| 09:19 | <Ms2ger> | "If we don't learn how to produce books from web content, there will soon be no books." |
| 09:21 | jgraham | blinks |
| 09:24 | <Ms2ger> | And you're right, getting tests into our automation has made me find bugs |
| 10:25 | <Ms2ger> | Hear, hear |
| 10:25 | <Ms2ger> | http://www.neowin.net/images/uploaded/screen%20109.jpg |
| 12:23 | <hsivonen> | Ms2ger: interesting choice to have rv: but no slash after "Gecko" |
| 12:23 | <hsivonen> | Ms2ger: still missing the innovation of syncing engine numbers with product numbers |
| 12:24 | <hsivonen> | Trident/7.0 rather than Trident/11.0 |
| 12:24 | <Ms2ger> | I would have gone with just copying Firefox's entirely, but they didn't ask me :) |
| 12:30 | <annevk> | still lots of redundant bytes |
| 12:37 | <hsivonen> | also interesting that browser-exposed .Net isn't dead enough to stop being exposed there |
| 12:37 | <hsivonen> | I'm glad we never put Java Web Start version info in the UA string |
| 12:40 | <SimonSapin> | hsivonen: can plugins have any influence on the UA string? |
| 12:42 | <MikeSmith> | hsivonen: speaking of UA strings you may have noticed I checked in a change to the validator sources to make the validator UA string configurable |
| 12:42 | <MikeSmith> | reason being, for all the W3C validator services we want to include a URL in the UA string |
| 12:43 | <MikeSmith> | due to us and MIT getting abuse reports for IP addresses that are the validator hosts |
| 12:44 | <MikeSmith> | so the W3C service now sends the string "Validator.nu/LV http://validator.w3.org/services"; |
| 12:44 | <zcorpan> | MikeSmith: no version number??????? |
| 12:45 | <MikeSmith> | heh |
| 12:45 | <MikeSmith> | "LV" is the version number |
| 12:45 | <MikeSmith> | "Modest traffic from these services does not consitute abuse against your website. Third parties using this service to review content you make publicly available is not substantially different from browsing your site. Web designers frequently evaluate techniques of other websites as a means to learn." |
| 12:45 | <jgraham> | "quite a lot like Gecko" |
| 12:45 | <jgraham> | Would actually be true for the validator :) |
| 12:46 | <zcorpan> | MikeSmith: ah, so 55 |
| 12:46 | <MikeSmith> | jgraham: yeah |
| 12:46 | <annevk> | does User-Agent: http://validator.w3.org/services violate HTTP? |
| 12:46 | <MikeSmith> | dunno |
| 12:46 | <MikeSmith> | do I have to put it in some delimiters? |
| 12:46 | <MikeSmith> | parens or something? |
| 12:47 | <jgraham> | It violates the expectations of the twitter generation who expect URLs to be < 10 characters |
| 12:47 | <annevk> | I would suggest to not care personally |
| 12:47 | <MikeSmith> | OK, that's easy |
| 12:47 | <MikeSmith> | because I actually don't care about violating HTTP |
| 12:47 | <jgraham> | Yeah, seriously |
| 12:48 | <jgraham> | If we all started caring about violationg HTTP now |
| 12:48 | <MikeSmith> | rather, given a good opportunity to violate HTTP, I'm going to violate, with pleasure |
| 12:48 | <jgraham> | We'd have to set up our own internet |
| 12:48 | <jgraham> | and it would only be us and Julian and Roy |
| 12:48 | <zcorpan> | http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.8 seems to me a url doesn't violate the production |
| 12:48 | <MikeSmith> | ok |
| 12:48 | <jgraham> | And how fun would that be? |
| 12:49 | <zcorpan> | "They MUST NOT be used for advertising or other non-essential information" |
| 12:49 | <jgraham> | So, does the additional character in "consume a character reference" actually do anything? It seems to me that that "Anything else" clause will have the same behaviour |
| 12:50 | <zcorpan> | MikeSmith: that seems like an obvious choice if you want to violate http |
| 12:50 | <annevk> | zcorpan: it does https://tools.ietf.org/html/rfc2616#page-17 |
| 12:50 | <annevk> | e.g. : is a separator and cannot be part of a token |
| 12:50 | <annevk> | but the requirements are nonsense anyway |
| 12:50 | <zcorpan> | annevk: ah |
| 12:51 | <zcorpan> | so everyone's User-Agent string are invalid because they contain colons and parens? |
| 12:53 | <annevk> | zcorpan: it's roughly token ["/" token] ["(" .... ")] |
| 12:54 | <MikeSmith> | jgraham: hey btw can you set me up with critic for https://github.com/w3c/web-platform-tests/pull/21 ? |
| 12:54 | <annevk> | zcorpan: might be easier to read https://tools.ietf.org/html/rfc2616#section-14.43 and search for "product =" and "comment =" |
| 12:54 | <MikeSmith> | jgraham: or remind me what I need to do |
| 12:55 | <Ms2ger> | MikeSmith, can't you log in with your github credentials? |
| 12:56 | <jgraham> | MikeSmith: Done |
| 12:57 | <zcorpan> | ok so url in parens is allowed |
| 12:57 | <zcorpan> | but advertising MikeSmith's hotline for buying viagra is not |
| 12:59 | <hsivonen> | SimonSapin: Not in non-IE browsers, AFAIK. Though prior to Firefox 4 (IIRC), extension could add cruft to the Firefox UA string |
| 13:00 | <hsivonen> | *extensions |
| 13:00 | <SimonSapin> | hsivonen: and Firefox does not allow this anymore? |
| 13:00 | <hsivonen> | SimonSapin: not anymore |
| 13:00 | <hsivonen> | SimonSapin: though Ubuntu puts "Ubuntu" there somehow anyway |
| 13:00 | <SimonSapin> | I see, thanks |
| 13:01 | <hsivonen> | perhaps by patching the source of Firefox itself |
| 13:01 | <jgraham> | MikeSmith: I can't add you as a reviewer unless you log into critic though :) |
| 13:01 | <jgraham> | Or rather unless you create an account, which means logging in at least once |
| 13:08 | <jgraham> | On the subject of those tests |
| 13:09 | <jgraham> | I note that they overlap a lot (entirely?) with Aryeh's reflection tests |
| 13:09 | <jgraham> | http://critic.hoppipolla.co.uk/r/5 |
| 13:10 | <MikeSmith> | jgraham: OK, I'm logged in |
| 13:12 | <Ms2ger> | Want me to submit Mozilla's reflection tests too? :) |
| 13:13 | <MikeSmith> | jgraham: so Yeah I thought about that |
| 13:13 | <MikeSmith> | but as far as I know, we don't have any policy for redundant tests |
| 13:15 | <MikeSmith> | and it would seem to be good to not discourage contributors by saying, Thanks for all your work but sorry we already have test cases that cover the same thing you're submitting |
| 13:18 | <jgraham> | Yeah, I don't know either |
| 13:19 | <jgraham> | So on the one hand I agree that tell ing people "thanks but no thanks" is bad |
| 13:19 | <jgraham> | On the other hand it does create a maintainance burden to have *totally* redundant tests |
| 13:19 | <jgraham> | I would of course err on the side of having redundant tests rather than inadequate tests though |
| 13:21 | <jgraham> | Ms2ger: Did bz write them? :p |
| 13:21 | <Ms2ger> | mounir |
| 13:23 | <jgraham> | Now I have to think of a serious reply :( |
| 13:23 | <jgraham> | I think the answer is "yes, if you want to convert them to testharness format and don't think they are already covered by existing tests" |
| 13:24 | <Ms2ger> | My issue is that Aryeh's give me a headache, so I'm not sure if they cover everything |
| 13:25 | <mounir> | Ms2ger: are you pushing my tests somewhere? |
| 13:25 | <jgraham> | Yeah, there is a certain amount of headache material there |
| 13:25 | <Ms2ger> | I've been thinking about it |
| 13:26 | <jgraham> | But they are quite good! |
| 13:26 | <jgraham> | Just needs someone to do an epic review job |
| 13:26 | <Ms2ger> | He's with Mozilla now, I can't do it |
| 13:27 | <jgraham> | Is he? I thought he was studying? And I thought he wrote the tests at Google. So I think you can :) |
| 13:27 | <Ms2ger> | Dammit |
| 13:28 | <hsivonen> | I wonder how much effort I should put into arguing about the fallback encoding for the Welsh locale |
| 13:28 | <Ms2ger> | Er, we're both studying, I can't do it :) |
| 13:28 | <jgraham> | Ms2ger: You seem to mostly be on IRC. He seems to be mostly not on IRC :p |
| 13:28 | <jgraham> | hsivonen: I think you may already have put in too much effort |
| 13:52 | <annevk> | FWIW, https://github.com/slightlyoff/NavigationController/ is now public |
| 13:52 | <hsivonen> | annevk: hooray |
| 13:52 | <hsivonen> | why is the Web and TV mailing list Member-confidential? |
| 13:52 | <jgraham> | Took a long time |
| 13:53 | <hsivonen> | I wonder what Opera is telling their TV customers about the Presto to WebKit move |
| 13:54 | <zewt> | a "web and TV" mailing list sure sounds like a place to discuss DRM, heh |
| 13:55 | <zewt> | which, since I'm a bit cynical, I'd suspect as a reason to keep it private (keep the public out = fewer objections) |
| 13:56 | <hsivonen> | so weird that some entities communicate with Liaison Letters instead of just filing bugs |
| 13:56 | <zewt> | sounds like the same sort of people who send email to lists with a company name in the from line and signature, with no name |
| 13:58 | Ms2ger | wonders who sent Liaison Letters |
| 13:58 | <zewt> | puffer fish |
| 13:58 | <zewt> | heh |
| 13:58 | <hsivonen> | sorry liaison statement |
| 13:58 | <hsivonen> | not letter |
| 13:59 | <hsivonen> | Ms2ger: the existence of the liaison statement is disclosed at https://www.w3.org/Bugs/Public/show_bug.cgi?id=21301 |
| 13:59 | <Ms2ger> | I see |
| 14:00 | <zewt> | reminds me of that guy (can't remember who--probably for the best) that was all "if you do this i will object", and didn't argue any further, as if he had total veto power and had no obligation to make technical arguments |
| 14:01 | <Ms2ger> | Weird Glenn? |
| 14:02 | <zewt> | not sure (and I guess I'd sooner not stick some random guy's name in a log somewhere for something he did a year ago, heh) |
| 14:02 | <zewt> | annevk: sadface @ window.event |
| 14:04 | <annevk> | yes |
| 14:04 | <Ms2ger> | Yay, webkit |
| 14:04 | <zewt> | annevk: i wonder if it can be restricted to events fired on DOM nodes, or on DOM nodes inside a document, or something like that |
| 14:05 | <zewt> | (maybe it already is; not very familiar with it) |
| 14:05 | <annevk> | I recommend advocating such a limitation to WebKit / IE people |
| 14:05 | <annevk> | I'm planning to define whatever they implement as just doing it slightly different seems like asking for more trouble going forward |
| 14:07 | <zewt> | i'll think more about how much it matters |
| 14:12 | <MikeSmith> | jgraham: thanks for the critic setup |
| 14:13 | <zewt> | heh, found a broken interop on the very first thing i tried with it (synthesized events don't show up in window.event in IE; they do in webkit) |
| 14:13 | <zewt> | (ie9) |
| 14:18 | Ms2ger | rolls his eyes |
| 14:24 | <hsivonen> | what's window.event? |
| 14:24 | <hsivonen> | some legacy Netscape event thing? |
| 14:25 | <Ms2ger> | IE |
| 14:25 | <jgraham> | Legacy IE event thing that WebKit supports |
| 14:25 | <hsivonen> | oh |
| 14:25 | <jgraham> | Creates an "event" variable in the global scope when handling an event |
| 14:26 | <sangwhan> | It's also a invitation for enterprising people to try to poll (and fail) instead of listen for events |
| 14:27 | <jgraham> | Wait, what? How would that work? |
| 14:28 | <jgraham> | Does it have a lifetime longer than the lifetime of the event handler? |
| 14:29 | <zcorpan> | Ms2ger: do you have an opinion on https://www.w3.org/Bugs/Public/show_bug.cgi?id=19388 ? |
| 14:30 | <Ms2ger`> | zcorpan, I don't |
| 14:30 | <zcorpan> | presto supports window.event as well, but maybe we made it undetectable (like document.all) |
| 14:31 | <sangwhan> | jgraham: I recall it being leaked globally if you time the access correctly - this might have been fixed though |
| 14:31 | <zewt> | it's only set synchronously, so the only theoretical danger is that it lets people access events from code that otherwise wouldn't have access to the variable (which is just a code scoping "who might be messing with my data" thing) |
| 14:32 | <zewt> | and that only matters for things that aren't right there in the document (where anyone can get to it anyway, with a capture listener) |
| 14:34 | <zewt> | so while random code down the stack having access to my otherwise-private event object in an XHR onload handler is lame, i guess it doesn't really matter much |
| 14:35 | <gsnedder1> | zcorpan: Became undetectable along with attachEvent |
| 14:36 | <sangwhan> | Nevertheless is a big wtf for proper scoping, wonder why so many browsers implemented it |
| 14:37 | sangwhan | scrummages through Presto history for clues |
| 14:37 | <zewt> | that's something we wonder often in here, heh |
| 14:38 | <sangwhan> | Not sure who's the best person to ask, but is there any reason why WebKit *only* implements ChildNode.remove() and none of it's friends? |
| 14:38 | <Ms2ger`> | Because it's easy, I guess |
| 14:38 | <Ms2ger`> | I've got a patch to add it to Gecko too |
| 14:39 | <zewt> | it's also one of the more convenient helpers--if I never have to type elem.parentNode.removeChild(elem) again, I'll be happy |
| 14:40 | <sangwhan> | If this starts picking up on live content, wonder what we're gonna do about Presto |
| 14:40 | <zewt> | still sad that we didn't fix the confusing "before" and "after" method names |
| 14:40 | <zcorpan> | seems webkit doesn't have attachEvent |
| 14:40 | <zcorpan> | and window.event is detectable |
| 14:40 | sangwhan | already experienced one instance on a *very* major website which eventually got "fixed" with the JQ equivalent |
| 14:41 | <miketaylr> | sangwhan: paul_irish might know, i seem to remember him saying something about some conflicts they ran into with the other stuff? |
| 14:41 | miketaylr | could be making this up |
| 14:42 | <sangwhan> | before()/after() is confusing, at first look I thought they were traverse helpers |
| 14:42 | <zewt> | yeah there's no verb, so they look like queries of some kind |
| 14:43 | <sangwhan> | miketaylr: about why WebKit only implements remove()? |
| 14:43 | <miketaylr> | sangwhan: yeah |
| 14:43 | <miketaylr> | but i bet he's sleeping right now |
| 14:43 | sangwhan | should start using twitter more |
| 14:44 | sangwhan | feels like a caveman in the 21st century. haven't tweeted in more than a year. |
| 14:44 | Ms2ger` | has never tweeted |
| 14:44 | <Ms2ger`> | Are you calling me a caveman?! |
| 14:44 | <zewt> | i think the real verb is "twitted" |
| 14:45 | <miketaylr> | i think cavemen would probably just tweet "Hungry." |
| 14:45 | <sangwhan> | Ms2ger`: You don't enjoy the simple life?! |
| 14:45 | <miketaylr> | (which is like 90% of all tweets anyways) |
| 14:45 | <zewt> | that's more content than most of twitter |
| 14:46 | <Ms2ger`> | miketaylr, "eating" |
| 14:46 | <miketaylr> | "Sabertooth tiger #YOLO" |
| 14:46 | <zewt> | twitter really needs to stop pretending people use it with SMS, and raise the character limit to something that allows expressing thoughts |
| 14:46 | <zewt> | it's particularly silly that it has the same character limit for japanese and english, even though japanese is far more dense per-character--so they can actually communicate |
| 14:47 | <sangwhan> | Yes, increase the character limit. Just in case they didn't have enough server load :P |
| 14:47 | <jgraham> | You mean you don't love the "everyone shouting at each other in a crowded bar" atmosphere it creates? |
| 14:47 | <miketaylr> | i've heard that VIPs get a higher char limit |
| 14:47 | <miketaylr> | like 255 |
| 14:47 | <zewt> | jgraham: everyone shouting at each other with their mouth full of food |
| 14:48 | <sangwhan> | zewt: Chinese is possibly the most efficient in character count, they don't have verb conjugation or connecting words |
| 14:48 | <sangwhan> | Japanese is surprisingly inefficient |
| 14:48 | sangwhan | hopes not to get stabbed in a Tokyo back alley for saying that in a logged chat |
| 14:49 | <zewt> | japanese is only inefficient in excessively-polite mode |
| 14:49 | <sangwhan> | LOL vs wwwwwwwwwwwwwwwwwwwwww |
| 14:49 | <sangwhan> | I rest my case |
| 14:49 | <zewt> | heh |
| 14:51 | <hsivonen> | looks like Philip TAYLOR (webmaster) is still active on www-validator even though he left the HTML WG |
| 14:52 | <zcorpan> | hsivonen: is www-validator a nice list or a noisy list? |
| 14:52 | <sangwhan> | Ah, a clue: http://my.opera.com/sitepatching/blog/2012/02/09/window-event-attachevent-detachevent-script-onreadystatechange |
| 14:52 | <hsivonen> | zcorpan: I don't read actively. I suspect the latter. |
| 14:52 | <zcorpan> | ok |
| 14:56 | <MikeSmith> | hsivonen: yeah seems so |
| 14:59 | <hsivonen> | MikeSmith: I kinda agree with Jukka that telling people to change the iso-8859-1 label to windows-1252 isn't particularly useful |
| 15:01 | <zewt> | really the world should just stop pretending they're different things, and add the win-1252 additions into 8859-1, heh |
| 15:02 | <zcorpan> | MikeSmith: hsivonen: IO Error: <data:text/html,%3cmeta%20charset=iso-8859-1%3e> Code: 1/PERCENT_ENCODING_SHOULD_BE_UPPERCASE in PATH: Percent-escape sequences should use uppercase. |
| 15:03 | <zcorpan> | now that's annoying after having looked up the ASCII table for "<" and ">" :-P |
| 15:03 | <MikeSmith> | hsivonen: I do too actually |
| 15:04 | <MikeSmith> | but that's not what the specs say currently |
| 15:04 | <MikeSmith> | HTML and Encoding |
| 15:04 | <MikeSmith> | so I suppose we need to chat with annevk and Hixie_ about that |
| 15:05 | <zcorpan> | Encoding says that anything but "utf-8" is invalid, so you could have the same message for both iso-8859-1 and windows-1252 |
| 15:05 | <MikeSmith> | zcorpan: we're mostly just passing that message on as-is from the Jena URI library |
| 15:05 | <MikeSmith> | zcorpan: yeah but the HTML spec doesn't exactly say that actualy |
| 15:06 | <zewt> | is this the "normalized encoding label" thing? |
| 15:06 | <MikeSmith> | my read on the HTML spec is that it says any "name" from the Encoding spec is valid |
| 15:10 | <zcorpan> | MikeSmith: "Authors should use UTF-8. Conformance checkers may advise authors against using legacy encodings." is only should and may, but gives you the option to give an error message for anything but "utf-8" |
| 15:10 | <MikeSmith> | yeah |
| 15:11 | <MikeSmith> | I don't know how hsivonen feels about error'ing on anything non-UTF8 |
| 15:11 | <annevk> | MikeSmith: the Encoding Standard gives the long term view |
| 15:11 | <MikeSmith> | ok |
| 15:11 | <annevk> | MikeSmith: I think I'd be fine with the validator not yet flagging all non utf-8 cases |
| 15:12 | <annevk> | MikeSmith: I think you want to flag cases where no encoding is declared and where the encoding declared is not a valid label per the Encoding Standard or maps to the replaced encoding |
| 15:12 | <MikeSmith> | yeah |
| 15:12 | <MikeSmith> | those are clearly errors |
| 15:13 | <zcorpan> | as far as usefulness goes, at this time i think it would be good to warn for anything but "utf-8" and not whine at all about non-preferred names (except for variants of utf-8) |
| 15:13 | <annevk> | MikeSmith: In addition you might want to flag certain crazy encodings, such as iso-2022-* and hz-gb-2312 |
| 15:13 | <annevk> | yeah, the "preferredness" stuff is noise |
| 15:13 | <MikeSmith> | for the ones that are non-UTF8 but that have names in the Encoding standard, we could Warn. And once we get the filtering feature deployed, people will have a lot more control about suppressing warnings that annoy them (or errors too, for that matter) |
| 15:16 | <zcorpan> | MikeSmith: does the filter feature log what people filter? that might be a useful communication channel for things we should change in the spec |
| 15:23 | <MikeSmith> | zcorpan: no it doesn't because the filtering is all client-side, and persisted only on the client side (using localStorage) |
| 15:24 | <zcorpan> | MikeSmith: do you agree that it could be a useful communication channel? |
| 15:24 | <MikeSmith> | yeah definitely |
| 15:25 | <MikeSmith> | there may be some way we could allow people to opt-in to sharing it back |
| 15:28 | <hsivonen> | zcorpan: I was thinking of adding a warning for non-UTF-8 |
| 15:29 | <hsivonen> | zcorpan: but still, it seems bogus to treat iso-8859-1, TIS-620 and such as errors if windows-1252 is only a warning |
| 15:29 | <zcorpan> | hsivonen: yeah. http://krijnhoetmer.nl/irc-logs/whatwg/20130325#l-920 |
| 15:30 | <hsivonen> | are all the non-preferred names supported in all browser that still matter? |
| 15:33 | <annevk> | not quite yet, WebKit does not support some of the iso-8859- labels which gives them compat issues |
| 15:37 | <zcorpan> | annevk: is there a list? http://wiki.whatwg.org/wiki/Web_Encodings doesn't include nice tables for safari and chrome |
| 15:38 | <annevk> | nah, I just know based on bug reports |
| 15:38 | <annevk> | I didn't document that other than by including the labels in the standard |
| 15:39 | <annevk> | bug report is mentioned in the commit message I think |
| 16:02 | <dglazkov> | good morning, Whatwg! |
| 17:20 | <MikeSmith> | hsivonen: btw, the W3C validator.nu backends are doing about 2.5 validations per second these days. With no noticeable performance problems, as far as I can see |
| 17:21 | <MikeSmith> | currently 4 hosts doing about 2.5 per second each, please one doing about .9 or so per second |
| 17:22 | <MikeSmith> | so a total of about 8.4 validations going on each second |
| 17:43 | <paul_irish> | sangwhan: no idea on why webkit only implements remove() and not its bretheren yet. http://trac.webkit.org/changeset/129400 i imagine because it's higher priority. |
| 17:44 | <annevk> | hmm, I wonder how it was added and did not conflict with HTMLSelectElement |
| 17:45 | <annevk> | paul_irish: see https://www.w3.org/Bugs/Public/show_bug.cgi?id=20720 |
| 17:45 | <Ms2ger> | Probably because their bindings suck :) |
| 17:47 | <annevk> | hmm |
| 17:47 | <annevk> | it seems it simply does not work on <select> |
| 17:47 | <annevk> | in WebKit |
| 17:47 | <annevk> | which does kinda suck |
| 17:47 | <annevk> | paul_irish: ^^ |
| 17:47 | Ms2ger | looks if he has a test for that |
| 17:49 | <odinho> | Hmm. Think some intern implemented all of that in Presto. |
| 17:49 | <odinho> | So there should be some tests for it somewhere at least :P |
| 17:57 | <Ms2ger> | Mm, JAM is back? |
| 17:58 | <Ms2ger> | odinho, https://github.com/w3c/web-platform-tests/pull/42 if you're interested |
| 18:01 | <odinho> | Ms2ger: No name for the first two tests? |
| 18:01 | <Ms2ger> | Knew it :) |
| 18:02 | <rcombs> | so, not sure if this has been discussed, but are there any plans for an API to combine EventSource and Web Notifications without actually having the page open, via use of a "push proxy"? |
| 18:02 | <Ms2ger> | odinho, want to suggest something? :) |
| 18:04 | <odinho> | Ms2ger: Uh. No. Something simple will work; select.remove(n) works, select.options.remove(n) works? :P |
| 18:05 | <odinho> | Anyway, other than that looks fine. Haven|t run it though, but guess you have :P |
| 18:05 | <odinho> | And we were going to add those author headers these days, no? :-) So prolly cool to add that too. |
| 18:06 | <odinho> | So everyone knowns who to bother. |
| 18:07 | <odinho> | lol, nice names. |
| 18:12 | <annevk> | rcombs: Web Notifications is just end-user notifications |
| 18:13 | <annevk> | rcombs: there are plans for push notifications |
| 18:13 | <rcombs> | annevk: cool, any draft specs out, or still in the works? |
| 18:13 | <annevk> | rcombs: as far as I know only Mozilla has something in that space |
| 18:13 | <annevk> | rcombs: https://wiki.mozilla.org/WebAPI/SimplePush and https://wiki.mozilla.org/WebAPI/PushAPI |
| 18:14 | <rcombs> | annevk: nice, thanks |
| 18:14 | rcombs | reads up |
| 18:16 | <odinho> | SSE has a document saying how telcos can implement their SSE with underlying SMS push, if I was not mistaken? |
| 18:16 | <odinho> | So far I've only seen SSE over HTTP though. |
| 18:17 | <odinho> | But the web browser abstracting away all work keeping the connection open and listening. |
| 18:17 | <annevk> | Yeah, there's a note, but nothing concrete. |
| 18:37 | <Ms2ger> | odinho, okay, want to merge now? :) |
| 18:38 | <odinho> | yee |
| 18:38 | <odinho> | lol, if I only could. Seems I have no rights to merge. darobin, jgraham, tobie |
| 18:50 | Ms2ger | wonders... |
| 18:51 | <Ms2ger> | odinho, you should be able to now |
| 20:21 | <annevk> | dglazkov: much more interesting than naming is answering the security questions |
| 20:21 | <annevk> | dglazkov: and the model you envision to build on top of rel=whatever |
| 20:22 | <annevk> | dglazkov: as that has much more interesting implications |
| 20:58 | <jgraham> | odinho: All the tests were selftests :( |
| 21:45 | <Ms2ger> | Hixie_, should have wontfixed that bug on its 15th birthday :) |
| 21:45 | <Hixie_> | heh |
| 21:59 | <Ms2ger> | jgraham, ta |
| 22:31 | <Hixie_> | GPHemsley: yt? |
| 22:33 | <Hixie_> | GPHemsley: (was wondering about the terminology "media type" vs "MIME type" and wondering if we could move mimesniff towards the latter and leave the former for CSS) |
| 22:44 | <annevk> | Hixie_: what's --CLEANUP--? |
| 22:45 | <Hixie_> | <!--CLEANUP--> is a comment |
| 22:46 | <Hixie_> | that says "CLEANUP" |
| 22:46 | <Hixie_> | so i know what i need to clean up |
| 22:46 | <Hixie_> | look for diffs with the checkin comment "Cleanup" |
| 22:46 | <annevk> | k |
| 22:47 | <Hixie_> | people complained that i was reflowing lines at the same time as making changes |
| 22:47 | <Hixie_> | so now i do them separately |
| 23:29 | <MikeSmith> | so I see the definition of valid e-mail address changed |
| 23:29 | MikeSmith | tries to remember if the validator is actually checking that |
| 23:49 | <Hixie_> | http://html5.clearboth.org/spec.html o_O |
| 23:52 | <TabAtkins> | That's... a lot of work. |
| 23:52 | <Hixie_> | i suspect it's machine-transalted. |
| 23:52 | <Hixie_> | translated. |
| 23:52 | <TabAtkins> | Likely. |
| 23:53 | <Hixie_> | why does http://www.w3.org/TR/webstorage/ have a disproportionate share of the source of bug spam? |
| 23:53 | <Hixie_> | weird |
| 23:53 | <MikeSmith> | Hixie_: http://html5.org/tools/web-apps-tracker?from=7771&to=7772 seems like a pretty bing change in behavior. Or not? |
| 23:53 | <MikeSmith> | the appcache change you just checked in |
| 23:53 | <Hixie_> | it's a change. |
| 23:53 | <Hixie_> | not a huuuge change. |
| 23:54 | <Hixie_> | all it does is switch the order of two steps in the navigate algorithm |
| 23:54 | <MikeSmith> | ok |
| 23:54 | <MikeSmith> | didn't come from a bug or comment? |
| 23:54 | <Hixie_> | should only affect people who have appcaches with a manifest saying they are in prefer-online mode, where some of the resources they explicitly cache happen to have URLs listed in their fallback namespaces |
| 23:54 | <Hixie_> | it was something i happened to come across when fixing another bug |
| 23:55 | <MikeSmith> | OK |
| 23:56 | <Hixie_> | seriously, why is /webstorage/ such a big source of spam |
