| 00:01 | <Hixie> | zewt: so long as nobody else did it :-) |
| 00:26 | <zewt> | Hixie: but a raw binary file containing an array of floats that happens to begin with 1.0 would match, heh |
| 00:26 | <zewt> | (eg. a list of vertices or whatever) |
| 00:29 | <benschwarz> | Hixie, did you get my note about renaming the dev spec? |
| 00:31 | <annevk> | Hixie: if you have a better idea for https://www.w3.org/Bugs/Public/show_bug.cgi?id=18780 it is welcome |
| 00:32 | <annevk> | Hixie: but given your comment, you would be okay with me drafting that and seeing where it will take you? |
| 00:48 | <Hixie> | benschwarz: i'm not aware of anything from you that i haven't replied to |
| 00:48 | <Hixie> | benschwarz: so either yes and i replied, or no :-) |
| 00:48 | <benschwarz> | lemme see… |
| 00:49 | <benschwarz> | https://github.com/benschwarz/developers.whatwg.org/issues/69 |
| 00:50 | <benschwarz> | @ Hixie |
| 00:50 | <Hixie> | didn't i do that already? |
| 00:50 | <Hixie> | i thought i did that a while back |
| 00:52 | <benschwarz> | maybbee you did |
| 00:52 | <benschwarz> | oh. |
| 00:52 | <benschwarz> | http://www.whatwg.org/specs/web-apps/current-work/dev-index look at that |
| 00:52 | <Hixie> | heh |
| 00:53 | <Hixie> | i think you asked me on irc and i did it then |
| 00:53 | <benschwarz> | right |
| 00:53 | <Hixie> | i don't think i'd seen the github issue |
| 00:53 | <benschwarz> | well. that was good |
| 00:53 | <benschwarz> | I'll rebuild / deploy then :) |
| 01:00 | <benschwarz> | Hixie: deployed :) |
| 02:00 | <erlehmann> | how would i put a link in a webvtt file? |
| 02:06 | <erlehmann> | hey rillian |
| 02:06 | <erlehmann> | how can i make annotated links in WebVTT? |
| 02:07 | <erlehmann> | i have this http://warumnicht.dieweltistgarnichtso.net/wn-15.html |
| 02:07 | <erlehmann> | and want to make it more accessible |
| 02:09 | <erlehmann> | maybe the whatwg is the wrong channel :/ |
| 02:23 | <erlehmann> | today i thought about targeting media elements with links |
| 02:24 | <erlehmann> | <a href="testfile.oga#t=20" target=audio>test</a><audio id=audio>this will be changed</audio> |
| 02:25 | <erlehmann> | <audio nam=audio> would it have to be |
| 02:25 | <erlehmann> | name |
| 04:23 | <erlehmann> | html5 table parsing annoys me |
| 06:07 | <MikeSmith> | Hixie: you got a typo in the "Unstyled XML documents" change |
| 06:07 | <MikeSmith> | http://html5.org/r/7525 |
| 06:08 | <MikeSmith> | "and thes a user agent following these requirements" |
| 06:08 | <Hixie> | thanks |
| 08:33 | <zcorpan> | erlehmann: if you read logs, webvtt doesn't support links, but you can use metadata kind and have custom captions. also <audio> doesn't have a name attribute, nor is it a browsing context so target="" doesn't work like that |
| 08:35 | <zcorpan> | you can have a script that changes src="" though. i guess you could use target that targets an iframe as no-script fallback, if you care |
| 09:55 | <annevk> | GPHemsley: font sniffing only happens in font context |
| 10:09 | hsivonen | wonders how sfnt support is actually implemented. |
| 10:09 | <hsivonen> | that is, how early does the dispatch to different code paths depending on the type of glyph data occur |
| 10:10 | <hsivonen> | i.e., does the sniffer even need to decide beyond 1) WOFF, 2) raw sfnt, 3) neither (reject) |
| 10:16 | <annevk> | the data I supplied came from Opera's font guy |
| 11:30 | <zcorpan> | i wonder if CfC to publish counts as "process discussions" in html wg |
| 12:04 | <Stevef> | zcorpan: yes and they should soon be posted to an admin mailing list see: http://lists.w3.org/Archives/Public/public-html/2012Nov/0079.html |
| 12:06 | <zcorpan> | Stevef: ah, so it's not in effect yet |
| 12:06 | <Stevef> | the CFC for the mailing list thing gas passed, so should happen soon |
| 12:07 | <Stevef> | has passed not gas passed :-) |
| 12:42 | <hsivonen> | what mechanism causes Chrome to decode the frame on this page as UTF-8? http://ht.hypertech.com.br/doutorja/ |
| 12:42 | <hsivonen> | the meta is beyond 1024 bytes |
| 12:42 | <hsivonen> | and WebKit isn't supposed to be doing reloading upon seeing a late meta |
| 12:43 | <hsivonen> | or do they reload these days? |
| 12:43 | <hsivonen> | or have they increased the lookahead from 1024 bytes? |
| 12:56 | <annevk> | they might have increased it, I saw there was discussion about that |
| 12:56 | <annevk> | no pointer, sorry |
| 13:24 | <annevk> | fyi http://annevankesteren.nl/2012/11/copyright-response |
| 13:24 | <annevk> | it's nothing new, just trying to be fair |
| 14:06 | <annevk> | Anyone opposed to me introducing the term "redispatching"? |
| 14:06 | <annevk> | You'll be able to use it to find dom.spec.whatwg.org if you forgot how to type that ;) |
| 14:09 | <smaug____> | AryehGregor: ping |
| 14:10 | smaug____ | tries to recall if we handle PI in ranges like comment/text on purpose |
| 14:10 | <smaug____> | (in Gecko) |
| 14:14 | <annevk> | I think the idea was to make PI just another CharacterData thingie |
| 14:14 | <smaug____> | yeah |
| 14:14 | <smaug____> | I can't find the right bug now |
| 14:14 | <annevk> | with some minor exceptions that is already largely the case |
| 14:14 | <annevk> | in most implementations |
| 14:14 | <smaug____> | Didn't help that I was trying to search for Mozilla bug using W3C bugzilla :) |
| 14:17 | <annevk> | I kinda wish I could just search through W3C/Mozilla/WebKit/radar:/Opera/IE/Chromium all at once |
| 14:17 | <annevk> | Unfortunately I can only search through some of those, and have to find the correct URL |
| 14:28 | <darobin> | annevk: I don't know if it's exposed systematically, but bugzilla has an API |
| 14:29 | <darobin> | it should be possible to address at least those collectively |
| 14:32 | <annevk> | http://dom.spec.whatwg.org/#dispatching-events is another reason why I favor imperative specifications. So easy to make surgical incisions. |
| 14:34 | <darobin> | bah, subjunctive specifications FTW! |
| 14:34 | <odinho> | annevk: What is radar:? |
| 14:35 | <smaug____> | odinho: Apple's internal bug tracker |
| 14:35 | <annevk> | odinho: URL scheme used for referring to issues in Apple's internal bug tracker, presumably used for Safari, but potentially much more |
| 14:35 | <annevk> | darobin: learned a new word, yay |
| 14:36 | <darobin> | it would be fun to write a spec in the subjunctive |
| 14:36 | <darobin> | if this attribute be in error |
| 14:37 | <darobin> | were the document invalid, the user agent would... |
| 14:38 | <annevk> | if I be as high as kite? |
| 14:42 | <JibberJim> | It may well be more readable than your normal specs darobin |
| 14:42 | <darobin> | annevk: you're tall, but not that tall |
| 14:42 | <darobin> | OMG there are ghosts here |
| 14:42 | <annevk> | the real JibberJim? |
| 14:42 | <darobin> | JibberJim: it's worth a shot |
| 14:43 | <annevk> | been a while |
| 14:43 | <darobin> | I wonder if someone has made an automatic subjunctiviser |
| 14:43 | <JibberJim> | I am the real me yes... I've found my way out of baby hood and re-entering my previous worlds. |
| 14:44 | <darobin> | so you've stopped being a big baby? |
| 14:44 | <JibberJim> | Something like that yes :) |
| 14:46 | <darobin> | how's Otter? |
| 14:46 | <annevk> | time to update http://www.jibbering.com/2002/4/httprequest.html ? :-) |
| 14:46 | <darobin> | and how old? like 2 or something by now? |
| 14:47 | <JibberJim> | 18months is all |
| 14:47 | <JibberJim> | What you mean that's not completely up to date and accurate? |
| 14:47 | <JibberJim> | anyone would think it was 10 years old |
| 14:47 | <darobin> | or http://jibbering.com/nauts/htmlnaut/ :) |
| 14:48 | <darobin> | that's some serious DHTML there |
| 14:48 | <darobin> | we so should reintroduce the usage of the word DHTML |
| 14:49 | <odinho> | darobin: It's been sorely missed (?) |
| 14:49 | <darobin> | odinho: I stumbled upon my old copy of the flamingo book last night; the feeling of nostalgia hasn't quite left me |
| 14:49 | <JibberJim> | I'll a book "The stupid persons guide to DHTML5" - that should do it |
| 14:50 | <odinho> | Oh, DTML5. Marketing will jump on it straight away ;D |
| 14:51 | <darobin> | the domain's free |
| 14:54 | <annevk> | smaug____: seems unlikely we'll get rid of initCustomEvent right? |
| 14:55 | JibberJim | wonders why he can't find darobin in htmlnaut |
| 14:55 | <darobin> | my foaf file probably went offline at some point in ~2006 |
| 14:56 | <darobin> | I'm waiting for the JSON-LD/microdata/RDFa return of FOAF! |
| 14:56 | <JibberJim> | I'm pretty sure I never scuttered since 2003 |
| 14:56 | <JibberJim> | For ages I didn't even think I had a cache of the data |
| 14:56 | <darobin> | in that I ought to be there, I used to be |
| 14:56 | <darobin> | *case |
| 14:56 | <zcorpan> | darobin: then how do people/non-people know who you are? |
| 14:57 | <darobin> | zcorpan: they have fond memories from seeing me dance topless? |
| 14:57 | <JibberJim> | You still do that? I may have to leave these circles again |
| 14:58 | <annevk> | darobin: if I'm elected to the TAG, I'll make DHTML top priority |
| 14:58 | <darobin> | YAY annevk FOR TAG! |
| 14:59 | <darobin> | JibberJim: I hadn't done it in a while, but I reinstated it at the last TPAC |
| 14:59 | <darobin> | in fact I think I hadn't done it since Enschede |
| 14:59 | <odinho> | annevk: You have to make buttons like erikjmoller have done for his employee rep campaign in Opera. |
| 14:59 | <zcorpan> | "TAG SUCKS"? |
| 15:01 | <annevk> | a whole lot of http://html5.org/temp/unicorn.svg |
| 15:05 | <darobin> | can you get D66 endorsement or something? |
| 15:05 | <darobin> | maybe we can start a W3C Pirate Party |
| 15:06 | <smaug____> | annevk: I think so |
| 15:07 | <smaug____> | (at least I'm not going to try it) |
| 15:07 | <darobin> | annevk: your TAG nomination has cleared, go fill out https://www.w3.org/2002/09/wbs/1/tagnom-201211/ |
| 15:11 | <annevk> | darobin: so the IE agreement is scoped to the group right? |
| 15:11 | <annevk> | darobin: e.g. I might end up being an IE in the TAG, but would not be automatically in say WebApps |
| 15:12 | <darobin> | annevk: yes, I thought of that, you're safe |
| 15:12 | <annevk> | k |
| 15:12 | <darobin> | W3C pwns all your copyrights for your contributions to the TAG |
| 15:12 | <darobin> | MUAHAHAHAHAHAHAHAHAHAHAHAHA |
| 15:13 | <darobin> | I'm *so* evil |
| 15:14 | <annevk> | I realise that, it's unfortunate but acceptable... http://www.w3.org/2004/10/27-tag-charter.html is kinda whacky though. |
| 15:14 | <annevk> | "W3C has adopted an architectural principle that XML should be used for the syntax of Web formats unless there is a truly compelling reason not to" |
| 15:15 | <darobin> | oh yeah the charter is a whole barrel laughs |
| 15:16 | <zewt> | is that quote from the mid-90s |
| 15:16 | <zewt> | heh |
| 15:16 | <darobin> | zewt: it feels that way, but actually we didn't have XML in the mid90s |
| 15:16 | <darobin> | annevk: it's all a matter of finding "compelling reasons not to" |
| 15:17 | <zewt> | simple, "xml is terrible" |
| 15:17 | <zewt> | i find that pretty compelling |
| 15:17 | <zewt> | i'm not sure i've hit a single legitimate use case for xml since json came around |
| 15:19 | <hsivonen> | I feel I have been baited into 386 on a list I don't normally follow: www-international |
| 15:19 | <annevk> | darobin: submitted |
| 15:19 | <zewt> | math coprocessor y/n |
| 15:19 | <darobin> | annevk: wicked cool |
| 15:19 | <darobin> | now go campaign! |
| 15:20 | <annevk> | hsivonen: my fault I think |
| 15:20 | <annevk> | darobin: when is the final day? |
| 15:20 | <darobin> | www-international — is that for Web Communism? |
| 15:20 | <darobin> | annevk: for nominations? the 30th |
| 15:20 | <darobin> | as things currently stand there are five candidates for four seats, so there'll be an election no matter what |
| 15:20 | <annevk> | darobin: and then elections until end of December? |
| 15:20 | <darobin> | yup |
| 15:21 | <darobin> | I forget when they close |
| 15:21 | <darobin> | maybe early January |
| 15:21 | <darobin> | you should blog about it like slightlyoff did |
| 15:21 | <darobin> | also, send some statement of your platform to www-tag |
| 15:21 | <darobin> | once you guys do that I can start sending some "Questions to the Candidates" there :) |
| 15:22 | <annevk> | will do in December then |
| 15:22 | <darobin> | good |
| 15:22 | <zcorpan> | hsivonen: which thread? |
| 15:23 | <zcorpan> | http://lists.w3.org/Archives/Public/www-international/2012OctDec/0078.html ? |
| 15:26 | <zcorpan> | is "In Internet Explorer 5.5 a BOM at the start of a file will cause the page to be rendered in quirks mode" true? http://www.w3.org/International/questions/new/qa-byte-order-mark-new |
| 15:26 | <zcorpan> | doesn't IE5.5 only have one mode anyway? |
| 15:27 | <zewt> | why is a document intended for this "intended audience" talking about IE5.5? heh |
| 15:30 | <JibberJim> | IE 5.5 was very fussy about exactly what it saw to avoid quirks mode - so I suspect the answer is yes |
| 15:30 | <JibberJim> | IE 5.5 is my era! I can still be useful! |
| 15:30 | <SimonSapin> | when is the vote? |
| 15:31 | <darobin> | SimonSapin: soon |
| 15:31 | <darobin> | after the nominations end, it should open pretty quickly |
| 15:31 | <darobin> | so early Dec |
| 15:32 | <annevk> | smaug____: thanks, guess I'll add a definition for initCustomEvent() then along with some guidelines for specifications defining their own events |
| 15:35 | <annevk> | SimonSapin: so for data:, what you want to do is to define a parser (see URL/HTML specs for examples of parsers) |
| 15:36 | <annevk> | SimonSapin: the input is "scheme data" + "?" + "query" (currently just "scheme data", but soon "query" will be separate) |
| 15:36 | <SimonSapin> | annevk: how does that parser hook into URL ? |
| 15:37 | <zcorpan> | JibberJim: i thought quirks/standards was introduced in ie *6* |
| 15:37 | <SimonSapin> | apparently the "?" and the query are both part of the "body" of data:,test?test |
| 15:37 | <annevk> | SimonSapin: I haven't entirely thought that through yet. What I think we want is that you have a parsed URL object. And then when you fetch it you do URL processing |
| 15:38 | <annevk> | SimonSapin: URL processing is scheme-specific and would run this parser, to obtain the resource |
| 15:39 | <MikeSmith> | zcorpan: http://www.w3.org/html/wg/wiki/DoctypeSwitching |
| 15:39 | <SimonSapin> | so this parser would first append "?" and the query to the string being parsed |
| 15:40 | <JibberJim> | zcorpan: I thought it had a seperate rendering mode that matched 4 and itself? but predates "quirks" and "standards" for sure? |
| 15:40 | <zcorpan> | JibberJim: [citation needed] :-P |
| 15:41 | <annevk> | SimonSapin: the URL processor takes a URL; if it sees the scheme is data it invokes the data URL processor with schemedata+"?"+query |
| 15:41 | <annevk> | SimonSapin: hmm, maybe not exactly like that, but to obtain the message body you need scheme data and query, yes |
| 15:42 | <annevk> | SimonSapin: the other thing that is somewhat different from http://tools.ietf.org/html/rfc2397 is that the default encoding is windows-1252 |
| 15:42 | <SimonSapin> | versus ascii |
| 15:42 | <annevk> | yeah |
| 15:42 | <annevk> | and things like data:text/html;,%E2%84%A2 do not fail |
| 15:42 | <hsivonen> | IE 5.5 had only the mode now known as quirks |
| 15:43 | <annevk> | (the specification requires a parameter after ";" which is not how implementations work apparently) |
| 15:44 | <annevk> | SimonSapin: you may want to coordinate MIME type parsing with GPHemsley though |
| 15:44 | <annevk> | SimonSapin: ideally that's going to be the same parser throughout the whole platform |
| 15:44 | <annevk> | SimonSapin: used for HTTP, data URLs, type attribute values, etc. |
| 15:44 | JibberJim | 's memory has let him down :( |
| 15:44 | <SimonSapin> | is WHATWG doing HTTP too? :) |
| 15:47 | <SimonSapin> | annevk: where should this parser live? |
| 15:48 | <annevk> | I rather not do HTTP, but I guess as long as we keep walking into stuff that's undefined or wrongly defined we might as well pick up things... |
| 15:49 | <annevk> | SimonSapin: MIME and a data URL spec? |
| 15:49 | <SimonSapin> | yes |
| 15:49 | <SimonSapin> | where should I send pull requests |
| 15:50 | <annevk> | I guess MIME type parsing in MIME (sniffing) and data URLs in their own spec or maybe in URL at some point. For now I'd just sketch it out somewhere, maybe the wiki or your own site |
| 15:51 | <SimonSapin> | is there a preprocessor for doing eg. cross-references? |
| 15:52 | <annevk> | http://wiki.whatwg.org/wiki/GitHub http://wiki.whatwg.org/wiki/Anolis |
| 15:54 | <SimonSapin> | oh, anolis too is using cssselect |
| 16:10 | <annevk> | heh, Chrome starts downloading if you enter data:x, |
| 16:10 | <annevk> | oh Opera too |
| 16:11 | <SimonSapin> | annevk: do they do that for any unknown mime type? |
| 16:11 | <annevk> | yeah, though Firefox fails to parse that as a MIME type... |
| 16:12 | <SimonSapin> | … and assumes text/plain |
| 16:12 | <annevk> | oh yeah, didn't notice that |
| 16:12 | <annevk> | browsers suck |
| 16:13 | <annevk> | I doubt much of that matters though |
| 16:17 | <SimonSapin> | annevk: MIME Sniffing is not about parsing MIME, is it? |
| 16:17 | <annevk> | just realised data URLs have a pretty high overhead, they're processed 2 times, and when found in HTML 3 times |
| 16:18 | <annevk> | SimonSapin: I think it should be in the end, then we'd just call it MIME |
| 16:18 | <SimonSapin> | annevk: I weasyprint I special-case value.startswith('data:') to skip some of the processing |
| 16:18 | <annevk> | SimonSapin: not sure how others feel about that |
| 16:18 | <annevk> | SimonSapin: case-insensitive match I hope? |
| 16:18 | <SimonSapin> | hum… :) |
| 16:19 | <SimonSapin> | nope |
| 16:19 | <SimonSapin> | should fix that |
| 16:20 | <SimonSapin> | annevk: what kind of case-insensitivity? :p |
| 16:20 | <annevk> | yeah, I guess you want optimised paths for data: in implementations |
| 16:21 | <annevk> | SimonSapin: heh, URL schemes can only be ASCII |
| 16:21 | <SimonSapin> | (kidding) |
| 16:21 | <annevk> | [dD][aA][tT][dD][:] |
| 16:22 | <darobin> | do we know who made Opera's submitted tests for <track>? http://w3c-test.org/html/tests/submission/Opera/media/interfaces/TextTrack/activeCues.html |
| 16:22 | <darobin> | I ask 'cause there's a bug being filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=20064 |
| 16:23 | <annevk> | prolly zcorpan |
| 16:24 | <darobin> | canPlayType() isn't exactly great for boolean tests :) |
| 16:25 | <SimonSapin> | I subscribed to whatwg⊙wo ~30min ago but did not get the confirmation email |
| 16:26 | <JibberJim> | Can I moan about the name of an API that uses CAN but doesn't return a Boolean |
| 16:27 | JibberJim | waits for someone to say "Probably" or "Maybe" in response to that request... |
| 16:27 | <darobin> | JibberJim.canIMoan("API") // "probably, maybe" |
| 16:29 | <darobin> | I still think that this should have been "possibly", "maybe"; it has a song to it http://www.youtube.com/watch?v=tE11_5Spq1I |
| 16:29 | <annevk> | SimonSapin: hmm, check spam maybe? otherwise you need to ask Hixie for help |
| 16:30 | <annevk> | JibberJim: you might have to get in line |
| 16:30 | <SimonSapin> | annevk: nothing in my mta’s logs |
| 16:31 | <annevk> | SimonSapin: meh, maybe try again? I don't really know how long it's supposed to take. I believe instant. Otherwise email Hixie |
| 16:32 | <SimonSapin> | "Depending on the configuration of this mailing list, your subscription request may have to be first confirmed by you via email, or approved by the list moderator." |
| 16:32 | <SimonSapin> | does it need confirmation? |
| 16:33 | <annevk> | maybe not, but http://lists.whatwg.org/listinfo.cgi/whatwg-whatwg.org states "Subscribe to whatwg by filling out the following form. You will be sent email requesting confirmation, to prevent others from gratuitously subscribing you." |
| 16:34 | <annevk> | at some point we should organise this a bit less centrally so people besides Hixie can help out |
| 16:38 | <GPHemsley> | annevk, SimonSapin: Is parsing a mime type really all that complex? (I mean, I understand reality doesn't match the HTTP spec.) It's just the notion of ignoring spaces and trailing semicolons, right? |
| 16:39 | <annevk> | GPHemsley: it's pretty basic |
| 16:40 | <annevk> | GPHemsley: you might also want to change it from a byte parser to a code point parser as especially when used in markup languages the input is going to be code points |
| 16:40 | <annevk> | (or code units) |
| 16:40 | <GPHemsley> | annevk: Are we still talking about the mime type parser? |
| 16:41 | <annevk> | uhuh |
| 16:41 | <GPHemsley> | I was actually just going to suggest a regex |
| 16:41 | <GPHemsley> | Is that not a good idea? |
| 16:41 | <zewt> | ew |
| 16:41 | <GPHemsley> | oh |
| 16:41 | <GPHemsley> | :) |
| 16:41 | <zewt> | then you have to normatively depend on the definition of regexes, which is orders of magnitude more complicated :) |
| 16:42 | <GPHemsley> | ah, fair point |
| 16:42 | <annevk> | regexp and BNF are equivalent in problems |
| 16:42 | <darobin> | zewt: bah, just do what XML Schema did and reference "Perl" |
| 16:42 | <GPHemsley> | annevk: BNF isn't simpler than regexp? |
| 16:42 | <annevk> | oh yeah, lets use XML Schema :) |
| 16:43 | <annevk> | GPHemsley: it's different, mostly |
| 16:44 | <GPHemsley> | So I should just spell out the parsing step by step? |
| 16:44 | <darobin> | <xs:simpleType name='mediaType'><xs:restriction base='xs:string'><xs:pattern value='\w+\/\w+'/></xs:restriction></xs:simpleType> |
| 16:44 | <darobin> | there! spec done |
| 16:44 | <darobin> | that wasn't so hard was it? |
| 16:44 | <annevk> | GPHemsley: yeah |
| 16:44 | <darobin> | you bunch of whiny losers |
| 16:44 | <annevk> | darobin: you can put that in your W3C copy :p |
| 16:45 | <GPHemsley> | darobin: But that doesn't handle parameters ;) |
| 16:45 | <darobin> | annevk: ooooh, and we'll *finally* get XML Schema in browsers! |
| 16:45 | <annevk> | darobin: i know right |
| 16:45 | <darobin> | we've been waiting so long |
| 16:45 | <darobin> | GPHemsley: bah |
| 16:45 | <darobin> | <xs:simpleType name='mediaType'><xs:restriction base='xs:string'><xs:pattern value='\w+\/\w+(;.*)?'/></xs:restriction></xs:simpleType> |
| 16:46 | <GPHemsley> | Well, OK... but you'll want to actually parse the parameter name and value... |
| 16:47 | <darobin> | you want to parse stuff with XML Schema? Srsly? |
| 16:47 | <darobin> | annevk: I wonder if we should use http://www.w3.org/TR/xmlschema-1/#key-null as an example for <dfn> |
| 16:47 | <GPHemsley> | (\w+)\/(\w+)(?:;\s*(?:(\w+)\s*(?:=\s*(\w*))?)?)* |
| 16:47 | <GPHemsley> | oh, actually, that's no good |
| 16:47 | <zewt> | :| |
| 16:47 | <GPHemsley> | (\w+)\/(\w+)(?:;\s*(?:(\w+)\s*(?:=\s*(.*))?)?)* |
| 16:47 | <zewt> | <zewt> ew |
| 16:48 | <GPHemsley> | yeah |
| 16:48 | <GPHemsley> | but that's roughly how I plan to do it |
| 16:48 | <zewt> | also handling errors with anything but "fail completely" is hard |
| 16:48 | <darobin> | GPHemsley: note that \w doesn't cut it |
| 16:48 | <GPHemsley> | darobin: I blame you |
| 16:49 | <darobin> | I don't know if they can start with anything other than [a-z], but they can certainly contain . and + |
| 16:49 | <GPHemsley> | this was a first approximation :) |
| 16:49 | <darobin> | GPHemsley: dude, if you take my XML Schema jokes and put them in your code, really, *really* don't come complaining when things don't work later |
| 16:50 | <GPHemsley> | in my original draft of the regex, I had . everywhere |
| 16:50 | <GPHemsley> | but just for that, I'm going to include it |
| 16:50 | <GPHemsley> | but only for you |
| 16:51 | <GPHemsley> | "If user agent is darobin, parse the Content-Type header according to the following XML Schema:" |
| 16:54 | <darobin> | haha |
| 16:54 | <darobin> | I reckon that's still a bit too simple |
| 16:54 | <darobin> | for one, it involves none of SOAP, WSDL, or ASN.1 |
| 16:55 | <annevk> | needs more WS-* |
| 17:28 | <annevk> | well look at that |
| 17:28 | <annevk> | back in the day |
| 17:28 | <annevk> | http://www.w3.org/Protocols/HTTP/AsImplemented.html timbl published in the public domain |
| 17:29 | <annevk> | ooh http://www.w3.org/Policy.html |
| 17:29 | <annevk> | "The definition of protocols such as HTTP and data formats such as HTML are in the public domain and may be freely used by anyone." |
| 17:30 | <annevk> | sad that got fucked along the way |
| 17:36 | <Ms2ger> | Robert Cailliau claims that he could take credit for that, fwiw |
| 17:42 | <annevk> | HTTP 0.9 is quite nice btw |
| 17:42 | <annevk> | you don't even tell the server it's about HTTP |
| 17:42 | <annevk> | just "GET /gimme" |
| 17:42 | <annevk> | don't even need a CRLF there |
| 17:43 | <Ms2ger> | <annevk> [dD][aA][tT][dD][:] |
| 17:43 | Ms2ger | hadn't heard of the datd: url scheme |
| 17:43 | <annevk> | oops |
| 17:44 | <SimonSapin> | and now decades of implementations reverse-engineering each other and adding support for datd: |
| 17:46 | <annevk> | nah, maybe for about:unicorn |
| 18:57 | <GPHemsley> | It's a real shame that nested comments aren't supported in HTML (or any commonly used programming language) |
| 18:59 | <zewt> | tricky with typical tokenization methods, iirc |
| 18:59 | <GPHemsley> | yeah... but it's still a shame |
| 18:59 | <GPHemsley> | Ocaml has it, though |
| 19:00 | <GPHemsley> | (which actually turned out to be somewhat problematic when I attempted to create a syntax highlighter for it, given the aforementioned issue) |
| 19:00 | <zewt> | vim gets upset with languages that have poor random access |
| 19:01 | <zewt> | (though you can increase the search distance, which isn't as big a deal today as it was 15 years ago) |
| 19:02 | <GPHemsley> | I was writing it for TextWrangler which, IIRC, used regex |
| 19:04 | <GPHemsley> | Of course, on top of the lack of support for nested comments in HTML is the prohibition on '--' within the comment |
| 19:08 | <annevk> | yeah, silly SGML legacy |
| 19:10 | <zewt> | don't know what the basis for the "--" thing is, since it parses fine and i'm sure is everywhere |
| 19:11 | <zewt> | i use dashes not infrequently in english text, so i've probably done it plenty of times myself, heh |
| 19:14 | <GPHemsley> | so... there are two types of TTF fonts |
| 19:14 | <GPHemsley> | one is an OpenType package, and the other isn't |
| 19:15 | <GPHemsley> | I think |
| 19:23 | <GPHemsley> | So, if fonts are not sniffed by the browser in the regular browsing context, then likely the only usecase for font sniffing is CSS Fonts, and that would be done specifically. |
| 19:23 | <GPHemsley> | Thus, the needs of CSS Fonts are what governs the list of fonts. |
| 19:23 | <GPHemsley> | (Just thinking out loud.) |
| 19:25 | <GPHemsley> | It's really sad that the first result for something like "css3-fonts" is w3schools instead of the spec itself |
| 19:30 | <GPHemsley> | ...and according to css3-fonts, in addition to EOT, I should also somehow accommodate SVG fonts |
| 19:30 | <GPHemsley> | hmm |
| 19:31 | <zewt> | w3schools was one of the first sites i blocked when google's site blocking turned up, heh |
| 19:31 | <GPHemsley> | zewt: Oh, where can I find this setting? |
| 19:31 | <zewt> | dunno, it's magic |
| 19:31 | <zewt> | sometimes it shows up if you view a site then return to the search results, often it just won't show up at all |
| 19:32 | <GPHemsley> | ah, found it |
| 19:32 | <GPHemsley> | just perform a search and click on the gear |
| 19:32 | <zewt> | which is annoying when you hit one of those results where you click the link and it shows a popover "pay us money to see all of the answers!" and you go back and google won't let you block them |
| 19:32 | <GPHemsley> | and navigate your way through |
| 19:33 | <GPHemsley> | Gear > Search settings > Blocking unwanted results > Block unwanted sites |
| 19:34 | GPHemsley | wonders who else he wants to block |
| 19:35 | <GPHemsley> | scumbag css3-fonts says "see Appendix A" but doesn't provide a link to it |
| 19:35 | <GPHemsley> | I mean, come on! You want me to scroll to it *manually*?! |
| 19:36 | <zewt> | first-world problem go |
| 19:36 | <zewt> | start a new youtube series, angry web spec nerd |
| 19:38 | <GPHemsley> | heh |
| 19:39 | <GPHemsley> | Why hasn't the CSS/Style working group set up an interface to filter mailing list messages by their spec tags? |
| 19:42 | <hober> | GPHemsley: that's what your mail client does |
| 19:43 | <hober> | GPHemsley: we also have a gmail account that has the entire history of www-style for easier searching |
| 19:43 | <hober> | GPHemsley: details are on w3c-css-wg |
| 19:49 | <GPHemsley> | hober: I'm not subscribed to the list; I'm just talking about a web interface for searching the archive. |
| 19:50 | <GPHemsley> | because otherwise, if I use the regular search, it will attempt to weight results based on relevance |
| 19:51 | <GPHemsley> | when all I want is the existing archive format, but with a filter to only show the messages related to a particular spec |
| 19:51 | <hober> | yeah, that's a general problem with the w3c mailing list archive ui |
| 19:51 | <GPHemsley> | (where existing = sort by thread, subject, author, date, etc.) |
| 19:51 | <hober> | same problem with, say, public-webapps |
| 19:51 | Ms2ger | decides against replying to weird Glenn |
| 19:52 | <GPHemsley> | hober: It's probably a problem with mailman itself, no? |
| 19:53 | <GPHemsley> | for some definitions of "problem" |
| 19:57 | GPHemsley | files a feature request against Mailman |
| 20:11 | <GPHemsley> | https://bugs.launchpad.net/mailman/+bug/1082495 |
| 20:18 | <GPHemsley> | gee, I sure hope the W3C actually uses Mailman |
| 20:21 | <Ms2ger> | Hypermail? |
| 20:21 | <GPHemsley> | I have no idea |
| 20:22 | <GPHemsley> | but I just used the www-style mailing list in my examples in that Mailman bug |
| 20:22 | <GPHemsley> | shoot, you're right |
| 20:22 | <GPHemsley> | argh |
| 20:23 | <GPHemsley> | oh, hmm |
| 20:23 | <GPHemsley> | " Hypermail is a free (GPL) program to convert email from Unix mbox format to html. " |
| 20:23 | GPHemsley | is suddenly realizing he knows very little about how mailing lists and mailing list archives work |
| 20:35 | <GPHemsley> | "NOTE: The Apple specification for TrueType fonts allows for 'true' and 'typ1' for sfnt version. These version tags should not be used for fonts which contain OpenType tables." |
| 20:35 | <GPHemsley> | Mozilla does have code that mentions those, but it's not clear to me how far they go to support them. |
| 20:35 | <GPHemsley> | I feel like they should probably be included, though. |
| 20:35 | <GPHemsley> | (They are what differentiate a TrueType font from an OpenType font with TrueType outlines, IIUC.) |
| 22:02 | <Hixie> | anyone know off-hand who implements script crossorigin? |
| 22:02 | <Hixie> | aha, webkit and gecko, according to bz |
| 22:33 | <Hixie> | there's really no way to do the "inner shadow" effect here using canvas, huh http://i.imgur.com/Sy4xM.png |
| 22:35 | <zewt> | not that i've ever thought of |
| 22:36 | <zewt> | not sure what photoshop's algorithm is |
| 22:36 | <SimonSapin> | re subscribing to whatwg⊙wo: the confirmation message took a few hours, but eventually arrived |
| 22:36 | <Hixie> | SimonSapin: excellent |
| 22:36 | <Hixie> | zewt: presumably, draw shadow on top and clip to shape |
| 22:40 | <zewt> | well that's how real inner shadow is, i think, but isn't that inner glow rather than inner shadow |
| 22:42 | <zewt> | yeah it is |
| 22:42 | <zewt> | not a shadow at all |
| 22:46 | <Hixie> | what's the difference? |
| 22:46 | <zewt> | https://zewt.org/~glenn/shadow.png FWIW these are the photoshop effects by name, which is probably the most common way people know them |
| 22:46 | <zewt> | outer shadow just takes the image, blurs it by some amount, then offsets it by some amount; trivial to do |
| 22:47 | <Hixie> | what's the difference between shadow and glow other than the offset and blur radius? |
| 22:48 | <zewt> | hmm, i'm not quite sure, to be honest |
| 22:48 | <zewt> | most of the time i'm using them i'm just fiddling with knobs in photoshop and i haven't tried reverse engineering their effects |
| 22:49 | <zewt> | outer shadow does look similar to outer glow if i set distance (the offset) to 0 and increase the spread (blur amount) by a lot |
| 22:50 | <gsnedders> | The difference is to do with light bleeding. |
| 22:50 | <Hixie> | i don't know how to interpret those words in this context |
| 22:51 | <gsnedders> | The light blurs over the darker part, not the other way around. |
| 22:52 | <gsnedders> | Typically approximated using a Gaussian blur |
| 22:52 | <Hixie> | what light |
| 22:52 | <Hixie> | i'm very confused |
| 22:53 | <gsnedders> | http://en.wikipedia.org/wiki/Glow_(shader_effect) |
| 22:55 | <zewt> | pretty sure ps glow and shadow are only based on the opacity, not brightness |
| 22:55 | <gsnedders> | Dunno, I tend not to use such things myself. :) |
| 22:55 | <zewt> | (what gets blurred over what being the blend mode, which is orthogonal) |
| 22:58 | <SimonSapin> | I don’t know about canvas, but CSS has inner box-shadow |
| 22:59 | <SimonSapin> | it’s described as a gaussian blur |
| 23:00 | <zewt> | i guess the basic math is to invert the opacity, so you get a layer that's opaque where the original image is transparent; blur that layer by some amount; then blit the result onto the image, clipped by the original image's opacity |
| 23:01 | <zewt> | dunno why that wasn't intuitive to me before |
| 23:03 | <zewt> | maybe the only actual difference between shadow and blur in PS is exposing different knobs |
| 23:17 | <zewt> | https://zewt.org/~glenn/shadow2.png yeah, it's a shadow of the inversion of the layer, not just putting the shadow on top |
| 23:36 | <zewt> | i guess that's equivalent to just inverting the shadow after blurring (why is this unintuitive to me? it's not like i haven't implemented shadows and blurs a zillion times) |
| 23:55 | <Hixie> | oh! |
| 23:55 | <Hixie> | well then you _can_ do it on canvas |
