| 00:52 | <Hixie> | TabAtkins: does an element htat is display:none ahve a computer 'color' ? |
| 00:52 | <Hixie> | er |
| 00:53 | <Hixie> | typo central |
| 00:53 | <Hixie> | TabAtkins: does an element that is 'display:none' have a computed value for its 'color' property? |
| 00:56 | <Hixie> | dbaron: you might know the answer too ^ |
| 01:37 | <Hixie> | hm, @font-face in a worker |
| 01:37 | <Hixie> | damn |
| 01:40 | <Hixie> | well that puts a damper on things, doesn't it |
| 01:40 | <Hixie> | time to go home i think; good luck figuring this one out, future hixie. |
| 02:11 | <dbaron> | Hixie, I don't see any reason it would have a computed value |
| 03:10 | <zewt> | i hate working in reference-counted GC environments; idioms that you just expect to work don't |
| 03:30 | <annevk> | zcorpan: that logo is awesome |
| 03:31 | <Hixie> | which one? |
| 03:31 | <annevk> | http://resources.whatwg.org/logo-xhr.svg |
| 03:32 | <Hixie> | data:text/html,<style>body { font: 35em sans-serif;}</style><center>o_O |
| 03:32 | <annevk> | heh |
| 03:33 | <annevk> | see http://en.wikipedia.org/wiki/AFC_Ajax for context |
| 03:33 | <Hixie> | lol |
| 09:04 | <hsivonen> | annevk: Do you know of anyone implementing Encoding Standard-compliant converters that output UTF-8 and that are implemented in C/C++? |
| 09:51 | <zcorpan> | annevk: thanks |
| 10:21 | <zcorpan> | i wanted to +1 Hixie's email about [hidden] but didn't |
| 11:44 | <smaug____> | am I allowed to hate Ranges a bit |
| 11:47 | <jgraham> | No |
| 11:47 | <jgraham> | Only a lot |
| 11:54 | <MikeSmith> | I wonder who is was who coined the term "drive-by Web" |
| 12:00 | <hsivonen> | MikeSmith: what does the term mean? |
| 12:00 | <hsivonen> | annevk: maybe the Encoding Standard should have a security considerations section about stuff like https://bugzilla.mozilla.org/show_bug.cgi?id=408457#c99 |
| 12:00 | <MikeSmith> | hsivonen: as far as I can tell it just means the Web |
| 12:01 | <MikeSmith> | but the people who use it do so in order to contrast the Web with something that's not the Web but that they don't actually have a name for |
| 12:02 | <MikeSmith> | so instead of naming that other thing, they call the Web the "drive-by Web" |
| 12:02 | hsivonen | wonders how the Web is drive-by |
| 12:02 | <MikeSmith> | this practice seems to be limited just to members of the Google Chrome engineering team |
| 12:02 | <hsivonen> | strange |
| 12:03 | <MikeSmith> | hsivonen: I suspect the term originally came from security discussion of "drive by" exploits on the Web |
| 12:03 | <jgraham> | So this means something like "web where SOP is the security model" |
| 12:03 | <MikeSmith> | jgraham: yeah, pretty much |
| 12:03 | <MikeSmith> | as far as I can tell |
| 12:04 | <jgraham> | As opposed to "non-web downloadable stuff that happens to reuse the web stack" |
| 12:04 | <MikeSmith> | yup |
| 12:04 | <MikeSmith> | anyway for the latest seen-in-the-wild usage of the term, see http://lists.w3.org/Archives/Public/public-webapps/2012OctDec/0470.html |
| 12:04 | <MikeSmith> | from Michael Nordman |
| 12:08 | <jgraham> | "drive-by web" makes me think of "drive-by shooting", which seems like the wrong image. It's more like the "drive-thru web". |
| 12:08 | <jgraham> | Or, if you don't like long names, "web" |
| 12:09 | <MikeSmith> | I have no idea what it has to do with driving at all. |
| 12:09 | <hsivonen> | should I start speaking about “drive-by UTF-16” instead of just “UTF-16”? |
| 12:09 | <zcorpan> | http://code.google.com/codesearch#search/&q=lang:%5Esgml$&type=cs |
| 12:09 | <MikeSmith> | hsivonen: heh |
| 12:10 | zcorpan | notices "HTML" isn't in the drop-down in google code search |
| 12:10 | <MikeSmith> | anyway I think they'd be a lot better off using "off Web" to their other thing |
| 12:10 | <jgraham> | Well it sounds like the metaphor is supposed to be something that is an point on a journey. |
| 12:11 | <MikeSmith> | ok |
| 12:11 | <zcorpan> | hsivonen: do we have data about how much utf-16 content there is? |
| 12:11 | <MikeSmith> | zcorpan: yeah I wonder what the rationale is for not providing HTML search |
| 12:11 | <hsivonen> | zcorpan: I don’t |
| 12:11 | <zcorpan> | MikeSmith: lang:html in the search box works, though |
| 12:12 | <MikeSmith> | oh |
| 12:12 | <zcorpan> | but SGML being in the drop-down is pretty surprising |
| 12:13 | <zcorpan> | http://w3techs.com/technologies/details/en-utf16/all/all |
| 12:14 | <hsivonen> | Gives a good impression about a security org to have "confidential" written in the footer of each page of a PDF available by simple HTTP GET from the organization’s site. |
| 12:15 | <zcorpan> | that page gives some URLs we can try to evang if we want to kill utf-16 |
| 12:16 | <hsivonen> | zcorpan: some of those sites say UTF-16 in <meta> but actually use something ASCII-compatible |
| 12:17 | <zcorpan> | though the data might be bogus. i checked app.com which uses <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" /> and later has <?xml version="1.0" encoding="UTF-16"?> |
| 12:17 | <MikeSmith> | Orcs.in ? |
| 12:18 | <hsivonen> | zcorpan: I checked six sites none of which actually ended up decoded as UTF-16 |
| 12:18 | <hsivonen> | zcorpan: they just have a bogus declaration that doesn’t get honored |
| 12:19 | <zcorpan> | hsivonen: yeah, so not the data we're looking for |
| 12:19 | <hsivonen> | right |
| 12:21 | <zcorpan> | reported the bogusness on twitter |
| 12:23 | <zcorpan> | http://trends.builtwith.com/topsites/UTF-16-UCS-2 also appears to have a bogus list of sites |
| 12:25 | <zcorpan> | also reported on twitter |
| 12:29 | <hsivonen> | said "confidential" document claims UTF-16 is popular and says “Taiwanese” when meaning “Korean” |
| 12:37 | <zcorpan> | $ grep -aPic "Content-Type:\s*text/html\s*;\s*charset\s*=\s*[\"']?utf-16(le|be)?" web200904 |
| 12:37 | <zcorpan> | 0 |
| 12:37 | <zcorpan> | (though maybe they excluded utf-16 content from web200904, intentionally or not) |
| 13:33 | <annevk> | zcorpan: btw, I'm making the stroke-width 1 |
| 13:34 | <annevk> | zcorpan: looks a lot better around the eye |
| 13:34 | <annevk> | GPHemsley: no idea what's wrong with that file |
| 13:38 | <annevk> | insertAdjacentElement / insertAdjacentText are kinda sucky imo given that we just added before/after |
| 13:38 | <annevk> | which are way way nicer |
| 14:09 | <annevk> | hsivonen: https://www.w3.org/Bugs/Public/show_bug.cgi?id=19961 (security considerations) |
| 14:14 | <annevk> | hsivonen: https://www.w3.org/Bugs/Public/show_bug.cgi?id=19962 (insertAdjacent... crap) |
| 14:15 | <zewt> | (i thought the "before/after" names were changed to something less confusing) |
| 14:17 | <annevk> | (no) |
| 14:17 | <zewt> | (lame) |
| 14:17 | <annevk> | (care) |
| 14:17 | <zewt> | (i hope you care about good api naming) |
| 14:18 | <annevk> | (of course) |
| 14:20 | <hsivonen> | crap’s supported by multiple browsers, though, right? |
| 14:20 | <hsivonen> | makes it a feature |
| 14:21 | <annevk> | dunno, haven't tested Opera |
| 14:22 | <annevk> | hsivonen: the other weird thing is that Gecko sometimes takes the lead in removing stuff, but maybe this is a case where that cannot be done |
| 14:26 | <zcorpan> | annevk: i think 2 is closer to the original. i made it thicker to be slightly consistent with the other whatwg logos |
| 14:26 | <annevk> | I tried 2, but it looks ugly |
| 14:28 | <annevk> | hsivonen: so yeah, Opera supports it too |
| 14:31 | <zcorpan> | i guess 1.5 is closer to the original |
| 14:35 | <annevk> | k, I'll change it to that |
| 15:11 | <annevk> | http://lists.w3.org/Archives/Public/public-iri/2012Nov/0021.html maybe if they stopped publishing in ASCII... |
| 15:13 | <smaug____> | zcorpan: what does grep -aPc "\.insertAdjacent(Text|Element)\s*\(" stevef-all mean? |
| 15:14 | <zcorpan> | smaug____: it's this data set http://www.paciellogroup.com/blog/2012/04/html5-accessibility-chops-data-for-the-masses/ |
| 15:14 | <smaug____> | also, I assume even 85 is very low number |
| 15:14 | <zcorpan> | yeah |
| 15:15 | <zcorpan> | that data set is about 600,000 pages iirc |
| 15:16 | <smaug____> | not too many |
| 15:17 | <zcorpan> | so at most it's 0.014% |
| 15:17 | <zcorpan> | but that data set is pretty old now |
| 15:17 | <smaug____> | 0.0.14 doesn't sound a good reason to add somewhat bad API to web |
| 15:22 | <zcorpan> | i guess there's also not so much incentive for opera/webkit/ie to remove it, since doing so can make some pages to stop working (they might work in firefox due to different code paths) |
| 15:23 | <zcorpan> | but we have removed old IEisms and continue to do so. maybe this is something that should be removed as well. |
| 15:24 | <zcorpan> | also these data sets don't have external scripts |
| 15:30 | <annevk> | zcorpan: any chance the IDNA implementor is going to deny/confirm what I posted? |
| 15:33 | <zcorpan> | annevk: i can ping him |
| 15:38 | <zcorpan> | oh, i didn't know brucel came up with "scooby doo". i thought jgraham made it up on the spot at tpac. |
| 15:41 | <darobin> | zcorpan: no, it's classic brucel :) |
| 15:47 | Ms2ger | thought he knew "scooby doo" from elsewhere |
| 15:49 | <darobin> | Ms2ger: I think they also made a TV show out of it |
| 15:50 | <jgraham> | Sorry, I should have made it clear that I'm not a comic genius |
| 15:50 | <jgraham> | Although frankly I thought it was already obvious |
| 15:52 | darobin | has often found jgraham quite amusing |
| 15:52 | <jgraham> | You've been laughing at me? :( |
| 15:54 | <darobin> | jgraham: awwwww |
| 15:54 | <GPHemsley> | I was actually wonder what Scooby Doo had to do with that algorithm... |
| 15:54 | <GPHemsley> | +ing |
| 16:10 | <odinho> | GPHemsley: Because he always does that kind of deducting. |
| 16:10 | <GPHemsley> | hmm... |
| 16:10 | <odinho> | GPHemsley: "It's not a, it's not b, --- it must be c!!!!112" |
| 16:18 | <darobin> | speaking of data sets with which to figure out compat issues, this just came up: http://www.w3.org/community/webdevdata/ |
| 16:21 | <annevk> | I wonder why the check for ":" in http://mxr.mozilla.org/mozilla-central/source/nsprpub/pr/src/misc/prnetdb.c#1575 is not an "else if" |
| 16:21 | <annevk> | I'm gonna treat as an oversight in that code |
| 16:22 | <divya> | TabAtkins: PING |
| 16:24 | jgraham | falls over |
| 16:24 | <jgraham> | That was a very loud ping |
| 16:24 | <TabAtkins> | divya: pong |
| 16:27 | <annevk> | zcorpan_: thanks |
| 16:31 | <dglazkov> | good morning, Whatwg! |
| 16:32 | <TabAtkins> | divya: For serious, I pong you right back, and you ignore me. SO RUDE. |
| 16:34 | <annevk> | TabAtkins: maybe it didn't work because you replied in lowercase |
| 16:34 | <TabAtkins> | That's what I'm thinking. |
| 16:34 | <TabAtkins> | DIVYA: POOOOOOOONG |
| 16:37 | zcorpan_ | sees jgraham fly outside his train |
| 16:43 | <divya> | TabAtkins: omg |
| 16:43 | <divya> | sorry |
| 16:43 | <divya> | TabAtkins: SORRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRY |
| 16:43 | <TabAtkins> | Why you hate jgraham's ears so much? |
| 16:43 | jgraham | no longer has ears |
| 16:44 | <divya> | ahahah |
| 16:50 | <odinho> | Hmm, so, public ping/pong, but private chat? Good to know that we get all the interesting signaling-bits without all the superflous data noise ;P |
| 16:51 | <TabAtkins> | odinho: You're welcome. |
| 16:52 | <divya> | ahahaha odinho |
| 16:52 | divya | reminds self to never use ahaha again |
| 17:13 | <annevk> | oh man |
| 17:13 | <annevk> | even IPv6 syntax is not interoperable |
| 17:14 | <annevk> | http://[0:0:0:0:0:FFFF:129.144.52.38]/ does not parse in Opera |
| 17:14 | <annevk> | and http://tools.ietf.org/html/rfc4291#section-2.5.5.1 seems to suggest this is allowed |
| 17:17 | <MikeSmith> | no zcorpan |
| 17:18 | <annevk> | which contradicts RFC 3986 |
| 17:18 | <MikeSmith> | wanted to ask about SVG path error reporting for the validator |
| 17:18 | <annevk> | MikeSmith: you want to write a parser for that I think |
| 17:18 | <MikeSmith> | I did |
| 17:18 | <annevk> | MikeSmith: I think in general SVG data types have weak checking |
| 17:18 | <MikeSmith> | well not really |
| 17:18 | <annevk> | MikeSmith: cool |
| 17:18 | <MikeSmith> | yeah |
| 17:19 | <MikeSmith> | I didn't write it |
| 17:19 | <MikeSmith> | I just stole some code from Batik |
| 17:19 | <MikeSmith> | real parsing code |
| 17:20 | <MikeSmith> | anyway I wanted to ask what error message Simon expected for the case he reported |
| 17:20 | <MikeSmith> | which is <path stroke="#000" d="M0,0,L5,5"/> |
| 17:21 | <MikeSmith> | not that I'm offering a choice |
| 17:21 | <MikeSmith> | just wondering if what it does report is what he expected or not |
| 17:23 | <annevk> | M coordinate expected, none found? |
| 17:23 | <MikeSmith> | hmm so I could probably change it to say that |
| 17:23 | <MikeSmith> | but what is says now is |
| 17:23 | <MikeSmith> | Unexpected character "L" |
| 17:23 | <annevk> | that seems okayish too |
| 17:23 | <MikeSmith> | OK |
| 17:23 | <annevk> | it's kinda weird path syntax allows both commas and whitespace as separators, but extra commas are problematic |
| 17:25 | <MikeSmith> | yeah |
| 17:25 | <MikeSmith> | but hey why is something like <path stroke="#000" d="M 0,0,1,2,3,4 L5,5"/> valid? |
| 17:26 | <MikeSmith> | you're allowed to have an unlimited list of coordinate pairs for each command? |
| 17:26 | <TabAtkins> | Unlimited duplicates of the arg list. |
| 17:26 | <MikeSmith> | wtf |
| 17:26 | <MikeSmith> | why? |
| 17:27 | <TabAtkins> | Because the path syntax is heavily optimized for length, so omitting the letter from subsequent commands lets you drop one character. |
| 17:27 | <MikeSmith> | ah |
| 17:27 | <MikeSmith> | I see |
| 17:27 | <TabAtkins> | I'm not a fan of the path syntax. |
| 17:28 | <MikeSmith> | I think anybody who's ever tried to write a parser for it is not a fan either |
| 17:28 | <TabAtkins> | Nah, parsing's easy. |
| 17:33 | <SimonSapin> | TabAtkins: seen the syntax3 stuff? |
| 17:33 | <TabAtkins> | SimonSapin: Yup, I'm in the middle of a patch for one of the issues on my home computer. |
| 17:34 | <SimonSapin> | no hurry |
| 17:34 | <annevk> | MikeSmith: that one doesn't make much sense though |
| 17:34 | <annevk> | MikeSmith: the validator could issue a warning because when you duplicate M or m arguments you could write it more concise |
| 17:35 | <TabAtkins> | annevk: Nope, duplicate M commands can be useful. They define zero-length path segments, which get markers. |
| 17:35 | <annevk> | what's a marker used for? |
| 17:36 | <TabAtkins> | It's a symbol placed at the vertexes between path segments. |
| 17:37 | <TabAtkins> | Used for, for example, arrowheads (placed on the last vertex), or circles-joints (placed at the internal vertexes). |
| 17:40 | <MikeSmith> | anyway I guess I could at least have the error be "Expecting a coordinate pair for the M command but instead found unexpected character 'L' ." and "Expecting second coordinate of a coordinate pair for the M command but instead found unexpected character 'L' ." |
| 18:44 | <zcorpan> | [tm]: i was expecting an error at all and a pointer to the first unexpected character in the value |
| 19:02 | <Ms2ger> | annevk, heh, thunderbird thinks your ipv6 email is a scam |
| 19:02 | <annevk> | really? not going to get much replies then :/ |
| 19:03 | <Ms2ger> | Well, it still ended up in my inbox |
| 19:05 | <svl> | Thunderbird's "scam" detection algorithm is rather laughable. Does it contain a link with an ip address => scam |
| 19:05 | <svl> | luckily it's only a warning message, not an actual filter |
| 19:28 | <annevk> | I guess we'll see what happens with that message |
| 19:28 | <annevk> | I suspect there's not much interest in it from implementors anyway, until it starts biting them |
| 19:34 | <zcorpan> | annevk: i'd say, just spec something that makes sense |
| 20:01 | <GPHemsley> | FTR, official IETF 85 minutes: http://www.ietf.org/proceedings/85/minutes/minutes-85-websec (PDF) |
| 20:01 | <GPHemsley> | "Gordon Hemsley (not present) had taken on writing |
| 20:01 | <GPHemsley> | a mime |
| 20:01 | <GPHemsley> | sniffing document at |
| 20:01 | <GPHemsley> | WHAT |
| 20:01 | <GPHemsley> | WG. This has been a charter item in WebSec, but we have not done any |
| 20:01 | <GPHemsley> | work on this for over a year. The W3C has documents referencing the mime |
| 20:01 | <GPHemsley> | sniffing document. Nobody in the group objected to having this move to WHAT |
| 20:01 | <GPHemsley> | WG, |
| 20:01 | <GPHemsley> | and according to Larry Manister, the W3C is also fine with referencing the |
| 20:01 | <GPHemsley> | WHAT |
| 20:01 | <GPHemsley> | WG document, so the work item will be removed from our charter. |
| 20:01 | <GPHemsley> | " |
| 20:01 | <GPHemsley> | (terrible formatting, eh?) |
| 20:35 | Ms2ger | has heard Larry say some questionable things before |
| 21:29 | <Smylers> | IETF minutes are published in PDF? |
| 21:31 | <jwalden> | that's...kind of funny |
| 21:32 | Ms2ger | wonders if they get to use non-ASCII characters |
| 21:39 | <gsnedders> | Ms2ger: What sort of madness are you talking about? |
| 21:40 | <gsnedders> | What is this non-ASCII of which you speak? |
| 21:40 | <Ms2ger> | Crossing from another IRC channel, what's the first browser you used? |
| 21:41 | <TabAtkins> | Whatever version of Netscape was built into AOL at the time. |
| 21:42 | <gsnedders> | IE4/Mac, I think. |
| 21:44 | <Ms2ger> | Wow, you had a mac at the time? |
| 21:44 | <gsnedders> | My father did. I didn't. |
| 21:44 | Ms2ger | guesses Rick Waldron is a troll |
| 21:45 | <gsnedders> | What's his response? |
| 21:45 | <Ms2ger> | That people use vendor prefixes because they're in a spec, AFAICT |
| 21:47 | <divya> | where is this from Ms2ger? |
| 21:47 | <Ms2ger> | https://github.com/whatwg/javascript/issues/4#issuecomment-10386022 |
| 21:53 | <jsbell> | I think he's implying: developers use prefixed APIs despite them not being in a spec. |
| 21:54 | <Ms2ger> | Isn't that the opposite of the point he's trying to make? |
| 21:55 | <divya> | rick waldron is saying that implementation requirements actually imply green light to using them |
| 21:55 | <divya> | or that is what i gathered Ms2ger |
| 21:56 | <Ms2ger> | Yeah |
| 21:56 | <Ms2ger> | I just can't square that with his latest comment |
| 21:56 | <Ms2ger> | (besides it being nonsense, of course) |
| 21:57 | <jsbell> | Yeah, I can't come up with a coherent stance for his comments. |
| 21:57 | <jsbell> | And: NCSA Mosaic on a Sun3 (I think). |
| 21:57 | <Ms2ger> | Can't beat that :) |
| 21:58 | Ms2ger | wanders off |
| 22:01 | <miketaylr> | funny i was just having this coversation with rwaldron in another chan |
| 22:01 | miketaylr | gives up too |
| 22:02 | <gsnedders> | MULTI-TASKING! |
| 22:02 | <miketaylr> | heh |
| 22:03 | GPHemsley | used AOL 3.0 and/or Netscape 4.x |
| 22:04 | <miketaylr> | s/coversation/conversation/ |
| 22:04 | <GPHemsley> | (depending on your definition) |
| 23:11 | <Hixie> | damn you past hixie for leaving me with this conundrum |
| 23:11 | <Hixie> | so, anyone have any hot ideas for how to do web fonts in a worker? |
| 23:11 | <Hixie> | is there some non-@font-face mechanism for loading web fonts? |
| 23:12 | <heycam> | http://lists.w3.org/Archives/Public/www-style/2012Sep/0101.html |
| 23:13 | <Hixie> | seems reasonable |
| 23:13 | <Hixie> | did it get any buy-in?> |
| 23:13 | <heycam> | not sure, can't remember where the thread ended up |
| 23:14 | <heycam> | TabAtkins probably knows |
| 23:24 | <TabAtkins> | Hixie: http://dev.w3.org/csswg/css3-fonts/#document-fontloader |
| 23:24 | <TabAtkins> | So, yes. |
| 23:25 | <TabAtkins> | At least among Moz, and I don't know any reason why we'd object. No objections on the mailing list. |
| 23:25 | <Hixie> | sweet |
| 23:26 | <TabAtkins> | document.fontloader.loadFont("font name", "optional text for figuring out unicode ranges needed") |
| 23:26 | <TabAtkins> | Ugh, jdaggett is still doing the indire tion through fontface for the event. |
| 23:27 | <TabAtkins> | That would make it hard in a Worker, because you'd have to instead return a neutered property bag or something. |
| 23:27 | <Hixie> | jdaggett: ping |
| 23:27 | <TabAtkins> | jdaggett: Yo, here's a stronger use-case for flattening the CSSFontFaceRule properties onto the CSSFontFaceLoadEvent interface. |
| 23:38 | <Hixie> | hm |
| 23:38 | <Hixie> | workers don't have a direction |
| 23:38 | <Hixie> | how do we pick a default writing direction |
| 23:38 | <Hixie> | workers suck. |
| 23:42 | <Hixie> | jdaggett, TabAtkins: would be good for canvas if there could be some conceptual "fonts in scope" which would be defined for each Element, Document, and worker |
| 23:56 | <TabAtkins> | Hixie: The same way that documents choose their default direction? Is that applicable? (I don't know what it is.) |
| 23:56 | <TabAtkins> | (Is it just ltr by default?) |
| 23:59 | <TabAtkins> | Hixie: "fonts in scope" should be pretty easy to define, I think. It's the set of local fonts available to the document, plus any webfonts after they trigger a "load" event. |
