| 00:10 | <gsnedders> | jgraham: Can you touch <https://github.com/html5lib/html5lib-python/issues/85>; yet? :P |
| 00:10 | <gsnedders> | jgraham: Or still too busy with this whole new-country thing? |
| 00:19 | <zewt> | wanted: firefox plugin to keep pages from being able to break password autocomplete |
| 00:59 | <Domenic_> | zewt: I use LastPass for that. |
| 01:03 | <MikeSmith> | IE supports IDNA2008 + TR46? http://lists.w3.org/Archives/Public/www-tag/2013Aug/0083.html |
| 01:20 | <JonathanNeal> | Does a rejection in a promise's .then() impact the next chained promise's .then()? |
| 01:37 | <Domenic_> | JonathanNeal: it is not entirely clear what you're asking, but I believe the answer is yes. |
| 01:38 | <Domenic_> | JonathanNeal: you can test all this behavior with any existing Promises/A+ library, by the way. |
| 01:38 | <Domenic_> | JonathanNeal: https://github.com/promises-aplus/promises-spec/blob/master/implementations.md |
| 01:39 | <MikeSmith> | kochi1: IME API is listed under http://www.w3.org/TR/#tr_Javascript_APIs |
| 01:40 | <kochi1> | MikeSmith: aha, Okay! I coundn't find it... |
| 01:41 | <MikeSmith> | yeah the usability of that page is not so great |
| 01:42 | <kochi1> | yeah, someone has to implement incremental search on the page! |
| 01:43 | <kochi1> | btw, "IME" matches many substrings like "Time" "Primer" "Multimedia" and so on |
| 01:43 | <kochi1> | we should choose more outstanding spec name :) |
| 01:44 | <MikeSmith> | ah yeah |
| 01:44 | <kochi1> | 0000IME API |
| 01:44 | <MikeSmith> | heh |
| 01:45 | <kochi1> | seriously, someone may argue "IME" is a microsoft term, and more generally it could be "input method" or "text input" |
| 01:46 | <kochi1> | e.g. OSX has a class NSTextInputClient |
| 01:47 | <kochi1> | MikeSmith: what is the moderation style in the meeting next week? |
| 01:50 | <MikeSmith> | kochi1: relatively informal |
| 01:50 | <MikeSmith> | I got your e-mail message |
| 01:50 | <MikeSmith> | you don't really need to be concerned much about the format |
| 01:51 | <MikeSmith> | 15 minutes is a pretty short time |
| 01:51 | <kochi1> | MikeSmith: ok, thanks. I'll prepare some 10-minute material and take some time for questions. |
| 01:51 | <MikeSmith> | I think it's sufficient to first just explain what problems we are trying to solve with the IM API |
| 01:51 | <MikeSmith> | kochi1: yup |
| 01:53 | <kochi1> | MikeSmith: agreed, thanks for the suggestion. |
| 09:16 | <jgraham> | TabAtkins: You know promises would make more sense if a) flatMap wasn't a silly word and b) flatMap wasn't the thing that doesn't flatten |
| 09:16 | <jgraham> | gsnedders: I'm not sure why you think that error messages shouldn't, in principle, be localisable |
| 09:17 | <jgraham> | But I also think that no one has ever shown any inclination to use the feature |
| 09:38 | <jgraham> | Hmm, I think if I am considering using metaclasses to solve a problem, I might be overengineering things |
| 09:38 | <Ms2ger> | https://twitter.com/davidbaron/status/369610549799026688 |
| 10:37 | <annevk> | Hmm, promises, IDNA, or encodings |
| 10:44 | <Ms2ger> | annevk, I assume you saw the rust encoding issue? |
| 10:47 | <annevk> | Ms2ger, no |
| 10:48 | <annevk> | Ms2ger: pointer? |
| 10:48 | <Ms2ger> | https://github.com/lifthrasiir/rust-encoding/issues/3 |
| 10:53 | <annevk> | Ms2ger: doesn't really look like a Rust issue |
| 10:54 | <Ms2ger> | Yeah, more like "implementation issue", I guess |
| 10:54 | <annevk> | Ms2ger: although maybe we should change the way encoder and decoder errors are written in the specification. At least this implementation seems suboptimal... |
| 10:54 | <annevk> | Ms2ger: "spec is hard" |
| 10:54 | <annevk> | lets do something simpler that comes close |
| 10:54 | <Ms2ger> | Encoding are hard, let's go shopping |
| 10:56 | <annevk> | I should actually do that one of these days. At least get a bicycle... |
| 11:36 | <zcorpan> | jgraham: is there something wrong here? critic has fewer commits looks like. https://github.com/w3c/web-platform-tests/pull/262 https://critic.hoppipolla.co.uk/r/238 |
| 11:43 | <jgraham> | zcorpan: It looks like it stopped tracking because the author did some kind of merge/rebase that critic didn't like |
| 11:45 | <jgraham> | In theory it should be possible to recover from this, but it seems to be broken. I don't know if it's a fixed bug, or an unknown bug, or something else |
| 11:45 | <jgraham> | Maybe I should try upgrading |
| 12:06 | <zcorpan> | jgraham: ouch :-/ i'll leave it to you and belem to sort out |
| 12:39 | <annevk> | I need someone to proofread my emails... Especially when replying to people who made up this internet thing. http://lists.w3.org/Archives/Public/www-tag/2013Aug/0085.html :/ |
| 12:41 | <GPHemsley> | TabAtkins: I was merely investigating what would have been a cool hypothesis if you had been left-handed. But if you're right-handed, it's far from being cool. |
| 12:41 | <GPHemsley> | TabAtkins: Nothing sinister, I promise. |
| 12:45 | <annevk> | http://people.mozilla.com/~roc/animated-SVG-glyphs.webm hah |
| 12:45 | <annevk> | jgraham: ^^ |
| 12:49 | <jgraham> | annevk: I hate you |
| 12:49 | <annevk> | jgraham: don't shoot the messenger, man |
| 12:50 | <Lachy> | jgraham, why? what's not to like about animated glyphs? |
| 12:51 | <wilhelm> | What is this I don't even. |
| 12:51 | <Lachy> | I can't wait for someone to make an animated, wiggly version of Comic Sans. That'd be awesome. |
| 12:53 | <SimonSapin> | animated glyphs seem like such a terrible idea |
| 12:53 | <jgraham> | annevk: roc is too far away to shoot |
| 13:02 | <annevk> | Animated glyphs seem great. I hope we can soon create glyphs that mask videos... |
| 13:02 | <jgraham> | annevk: Now you can't use your "only the messanger" defence |
| 13:03 | <annevk> | jgraham: as long as you keep staring at your screen I think I'm safe |
| 13:10 | <GPHemsley> | Ms2ger: " Do you have a FfOS Mobile Device? " |
| 13:25 | <Ms2ger> | GPHemsley, "No" |
| 13:26 | <GPHemsley> | Ms2ger: Note it says FfOS, not FxOS |
| 13:29 | <SimonSapin> | annevk: "iso-8859-8 and iso-8859-8-i are distinct encoding names, because iso-8859-8 has influence on the layout direction." What does this mean? |
| 13:30 | <annevk> | SimonSapin: what do you think it means? |
| 13:31 | <SimonSapin> | Is this about CSS writing modes? I think there should be a link there |
| 13:32 | <SimonSapin> | or is it in HTML? |
| 13:33 | <annevk> | Once someone has defined it I might add a link I guess |
| 13:33 | <annevk> | There's a bug somewhere |
| 13:33 | <annevk> | It's mentioned in the source: https://www.w3.org/Bugs/Public/show_bug.cgi?id=19505 |
| 13:35 | <SimonSapin> | annevk: I think it’s bad to have valuable info hidden in HTML comments. <p class=note> or <p class=issue> would be better |
| 13:36 | <annevk> | Yeah, nowadays I usually do that |
| 14:09 | <gsnedders> | jgraham: I'd argue they're developer-facing messages, which others often don't localize. |
| 14:13 | <jgraham> | gsnedders: What kind of developers? |
| 14:14 | <jgraham> | gsnedders: If, for example, you wanted to write a python-based validator with html5lib as the parser layer, you might well want to localise these messages |
| 14:31 | <gsnedders> | jgraham: Does any validator do this or have any intension of every doing this? |
| 14:32 | <jgraham> | Localised error messages or use html5lib? |
| 14:33 | <gsnedders> | The former. |
| 14:34 | <jgraham> | http://www.w3.org/wiki/CssValidator/Localization |
| 14:46 | <annevk> | Hixie_: https://www.w3.org/Bugs/Public/show_bug.cgi?id=22496#c11 |
| 14:46 | <annevk> | Hixie_: and you were right, my bad :/ |
| 15:05 | <annevk> | Hixie_: might also want to review https://www.w3.org/Bugs/Public/show_bug.cgi?id=21976#c6 as it seems the single host concept might be obsolete soon |
| 15:07 | <MikeSmith> | win 27 |
| 15:12 | <annevk> | I wonder to what extent people have reviewed this HTML imports stuff... |
| 15:13 | <annevk> | Seems https://dvcs.w3.org/hg/webcomponents/raw-file/tip/spec/imports/index.html should just be folded into HTML |
| 15:18 | <Domenic_> | what is the use case of html imports again? :-/ |
| 15:20 | <annevk> | External templates, afaict |
| 15:20 | <Domenic_> | hmm. most people seem to do that via require('./myTemplate.html') at least in my circles. |
| 15:20 | <annevk> | Domenic_: it seems return in then()'s callbacks doesn't need to do anything either and can just call resolve() |
| 15:21 | <Domenic_> | annevk: yup. step 7 at https://github.com/promises-aplus/promises-spec#the-then-method |
| 15:21 | <annevk> | Domenic_: at which point the question is what kind of checks then() does |
| 15:21 | <Domenic_> | annevk: resolve(), you mean? |
| 15:22 | <annevk> | Domenic_: I thought we were going to spec the model where then() does unwrapping |
| 15:22 | <Domenic_> | annevk: yes, so what exactly resolve() does is different in that model, but the fact that return <-> resolve (and throw <-> reject) stays the same |
| 15:23 | <annevk> | Domenic_: afaict in that model resolve() just sets promise's value |
| 15:24 | <Domenic_> | annevk: oh I see, yeah you're right, it's then()s initial behavior that becomes the new question. |
| 15:28 | <annevk> | this model seems kinda weird, in that the promise has a value that's a promise... and then if you do .then() it needs to be chained all the way through |
| 15:29 | <Domenic_> | well, it's not the promise's value, it's the promise's resolution... but yes, the fact that value/reason/state are all now emergent makes it a bit trippy. |
| 15:29 | <Domenic_> | Mark and I are planning to nail down an exact algorithm later tonight-ish. The plan is to leave what "isPromiseLike" means exactly unspecified for now, to be argued about after we have a solid algorithm in place. Like http://esdiscuss.org/topic/promises-consensus-with-a-terminology but with more detail. |
| 15:32 | <annevk> | thanks |
| 15:51 | <annevk> | TabAtkins: I referenced the "earlier X" quite explicitly |
| 16:54 | <JonathanNeal> | Promises, Promises. |
| 16:54 | <Ms2ger> | Futures. |
| 16:55 | <JonathanNeal> | Is futures a more accurate term, just less aesthetic to me? |
| 17:06 | <Domenic_> | it is what some other languages use |
| 17:06 | <Domenic_> | mostly i hear java people using it |
| 17:31 | <Hixie_> | hsivonen: re https://www.w3.org/Bugs/Public/show_bug.cgi?id=12400 - what do you think is wrong in the specs, exactly? |
| 17:40 | <Hixie_> | i'm amused that all the people who wanted their changes done to the whatwg html spec 5+ years ago, like having cite allow people's names, having blockquote ambiguously allow citations to be inlined, etc, are now just asking the htmlwg to make the changes instead |
| 17:58 | <zcorpan> | TabAtkins: pls see https://www.w3.org/Bugs/Public/show_bug.cgi?id=22500 |
| 18:21 | <Hixie_> | if you want a constructor that takes arguments and does something special to them, and you have several of these, is NamedConstructor the preferred solution or are factory methods preferred? |
| 18:22 | <Hixie_> | e.g. something like (but obviously not actually this): new IntegerByAddition(1,2,3) => an Integer representing 6; new IntegerByConcatenation(1,2,3) => an Integer representing 123 |
| 18:22 | <Hixie_> | vs createIntegerByAddition(1,2,3) and createIntegerByConcatenation(1,2,3) |
| 18:26 | <jsbell> | I can think of examples of the latter, but not the former, assuming the resulting objects are indistinguishable |
| 18:31 | <jsbell> | FWIW, I only find two NamedConstructor examples in blink's IDLs, and they are one per interface (Option for HTMLOptionElement and Audio for HTMLAudioElement), not even multiple per interface let alone different names |
| 18:31 | <Hixie_> | there should also be one for Image |
| 18:31 | <Hixie_> | which has three |
| 18:31 | <Hixie_> | but they're all just Image |
| 18:31 | <Hixie_> | certainly right now i'm not aware of anyone doing the named constructor thing |
| 18:32 | <Hixie_> | but people hate factory methods |
| 18:32 | <Hixie_> | so i don't know which is preferred |
| 18:35 | <jsbell> | Yeah, looks like Blink's HTMLImageElement constructor has custom code rather than being expressed through IDL; may be crufty code that can be updated, dunno. (And why not Video, for symmetry?) |
| 18:38 | <jsbell> | The third alternative is new Integer({add: [1, 2, 3]}) vs. new Integer({concat: [1,2,3]}). With argument destructuring that might be considered more ES6-ish |
| 18:40 | <jsbell> | i.e. as an implementer in ES6 you can write |let {add, concat} = options;| and decide what to do from there; supposedly that will be the new hotness; since they're exclusive it's icky, though |
| 18:50 | <annevk> | Hixie_: Integer.create(...) is what ES uses these days |
| 18:53 | <TabAtkins> | jgraham: flatMap *does* flatten. It flattens *once* - when you return a promise from the callback, rather than making a promise for a promise, it flattens it. |
| 18:53 | <TabAtkins> | jgraham: The name comes from the fact that it's identical to a .map() method, except it adds the flattening at the end. |
| 18:54 | <TabAtkins> | jgraham: I think it's from Scala, and it's at least better than .bind() (or >>=). |
| 18:55 | <TabAtkins> | annevk: What do you mean by "Integer.create() is what ES uses"? |
| 18:56 | <TabAtkins> | Oh, didn't see the context. You meant when there are multiple constructors. |
| 18:56 | <hsivonen> | Hixie_: maybe there isn't anything wrong with the specs. Basically, if browsers don't support marking up combining marks separately, it should be an authoring error to do so |
| 18:56 | <hsivonen> | Hixie_: requiring each text node to be in NFC takes care of this, AFAICT |
| 19:24 | <annevk> | hsivonen: roc at one point touted this as a feature |
| 22:14 | <Hixie_> | jsbell: Audio used to be a non-element object, the Audio constructor was added to <audio> when we made <audio> and removed that object. |
| 22:14 | <Hixie_> | hsivonen: does NFC require no leading combining characters? |
| 22:15 | <Hixie_> | hsivonen: and do we require NFC? |
| 22:17 | <Hixie_> | hsivonen: (see in particular https://www.w3.org/Bugs/Public/show_bug.cgi?id=13502#c22) |
| 22:49 | <MikeSmith> | Hixie_: I'm pretty sure that neither the HTML spec nor any of its depedencies require NFC. That's why I changed the validator (a couple years or so ago) to emit warnings about NFC problems, instead of errors |
| 22:50 | <Hixie_> | yeah, that was my conclusion too |
| 22:52 | <MikeSmith> | I think http://www.w3.org/TR/charmod-norm/ was the only document that said anything about NFC, and HTML doesn't reference that |
| 22:57 | <Hixie_> | MikeSmith: yeah, there's even a comment in the source saying i should do something to allow isolated combining characters if we ever _do_ reference that |
| 22:59 | <MikeSmith> | Hixie_: ah oK |