| 05:37 | <zcorpan> | "My same thoughts on "break the web" ... I think whoever put "use strict" in there would eventually never expect the `this` to be the global context." (on es-discuss) |
| 07:11 | <annevk> | zcorpan: that guy routinely suggests breaking the web |
| 07:13 | <hsivonen> | jgraham: Web Platform Test integration with Gecko testing seems awesome. Thank you! |
| 07:22 | <annevk> | I wonder what simon.html5.org hosts that is so popular. 20GiB of bandwidth. philip.html5.org reaches 80GiB |
| 07:22 | <annevk> | html5.org itself meanwhile is at 23 |
| 07:24 | <annevk> | dom.spec.whatwg.org is at 4 |
| 08:33 | <jgraham> | hsivonen: Great! |
| 09:15 | <Domenic> | zcorpan: yes, every message from that poster makes me question my subscription :( |
| 09:26 | <annevk> | krijnhoetmer: can haz /irc-logs in TLS? |
| 09:41 | <annevk> | mounir: did you just go back to my original strawman for sensor API design? |
| 09:41 | <annevk> | mounir: that thread went full circle fast |
| 09:42 | <jgraham> | annevk: It seems my blog.whatwg.org account is still jgraham⊙oc Can you fix it (or vend me a new account?) It seems like a good time to do a testing status update. For all 6 readers of the whatwg blog. |
| 09:43 | <Domenic> | it'd be cool to have more posts on the blog |
| 09:43 | <annevk> | jgraham: I'll have a look |
| 09:45 | <mounir> | annevk: not really |
| 09:45 | <mounir> | annevk: I just pointed an alternative |
| 09:45 | <mounir> | annevk: I thought the Rick W. was championing your proposal actually |
| 09:46 | <annevk> | he mocked it on Twitter |
| 09:46 | <mounir> | annevk: I think your proposal or Tim's are sensitive |
| 09:46 | <jgraham> | Domenic: Yeah, I guess we just need to wait until blogging becomes retro-cool |
| 09:46 | <mounir> | annevk: he did a mock of the proposal in 140 characters? |
| 09:47 | <annevk> | mounir: https://twitter.com/rwaldron/status/504355208315830272 |
| 09:48 | <mounir> | annevk: oh, I thought that he mocked the fact that I pointed to the same thing as an alternative |
| 09:48 | <mounir> | annevk: note that without the mocking part I agree with him - as pointed in my email - |
| 09:49 | <annevk> | I'm no longer following |
| 09:49 | <mounir> | annevk: it's probably a bit painful to only expose an async api |
| 09:49 | <annevk> | But I'm not sure I need to |
| 09:49 | <annevk> | Oh man, WordPress updates... |
| 09:49 | <mounir> | annevk: did you look at Tim's proposal? |
| 09:49 | <annevk> | Yeah, but it didn't seem so nice |
| 09:50 | <annevk> | It has a factory design |
| 09:50 | <mounir> | how so? |
| 09:50 | <annevk> | getDeviceOrientationSensor() |
| 09:51 | <mounir> | annevk: it's not really a factory design if you return a very specific type out of it ;) |
| 09:51 | <mounir> | annevk: I hear your concern but it's only cosmetic to me |
| 09:52 | <annevk> | That's what C++ API designers say about JavaScript APIs too :-P |
| 09:53 | <annevk> | jgraham: hoppipolla? |
| 09:54 | <annevk> | jgraham: I went with that |
| 10:01 | <darobin> | wow, that became a meme fast! |
| 10:03 | <mounir> | darobin: a meme feast? |
| 10:03 | <darobin> | heh |
| 10:03 | <darobin> | there clearly is someone sitting on this channel and bored at a meeting — I suspect a CSS meeting :) |
| 10:03 | <mounir> | it can't be sgalineau |
| 10:04 | <darobin> | it can't be either sgalineau nor hober |
| 10:04 | <jgraham> | annevk: Yeah |
| 10:05 | hober | whistles tunelessly |
| 10:06 | <annevk> | heh |
| 10:06 | <hober> | hello from sophia antipolis! |
| 10:07 | <TabAtkins> | hober: Your meme unfairly credits that line to annevk, when it was mounir. |
| 10:07 | <annevk> | TabAtkins: you're not reading it correctly |
| 10:07 | <mounir> | TabAtkins: I think anne is credit |
| 10:07 | <mounir> | ... is credited what he should be credited |
| 10:08 | <TabAtkins> | annevk: Hm, I guess so. It's not super-clear, but I get it when I read it sufficiently closely. |
| 10:13 | <darobin> | you can't just read w3cmemes quickly like that TabAtkins |
| 10:13 | <darobin> | it requires exegesis |
| 10:13 | <TabAtkins> | I REFUSE TO EXPEND CRITICAL THOUGHT ON MEMES |
| 10:14 | <Domenic> | this article on google's sha-1 deprecation is really good https://konklone.com/post/why-google-is-hurrying-the-web-to-kill-sha-1 |
| 10:23 | <Philip`> | annevk: 80GB over what time period? |
| 10:27 | <eto> | hello |
| 10:28 | <annevk> | Philip`: monthly |
| 10:28 | <annevk> | Philip`: it's not really a big deal btw, was mostly surprised |
| 10:29 | <Philip`> | annevk: Seems a lot higher than I would expect too |
| 10:29 | <Philip`> | Maybe someone found the pirated movies I stored on that domain |
| 10:39 | <mathiasbynens> | for simon.html5.org it’s probably due to people ironically hotlinking http://simon.html5.org/valid-html5.png |
| 11:41 | <hsivonen> | is there a newbie guide for submitting html5lib-tests pull requests the Right Way? |
| 11:43 | <annevk> | jgraham: ^^ |
| 11:43 | <Ms2ger> | hsivonen, I assume the same process as web-platform-tests |
| 11:45 | <jgraham> | Yeah, same process, but to html5lib-tests rather than web-platform-tests mostly |
| 11:46 | <hsivonen> | jgraham: so 1) fork repo on github, 2) pull your own fork, 3) create a branch in the local repo, 4) commit changes to branch, 5) ??? |
| 11:46 | <aleray> | hi, is this valid html? http://jsfiddle.net/pef48j5e/ |
| 11:46 | <jgraham> | hsivonen: 5) Press the "create pull request" button in the GitHub UI |
| 11:47 | <aleray> | I want some content to be "centered and sized", and some other content to take the full width of the viewport |
| 11:47 | <hsivonen> | jgraham: OK. Should there be squashing before that? |
| 11:48 | <jgraham> | hsivonen: Starting with a single commit is better, unless it's clearly logically seperate commits. Please don't squash after the initial submission though |
| 11:48 | <jgraham> | (it upsets the code review tool) |
| 11:48 | <hsivonen> | my understanding of the lifetime of feature branches in the PR-based workflow is still hazy |
| 11:48 | <hsivonen> | jgraham: ok. that differs from the Gaia guideline |
| 11:49 | <jgraham> | hsivonen: If you don't want to fork, you are welcome to have access to the main repo to push your branch |
| 11:49 | <jgraham> | This is better in several ways (notably that other people can more easilly push commits to fix review issues) |
| 11:49 | <hsivonen> | jgraham: thanks. that kind of thing is why I haven't learned the PR-based workflow :-) |
| 11:50 | <jgraham> | hsivonen: It's still a PR-based workflow |
| 11:50 | <hsivonen> | jgraham: oh |
| 11:50 | <jgraham> | You just push the branch to the main repo rather than to your fork |
| 11:51 | <hsivonen> | so far, I've had too much write access to learn how the real workflow works |
| 11:51 | <jgraham> | (under the covers it's actually even more similar than you would think given the UI) |
| 11:52 | <hsivonen> | jgraham: will the branch stay around forever or does it get flattened out somehow after the PR has been accepted? |
| 11:52 | <hsivonen> | jgraham: will future generations be able to laugh at my intermediate steps? |
| 11:53 | <jgraham> | hsivonen: You can rebase+squash once the review is accepted and delete the review branch |
| 11:54 | <hsivonen> | jgraham: ok. |
| 11:54 | hsivonen | still doesn't grok how "everyone has a copy of all data" and "you can delete a branch" interact |
| 11:57 | <jgraham> | hsivonen: Well the branch is only a pointer to the commits. So you can delete the pointer without deleting the commits. Then after some time (90 days by default) any commits that aren't reachable from a pointer are discarded. Of course if someone else has cloned the repo, they might still have a pointer to your commits locally and they could push those commits back upstream. So you can't guarantee that they have been expunged forevermore, but it's n |
| 11:57 | <hsivonen> | jgraham: you got cut off at "but it's no" |
| 11:58 | <smaug____> | oh, git exposes gc behavior. boo |
| 11:58 | <jgraham> | "but it's not a scenario that happens much in practice becase there usually isn't any reason to push old commits that have already been merged." |
| 12:01 | <jgraham> | smaug____: Well that's a feature I think. The mercurial equivalent is either strip files or not actually deleting commits but marking them as hidden |
| 12:02 | <smaug____> | I would expect gc-able stuff to not show up anywhere, certainly not in clones |
| 12:02 | <smaug____> | so gc could run during cloning |
| 12:04 | <jgraham> | I don't know if gc-able stuff does show up in clones, actually. I was thinking of the situation where someone had already cloned and the commits were de-referenced upstream. |
| 12:19 | <hsivonen> | does anyone want to bet whether Blink gets svg.innerHTML = "<frameset>" right? |
| 12:20 | <hsivonen> | does anyone want to guess if the spec gets it right? |
| 12:20 | <hsivonen> | or what the spec says about it |
| 12:20 | hsivonen | hasn't read the spec on that point, yet |
| 12:26 | <TabAtkins> | hsivonen: Dunno what the spec says, but Blink just creates an <svg|frameset> element. |
| 12:27 | <Domenic> | hsivonen: out of curiousity why are you looking into this stuff? For Servo? Ran into similar stuff on jsdom recently. |
| 12:29 | <hsivonen> | Domenic: for Gecko. |
| 12:30 | <Domenic> | hsivonen: kk. is there a bug I can watch? |
| 12:30 | <hsivonen> | jgraham: I accepted your invitation. Thanks |
| 12:30 | <hsivonen> | Domenic: https://bugzilla.mozilla.org/show_bug.cgi?id=886390 |
| 12:30 | TabAtkins | fantasai, your latest changes to default.css give all the links a double underline |
| 13:18 | <ondras> | so, web components/shadow dom |
| 13:18 | <ondras> | is this channel suitable or is there a better one. |
| 13:18 | <jgraham> | It depends on what you want to talk about, I guess |
| 13:18 | <ondras> | I am having an issue with createdCallback not called when the custom element in question is created inside a shadow root of another element |
| 13:20 | <ondras> | but this is for the first time I am digging deeper into this stuff, so the chance is the problem is somewhere between the chair and the monitor. |
| 13:22 | <TabAtkins> | ondras: Maybe ask in #polymer? |
| 13:22 | <ondras> | TabAtkins: well I would like to have this working without polymer at first, adding polyfills later |
| 13:28 | <TabAtkins> | ondras: I just meant that people in that room know quite a bit more about this specific topic. |
| 13:28 | <TabAtkins> | (I also suspect it's PEBCAK, but I can't help you any further than that.) |
| 13:28 | <ondras> | .) |
| 13:29 | <ondras> | okay, thanks, will try. |
| 13:31 | <TabAtkins> | Lolwut http://www.w3.org/community/w3process/track/issues/124 |
| 13:33 | <ondras> | thou shalt be blacklistored. |
| 14:12 | <annevk> | TabAtkins: RaA? |
| 14:12 | <annevk> | jgraham: did it work out with the blog? |
| 14:12 | <TabAtkins> | Reductio ad absurdum |
| 14:13 | <annevk> | I see |
| 14:13 | <annevk> | The situation seems fairly logical to me, and the person calling FUD is actually doing that himself |
| 14:14 | <annevk> | film at 11? |
| 14:15 | <darobin> | annevk: huh? |
| 14:16 | <zcorpan> | foolip: interested in investigating the compat impact of changing (min-|max-)?device-(width|height) media features so they are equivalent to width/height? |
| 14:16 | <Domenic> | ^ lol, the web |
| 14:17 | <annevk> | Didn't the "mobile web" ruin that? |
| 14:18 | <jgraham> | annevk: Uh, I got distracted by actual bugs. And an inability to write through lack of practice. But I managed to reset my password, so yeah |
| 14:18 | <annevk> | jgraham: sounds good |
| 14:20 | <jgraham> | annevk: Which situation? |
| 14:21 | <jgraham> | I'm pretty sure that any situation that's logical is offtopic, so we need to correct this as fast as possible |
| 14:21 | <annevk> | jgraham: heh |
| 14:21 | <annevk> | jgraham: https://twitter.com/domenic/status/508964136366391297 |
| 14:26 | <hober> | Domenic: you keep referring to something the Director said, but i missed the relevant actual quote from him. |
| 14:26 | <Domenic> | hober: Art was referencing a conversation which I presume he did not take notes during. |
| 14:26 | <annevk> | hober: "However, based on my conversations with Consortium staff last week, the Director will NOT permit a Proposed Recommendation to include a normative reference to a WHATWG spec." from http://www.w3.org/community/w3process/track/issues/124 |
| 14:27 | <hober> | annevk: ahh, thanks. so not only is there not a quote from the director, the conversation didn't even include him? |
| 14:28 | <annevk> | hober: depending on what the actual the Director said, a conversation with the Director does not have to include the actual the Director |
| 14:28 | <hober> | annevk: :) |
| 14:28 | <annevk> | hober: so I've no idea :-) |
| 14:29 | <hober> | i guess what i'm saying is that https://twitter.com/domenic/status/508977165028057088 is overstating the case |
| 14:30 | <jgraham> | Can we get a journalist to phone TimBL and get a quote |
| 14:31 | <jgraham> | Everyone knows that the best medium for relationship counselling is through the media |
| 14:38 | <darobin> | jgraham++ |
| 14:39 | <darobin> | that is, indeed, not an actual quote from the Director |
| 14:40 | <darobin> | who, as far as I know, was not on the call being referred to |
| 14:40 | <darobin> | and therefore was unable to make quotes either way, irrespective of his opinion |
| 14:41 | <darobin> | but hey, it would be ridiculous to let facts get in the way of some good old fashioned standards drama |
| 14:42 | <hober> | s/standards drama/game of telephone/ |
| 14:43 | <jgraham> | More worrying, tracker seems to think that reverse chronological order is a sensible way to sort email |
| 15:04 | <annevk> | Is TLS the new URI or ECMAScript, or actually a term we should favor over SSL? |
| 15:05 | <sgalineau> | darobin: well, the most recent game of standards drama worked out so well, why not have another round? |
| 15:12 | <annevk> | sgalineau: darobin will solve it soon by introducing something that is neither W3C nor WHATWG |
| 15:13 | <hober> | ... and combines the problems of both |
| 15:13 | <sgalineau> | annevk: a third group could be the only way to make WHATWG and W3C agree on something "OMG NO" |
| 15:14 | darobin | soldiers on! |
| 15:14 | sgalineau | cue 'Object to this' meme |
| 15:16 | <annevk> | sgalineau: contrasted with the IETF, at least we can agree on Unicode and text/html (these days) |
| 15:17 | <sgalineau> | annevk: whoa. WHATWG agrees on *two* things with a standards body? Jesus. You guys are practically selling out. |
| 15:18 | <annevk> | sgalineau: did you hear we're publishing snapshots? http://blog.whatwg.org/make-patent-commitments |
| 15:19 | <sgalineau> | annevk: you can't fool me. you are the Tools Of Evil. |
| 15:27 | <Domenic> | TabAtkins: you may enjoy https://github.com/domenic/ecmarkdown. For progress see https://github.com/domenic/ecmarkdown/tree/master/test/cases |
| 15:35 | <TabAtkins> | Domenic: You should match CMD's list handling. |
| 15:35 | <Domenic> | TabAtkins: this isn't meant to be a markdown dialect; much more restrictive. |
| 15:35 | <Domenic> | for algorithms only |
| 15:35 | <TabAtkins> | Right, I know, it's a subset and redefinition. |
| 15:36 | <TabAtkins> | But insofar as you're copying MD, you should copy CMD. |
| 15:36 | <Domenic> | Part of the goal is uniformity. We don't want some parts of the spec using 0., 0., 0. and others using 1., 1., 1. and others using 1., 2., 3. |
| 15:41 | <TabAtkins> | ??? |
| 15:42 | <Domenic> | This is for ECMAScript which will have many contributors |
| 15:43 | <TabAtkins> | What I'm talking about is the indentation. |
| 15:43 | <Domenic> | Oh! |
| 15:44 | <Domenic> | I haven't read that, so perhaps I will. |
| 15:45 | <annevk> | mathiasbynens: seen a good DNSSEC guide? TransIP seems to offer it, but I've no idea how to configure it |
| 15:45 | <mathiasbynens> | annevk: nope |
| 15:45 | <annevk> | Also, it's a bit unclear to me what DNSSEC actually offers |
| 15:48 | <TabAtkins> | Domenic: Each list-item establishes a required indentation by the number of characters before it starts. |
| 15:49 | <annevk> | http://en.wikipedia.org/wiki/Domain_Name_System_Security_Extensions seems fairly clear as to what it offers... |
| 15:54 | <jgraham> | These things annoy me: emails sent as both HTML and plain text where the plain text is missing substantial information compared to the HTML |
| 16:00 | <Hixie_> | annevk: did you point tantek to that? |
| 16:00 | <annevk> | Hixie_: context? |
| 16:00 | <Hixie_> | http://www.w3.org/community/w3process/track/issues/124 |
| 16:00 | <Hixie_> | and yeah, if anyone ever wondered what the w3c's position on the whatwg was, the phrase "explicitly blacklist WHATWG" should pretty much put it to rest |
| 16:01 | <annevk> | I saw on Twitter the guy just lost his sandwhich, not sure I want to bring more bad news |
| 16:02 | <Hixie_> | lol |
| 16:03 | <Hixie_> | wait, where are you seeing that he lost his sandwich! |
| 16:03 | <annevk> | hober: http://googlefight.com/index.php?lang=en_GB&word1=ssl&word2=tls |
| 16:04 | <annevk> | Hixie_: https://twitter.com/adactio/status/508982320553283584 |
| 16:04 | <Hixie_> | lol |
| 16:05 | <jgraham> | Hixie_: To be fair the phrase "explicitly blacklist WHATWG specs" came from Art; he pointed out that he disagreed with it as policy but believed it to be W3C's position, and further discussion suggested that it was not, in fact, W3C |
| 16:05 | <Domenic> | annevk: interesting, i did not expect that tls. |
| 16:05 | <jgraham> | 's position (or hasn't yet been confirmed as such) |
| 16:05 | <jgraham> | So, all in all, it seems like a terrible way of infering what W3C's actual position is |
| 16:06 | <Domenic> | annevk: lots of non-web related results on first page. "Top law schools" "times literary supplement" other orgs with that acronym. |
| 16:06 | <Hixie_> | jgraham: yeah, their official position is "I think it would be inappropriate for the Staff to publicly speculate" |
| 16:07 | <Domenic> | annevk: SSL has a couple of those too though |
| 16:07 | <jgraham> | Right. I agree "we aren't going to tell you our position" is a pretty terrible position to take |
| 16:07 | <jgraham> | Hopefully it's shortlived |
| 16:08 | <Hixie_> | it's already not short-lived :-) |
| 16:09 | <jgraham> | compared to that email |
| 16:10 | <Hixie_> | btw i was earlier going to say that we should adopt a policy of not referencing stale specs |
| 16:10 | <Hixie_> | but then i realised we already have that policy |
| 16:10 | <Hixie_> | since we just reference ed drafts |
| 16:13 | <annevk> | Domenic: URI <> URL; Atom <> RSS are much more clearcut |
| 16:13 | <Hixie_> | Domenic, annevk: what's the setup you want on streams.spec.whatwg.org |
| 16:14 | <annevk> | Hixie_: sounds like Domenic wants an SSH account |
| 16:14 | <Hixie_> | basically there's two options. Domenic has an ssh account, annevk has an ssh account. :-) |
| 16:14 | <Hixie_> | k |
| 16:15 | <Domenic> | ^ sounds good |
| 16:15 | <Hixie_> | username? |
| 16:15 | <Domenic> | domenic |
| 16:15 | <Hixie_> | you don't have any weird preferences for shell, do you |
| 16:15 | <Hixie_> | bash is ok? |
| 16:16 | <Domenic> | bash is superb |
| 16:16 | <Hixie_> | hm, looks like domenic is taken |
| 16:16 | <Hixie_> | (sorry, shared host) |
| 16:16 | <Domenic> | aw |
| 16:16 | <Domenic> | I think ddenicola will be easiest to remember |
| 16:16 | <Domenic> | (but nobody spells my name with an "e"... craziness!) |
| 16:17 | <Hixie_> | k... now the interesting part |
| 16:17 | <Hixie_> | how do i get this password to you with a theoretically secure chain of trust |
| 16:17 | <Domenic> | i was wondering that myself |
| 16:18 | <jgraham> | Write it down and go visit him? |
| 16:18 | <Hixie_> | you on hangouts? that's encrypted end-to-end, and i trust google not to steal these credentials since i already trust them not to use my e-mail to break into my dreamhost host |
| 16:18 | <Hixie_> | (and also, i am part of google.) |
| 16:19 | <Domenic> | Hangouts should work yeah |
| 16:20 | <Hixie_> | send one to ian.hickson⊙gc |
| 16:20 | <Hixie_> | annevk: while i'm at the dreamhost panel, what's out story with this tls nonsense |
| 16:21 | <annevk> | Hixie_: we have two options, either you get StartSSL verified, or I you hand over StartSSL WHATWG business to me somehow as I'll be StartSSL verified (class 2) within 10 business days |
| 16:21 | <Hixie_> | how do i do that second one? |
| 16:21 | <Domenic> | For future reference https://www.npmjs.org/package/cipherhub seems like a feasible solution for key exchange |
| 16:21 | <annevk> | Hixie_: we need to be StartSSL class 2 for whatwg.org as we have a complex subdomain setup |
| 16:21 | <annevk> | Hixie_: I guess I need to ask them |
| 16:23 | <annevk> | Hixie_: now that I know you're okay with that, I'll look into it |
| 16:24 | <Hixie_> | i'm definitely ok with offloading as much work as possible :-) |
| 16:24 | <Hixie_> | Domenic: ok, your account is set up on streams.spec.whatwg.org |
| 16:24 | <Hixie_> | might take a few minute for dns to propagate |
| 16:24 | <Hixie_> | let me know when to update the spec index |
| 16:25 | <Domenic> | Hixie_: awesome, thanks. And yeah, no need to update the spec index for a couple weeks probably; I just want to kill the URL I am currently using so nobody links to it. |
| 16:26 | <Hixie_> | i wonder how i do the wildcard stuff on dreamhost |
| 16:26 | <Hixie_> | do i have to manually add the same cert to each of these damn subdomians? |
| 16:26 | <Hixie_> | subdomains |
| 16:27 | <annevk> | Yeah you might have to, apparently if you have direct access to the server it's much more convenient |
| 16:27 | <Hixie_> | i technically do have root but i think the panel happily nukes my settings every now and then |
| 16:27 | <mathiasbynens> | annevk, Hixie_: btw iirc even class 2 only allows *.domain.ext, not *.*.domain.ext |
| 16:27 | <mathiasbynens> | re: *.spec.whatwg.org |
| 16:27 | <annevk> | mathiasbynens: we don't need *.*, just *.spec.whatwg.org |
| 16:28 | <Hixie_> | i guess we'd need *.spec.whatwg.org and *.whatwg.org |
| 16:28 | <annevk> | right |
| 16:28 | <mathiasbynens> | and whatwg.org |
| 16:28 | <Hixie_> | that we actually have already |
| 16:28 | <mathiasbynens> | Hixie_: did you see http://krijnhoetmer.nl/irc-logs/whatwg/20140907#l-257? |
| 16:29 | <Hixie_> | yeah that's why we're doing this |
| 16:29 | <Hixie_> | amongst other reasons |
| 16:30 | <mathiasbynens> | Hixie_: the mixed content issues can be fixed already, though |
| 16:30 | <Domenic> | Yeah I think just change the <base> to // instead of http://? |
| 16:30 | <mathiasbynens> | for example, yeah |
| 16:30 | Hixie_ | discovers that dreamhost actually does support splitting one domain across multiple users after all |
| 16:30 | <mathiasbynens> | for that particular page |
| 16:31 | <Hixie_> | oh, i see |
| 16:31 | <Hixie_> | wait, what does that page reference that's explicitly by http://? |
| 16:31 | <Domenic> | the <base> tag |
| 16:31 | <mathiasbynens> | <base href="http://whatwg.org/specs/">; <!-- because it's also used on http://spec.whatwg.org/ --> |
| 16:31 | <Domenic> | fun fact, browsers let you type //example.com into the URL bar |
| 16:31 | <Hixie_> | ohh |
| 16:31 | <Hixie_> | right |
| 16:32 | <Hixie_> | actually whatwg.org is just a redirect to www.whatwg.org |
| 16:33 | <Hixie_> | so i'm not sure we can secure that one technically (given how dreamhost does things) |
| 16:33 | <Hixie_> | anyway. spec list fixed. |
| 16:33 | <Domenic> | boooo no www |
| 16:33 | <mathiasbynens> | isn’t whatwg no-www friendly? |
| 16:33 | <mathiasbynens> | Hixie_: for other documents, it’s probably easiest to grep for 'http://whatwg.org'; and go from there |
| 16:33 | <Hixie_> | it's www.whatwg.org because it's the domain for the world wide web's specs! |
| 16:33 | <Hixie_> | not because www. is our web server |
| 16:34 | <jgraham> | No, then it would be www.spec.whatwg.org Or sepc.www.whatwg.org |
| 16:34 | <jgraham> | *spec |
| 16:34 | <Hixie_> | url.spec.www.whatwg.org would have been too long |
| 16:35 | <Hixie_> | and dreamhost won't let us host www.spec.whatwg.org separate from spec.whatwg.org |
| 16:35 | <jgraham> | so which non www specs are you planning on writing there? |
| 16:35 | <Hixie_> | i tried to set that up for hixie.ch vs www.hixie.ch literally over a decade ago |
| 16:35 | <Hixie_> | and they wouldn't have any of it |
| 16:35 | <Ms2ger> | www.spec.whatwg.org is the HTML spec, right? |
| 16:36 | <jgraham> | (it seems like any non-www specs are also non-*w*hatwg-specs) |
| 16:36 | <Hixie_> | should be. sadly it's html.spec.whatwg.org |
| 16:36 | <Hixie_> | jgraham: the existence of a label doesn't imply the existence of other labels :-) |
| 16:37 | <jgraham> | Hixie_: No, but that situation does suggest redundancy :) |
| 16:37 | <Hixie_> | possibly |
| 16:37 | <Domenic> | lazyirc: i assume a bunch of tests exist in web-platform-tests for document.write? |
| 16:37 | <jgraham> | Domenic: "a bunch" |
| 16:37 | <Hixie_> | but redundancy in web hosting is good, it makes us more reliable! :-P |
| 16:38 | <Domenic> | jgraham: enough that we feel good about it, as opposed to most of the coverage in WPT? |
| 16:38 | <Domenic> | jgraham: I feel like document.write is such a hairy area but also a pretty interoperable one that someone must have tested it a lot.... |
| 16:39 | <jgraham> | https://github.com/w3c/web-platform-tests/tree/13bff083fba249ed260966bca65319b1b35d3f34/html/dom/dynamic-markup-insertion |
| 16:39 | <Hixie_> | document.write is less interoperable than you might imagine |
| 16:39 | <Domenic> | jgraham: that looks very nice, thanks :D |
| 16:41 | <caitp> | it's interoperable enough for a react.js slide deck which will only ever be opened on chromium |
| 16:41 | <Hixie_> | Ms2ger: re https://www.w3.org/Bugs/Public/show_bug.cgi?id=26744 |
| 16:42 | <mathiasbynens> | Hixie_: fwiw there are some hardcoded `http://whatwg.org`s in https://www.whatwg.org/specs/web-apps/current-work/*.js |
| 16:42 | <Hixie_> | Ms2ger: the text you're talking about is three paragraphs from a heading that introduces the disabled attribute |
| 16:42 | <Hixie_> | Ms2ger: and the first list item there explicitly lists the elements |
| 16:42 | <Hixie_> | Ms2ger: what more can i do? |
| 16:42 | Ms2ger | looks |
| 16:42 | <jgraham> | caitp: So you're saying it's interoperable enough that Blink behaviour never regresses? I wonder if you have tested that when they actually made changes to the document.write code :p |
| 16:43 | <caitp> | well, being a slide deck, it stops mattering after a few weeks :> |
| 16:43 | <Ms2ger> | Oh, hm |
| 16:44 | <Hixie_> | annevk: the text you cite in https://www.w3.org/Bugs/Public/show_bug.cgi?id=25099#c33 is exactly what i want to write, but that text doesn't work with the way you've defined events. |
| 16:44 | <Hixie_> | annevk: (because it doesn't invoke the constructor, which would be a very weird thing to do in prose) |
| 16:44 | <Ms2ger> | Hixie_, okay, wfm |
| 16:44 | <caitp> | the joke is that it's interoperable enough to be used in situations where you don't care about interoperability |
| 16:45 | <Hixie_> | Ms2ger: cool |
| 16:45 | Ms2ger | was going to complain at Hixie_ about something, but can't remember what |
| 16:45 | <Ms2ger> | You get off this time! |
| 16:45 | <annevk> | Hixie_: yeah fair, I wonder how we resolve the situation of not defining the global object for many objects |
| 16:46 | <Hixie_> | annevk: ? |
| 16:46 | <Hixie_> | annevk: how do you mean? |
| 16:47 | <jgraham> | Ms2ger: Come on? You can't think of anything to complain to Hixie_ about? localStorage? The ugly WHATWG green? His belief that cats make good pets? |
| 16:47 | <Hixie_> | cats make terrible pets |
| 16:47 | <Hixie_> | they make great housemates though |
| 16:47 | <annevk> | Hixie_: https://www.w3.org/Bugs/Public/show_bug.cgi?id=24652 |
| 16:47 | <Ms2ger> | WHATWG green isn't too bad |
| 16:47 | jgraham | finds something to complain at Ms2ger about |
| 16:48 | <Ms2ger> | But I was thinking something he didn't already know |
| 16:48 | mathiasbynens | randomly blames Ms2ger for the low-res WHATWG favicon.ico |
| 16:48 | <Hixie_> | annevk: i don't understand the relevance of that to the event thing |
| 16:48 | <annevk> | Hixie_: if we define things in terms of a constructor that we invoke, there'd be an associated realm |
| 16:49 | <Hixie_> | annevk: why do we care about associated realms here |
| 16:49 | <annevk> | because otherwise it's unclear where the prototype comes from |
| 16:49 | <Hixie_> | annevk: we create an object. creating an object is something we do all over the place in the specs. |
| 16:49 | <annevk> | yeah, this is a pretty big problem |
| 16:49 | <Hixie_> | annevk: well we're not going to create constructors for everything we create objects for |
| 16:50 | <Hixie_> | annevk: so whatever our solution is, it has nothing to do with constructors |
| 16:50 | <Hixie_> | annevk: so again, i don't see how this affects the events issue |
| 16:50 | <annevk> | well for events constructors are nice because of the dictionary design |
| 16:50 | <annevk> | as spec authors have difficulty getting the prose right |
| 16:50 | <annevk> | IDL is much easier |
| 16:51 | <Hixie_> | if spec authors have problems writing prose, this is going to be the least of their issues |
| 16:56 | <annevk> | Is this issue high priority for you? I'd prefer to wait until IDL is revamped. Hopefully then we have a better idea of how to define objects. Provided real effort is put into that |
| 16:57 | <Hixie_> | given that i think that the resolution is for me to do nothing and you to revert the changes to DOM, no, not particularly urgent. :-) |
| 17:00 | <Hixie_> | annevk: while i have you here, see my response on https://www.w3.org/Bugs/Public/show_bug.cgi?id=26103#c10 |
| 19:55 | <jwalden> | peoples! can someone explain to me why the searchParams serialization algorithm detailed for URLSearchParams would be totally unaware of anything but ASCII? seems like that nonsense just shouldn't be, for new APIs |
| 20:02 | <jsbell> | jwalden: You mean http://url.spec.whatwg.org/#concept-urlencoded-serializer ? |
| 20:02 | <jwalden> | jsbell: yes, that and http://url.spec.whatwg.org/#concept-urlencoded-byte-serializer |
| 20:04 | <okj579> | that algorithm can use any encoding but defaults to utf8 |
| 20:11 | <jwalden> | oh, am I misreading? step 1 converts Unicode stuffs into byte arrays, then serialization individually converts each byte into +/percent-encoding/that code point? |
| 20:16 | <okj579> | yeah, i think so |
| 20:22 | <arunranga> | hi Domenic! heads up that domenic.me/2014/02/14/the-revealing-constructor-pattern/ results in 404 |
| 20:23 | <Hixie_> | jorendorff: bz says that Firefox is going to "progressively parse scripts, and possibly progressively bytecode-compile, and possibly progressively native-code-compile". Do you know how this will interact with the ES6 moduel loader? |
| 20:44 | <jwalden> | Hixie_: what does "progressively" mean in these contexts? |
| 20:45 | <Hixie_> | did i delete the e-mail that explained that |
| 20:45 | <Hixie_> | ah, no, yoav explains it: |
| 20:45 | <Hixie_> | > If receiving multiple resources progressively provides better performance |
| 20:45 | <Hixie_> | > then having the server sending them one after the other, then the hint |
| 20:45 | <Hixie_> | > should be sent. |
| 20:51 | <jorendorff> | Hixie_: Hmm. It seems like there is a problem in the ES6 Loader design: that design requires parsing (the full module body) to succeed before starting work on dependencies |
| 20:53 | <jorendorff> | Hixie_: the normalize hook is called, for each import-declaration, after parsing -- |
| 20:53 | <jorendorff> | seems like it is specified to NOT eagerly go after those dependencies before the first one is done loading |
| 20:54 | <Hixie_> | jorendorff: yeah, i've been trying to tell es-discuss this for like a month now |
| 20:56 | <jorendorff> | Hixie_: probably because if you *really* want that Loader to perform, you're not waiting around for all these round trips anyway |
| 20:56 | <jorendorff> | the server has to tell the app what it needs and push it -- which the model does kind of support |
| 21:03 | <Hixie_> | jorendorff: not really, unless you mean it supports it because the cache gets primed. |
| 21:03 | <Hixie_> | jorendorff: but in practice i don't think we can rely on servers like that |
| 21:03 | <jorendorff> | it "supports" it only in that it has appropriate hooks |
| 21:03 | <Hixie_> | jorendorff: i mean, it's been 24 years and people still can't set their mime types correctly |
| 21:03 | <jorendorff> | Hixie_: This is where I have trouble understanding what you're saying though |
| 21:04 | <jorendorff> | Hixie_: because if the client is to have any understanding beforehand of what is to be loaded... then it has to get that somewhere |
| 21:04 | <jorendorff> | from a manifest or "this needs that" declarations of some kind |
| 21:04 | <Hixie_> | sure. people are really good at making their servers serve up static files. |
| 21:04 | <Hixie_> | url -> file |
| 21:04 | <Hixie_> | or url -> cgi script -> file |
| 21:05 | <Hixie_> | it's when you have to configure the server that things get dicey |
| 21:05 | <Hixie_> | url -> file + configured metadata |
| 21:05 | <Hixie_> | url -> multiple files |
| 21:05 | <Hixie_> | url -> redirect, even, in many cases |
| 21:07 | <jorendorff> | "asset pipeline" stuff exists though |
| 21:08 | <jorendorff> | it's not seen as configuration by the people who have to put up with it |
| 21:08 | <jorendorff> | it's just the way their platform does things, and they play along with a few rules |
| 21:10 | <Hixie_> | for some people, sure |
| 21:10 | <Hixie_> | most people, though, i doubt will ever get this working |
| 21:10 | <Hixie_> | just like now some people have CDMs that manage mime types correctly, but most people just ship their files with bogus default types that apache ships with |
| 21:11 | <Hixie_> | but unlike with mime types, where browsers just ignore them, if we don't provide a way around this, that long tail will either not be able to use es6, or they'll have terrible performance. |
| 21:16 | <Hixie_> | jorendorff: in any case, this is all moot relative to the point i was talking about above. Even if you only have one module, how can you parse it like bz describes if it has to go through the ES6 module loader pipeline? |
| 21:17 | <jorendorff> | Hixie_: oh, it can speculatively parse all it wants to, just like HTML, right? |
| 21:17 | <Hixie_> | not if the browser doesn't see a single byte until the translate hook has returned |
| 21:17 | <Hixie_> | which happens after the entirety of the fetch hook |
| 21:18 | <jorendorff> | Hixie_: sure, if you play by the rules |
| 21:18 | <Hixie_> | ... |
| 21:18 | <jorendorff> | Hixie_: you can either speculatively parse or not. if you do speculatively parse, you have to not get caught |
| 21:18 | <jorendorff> | like every other optimization |
| 21:18 | <Hixie_> | so basically anyone who uses an ES6 loader is going to see their performance drop off a cliff? |
| 21:19 | <Hixie_> | we should design this so that using this correctly and fully is completely compatible with the optimisations |
| 21:19 | <Hixie_> | e.g. using streams instead of promises |
| 21:19 | <Hixie_> | imho |
| 21:20 | <jorendorff> | that's fine with me |
| 21:25 | <Hixie_> | jorendorff: you or bz or something should tell es-discuss :-) |
| 21:26 | <Hixie_> | i don't think people there are listening to me anymore |
| 21:46 | <caitp> | is it actually specified anywhere that SVG elements in an HTML document are supposed to have the `innerHTML` property? |
| 21:46 | <caitp> | the interaction between svg and html is just super confusing and hard to keep track of |
| 21:47 | <Hixie_> | it's currently defined that they should not, iirc |
| 21:48 | <caitp> | well, blink and gecko both expose it, and applications depend on it |
| 21:48 | <caitp> | but unfortunately those applications break in safari |
| 22:12 | <Hixie_> | are there any methods that take 'any...' arguments and then later try to coerce that to a dictionary? |
| 22:15 | <Ms2ger> | None that I know of |
| 22:18 | <Domenic> | I am afraid to ask why you're asking that... |
| 22:39 | <Hixie_> | Domenic: to spec getContext()'s {alpha:false} feature which multiple browsers implement |
| 22:43 | <Hixie_> | Domenic: btw, is it intended that the WebIDL conversion algorithm for jsval to dict doesn't throw if passed a promise? |
| 22:43 | <Hixie_> | (it does if passed a Date or RegExp) |
| 22:44 | <Domenic> | Hixie_: It's not intentional in that I don't believe anyone made a conscious decision |
| 22:44 | <Hixie_> | k |
| 22:45 | <Domenic> | IMO Date and RegExp are the weird ones out ... Map, Set, Array, Error, etc. all don't throw either. |
| 22:47 | <Hixie_> | true |
| 22:48 | <Hixie_> | Array and Error are less magical than Map, Set, and Promise though |
| 22:48 | <Hixie_> | and those are the only old ones there |
| 22:48 | <Domenic> | Eh, unclear. Array has some fun magic going on. |
| 22:50 | <Domenic> | argggh why did the ES spec suddenly get big enough to hang Firefox like the HTML spec |
| 22:52 | <Hixie_> | yay, it's not just me! |
| 22:54 | <caitp> | jorendorff's mirror, or some exotic secret future version with es7 stuff? |
| 22:54 | <jorendorff> | Domenic: ? |
| 22:55 | <Domenic> | jorendorff: filing an issue now |
| 22:56 | <Domenic> | https://github.com/jorendorff/es-spec-html/issues/78 |
| 22:57 | <jorendorff> | hmm. current release ffox or Nightly? |
| 22:58 | jorendorff | hadn't noticed anything |
| 22:58 | <jorendorff> | thanks for the bug |
| 22:58 | <jorendorff> | probably it's just time to split it across some pages |
| 22:59 | <Hixie_> | nooooooo |
| 23:00 | <jorendorff> | orrrrrr, we could, like, fix whatever is busted in firefox |
| 23:00 | <jorendorff> | i don't think the ES spec got a lot bigger all of a sudden. |
| 23:05 | <Domenic> | I was thinking maybe the HTML got bigger even if the contents didn't |
| 23:05 | <Domenic> | jorendorff: I am only 50% sure I tested on Nightly, but definitely on current (and probably last?) stable. |
| 23:40 | <terinjokes> | i wish shims were more realible |
