| 00:04 | <zewt> | i dream of gmail undo that doesn't lag behind by ten seconds when I hold ^Z for a second |
| 00:11 | <ubitux> | anyone has an example of a webvtt file with one or more "cue identifier"? |
| 00:12 | <ubitux> | i've a hard time figuring out what it is useful for, and i'm wondering if i'm not misunderstanding the spec |
| 00:13 | <zewt> | hmm, "not a descendant" isn't right, since that would include nodes in different trees |
| 00:14 | <Hixie> | ubitux: the cue identifier is the line before the line with the --> |
| 00:14 | <Hixie> | ubitux: it's not very useful except for styling the cue from CSS |
| 00:15 | <ubitux> | Hixie: right, so it's to keep compat with the number in the old subrip and first versions of the webvtt? |
| 00:15 | <ubitux> | oh you can access it from the css? fear. |
| 00:15 | <zewt> | (and from scripts, but standalones don't care about that) |
| 00:34 | <zewt> | anyone recall off-hand where the MessagePort GC issue was explained? (not sure if it was on a list or the BTS) |
| 04:32 | <Hixie> | aha, another checkin from the w3c html editors! this one removing the links to the whatwg spec that say where you can find roughly the same text under a more liberal license |
| 04:33 | <deane> | Just tell them you want it all back now. |
| 04:34 | <zewt> | Hixie: replaced with more /TR/ links, just to avoid confusion |
| 04:35 | <Hixie> | zewt: ? |
| 04:35 | <zewt> | jooooke |
| 04:38 | <Hixie> | ah |
| 04:38 | <Hixie> | didn't seem implausible :-) |
| 04:38 | <Hixie> | i was just confused cos i didn't see it in the diff :-) |
| 04:39 | <zewt> | the best satire--things that are completely absurd and completely plausible at the same time :) |
| 07:28 | jgraham | wonders what a "Matrix Maker" is, fears it is nothing good |
| 07:41 | <MikeSmith> | jgraham: yeah I wondered about that when I saw it too |
| 07:42 | <MikeSmith> | but I can't remember where I saw it |
| 07:42 | <MikeSmith> | bugmail? |
| 07:42 | <MikeSmith> | e-mail? |
| 07:42 | <jgraham> | coremob |
| 07:42 | <MikeSmith> | ah yeah |
| 07:43 | <MikeSmith> | darobin: so what is Matrix Maker? |
| 07:46 | <Ms2ger> | Is it darobin's new job description? |
| 07:47 | <darobin> | Matrix Maker? where? |
| 07:47 | <darobin> | oh, in the TF |
| 07:48 | <darobin> | it's a tool that allows you to list specs vertically, and device types or UAs horizontally, and shows you a pretty report of the passing stuff |
| 07:49 | <MikeSmith> | darobin: ah ok |
| 07:49 | <darobin> | ah, I knew it would be fun to select Julian for my first bug |
| 07:49 | <jgraham> | I don't like it when people start doing that/ |
| 07:50 | <jgraham> | It gives people all the wrong incentives for sontributing tests |
| 07:50 | <MikeSmith> | darobin: I thought you said "hug" there for a second |
| 07:50 | <jgraham> | *contributing |
| 07:50 | <darobin> | rofl |
| 07:50 | <MikeSmith> | Ms2ger: pretty sure his new title is "The Freshmaker" |
| 07:50 | <jgraham> | Maybe all he needs is a hug |
| 07:51 | <MikeSmith> | jgraham: what you mean about giving the wrong incentives? |
| 07:51 | <darobin> | mix me up with some coke! |
| 07:52 | <jgraham> | MikeSmith: I mean that people tend to start making tests in order to look good rather than in order to find bugs |
| 07:53 | <MikeSmith> | ah |
| 07:53 | <MikeSmith> | yeah |
| 07:53 | <darobin> | you only good if your tests catch problems with the other implementations |
| 07:53 | <jgraham> | Or with your own implementation :) |
| 07:54 | <MikeSmith> | I thought the goal was to construct your test suites such that your browser passes all the tests but other browsers tank on it |
| 07:54 | <darobin> | so long as everyone does that, we all win |
| 07:54 | <MikeSmith> | then you do a blog posting with a nice graphic showing that |
| 07:55 | <jgraham> | But there is a history of people cherry-picking tests to release based on "we get 100% and other people get 0% due to a single bug" |
| 07:55 | <jgraham> | That is, you intentionally find something that breaks in other browsers and make each test depend on it even if it isn't really relevant |
| 07:55 | <darobin> | speaking of tests and stuff, any suggestions on what would be most useful to cover at Test The Web Forward Paris are welcome |
| 07:56 | <jgraham> | Ms2ger, annevk Someone (TM) suggested writing DOM4 tests at TTWF. Does that sound sensible? |
| 07:57 | <annevk> | http://testsuite.org/ |
| 07:57 | <jgraham> | (actually I have no idea who the someone was) |
| 07:57 | <darobin> | I thought it was you jgraham :) |
| 07:57 | <jgraham> | darobin: I thought I just agreed that it might be a good idea |
| 07:57 | <darobin> | heh |
| 07:58 | <darobin> | "It wasn't me" |
| 07:59 | <jgraham> | annevk: Well yeah, but do you have any feeling for whether it is well tested or not? |
| 07:59 | <annevk> | jgraham: parts are well tested by tests from Ms2ger and AryehGregor |
| 07:59 | <annevk> | jgraham: those tests are somewhere in W3C dcvs |
| 07:59 | <annevk> | dvcs? |
| 07:59 | <jgraham> | OK, I guess I will need to look in more detail |
| 08:00 | <jgraham> | dcvs -> like cvs but with more "d" |
| 08:00 | <jgraham> | darobin: DOM APIs for form elements might also be good |
| 08:01 | <jgraham> | HTMLOptionCollection and so on |
| 08:01 | <darobin> | ooh yeah, that's a good one |
| 08:01 | <darobin> | I might actually take that one for my group |
| 08:01 | darobin | bbiab |
| 08:03 | <annevk> | Hixie: with seamless there's guaranteed sync access between the two trees right? |
| 08:04 | <annevk> | Hixie: why does the object need to be duplicated? |
| 08:04 | <annevk> | Hixie: why can't you just expand the dispatch tree |
| 08:04 | <annevk> | Hixie: as you've done for Window already |
| 08:10 | <smaug____> | annevk: mouseevent coordinates wouldn't make sense |
| 08:11 | <smaug____> | event retargeting can't be used with seamless, but there needs to be separate events per document |
| 08:11 | <annevk> | unless they are against the outer document when dispatched on a seamless frame |
| 08:12 | <smaug____> | (event target chain can be created before dispatching either one) |
| 08:12 | <annevk> | but I guess now browsers have seamless without this event model it's too late for that |
| 08:12 | <smaug____> | I don't think seamless is in any way stable thing yet |
| 08:12 | <Ms2ger> | jgraham, events, maybe |
| 08:13 | <annevk> | jgraham: actually not, dvcs.w3.org |
| 08:13 | <annevk> | jgraham: oh, maybe you were joking :) |
| 08:14 | <Ms2ger> | I was thinking about boxes with links to the tests in the spec, btw |
| 08:20 | <jgraham> | Little boxes in the margin, little boxes filled with links-to-tests? |
| 08:20 | <annevk> | I need to download some new Weeds episodes |
| 08:20 | <jgraham> | Ms2ger: But seriously, please continue. I want something like that too |
| 08:20 | <Ms2ger> | Yeah, something like that |
| 08:20 | <annevk> | Ms2ger: we should Hixie to abstract his system |
| 08:21 | <annevk> | Ms2ger: get*; so we can use it for DOM / Encoding / Fullscreen / ... |
| 08:21 | <Ms2ger> | I wrote a script to pull out the rel=help links from tests, and put the data into a json file |
| 08:21 | <jgraham> | The main problem with links to tests is working out which test should be linked where |
| 08:21 | <Ms2ger> | Now I just need to coerce someone into doing the frontend :) |
| 08:23 | <jgraham> | I don't like rel=help |
| 08:24 | <jgraham> | For a start it doesn't work |
| 08:24 | <jgraham> | Because a test is not a file |
| 08:24 | <jgraham> | e.g. the html5lib tests run in 3 different modes depending on the query string |
| 08:25 | <Ms2ger> | Let's leave the html5lib tests out of this ;) |
| 08:25 | <jgraham> | Let's not |
| 08:25 | <jgraham> | Making tests that depend on the query string isn't exactly uncommon |
| 08:27 | <jgraham> | Opera seem to have ~4500 tests that require a query string (and ~20 that require a fragment id) |
| 08:28 | <Ms2ger> | OH: "Do you have a use case that does not involve the vanity of the library's authors?" |
| 08:29 | <Ms2ger> | jgraham, do those usually test different features when used with different query strings? |
| 08:31 | Ms2ger | goes off for a bit |
| 08:31 | <jgraham> | No idea |
| 08:32 | <jgraham> | In any case my point is that linking tests to files is the wrong data model |
| 08:32 | <jgraham> | Tests are URIs |
| 08:33 | <jgraham> | You could of course make each test that can be loaded with > 1 URI has to list that in the test file, or something |
| 08:33 | <smaug____> | I need chromium nightly. where can I find the .zip or .tar.gz file for linux |
| 08:33 | <smaug____> | (always difficult to find the nightly chromium build for linux) |
| 08:38 | <smaug____> | hmm, now I found it, but it doesn't run on this machine |
| 08:48 | <annevk> | Firefox does not allow dynamically setting the base URL? |
| 08:48 | <annevk> | That makes testing annoying |
| 08:55 | <annevk> | fwiw |
| 08:55 | <annevk> | there is no interop for <base href=about:blank><script src=local.js></script> |
| 08:56 | <annevk> | most browsers load local.js relative to the current file, Gecko does not |
| 08:57 | <annevk> | with a twist of course, Firefox view source resolves it relative too |
| 08:57 | <annevk> | that's some confusing mismatch |
| 09:02 | <darobin> | does Firefox's view source respect <base> in general? |
| 09:03 | <annevk> | not for http://example.org/ either |
| 09:03 | <annevk> | so yeah, might be a more generic bug |
| 09:07 | <smaug____> | annevk: file a bug, please. CC hsivonen |
| 09:07 | <smaug____> | (although this might be a UI problem) |
| 09:07 | <annevk> | I suspect this to be filed already |
| 09:10 | <smaug____> | annevk: ah, perhaps https://bugzilla.mozilla.org/show_bug.cgi?id=617789 |
| 09:19 | <annevk> | looks like it |
| 09:19 | <annevk> | dynamic base is prolly filed too |
| 09:31 | <annevk> | heh, hsivonen is on air: http://futuradio.fi |
| 09:36 | <jgraham> | The volume control on the radio image doesn't actually control the volume :( |
| 09:39 | <darobin> | that's how the future of radio is! |
| 09:40 | <Ms2ger> | Clearly we need a volume content attribute |
| 09:42 | <jgraham> | darobin: Excessively loud? |
| 09:43 | <Ms2ger> | jgraham, no, that's the present |
| 09:43 | <darobin> | jgraham: yeah, you know, Finnish radio is special |
| 09:43 | <darobin> | I think they still have the one and only channel broadcasting in Latin |
| 09:44 | <darobin> | they are also known for having broadcasted computer programs over FM radio (you could tape it, and then just load the tape) |
| 09:44 | <darobin> | plus, they have hsivonen talking about browsers and mobile networks |
| 09:45 | <darobin> | he's making me all sad about the Android update situation |
| 09:50 | <smaug____> | darobin: it is not a channel, but just one program in Latin |
| 09:50 | <smaug____> | basically news in Latin |
| 09:51 | <darobin> | smaug____: my version of reality sounds better |
| 09:51 | <smaug____> | sure |
| 09:51 | <smaug____> | :) |
| 09:52 | <darobin> | next you're going to tell me it's just one program with hsivonen and not a whole 24/7 broadcast of his thoughts |
| 09:52 | <smaug____> | no no, hsivonen has his own channel and talking 24/7/365 |
| 09:53 | <jgraham> | Oh, I assumed he had some sort of helmet to read his thoughts in realtime |
| 10:48 | <deane> | I wish I had heard all of that interview with hsivonen. |
| 11:25 | <annevk> | test |
| 11:26 | <Ms2ger> | failed |
| 11:26 | <annevk> | there's red all over the floor |
| 11:26 | <annevk> | aaaah |
| 11:26 | <annevk> | you're being eaten by a grue |
| 11:27 | <Ms2ger|waseatenb> | Dammit |
| 11:27 | <annevk> | haha |
| 11:29 | <jgraham> | I wish hsivonen had done the whole interview in LAtin |
| 11:49 | <annevk> | Chrome has something weird |
| 11:49 | <annevk> | about:blank as base with "unknown :test" as link creates "about:unknown%20:test" |
| 11:50 | <annevk> | no other browser tries to resolve URLs against about:blank |
| 11:52 | <jgraham> | Yeah, resolving against about:blank seems weird |
| 11:56 | <annevk> | kinda wish URLs were simpler |
| 11:56 | <annevk> | but then somebody might have defined them already I guess |
| 12:02 | <jgraham> | Ms2ger: (I just came up with another test that can run with a fragment ID on the URL) |
| 12:02 | <jgraham> | (or not, but to test slightly different things) |
| 12:02 | <annevk> | heh, in Gecko a data URL is only valid if it matches the whole data URL production |
| 12:02 | <jgraham> | (same part of the spec still) |
| 12:02 | <jgraham> | (except there is no spec yet) |
| 12:03 | <Ms2ger> | annevk, that sounds... sane? |
| 12:04 | <Ms2ger> | Though Julian did something with base64 and ; |
| 12:04 | <annevk> | Ms2ger: well it means data URL parsing becomes part of URL parsing |
| 12:04 | <annevk> | Ms2ger: if you want every URL library to behave the same way anyway |
| 12:04 | <Ms2ger> | Pff ;) |
| 12:05 | <annevk> | I mean if you stop caring about the details all of this is very simple |
| 12:05 | <annevk> | but if you do care about whether "DATA:test" becomes normalized and of which .protocol returns "data:" rather than the empty string (that's for invalid) you suddenly need to define this |
| 12:06 | <annevk> | well |
| 12:06 | <annevk> | invalid might also return ":" |
| 12:06 | <darobin> | piffle |
| 12:06 | <annevk> | browsers are in dispute of course |
| 12:06 | <darobin> | only quiche eaters process URLs with an API |
| 12:06 | <Ms2ger> | I only care about browsers :) |
| 12:07 | <annevk> | darobin: just had some quiche |
| 12:07 | <darobin> | real hackers do it with a twenty page long regex |
| 12:07 | <darobin> | my point exactly |
| 12:07 | <annevk> | and I'm using a 200 line incomplete tokenizer |
| 12:07 | <darobin> | bah, kids these days |
| 12:09 | <annevk> | so I take it URLs becomes a "this is implementation-dependent" in "HTML5"? |
| 12:11 | <Ms2ger> | No |
| 12:11 | <Ms2ger> | They become IRAs |
| 12:11 | <Ms2ger> | ARIs? |
| 12:11 | <Ms2ger> | ARAs? |
| 12:11 | <darobin> | URX I think |
| 12:12 | <annevk> | so ™ in the authority part becomes normalized to tm... |
| 12:13 | <annevk> | if somebody could explain authority parsing to me that would be awesome |
| 12:13 | <darobin> | heh, I wonder how well http://search.cpan.org/~gaas/URI-1.60/URI.pm#PARSING_URIs_WITH_REGEXP actually fares in the real world |
| 12:13 | <annevk> | unfortunately that also seems to differ wildly |
| 12:14 | <annevk> | e.g. "http:test test" yields invalid in Gecko/Safari, "http://test%20test/"; in Chrome |
| 12:15 | <annevk> | maybe invalid is okay |
| 12:15 | <darobin> | you'd sort of hope so |
| 12:18 | <annevk> | well |
| 12:19 | <annevk> | http:annevankesteren.nl "works" |
| 12:19 | <annevk> | in most browsers |
| 12:19 | <annevk> | giving http://annevankesteren.nl/ |
| 12:19 | <annevk> | unless |
| 12:19 | <annevk> | there's a base URL |
| 12:19 | <annevk> | in which case it's a relative path and you get something like http://example.org/annevankesteren.nl (scheme has to match) |
| 12:20 | <annevk> | but that part I got covered already |
| 12:20 | <annevk> | I'm wondering more about what characters to allow and to reject in host |
| 12:20 | <annevk> | well, authority |
| 12:29 | <annevk> | man |
| 12:30 | <annevk> | http:test -> http://test/ in Safari, but http:maraña.org is invalid, only works with // |
| 12:30 | <annevk> | Chrome's URL code is a little better there |
| 12:30 | <annevk> | Gecko too, but doesn't use punycode |
| 12:30 | <annevk> | because, why would anything here match each other... |
| 12:33 | <zcorpan> | annevk++ for speccing this |
| 12:34 | <Ms2ger> | annevk++ just because |
| 12:40 | <jgraham> | annevk++ to increase hit points agianst attacks from RFC-wonks |
| 12:40 | <karlcow> | annevk-- for destroying Poney and Rainbow tales ;) |
| 12:42 | <darobin> | annevk: re the HTML5 thing with MikeSmith, are you only doing it in .nl or are you travelling a bit too? |
| 12:42 | <darobin> | specifically, to the Paris one |
| 12:58 | <Ms2ger> | http://i.imgur.com/jxBZG.jpg |
| 13:11 | <zcorpan> | aaah, the browser spec. https://www.w3.org/Bugs/Public/show_bug.cgi?id=14689 |
| 13:12 | <zcorpan> | gotta monkey patch xslt in the browser spec |
| 13:18 | <Ms2ger> | There is no scope, dammit |
| 13:19 | Ms2ger | kicks zcorpan out of the secret tree house |
| 13:21 | <zcorpan> | i just find it amusing that the *xml core wg* who refused to have me specify this in the *xml-stylesheet PI spec* think that it is most appropriate to spec it in HTML |
| 13:21 | <Ms2ger> | "We don't have to deal with it" |
| 13:26 | <zcorpan> | (UI requirements for the xml-stylesheet PI is still absent, except i think CSSOM has rules for CSS) |
| 13:38 | <annevk> | darobin: I do not know about a Paris one, so presumably just here |
| 13:38 | <darobin> | annevk: ah, shame, it could have involved beer |
| 13:39 | <annevk> | darobin: I also only plan to put travel money into vacation for now |
| 13:39 | <darobin> | or wine, since they seem to be paying for wine and pizza |
| 13:39 | <annevk> | does kinda sound like vacation :p |
| 13:39 | <darobin> | you're going to take even more vacations??? ;-) |
| 13:40 | <Ms2ger> | Oh dear |
| 13:48 | <annevk> | haha, no, unfortunately not |
| 13:48 | <annevk> | apart from some weekends |
| 13:51 | <annevk> | so in Gecko http://\0example.org/ yields http://example.org/ |
| 13:51 | <annevk> | is that safe? |
| 13:53 | <annevk> | Safari treats the zero byte as eof |
| 13:53 | <annevk> | same for Opera |
| 13:54 | <annevk> | Chrome converts to %00 |
| 13:58 | <zcorpan> | %00 or treating the whole url as invalid seem like the sane options |
| 13:59 | <annevk> | during authority -> invalid otherwise %00 is what I'm going with I guess |
| 13:59 | <annevk> | in the end |
| 14:00 | <zcorpan> | sounds good |
| 14:01 | <annevk> | it seems in the host name you need to unescape %.. too |
| 14:01 | <annevk> | in the path only Chrome does it |
| 14:01 | <annevk> | which I think might result in some bugs |
| 14:02 | <annevk> | some sites expect paths to be pretty much literals |
| 14:13 | <annevk> | hsivonen: why does Gecko give the unbalanced tree message if all I write is <br>? |
| 14:13 | <darobin> | there's behaviour that I can't find to be specified anywhere, I'd appreciate a pointer if any |
| 14:13 | <annevk> | ask, don't ask to ask |
| 14:14 | <darobin> | I wasn't asking to ask, I was introducing my question :) |
| 14:14 | <darobin> | if you serve something as XML containing <foo><p>foo</p></foo>, it consistently gets rendered as a source tree |
| 14:14 | <darobin> | if however you add the namespace to an inner element no matter how deep, as in <foo><p xmlns='http://www.w3.org/1999/xhtml'>foo</p></foo>; |
| 14:14 | <darobin> | it gets upgraded to an HTML document |
| 14:15 | <darobin> | (even if the namespace happens after megabytes of useless XML |
| 14:15 | <darobin> | everyone seems to do the same thing (at least given a cursory look) but I can't find it specified |
| 14:15 | <annevk> | if http://www.whatwg.org/specs/web-apps/current-work/multipage/history.html#read-xml does not have it, prolly a bug |
| 14:16 | <darobin> | yeah, that's what I was looking at |
| 14:16 | <annevk> | or maybe that's more about rendering XML |
| 14:16 | <annevk> | dunno if we have a section for that |
| 14:17 | <darobin> | I investigated because of https://www.w3.org/Bugs/Public/show_bug.cgi?id=15180 |
| 14:18 | <darobin> | I don't believe there's anything about rendering XML |
| 14:18 | <darobin> | I'll figure something out, cheers |
| 14:19 | <jgraham> | Is there a shorter term for "documents without a browsing context"? |
| 14:21 | <annevk> | cookie-averse Document object, though that also means something else |
| 14:27 | <annevk> | man, IDNA is some vague shit |
| 14:27 | <jgraham> | I was hoping for "detached documents" or something. |
| 14:29 | <zewt> | "contextless document" seems accurate, but ... it's sort of a tonguetwister, heh |
| 14:30 | <darobin> | it's really funny, <foo><zorglub xmlns='http://www.w3.org/1999/xhtml'>foo</zorglub></foo>; and even <foo><x:zorglub xmlns:x='http://www.w3.org/1999/xhtml'>foo</x:zorglub></foo>; work, but the namespace declaration on its own is not enough |
| 14:31 | <darobin> | jgraham: how about aloof documents |
| 14:32 | <annevk> | darobin: if you want to define this I would just say that the document needs to be rendered per CSS rules if there's at least one element in the HTML namespace |
| 14:32 | <annevk> | darobin: and maybe MathML namespace |
| 14:32 | <darobin> | yeah, that's what I was thinking of |
| 14:32 | <darobin> | lemme do a few more tests |
| 14:34 | <darobin> | works for MathML and SVG |
| 14:35 | <annevk> | I guess SVG gives you a CSS context as well |
| 14:36 | <annevk> | be careful not to make Rendering and XML Rendering sound exclusive |
| 14:41 | <darobin> | I'll put it on GitHub, nothing's exclusive there! |
| 14:44 | <annevk> | this channel did need more non sequitur |
| 14:44 | <annevk> | or does |
| 14:44 | <annevk> | hmm |
| 14:47 | <smaug____> | jgraham: just curious, and I think the answer is that you aren't allowed to say anything... how does Opera handle cycles in C++ objects |
| 14:47 | <smaug____> | well, C++ -> JS ->C++ cycles |
| 14:56 | <jgraham> | smaug____: The answer is honestly that I don't know |
| 14:56 | <jgraham> | I just asked if that kind of a cycle was a problem for us ans was told by the relevant developers that it wasn't |
| 14:57 | <annevk> | zero bytes on the other hand :p |
| 14:57 | <jgraham> | Yeah, excuse me while I list the zero websites that affects :p |
| 14:58 | <annevk> | well I have this file here on my local hard drive... |
| 14:58 | <jgraham> | Well, sometimes it is annoying |
| 14:59 | <jgraham> | I wrote a fuzzer once where it truncated the debug output |
| 14:59 | <jgraham> | But it's not like it makes sense to prioritise fixing it |
| 15:34 | <Hixie> | darobin: can you elaborate on your reasoning for https://www.w3.org/Bugs/Public/show_bug.cgi?id=18067 ? I was thinking about it myself but wasn't convinced it wouldn't be a security problem. Can you elaborate on why you think it's safe? |
| 15:35 | <darobin> | Hixie: I looked at the gopher protocol and the sort of content that it supports, and I could find no reason to believe that it would be more unsafe than any of the others |
| 15:36 | <darobin> | if there are rules for assessing this more systematically, I'd be happy to hear |
| 15:36 | <Hixie> | wouldn't it have the same risks as http:? |
| 15:36 | jgraham | checks his watch |
| 15:36 | <jgraham> | No, doesn't seem to be 1992 |
| 15:37 | <darobin> | you mean highjacking all those gopher clients out there? |
| 15:37 | <Hixie> | jgraham: yeah, i'm not saying it would be useful in any way in the first place, but... |
| 15:37 | <zcorpan> | how is one supposed to discover which *.specs.whatwg.org exist? |
| 15:37 | <Hixie> | zcorpan: one is currently not, sadly |
| 15:38 | <Hixie> | darobin: well not just that, but in theory, e.g. in-place mitm in an iframe, or similar |
| 15:38 | <jgraham> | Hixie: "i'm not saying it would be useful in any way" no usecases -> not allowed, seems more reasonable |
| 15:38 | <Hixie> | darobin: to the extent that anyone supports it, it seems bad to support it inline |
| 15:39 | <Hixie> | jgraham: yeah, that's kinda my thinking |
| 15:40 | <darobin> | frankly I couldn't think of any reasonably mitm that would do anything useful — at the same time I can see this as useful for whoever wants to maintain historical gopher connectivity to implement it that way |
| 15:42 | <Hixie> | to the extent that it can be useful, it seems like it could be used in an attack. |
| 15:42 | <Hixie> | to the extent that it's not useful, it seems pointless to support. |
| 15:42 | <darobin> | the UC that I have in mind is exposing historical document retrieval systems |
| 15:44 | <darobin> | that said I can't say I am overwhelmed with desire for gopher, so I won't stand in the way of reversal |
| 15:44 | <Hixie> | i think the whatwg spec will be erring on teh side of abundant caution on this one, fwiw |
| 15:45 | darobin | has a fleeting image of an "I'll die for gopher" t-shirt |
| 15:46 | <annevk> | gopher is one of the schemes at least WebKit-based browsers use to trigger their relative URL handling |
| 15:46 | <jgraham> | If the usecase is "historical curiosity" it doesn't seem like integration with the web is important so you could require urls to be rewritten as web+gopher (and do so automatically where you have access to the URL) |
| 15:46 | <Hixie> | yeah |
| 15:46 | <annevk> | gopher implies a hierarchical URL, so to say |
| 15:47 | <darobin> | can one of the logbots here give a URL pointer to the logs? |
| 15:48 | <darobin> | that's a good point jgraham |
| 15:48 | <Ms2ger> | http://krijnhoetmer.nl/irc-logs/whatwg |
| 15:48 | <hober> | darobin: see the topic |
| 15:48 | <Ms2ger> | Also, now I need to review D3E again |
| 15:48 | <darobin> | no, that much I know, I meant a pointer to the current entry |
| 15:48 | <darobin> | (but I'll live without) |
| 15:48 | <jgraham> | At least krijn isn't really a logbot |
| 15:49 | <jgraham> | It's more like a radiator that also saves IRC conversations |
| 15:49 | <darobin> | unless you count mIRC as a logbot :) |
| 15:51 | <jgraham> | In a conversation about gopher mIRC is almost modern |
| 15:52 | <Hixie> | bbiab |
| 15:52 | <Ms2ger> | Zing |
| 15:53 | <karlcow> | Go for ever |
| 15:53 | <jgraham> | You want Hixie killed? |
| 15:53 | <Ms2ger> | No, just gone |
| 15:53 | <jgraham> | Oh, not "gone" |
| 16:11 | <darobin> | Hixie: re https://github.com/w3c/html/commit/c953c8da7b2363af2ec119f14c9383f0cee1b13b it confused someone, I could see how "this" may be slightly unclear, I figured that a two word non-normative change that had confused someone wouldn't hurt |
| 16:16 | <dglazkov> | good morning, Whatwg! |
| 16:31 | <say2joe> | good morning, Dimitri (and fellow west coaster) |
| 16:39 | <Ms2ger> | Hah: https://github.com/apache/httpd/commit/a381ff35fa4d50a5f7b9f64300dfd98859dee8d0 |
| 16:39 | darobin | thinks you'd have to be mad to steal a cookie from a gopher |
| 16:39 | <Hixie> | darobin: you'll notice i made the change also :-) |
| 16:39 | <darobin> | Hixie: yes, I was just reacting to your comment :) |
| 16:40 | <Hixie> | darobin: there are definitely cases (not this particular one) where people are like "this is confusing" and despite my best efforts at working out why, i end up concluding it's just because they were tired or don't know how to read |
| 16:40 | <darobin> | yeah, I clearly saw a few of those in the bugs |
| 16:40 | <say2joe> | lol |
| 16:41 | <darobin> | in this case, I could see how it may be confusing (if I tilted my head right) |
| 16:41 | <darobin> | I sometimes wish the comment thing had a name field |
| 16:43 | <say2joe> | In Chrome Developer Tools, I'm using the Resources panel on the LATimes.com web site. I've got about 60 cookies from that site (and others) being utilized on LATimes.com — if I select all of them and try to delete them, nothing happens. However, I can remove 1-2 at a time it seems… is anyone here familiar with this being a bug or simply the way it works for some reason? |
| 16:49 | <Hixie> | darobin: if they're logged in on the spec, it includes their e-mail address on the line with the IP address |
| 16:50 | <karlcow> | Ms2ger this is all insane the apache thing |
| 16:50 | <darobin> | Hixie: yeah I saw, but the weirder ones aren't the logged in ones ;) |
| 16:50 | <Hixie> | yeah |
| 16:51 | <Ms2ger> | karlcow, how about we call it "the IE thing" ;) |
| 16:51 | <karlcow> | not going into that discussion again :) |
| 16:51 | <Ms2ger> | Me neither |
| 16:51 | <Hixie> | darobin: it's a natural trade-off between reducing the barrier to getting feedback, and getting more info |
| 16:52 | <Hixie> | darobin: the bugs filed on the w3c copy require that they get a login, so you get a name there (but less feedback overall) |
| 16:52 | <darobin> | I know, I wasn't suggesting requiring logins for all, au contraire |
| 16:52 | <darobin> | I was just wondering if adding a name field there might make sense — just mulling it over |
| 16:53 | <Ms2ger> | darobin, hey, shouldn't you be removing references to the WHATWG from HTML5 instead of hanging around here? :) |
| 16:55 | <darobin> | Ms2ger: nah, that was fun while outside W3C |
| 16:56 | <darobin> | now that I'm on the inside, I seem to mostly focus on pissing off XML Core WG and I18N |
| 16:56 | <Ms2ger> | Not a11y? |
| 16:56 | <darobin> | haven't gotten around to that yet :-) |
| 16:57 | Ms2ger | prepares some popcorn |
| 16:57 | <othermaciej> | I've never seen them in a non-pissed-off state so I'm not sure how you would know the difference |
| 16:58 | <Ms2ger> | Zing |
| 17:00 | <Hixie> | ouch |
| 17:03 | <Ms2ger> | Hixie, you got your HTMLCollection test wrong, btw |
| 17:03 | <Hixie> | no surprise there |
| 17:03 | <Hixie> | what specifically did i get wrong? |
| 17:03 | <Ms2ger> | ToUint32('test') == 0 |
| 17:03 | <Ms2ger> | So it didn't prove anything |
| 17:04 | <Ms2ger> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=18789 |
| 17:04 | <Hixie> | d'oh |
| 17:05 | <Hixie> | i want strict typing damnit |
| 17:05 | <Hixie> | anyway |
| 17:05 | <Hixie> | bbl again |
| 17:06 | <darobin> | I find that a spellchecker can help lots with strict typing |
| 17:06 | darobin | gets his coat |
| 17:07 | Ms2ger | kicks darobin out |
| 17:11 | <annevk> | miketaylr: fullscreen is a single word |
| 17:11 | <annevk> | miketaylr: I blame Mozilla for realizing that too late |
| 17:12 | <miketaylr> | annevk: yeah my intuition says that too |
| 17:12 | <annevk> | miketaylr: in a couple of years from now people will be thankful not to have to write full-dash-screen |
| 17:12 | <miketaylr> | too lazy to look it up |
| 17:12 | <annevk> | well it used to be full screen |
| 17:12 | <annevk> | then full-screen, and nowish fullscreen |
| 17:12 | <annevk> | the English language sucks that way, so you better just remove that hyphen right away |
| 17:14 | <miketaylr> | can't wait to find all the broken sites where people create an analogous requestFullScreen from the prefixed guys |
| 17:14 | <miketaylr> | (html5test is the first i know of) |
| 17:14 | <paul_irish> | miketaylr: got spec link handy? |
| 17:14 | <paul_irish> | or list discussion? |
| 17:15 | <miketaylr> | paul_irish: here's a PR w/ a spec link inside https://github.com/NielsLeenheer/html5test/pull/196 |
| 17:16 | <paul_irish> | thx |
| 17:18 | <annevk> | paul_irish: spec is hosted here for now: http://dvcs.w3.org/hg/fullscreen/raw-file/tip/Overview.html |
| 17:19 | <annevk> | will prolly be fullscreen.spec.whatwg.org at some point |
| 17:19 | <annevk> | have to sort all that out still |
| 17:21 | <paul_irish> | oh man that's a hot URL. looking forward to it |
| 17:22 | <paul_irish> | annevk: im going to update the MDN and html5rocks tutorials with their casing and point this out more clearly. |
| 17:24 | <annevk> | thanks paul_irish! |
| 17:25 | <miketaylr> | paul_irish++ |
| 18:25 | <annevk> | very slowly approaching more followers than commits on @WHATWG |
| 18:25 | <Hixie> | oh dear |
| 18:25 | <Hixie> | i'd better commit more |
| 18:25 | <annevk> | per character commits! |
| 18:26 | <Hixie> | actually since i don't have to deal with the htmlwg process any more i've been grouping typo commits again |
| 18:28 | <Hixie> | sicking: ping |
| 19:34 | <Hixie> | hober: looks like some bugs in the html5 spec component are getting assigned to me? |
| 19:34 | <Hixie> | not sure why |
| 20:21 | <sicking> | Hixie: pong |
| 21:03 | <hober> | Hixie: link(s)? |
| 21:04 | <hober> | Hixie: nvm, i see the mail on p-h-b |
| 21:33 | <zcorpan> | Hixie: maybe the live dom viewer bug thing should have its own bugzilla component? also, maybe prompt() for a description? |
| 21:36 | <TabAtkins> | Hixie: Why is the "DOM view" link a data:text/plain, when its contents are clearly HTML? |
| 21:37 | <TabAtkins> | Also: it would be nice to have the data url include the styles that it's normally rendered with. |
| 21:38 | <zcorpan> | hmmmmmmmmmmmmmmmmmmm. http://www.w3.org/community/texttracks/wiki/608_to_WebVTT |
| 22:07 | <zcorpan> | the style sheet there also has bogus selectors |
| 22:16 | <TabAtkins> | Indeed - <track> is never rendered, and ::cue-part() should hang off the <video> element instead. |
| 22:36 | <zcorpan> | and the /* default values */ should probably use ::cue, but it also includes properties that don't apply to cues |
| 23:12 | <Hixie> | TabAtkins: i use it to copy and paste the source into the spec |
| 23:12 | <Hixie> | zcorpan: yeah, prompting for a desc might make sense. is it still being abused? |
| 23:12 | <Hixie> | if not, then people can just put the desc in a comment... |
| 23:12 | <TabAtkins> | Hixie: Ah, kk. |
| 23:13 | <Hixie> | sicking: archives api -- is anyone taking care of speccing that? it was raised in the whatwg list and i want to make sure i'm not ignoring something that was intended to be work for me |
| 23:13 | <Hixie> | d'oh |
| 23:13 | <zcorpan> | Hixie: i don't know if it's abused, but apparently they don't have descs in a comment, so it's hard to tell if it's spam or worthwhile or what it's about at all |
| 23:13 | <zewt> | "run, coward, run" |
| 23:13 | <zewt> | Hixie: (you mean that ZIP API?) |
| 23:14 | <sicking> | Hixie: no one is writing drafts for it right now |
| 23:14 | <zcorpan> | Hixie: the other concern was that most seem to be about css browser bugs, not html browser bugs |
| 23:14 | <sicking> | Hixie: we've been trying to get input from authors, but so far not very successfully |
| 23:14 | sicking | looks at paul_irish ;-) |
| 23:14 | <zcorpan> | Hixie: i was just reading minutes, i haven't really looked at the bugs myself |
| 23:15 | <sicking> | Hixie: i wasn't counting on you writing it up though. But if you are so inclined that'd be lovely :) |
| 23:17 | <Hixie> | zcorpan: yeah, true |
| 23:18 | <Hixie> | sicking: i wish i had the bandwidth |
| 23:18 | <Hixie> | sicking: basically i'm trying to work out if i should ignore the thread as if there was another editor tracking that feedback, or reply with "that's nice, but i don't have the bandwidth to do it now" |
| 23:18 | <Hixie> | zcorpan: k |
| 23:19 | <Hixie> | zcorpan: i'm happy to change it |
| 23:19 | <Hixie> | zcorpan: ideally not right now though |
| 23:19 | <Hixie> | send mail? |
| 23:19 | <sicking> | Hixie: up to you. I wasn't expecting you to do it, so no comment needed |
| 23:20 | <Hixie> | k. |
| 23:20 | <Hixie> | i shall pretend it never happened :-) |
| 23:22 | <zcorpan> | sent |
