browser A will support remote controlling your toaster while browser B will only support watering your lawn and making coffee |
| 04:56 | <hemanth> | :D |
| 05:02 | <Domenic> | Thanks hemanth :) |
| 05:02 | <hemanth> | :) |
| 05:05 | <roc> | scheib: you know about the FirefoxOS Bluetooth work I assume? |
| 05:05 | <roc> | and are in touch with the people doing it? |
| 06:48 | <hemanth> | BuzzFeedJS is �ber funny :D Criticisms at it's best. |
| 08:05 | <foolip> | annevk: I'll take a look at your spec change today. are you still stuck without something like a storage mutex? |
| 08:14 | <annevk> | foolip: it doesn't do full protection against multiple tabs trying to go fullscreen |
| 08:14 | <annevk> | foolip: the "can activate a popup" helps with that of course, but it's not a perfect system |
| 08:19 | <foolip> | you mean because there's no way to actually click in the tabs fast enough? |
| 08:19 | <foolip> | are there any races involving exitFullscreen() |
| 08:20 | <foolip> | s/tabs/frames/ |
| 08:20 | <foolip> | if it's only requestFullscreen(), then some vague nonsense about waiting until other requests in the document tree have either failed or completed could be added I guess |
| 08:33 | <annevk> | foolip: I think exitFullscreen() is fine |
| 08:33 | <annevk> | foolip: I moved most of the logic for both as part of the animation frame task, with the resizing happening before for request and after for exit, asynchronously |
| 08:34 | <annevk> | foolip: at least it no longer tries to access document variables from a potentially parallel thread |
| 08:51 | <foolip> | annevk: I'll have a look at file a ton of bugs :) |
| 08:52 | <hemanth> | Broken link? https://github.com/whatwg/streams/issues/163 |
| 09:17 | <annevk> | foolip: that's why we love you |
| 09:25 | <Ms2ger> | Hey foolip, you like reviewing tests, right? ;) |
| 09:25 | <Ms2ger> | Oh look: http://httpwg.github.io/specs/rfc7230.html |
| 09:46 | <annevk> | position:fixed and fragment identifiers don't go well together |
| 10:10 | <foolip> | Ms2ger: well, is it media-related? |
| 10:10 | <Ms2ger> | No |
| 10:10 | <foolip> | what is it then? |
| 10:10 | <foolip> | "plz review everything in w-p-t"? |
| 10:11 | <Ms2ger> | Nah |
| 10:11 | <Ms2ger> | Please review all *my* PRs ;) |
| 10:11 | <Ms2ger> | Or any of them |
| 10:11 | <Ms2ger> | https://github.com/w3c/web-platform-tests/pulls/Ms2ger |
| 10:12 | <foolip> | Ms2ger: you don't like semicolons? |
| 10:13 | <foolip> | https://critic.hoppipolla.co.uk/39bb75ba?review=2192 |
| 10:13 | <Ms2ger> | I followed file style |
| 10:13 | <Ms2ger> | I think annevk's fault :) |
| 10:13 | <foolip> | of course, I assumed the test was originally yours as well |
| 10:13 | <annevk> | "fault" I say |
| 10:14 | <annevk> | "Just" learn ASI |
| 10:14 | <Ms2ger> | commit ddfe31fcc58bf52fab979d0df20fda121a6f8bda |
| 10:14 | <Ms2ger> | Author: Anne van Kesteren <annevk⊙oc> |
| 10:14 | <Ms2ger> | Date: Tue Sep 7 09:45:24 2010 +0200 |
| 10:14 | <Ms2ger> | adoptNode/importNode should not throw for invalid names as per zcorpan's original draft and implementations |
| 10:14 | <foolip> | I shouldn't complain, I like to use == instead of === |
| 10:15 | Ms2ger | gasps |
| 10:15 | foolip | isn't sure what the cool stance is |
| 10:15 | <annevk> | SameValueZero of course |
| 10:15 | <Ms2ger> | assert_equals :) |
| 10:15 | <foolip> | well yeah, but when not writing tests I mean :) |
| 10:16 | <Ms2ger> | I guess === is most correct for the effort required |
| 10:16 | <annevk> | Of course, SameValueZero only exists for Map/Set, there's no primitive |
| 10:16 | <foolip> | what's SameValueZero? google denies knowledge |
| 10:16 | <annevk> | foolip: http://people.mozilla.org/~jorendorff/es6-draft.html#sec-samevaluezero |
| 10:17 | <annevk> | foolip: it's like Object.is(), except that -0 and +0 are also equal |
| 10:18 | <foolip> | but... 1/Infinity === -1/Infinity |
| 10:18 | <foolip> | but apprently !Object.is(1/Infinity, -1/Infinity) |
| 10:18 | <foolip> | cute |
| 10:19 | <annevk> | foolip: https://github.com/domenic/Array.prototype.contains#why-samevaluezero |
| 10:20 | <foolip> | Object.is(NaN, NaN) is true, that's refreshing |
| 10:20 | <annevk> | Yeah, that's the difference with === |
| 10:21 | <foolip> | well, I'm glad there's no complexity on the Web! |
| 10:21 | <annevk> | They should just add Object.isz() and be done with it |
| 10:21 | <annevk> | Or ==== |
| 10:21 | <annevk> | foolip: at least there's no racing :p |
| 10:21 | <foolip> | what's wrong with !Object.is(1/Infinity, -1/Infinity) |
| 10:21 | <foolip> | ? |
| 10:22 | <foolip> | they are different, after all |
| 10:22 | <annevk> | I'm not sure, but reportedly you want -0 and +0 to be treated identically commonly |
| 10:26 | <foolip> | Ms2ger: https://critic.hoppipolla.co.uk/showcomment?chain=6942 |
| 10:50 | <Ms2ger> | Landed as "Sumbit progress and range tests from IE10 development" |
| 11:19 | <foolip> | Ms2ger: https://critic.hoppipolla.co.uk/r/33 looks like a bucketload of fun, is that completely unreviewed or just in need of rubber stamping? |
| 11:24 | <sangwhan> | Does anyone know why this test passes in particular on Gecko? http://tests.sangwhan.com/tests/generated_keyevent.html |
| 11:24 | <sangwhan> | seems to fail on everything else i tried (which excludes IE, which I don't have access to) |
| 12:23 | <Ms2ger> | "W3C Invites Implementations of #HTML5" |
| 12:23 | <Ms2ger> | Why thank you, I'll pass |
| 12:33 | <jgraham> | Ms2ger: Do you have time to look at the ServiceWorker PR for testharness.js |
| 12:33 | <jgraham> | I have looked through most of it, but didn't mark it reviewed so that you could |
| 12:33 | <jgraham> | [Service]Worker |
| 12:34 | <Ms2ger> | I guess I'll try to do that today |
| 12:34 | jgraham | wonders if #HTML5 is for people who are high as a kite |
| 12:35 | <jgraham> | (that works better if you prnounce # as hash rather than in one of the 26 other ways…) |
| 12:40 | <MikeSmith> | sangwhan: suggest asking bz on #developers on irc.mozilla.org |
| 12:44 | <MikeSmith> | Ms2ger: your super-important hspace/vspace test is still awaiting responses to my extensive review comments |
| 12:46 | <Ms2ger> | I'm more into review-without-comment ;) |
| 12:46 | <Ms2ger> | MikeSmith, will do |
| 12:50 | <Ms2ger> | MikeSmith, r? https://critic.hoppipolla.co.uk/r/2016 |
| 12:52 | <MikeSmith> | Ms2ger: buttons pushed |
| 12:53 | Ms2ger | hurries up and close it before anyone has a change of heart |
| 12:53 | <Ms2ger> | *es |
| 12:54 | <MikeSmith> | heh |
| 12:59 | <jorendorff> | foolip: It would be weird for map.get(Math.round(-0.1)) not to find an entry with key 0 |
| 13:00 | <sangwhan> | MikeSmith: i'll try that, lingering in irc.mozilla might raise a eyebrow or two but no harm done i guess |
| 13:03 | <MikeSmith> | I linger there, so the atmosphere's already tainted |
| 13:05 | <foolip> | annevk: a colleague of mine just found https://www.w3.org/Bugs/Public/show_bug.cgi?id=26480 if you want something easy to fix today :) |
| 13:09 | <annevk> | Hah, going to defer to roc |
| 13:10 | <annevk> | I already fixed an easy bug today by removing a feature from Fetch |
| 13:10 | <foolip> | can't overrun the quota I guess |
| 13:15 | <annevk> | JakeA: thanks for going through the matrix |
| 13:15 | <annevk> | JakeA: I guess I need to put some time aside to see it makes sense |
| 13:16 | <JakeA> | annevk: no worries, will get up to speed on that headers stuff to |
| 13:16 | <JakeA> | too* |
| 13:21 | <annevk> | JakeA: so the idea is that Host, cache headers, etc. are all determined post SW |
| 13:21 | <annevk> | JakeA: because that's how we keep talking about the layering and is in fact how the layering works |
| 13:21 | <JakeA> | Fine with those being post, especially cache |
| 13:21 | <JakeA> | cache needs to be as it could be a security issue |
| 13:21 | <annevk> | JakeA: headers visible to SW would be Accept / Accept-Language, and headers set by developers |
| 13:22 | <annevk> | JakeA: referrer/origin would be exposed as property on Request in due course |
| 13:22 | <annevk> | JakeA: Host etc. is exposed already through Request's url |
| 13:22 | <JakeA> | annevk: yeah, that's great |
| 13:23 | <JakeA> | annevk: "Accept" & "Accept-Language" were the ones I was concerned about |
| 13:23 | <annevk> | The downside with this approach is that each spec that does a fetch needs to discuss those headers |
| 13:23 | <annevk> | But that might not be a bad idea anyway |
| 13:24 | <JakeA> | Well, it's kinda magic at the moment that <img> sends different Accept headers |
| 13:24 | <annevk> | Yeah, we never really defined those things in detail |
| 13:44 | <annevk> | JakeA: http://fetch.spec.whatwg.org/#http-header-layer-division |
| 13:48 | <JakeA> | annevk: that works for me |
| 15:10 | <jgraham> | Anyone who has a better feel for promises than me want to comment on promise_test in https://critic.hoppipolla.co.uk/showcommit?review=2005&filter=files&file=165 (line 465 on the right)? |
| 15:26 | <caitp> | how do you even look at a diff in critic anyways |
| 15:27 | <jgraham> | Click on the name of the file you are interested in |
| 15:39 | <sangwhan> | or press 'e' to see everything |
| 15:50 | <caitp> | anyways, the promise_test seems a bit weird to me, since it's pretty much a synchronous operation happening there |
| 15:50 | <caitp> | unless you returrn a promise from the test function I guess |
| 15:51 | <caitp> | which fair enough, but then it's not really abstracting the operation :< |
| 15:57 | <jgraham> | caitp: Please comment on the patch |
| 15:57 | <jgraham> | I don't want to end up with a crappy API here |
| 15:57 | <foolip> | annevk: I like the new "fully exit fullscreen"! |
| 15:58 | <caitp> | I don't have anything really substantive to add jgraham, maybe if I saw a test case which is using that api |
| 15:59 | <caitp> | if the expectation is that the function passed should return a promise then it's all good I guess |
| 15:59 | <jgraham> | caitp: The apisampleX.htm files in the review should be a guide |
| 15:59 | <jgraham> | See https://critic.hoppipolla.co.uk/showcommit?review=2005&filter=files&file=55848,55849,55189,55850,55851,55852,55853,55854,165 |
| 16:00 | <caitp> | right, so each of those test cases is returning a promise, so I guess that's the expected behaviour |
| 16:28 | <caitp> | annevk, if you overwrite the timeout attribute of an XMLHttpRequest after sending, with the value 0, would that mean "cancel the timeout", or would that mean "timeout is going to happen right now" |
| 16:29 | <caitp> | that's not totally clear from the spec to me |
| 16:31 | caitp | looks at gecko's implementation |
| 16:37 | <caitp> | it looks like in gecko, if you set the timeout attribute to 0, the timeout timer is cancelled? |
| 16:41 | smaug____ | is pretty sure gecko does what the spec at least used to say |
| 16:43 | <smaug____> | but indeed, interpreting the current spec is hard |
| 16:44 | <smaug____> | ah, the note |
| 16:44 | <smaug____> | caitp: "When set to a non-zero value will cause fetching to terminate after the given time has passed" |
| 16:46 | <caitp> | to me, that makes sense when we're talking about before the request has been sent, but following the note, setting timeout to 0 and updating the timeout timer relative to the start of the request means, 0 ms since the start of the request |
| 16:46 | <caitp> | oh it would be so nice to clarify this |
| 16:46 | <caitp> | but I guess I'll match gecko's behaviour for now |
| 16:59 | <SamB> | so what does -1 do? |
| 17:00 | <caitp> | assuming the timer is clever, times out immediately I guess |
| 18:58 | <annevk> | caitp: there's an open bug |
| 18:58 | <annevk> | caitp: we should actually evaluate timeout each time something is pushed from the network I guess |
| 19:01 | <caitp> | is that the one I reported? if not I'd like to CC myself to it |
| 19:06 | <annevk> | caitp: list of open bugs is available per link at top of the spec |
| 19:10 | <Hixie> | roc: do you have any advice regarding the second paragraph of https://www.w3.org/Bugs/Public/show_bug.cgi?id=23515#c9 ? |
| 19:11 | <caitp> | huh, I reported one of those as well, but i guess someone closed it or something |
| 19:11 | <Hixie> | foolip: ping https://github.com/w3c/web-platform-tests/pull/970 |
| 19:13 | <annevk> | caitp: you sure we didn't just talk about it or it was elsewhere? |
| 19:14 | <annevk> | caitp: I remember you filed a bug on Fetch with regards to Content-Length |
| 19:14 | <caitp> | i've filed lots of bugs, but for w3 and whatwg bugs I rarely manage to keep track of them :> |
| 19:17 | <Hixie> | do you file them from your own account? |
| 19:18 | <annevk> | caitp: WHATWG and W3C share a Bugzilla instance, although both also use GitHub and mailing lists, and W3C also has a variety of other tools |
| 19:18 | <annevk> | Hixie: you're ok with defining the HTML vocabulary in several drafts now? (re nonce) |
| 19:18 | <Hixie> | zcorpan: so how does <picture> get around the problem where the preloader can't evaluate media queries? did broser vendors relent on this restriction, or what? |
| 19:18 | <Hixie> | annevk: ideally i'd have every web spec in one single document. |
| 19:19 | <caitp> | yeah, the issue is that if I don't receive mail on a bug it usually vanishes, since I don't frequently use the w3's bugzilla |
| 19:19 | <annevk> | Hixie: all in en-GB? |
| 19:19 | <caitp> | that's the way the cookie crumbles |
| 19:19 | <Hixie> | annevk: i don't really care what language :-) |
| 19:19 | <annevk> | Hixie: you broke a bunch of "permanent" links with that change |
| 19:19 | <Hixie> | which change? |
| 19:20 | <annevk> | en-US -> en-GB |
| 19:20 | <Hixie> | nobody seems to have noticed, if so. that happened months ago. |
| 19:20 | <annevk> | fragment-links.js change also caused xref to burn |
| 19:20 | <Hixie> | o_O |
| 19:20 | <Hixie> | how so |
| 19:20 | <annevk> | Hixie: maybe it happened a month ago or so? |
| 19:21 | <annevk> | Hixie: it tries to parse the first line as JSON (after eliminating the first seventeen code points) |
| 19:21 | <annevk> | Hixie: however, there's more on the first line now |
| 19:21 | <annevk> | Hixie: I "fixed" it |
| 19:21 | <Hixie> | if it's a JS file, parsing it as JSON seems unwise :-) |
| 19:21 | <annevk> | Hixie: but e.g. my links to "ASCII serialization of origin" broke |
| 19:22 | <Hixie> | ah, yeah. i can add legacy IDs if that would help. |
| 19:22 | <annevk> | Hixie: you also have colour and serialization somewhere, both variants in a single term |
| 19:22 | <annevk> | Hixie: too late now I guess |
| 19:22 | <Hixie> | i think the only mention of "serialization" is in en-US <cite> elements |
| 19:23 | <annevk> | Hixie: but I'm not sure it's helping that we have one spec in en-GB and everything else in en-US for non-native speakers |
| 19:23 | <Hixie> | (i fixed that much more recently) |
| 19:23 | <annevk> | Hixie: maybe some other term with z then |
| 19:23 | <Hixie> | (in fact that might not have been checked in yet) |
| 19:23 | <Hixie> | annevk: feel free to move them all to en-GB :-) |
| 19:23 | <annevk> | Hixie: yeah, it's "rules-for-serializing-simple-colour-values" |
| 19:24 | <Hixie> | the ID? |
| 19:24 | <Hixie> | i can add that |
| 19:24 | <annevk> | Hixie: no, that's the one you have in the spec |
| 19:24 | <Hixie> | that's fixed, just not checked in yet |
| 19:24 | <Hixie> | still trying to pin down a really subtle bug in the new pipeline |
| 19:24 | <annevk> | mkay |
| 19:25 | <annevk> | Moving everything to en-GB is not a credible suggestion |
| 19:25 | <JakeA> | racist |
| 19:25 | <Hixie> | (it takes like 30 minutes for me to run the pipeline each time i change the source a bit because i'm running it in an emulator, sigh) |
| 19:26 | <annevk> | JakeA: you're happily using z's in SW |
| 19:26 | <JakeA> | ServiceWorkerz |
| 19:26 | <annevk> | JakeA: clearly you spent too much time in the colonies |
| 19:26 | <Hixie> | annevk: anyway i added "rules-for-serializing-simple-colour-values" as an ID for the paragraph that will soon say "serialising" |
| 19:26 | <JakeA> | haha |
| 19:26 | <annevk> | Hixie: please don't, that was just an example |
| 19:26 | <Hixie> | annevk: oh |
| 19:26 | <Hixie> | annevk: i thought you were saying it was a broken link |
| 19:27 | <annevk> | Hixie: the things that actually broke for were around serializing origins |
| 19:27 | <Hixie> | ah |
| 19:27 | <annevk> | Hixie: well maybe, not sure what spec references that term though |
| 19:27 | <annevk> | Hixie: but I've updated xref |
| 19:27 | <Hixie> | k well either way, let me know what IDs I should add if you see other breakage |
| 19:27 | <annevk> | Hixie: so everyone that generates from whatwg/xref will get the new links |
| 19:29 | <annevk> | Hixie: you broke this btw https://github.com/whatwg/xref/commit/baa85a61896e7c044b17f000fda1c73345e8a437#diff-38676f2ddfbfc15a86de3811c50d83c2R5 |
| 19:29 | <annevk> | Hixie: as you can tell the fix is ugly |
| 19:29 | <Hixie> | well parsing that file is a lost cause |
| 19:29 | <Hixie> | that's not a data file |
| 19:29 | <Hixie> | it's a script |
| 19:30 | <Hixie> | i don't guarantee it won't change dramatically again |
| 19:30 | <Hixie> | in particular, i expect it to get a lot smaller |
| 19:30 | <Hixie> | maybe using a trie or something |
| 19:30 | <Hixie> | to make the data shorter |
| 19:31 | <Hixie> | if you want though i can definitely output a dedicated file that's actually in a defined format that you can parse |
| 19:31 | <Hixie> | bikeshed is asking for something similar |
| 19:31 | <Hixie> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=26335 |
| 19:45 | <annevk> | Hixie: I don't really care about maintaining any of these various things |
| 19:45 | <annevk> | Hixie: I'd like to switch to Bikeshed I guess if that's going to be maintained |
| 19:46 | <annevk> | Hixie: writing and coordinating spec writing is more fun |
| 20:19 | <caitp> | Hixie: it would be really neat if there were a section in http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html which enumerates all of the different ways a "submit" event would be dispatched --- primarily so that I have a nice snippet to link to on documentation for other projects |
| 20:19 | <caitp> | do you think such a thing could be possible |
| 20:20 | <Hixie> | there are different ways 'submit' can be dispatched? |
| 20:20 | <Hixie> | you mean other than a form submission and dispatchEvent() ? |
| 20:20 | <caitp> | particularly what would cause a form to be submit |
| 20:21 | <Hixie> | oh how to submit a form |
| 20:21 | <Hixie> | that's an entirely different question |
| 20:21 | <Hixie> | submit buttons, implicit submit buttons, .submit()... anything else? |
| 20:22 | <Hixie> | (form.submit() doesn't fire the 'submit' event but does submit the form) |
| 20:22 | <caitp> | well, I feel like submit should only be dispatched during form submission, but that's getting into semantics :p someone filed a bug about us invoking a submit event handler when an implicit submit button was clicked, and this resulted in confusion for them |
| 20:22 | <caitp> | so it would be handy to just have a nice collection of these to link people to from docs |
| 20:22 | <caitp> | and it might make writing tests and whatever easier for implementors, too |
| 20:23 | <Hixie> | click on "submitted" here: http://www.whatwg.org/specs/web-apps/current-work/#concept-form-submit |
| 20:23 | <annevk> | roc: do you still want to be copied on all those Fullscreen bugs? |
| 20:24 | <Hixie> | that links to all the places that call that algorithm in the HTML spec |
| 20:24 | <annevk> | roc: I sometimes wonder whether I bother you too much |
| 20:24 | <caitp> | it's extremely dangerous to open the single-page version on a laptop while building chromium hixie :( if this starts a fire I'm totally blaming you |
| 20:25 | <annevk> | Hixie: remember how I once suggested that we should change the markup around elements and such, such that indexes could be generated rather than written by hand? |
| 20:25 | <annevk> | Hixie: https://twitter.com/codylindley/status/494155769307078662 |
| 20:26 | <Hixie> | file a bug, put 'tools' in the whiteboard, make sure the bug report describes exactly what you want to get out |
| 20:26 | <Hixie> | right now there are so many exceptions that i don't know how to do it |
| 20:26 | <Hixie> | i mean, some of the indexes and stuff are autogenerated |
| 20:26 | <Hixie> | but in general what is in the table and what is in the element definitions need to be different in subtle ways |
| 20:32 | <Ms2ger> | Does <script async defer src=...> even make sense? |
| 20:32 | <Ms2ger> | Hixie, ^ |
| 20:35 | <Hixie> | right now async and defer are mutually exclusive iirc |
| 22:25 | <TabAtkins> | Hixie: How is your WebIDL processed? Does your pipeline parse it and add markup or something, or is it all done manually? |
| 22:26 | <Hixie> | right now i just have some poor man's regexp |
| 22:26 | <Hixie> | but my plan is to actually parse it |
| 22:28 | <TabAtkins> | Okay. |
| 22:28 | <TabAtkins> | You're not using Python, so Peter's widlparser project won't help you, but it might be useful to look at for reference. |
| 22:28 | <TabAtkins> | I was wondering because it's the main thing that needs to be marked up to make Bikeshed happy. |
| 22:29 | <Hixie> | marked up how? |
| 22:47 | <TabAtkins> | With a data-dfn-type attribute, so Shepherd knows whether the dfns are for interface names, method names, attributes, dicts, etc. |
| 22:48 | <TabAtkins> | And for things like attributes, a data-dfn-for giving the name of the interface they belong to, to allow collision resolution. |
| 22:52 | <roc> | annevk: fullscreen bugs aren't the highlight of my day, but I do like to be helpful |
| 23:05 | <Hixie> | TabAtkins: i don't expect to add any attributes that aren't useful for the end user, since those are just going to use up bytes on the wire. But I could make a dedicated file in the appropriate format. |
| 23:05 | <Hixie> | TabAtkins: anyway, file a bug with precise details on what you want |
| 23:05 | <TabAtkins> | Yeah, filing it now. |
| 23:05 | <Hixie> | thanks |