| 01:20 | <SamB> | Hixie: I just got the "too slow?" note *on* the multi-page version ... |
| 01:20 | <SamB> | or so it seemed |
| 01:22 | <SamB> | Hixie: also isn't blink supposed to be non-defined at #the-blink-element rather than just #blink ? |
| 08:12 | <mathiasbynens> | https://code.google.com/p/v8/issues/detail?id=2218 + https://code.google.com/p/v8/issues/detail?id=2217 interoperability++ |
| 08:12 | <mathiasbynens> | took over 2 years |
| 08:43 | <annevk> | mathiasbynens: writing the spec, the tests, and the implementation is a shame though, but kudos! |
| 08:44 | <annevk> | mathiasbynens: are these methods in ES6 now? |
| 08:45 | <annevk> | looks like it |
| 08:45 | <annevk> | you can search for "blink" :-) |
| 08:53 | <annevk> | mathiasbynens: are you waiting with updating http://javascript.spec.whatwg.org/ until ES6 is published? |
| 08:57 | <mathiasbynens> | yeah, thanks to spec bug reports i filed two years ago |
| 08:57 | <mathiasbynens> | they are in ES6 now |
| 08:58 | <mathiasbynens> | re: waiting, i’m not sure. is https://people.mozilla.org/~jorendorff/es6-draft.html official enough to use as a reference in a spec? |
| 09:14 | <tobie> | mathiasbynens: that's the one we reference in specref: http://specref.jit.su/search-refs?q=es6 |
| 09:14 | <tobie> | http://specref.jit.su/bibrefs?refs=ECMASCRIPT |
| 09:19 | <annevk> | mathiasbynens: definitely for WHATWG |
| 09:20 | <annevk> | mathiasbynens: it's as close to Living Standard as you can get, and it's what implementers are targeting |
| 09:21 | <mathiasbynens> | ok, cool |
| 10:20 | <tobie> | annevk: are you using the enhancement tag to denote post v1 features in SW repo? |
| 10:20 | <annevk> | tobie: I guess |
| 10:20 | <annevk> | tobie: post Q3 features |
| 10:20 | <tobie> | just added a v2 milestone to do the same. :/ |
| 10:20 | <annevk> | tobie: that seems a bit clearer actually |
| 10:20 | <annevk> | tobie: can we rename labels? |
| 10:21 | <tobie> | yup. will do. |
| 10:21 | <tobie> | a preference for using a milestone, or a tag? |
| 10:29 | <annevk> | tobie: tag is what we did so far |
| 10:30 | <tobie> | annevk: oh, well. added a version 2 milestone (we can always replace by a tag instead) |
| 10:31 | <tobie> | currently triaging enhancements as pertaining to required enhancements for v1 or feature requests for the future |
| 10:34 | <tobie> | annevk, JakeA: status of https://github.com/slightlyoff/ServiceWorker/issues/224 |
| 10:34 | <tobie> | ? |
| 10:37 | <JakeA> | tobie: I'd rather we just did this in CSP, but I think we're going to end up sending some kind of header as slightlyoff says |
| 10:39 | <tobie> | JakeA: still something we need for v1, right? |
| 10:39 | <JakeA> | yep |
| 10:53 | <tobie> | annevk: added a version 1 milestone to all issues tagged with impacts MVP |
| 10:53 | <tobie> | annevk: you can look at unassigned issues here: https://github.com/slightlyoff/ServiceWorker/issues?q=is%3Aopen+is%3Aissue+no%3Amilestone |
| 11:17 | <zcorpan> | what's the reason we can't always autorotate like mobile safari? https://www.w3.org/Bugs/Public/show_bug.cgi?id=25508 |
| 11:19 | <annevk> | zcorpan: does it do that even for <img>? |
| 11:19 | annevk | thought that was not going to be compatible |
| 11:21 | <zcorpan> | i haven't tested but i see claims that it does |
| 11:28 | <zcorpan> | https://code.google.com/p/chromium/issues/detail?id=56845#c52 |
| 11:40 | <zcorpan> | https://raw.githubusercontent.com/recurser/exif-orientation-examples/master/Landscape_6.jpg looks like gecko/blink/webkit honor exif for the top-level img page but not for <img> |
| 11:47 | <zcorpan> | http://www.browserstack.com/screenshots/5864b36445508e27c01ccb333d17b979c27cdfb7 |
| 11:48 | <annevk> | Somewhat weird as both use <img> |
| 11:55 | <zcorpan> | who should i cc from apple/webkit? |
| 12:35 | <annevk> | zcorpan: hober? |
| 12:37 | <zcorpan> | thx |
| 13:39 | <tobie> | annevk: why isn't QuotaExceededError not a subclass of DOMException? |
| 13:40 | <tobie> | annevk: why isn't QuotaExceededError a subclass of DOMException? |
| 13:41 | <tobie> | (sorry for !!) |
| 13:42 | <tobie> | annevk: I thought things went something like QuotaExceededError < DOMException < Error |
| 13:47 | <zcorpan> | i've caught up with email now, except for critic |
| 13:48 | zcorpan | marks all as read |
| 13:52 | <annevk> | tobie: we haven't sorted exceptions |
| 14:08 | <Domenic> | tobie: there are no subclasses of DOMException; just DOMExceptions with different `name` properties |
| 14:36 | <annevk> | So yeah Jake, what about your data? http://jakearchibald.com/ |
| 14:36 | <annevk> | JakeA: ^^ |
| 14:38 | <JakeA> | annevk: haha, I was wondering when someone would point that out. It's on my todo list. I should go through the pain of it |
| 14:40 | <annevk> | I need to do it too, but it's going to be a lot of hassle I'm afraid. DreamHost also uses separate domains for MySQL so... |
| 14:41 | <JakeA> | I'm on Amazon so it might not be too bad, although I host multiple sites from that box so I'll have to work out how to do the SNI stuff |
| 14:42 | <annevk> | If you use SNI, how many browsers do you cut off? |
| 14:42 | <annevk> | Back in the day there were browsers without Host... |
| 14:47 | <JakeA> | IE < 7 |
| 14:47 | <JakeA> | Oh & any version of IE on XP |
| 14:47 | <annevk> | I think NN4 didn't do Host, but not really sure |
| 14:48 | <annevk> | JakeA: seems acceptable |
| 14:48 | <JakeA> | Oh, maybe Android 2 too, ugh. |
| 14:49 | <JakeA> | The default browser anyway. Opera/Firefox will be fine |
| 14:49 | <annevk> | Personally, don't really care about that |
| 14:49 | <JakeA> | Same, but I might if I was making a site for "the public" |
| 14:49 | <annevk> | The main problem with TLS is the maintenance |
| 14:50 | <annevk> | I already authenticated with my registrar somehow, it would be nice if they handled the setup... |
| 14:54 | <Domenic> | CloudFlare seems to have some pretty simple TLS setup |
| 14:54 | <Domenic> | Extremely user-friendly |
| 14:58 | <annevk> | I just noticed I can do it almost automatically with DreamHost too for free |
| 14:58 | <annevk> | Provided I don't want my own IP, which has this caveat: http://wiki.dreamhost.com/Secure_Hosting#SNI_Caveat |
| 14:59 | <annevk> | Oh wait, that's for self-signed. A signed certificate is USD 15 |
| 15:46 | <Ms2ger> | In other news: http/2 breaks the web |
| 15:49 | <caitp> | you say that like it's a tremendous and avoidable failure =) |
| 15:50 | <jgraham> | the web, or just twitter? |
| 15:50 | <Ms2ger> | At least twitter |
| 16:45 | <smaug____> | Hixie_: how long is a BroadcastChannel supposed to stay alive? |
| 16:46 | <smaug____> | like, does 'new BroadcastChannel('foobar').onmessage = function(){};' in a worker keep the BroadcastChannel and the relevant worker alive forever (assuming the relevant Window object stays alive) |
| 16:48 | <tobie> | Domenic: figured (re DOMExceptions). A little surprising, though. Do you know why that is the case? |
| 16:48 | <Domenic> | tobie: history i guess |
| 16:49 | <tobie> | C heritage? |
| 16:49 | <smaug____> | Hixie_: I guess so, given the note |
| 16:50 | <smaug____> | ...about garbage collecting |
| 16:53 | <Ms2ger> | tobie, used to be DOMException with .code == DOMException.SYNTAX_ERR, remember ;) |
| 16:53 | <tobie> | sigh |
| 16:55 | <tobie> | Ms2ger: I remember how we were hard coding all of those constants for nodetypes. |
| 16:56 | <tobie> | if element.nodeType == 3, etc. |
| 16:57 | <Ms2ger> | I don't think I ever had to do that, but then again I've mostly written tests |
| 16:58 | <tobie> | node type constants weren't implemented in IE < 7 iirc. Hence hard coding their value. |
| 16:59 | <tobie> | Which I believe is one of the main reasons why front-end devs prefer strings over int + const. |
| 17:01 | <tobie> | We just got used to the latter being unreliable across browsers. |
| 17:01 | <Ms2ger> | Yeah |
| 17:02 | <tobie> | it's a shame, though. const help with readability plus catching typos upfront. |
| 17:06 | <jgraham> | Maybe ES10 will be the one where we get strong typing and enums |
| 17:10 | <gsnedders> | jgraham: nah, that's ES4 |
| 17:29 | <annevk> | Even JavaScript itself uses strings for enums |
| 17:32 | <SamB> | annevk: it's not uncommon in C to #define strings used in that manner |
| 17:32 | <SamB> | for typo-avoidance alone |
| 17:45 | <Domenic> | typo avoidance is not achieved by the int-enums |
| 17:45 | <Ms2ger> | Well, not in JS |
| 17:45 | <Domenic> | right |
| 17:45 | <tobie> | Domenic: well, unless they're globals. |
| 17:45 | <Domenic> | O_O |
| 17:46 | <Domenic> | with proxies I think we will start seeing experiments |
| 17:46 | <Domenic> | with things that throw on undefined properties |
| 17:46 | <Domenic> | there will be chicken-egg problems with perf (nobody uses them because they're slow; nobody optimizes them because nobody uses them) |
| 17:47 | <Domenic> | but if we can overcome that, especially for simple proxies like "pass-through everything, except throw for undefined gets", then I can see a standards-track typo-preventing enum |
| 17:47 | <tobie> | Yeah, that's what I meant by globals (not properties of something). |
| 17:48 | <tobie> | as in .code == DOM_EXCEPTION_SYNTAX_ERR |
| 17:49 | <SamB> | Domenic: I wasn't suggesting that it was |
| 17:50 | <SamB> | Domenic: I was merely mentioning a thing that can be helpful in dealing with typos even if you do want to use string values for this |
| 17:51 | <Domenic> | SamB: sorry, I was replying to tobie |
| 17:53 | <tobie> | Domenic: agree the normal, hanging off an object enum won't help you pick up typos. |
| 20:34 | <Hixie_> | SamB: file bugs |
| 20:34 | <Hixie_> | smaug____: yeah |
| 20:36 | <Hixie_> | jorendorff: your day any better today? :-) |
| 20:37 | <jorendorff> | phew |
| 20:37 | <jorendorff> | Hixie_: it's honestly not great... today could work, but tomorrow would be better if it would be OK |
| 20:38 | <Hixie_> | jorendorff: happy to get your help whenever it's available. I'll try to forge on without it in the meantime. :-) |
| 20:39 | <jorendorff> | Hixie_: Please ping me tomorrow as soon as possible. I'm sorry about today. Fires. |
| 20:39 | <jorendorff> | Fires and interns. |
| 20:39 | <Hixie_> | jorendorff: good luck with your fires! |
| 20:40 | <crankharder> | wondering if this is a good place to ask about the Content-Security-Policy-Report-Only header, seen here: http://www.w3.org/TR/CSP/#content-security-policy-report-only-header-field |
| 20:40 | <jorendorff> | thank you |
| 20:43 | <Hixie_> | abarth: ping ^ |
| 21:37 | <crankharder> | how can something violate this Content-Security-Policy? "default-src *" |
| 21:37 | <crankharder> | getting lots of warnings with just that set |
| 22:33 | <astearns> | crankharder: default-src: * will disable inline script and style, and eval |
| 23:30 | <Hixie_> | Domenic: you around? |
| 23:31 | <Hixie_> | Domenic: what is PromiseBuiltinCapability() ? |
| 23:36 | Hixie_ | stares at ProceedToTranslate(loader, load, p) Abstract Operation in confusion |
| 23:39 | Hixie_ | wonders what PromiseCatch() does and what returning PromiseCatch() is supposed to mean |
| 23:41 | <jtcranmer> | annevk: I've got a tough UTR#46 question for you |
| 23:42 | <jtcranmer> | should I say faß.de and xn--fa-hia.de are "equivalent" domains? |
| 23:42 | <Hixie_> | how does PromiseThen(PromiseOf(undefined), F) differ from just calling F? |
| 23:48 | <Hixie_> | man, i think the people who complain that the HTML spec is "too imperative" really need to read the ES6 spec to get some perspective. |
| 23:56 | <Hixie_> | kittens, it's hard to follow promise-based code |
| 23:56 | <Hixie_> | how do you even draw a flow chart for it |
| 23:56 | <Hixie_> | do you draw the line from the place that does the call, or from the place that's going to trigger the promise? |
