| 02:17 | <MikeSmith> | Hixie: does the spec anywhere actually state document-conformance requirements on what a valid e-mail address is a source document? |
| 02:17 | <MikeSmith> | unparsed document |
| 02:18 | <MikeSmith> | well, not unparsed exactly |
| 02:18 | <MikeSmith> | but you know what I mean |
| 02:19 | <MikeSmith> | that is, requirements for authors/documents as opposed to UA requirements |
| 03:02 | <MikeSmith> | good golly the e-mail address regexp twitter talk is still going |
| 03:03 | <MikeSmith> | that tweet should have stating things more precisely |
| 03:03 | <MikeSmith> | b |
| 03:03 | <MikeSmith> | but hard to do in a 140 characters |
| 03:04 | <MikeSmith> | especially when 70 of them are a regular expression patterm |
| 03:06 | <MikeSmith> | oh, paul_irish retweeted it just a little while ago |
| 03:07 | MikeSmith | checks to see if http://html5.org/tools/web-apps-tracker is surviving |
| 03:10 | <Hixie> | miketaylr: btw the best thing about that tweet is that it's wrong ;-) |
| 03:10 | <Hixie> | er |
| 03:11 | <Hixie> | s/miketaylr/MikeSmith/ |
| 03:11 | <Hixie> | but i see he's left |
| 03:11 | <Hixie> | oh well |
| 06:26 | <paul_irish> | Yeah. realized that late, as well. :/ /me retracts retweet |
| 06:27 | <Hixie> | a stupid mistake on my part, sadly |
| 06:27 | <Hixie> | gerv pointed it out to me in e-mail |
| 06:30 | <abarth> | Hixie: quick question about https://bugs.webkit.org/show_bug.cgi?id=57185 . should we be firing the drop event in content editable? /me is looking through the spec now, but I thought you might know off the top of your head |
| 06:30 | <Hixie> | can't see why we wouldn't... |
| 06:30 | <Hixie> | though scripts might be disabled entirely? |
| 06:30 | <Hixie> | i forget what the state of that stuff is |
| 06:31 | <Hixie> | AryehGregor, do you know? |
| 06:31 | <abarth> | the bug claims the spec used to not fire the event |
| 06:31 | <abarth> | but IE and Firefox do |
| 06:31 | <rniwa> | abarth: I think the spec doesn't mention this case |
| 06:32 | <Hixie> | the spec is trying to copy IE, so if the spec really says that it,s a bug, especially if Firefox does it |
| 06:32 | <abarth> | ok |
| 06:32 | <Hixie> | if it doesn't say that an event doesn't fire, then the place where it says the event does fire would mean it does |
| 06:32 | <abarth> | thanks |
| 06:32 | <Hixie> | i'd ask the reporter for a spec citation |
| 06:32 | <rniwa> | Hixie: I think the spec is currently lacking that case |
| 06:32 | <Hixie> | the spec has a fallback for all cases, no? |
| 06:32 | <Hixie> | so it can't lack a case |
| 06:32 | <abarth> | that's pretty elliptical, but i'll assume you mean we should try to fix webkit to match IE and Firefox :) |
| 06:33 | <Hixie> | abarth: well we should fix everything to match, i don't really mind if it means fixing IE, webkit, firefox or the spec, or some combination thereof |
| 06:34 | <Hixie> | abarth: i'm just saying that if the spec says "in case 1 do X, otherwise do Y", then that means that in case 2, you do Y, even though case 2 isn't mentioned explicitly |
| 06:35 | <abarth> | understood. thanks |
| 06:36 | <rniwa> | oops, I think the spec does say we should fire drop event in this case |
| 06:36 | <Hixie> | oh? |
| 06:36 | <rniwa> | 8.6.5 |
| 06:36 | <rniwa> | Hixie, abarth: http://www.whatwg.org/specs/web-apps/current-work/multipage/dnd.html#dragevent |
| 06:36 | <Hixie> | that's a big section |
| 06:36 | <Hixie> | cite the text :-) |
| 06:37 | <Hixie> | quote the text, rather |
| 06:37 | <rniwa> | If the current drag operation is "none" (no drag operation), or, if the user ended the drag-and-drop operation by canceling it (e.g. by hitting the Escape key), or if the current target element is null, then the drag operation failed. Run these substeps: |
| 06:37 | <rniwa> | Otherwise, the drag operation might be a success; run these substeps: |
| 06:38 | <rniwa> | Let dropped be true. |
| 06:38 | <rniwa> | If the current target element is a DOM element, fire a DND event named drop at it; otherwise, use platform-specific conventions for indicating a drop. |
| 06:38 | <rniwa> | If the event is canceled, set the current drag operation to the value of the dropEffect attribute of the DragEvent object's dataTransfer object as it stood after the event dispatch finished. |
| 06:38 | <rniwa> | Otherwise, the event is not canceled; perform the event's default action, which depends on the exact target as follows: |
| 06:38 | <rniwa> | If the current target element is a text field (e.g. textarea, or an input element whose type attribute is in the Text state) or an editing host or editable element, and the drag data store item list has an item with the drag data item type string "text/plain" and the drag data item kind Plain Unicode string |
| 06:38 | <rniwa> | Insert the actual data of the first item in the drag data store item list to have a drag data item type string of "text/plain" and a drag data item kind that is Plain Unicode string into the text field or editing host or editable element in a manner consistent with platform-specific conventions (e.g. inserting it at the current mouse cursor position, or inserting it at the end of the field). |
| 06:38 | <rniwa> | Otherwise |
| 06:38 | <rniwa> | Reset the current drag operation to "none". |
| 06:38 | <rniwa> | abarth, Hixie: so drop event must fire immediately before we modify the text in the editable region |
| 06:38 | <Hixie> | i meant quote the specific text you think means something should happen with an editable element, not the whole thing :-P |
| 06:38 | <rniwa> | Hixie: this whole thing is relevant |
| 06:38 | <rniwa> | Hixie: unfortunately I didn't have a time to paraphrase the section |
| 06:39 | <Hixie> | i just needed something to find the place you needed using search :-) |
| 06:39 | <rniwa> | Hixie: actually, this just 4.1 :( |
| 06:39 | <rniwa> | there's also 4.2 & 4.3 |
| 06:40 | <Hixie> | i don't see anything here that would stop us firing 'drop' in an editable element |
| 06:40 | <Hixie> | if anything, i only see stuff that would make us fire 'drop' more easily in an editable element than anywhere else |
| 06:41 | <rniwa> | Hixie: it doesn't say so |
| 06:41 | <Hixie> | because editable elements are special-cased to not need the event handler to explicitly cancel the events to receive the next one, etc |
| 06:41 | <rniwa> | Hixie: look at step 2 after "Otherwise, the drag operation might be a success; run these substeps:" |
| 06:41 | <Hixie> | right, that fires drop |
| 06:41 | <rniwa> | Hixie: yup |
| 06:42 | <Hixie> | before you get to anything that special cases content editable |
| 06:42 | <rniwa> | Hixie: yeah |
| 06:42 | <Hixie> | so... we would fire drop |
| 06:42 | <rniwa> | Hixie: so we need to drop even |
| 06:42 | <rniwa> | fire* drop event |
| 06:42 | <Hixie> | oh you're saying we _should_ drop the event |
| 06:42 | <Hixie> | ok |
| 06:42 | <Hixie> | i misunderstood |
| 06:42 | <Hixie> | yes, i agree we should fire the event |
| 06:43 | <Hixie> | the only reason we would need to _not_ fire it is if there was something that overrode the default behaviour here |
| 06:43 | <Hixie> | which there isn't |
| 06:43 | <rniwa> | yup |
| 06:51 | <MikeSmith> | paul_irish: btw, wasn't meaning to give you grief about the retweet |
| 06:52 | <MikeSmith> | I was just commenting on the fact that the twitter discussion about it seemed to have died down but then came back again |
| 06:52 | <MikeSmith> | and I realize that was happened after your retweet |
| 06:53 | <MikeSmith> | based on observing this I think I may adopt a best practice of not tweeting regular expressions |
| 06:53 | <MikeSmith> | not that I can think of another occasion where I'd want too anyway |
| 06:55 | <MikeSmith> | Hixie: about it being wrong, I actually kind of wondered about that when you first added it |
| 06:55 | <Hixie> | shoulda told me ;-) |
| 06:55 | <MikeSmith> | well, my faith in you is so solid that I never hesitated to question it :) |
| 06:55 | <Hixie> | boy is that misplaced faith :-P |
| 06:57 | <MikeSmith> | heh |
| 06:58 | <MikeSmith> | Hixie: did you see my earlier question |
| 06:58 | <MikeSmith> | http://krijnhoetmer.nl/irc-logs/whatwg/20120112#l-88 |
| 06:58 | <Hixie> | you mean like for <input type=email value=...>? |
| 06:59 | <Hixie> | if so, "The value attribute, if specified and not empty, must have a value that is a single valid e-mail address." (for when multiple="" isn't set), and there's a similar line for when multiple="" is set |
| 06:59 | <Hixie> | "The value attribute, if specified, must have a value that is a valid e-mail address list." |
| 06:59 | <MikeSmith> | yeah |
| 07:00 | <MikeSmith> | but the spec also say the UA should change IDNs to punycode |
| 07:00 | <Hixie> | in user input, yeah |
| 07:00 | <MikeSmith> | IDN e-mail addresses |
| 07:00 | <MikeSmith> | ah |
| 07:00 | <MikeSmith> | yeah |
| 07:00 | <Hixie> | there's no current support for IDN in value="" |
| 07:00 | <MikeSmith> | yeah? |
| 07:00 | <MikeSmith> | why not? |
| 07:00 | <Hixie> | what's the use case? the value in the database would be punycoded |
| 07:01 | <Hixie> | since that's all the server will ever send |
| 07:01 | <Hixie> | er |
| 07:01 | <Hixie> | all the client |
| 07:01 | <Hixie> | will ever send |
| 07:01 | <Hixie> | to the server |
| 07:01 | <MikeSmith> | ah yeah |
| 07:01 | <MikeSmith> | man, I need to think better |
| 07:01 | <MikeSmith> | sorry for that |
| 07:01 | <Hixie> | join the club man |
| 07:01 | <MikeSmith> | heh |
| 07:01 | <Hixie> | we have t-shirts |
| 07:02 | <Hixie> | (they're probably misprinted) |
| 07:02 | <MikeSmith> | t-shirts? |
| 07:02 | <Hixie> | for our club of people who need to think better |
| 07:04 | <MikeSmith> | aha |
| 07:04 | <MikeSmith> | ah, misprinted |
| 07:04 | <MikeSmith> | yeah :) |
| 07:06 | <MikeSmith> | when thinking about stuff, I find it is sometimes optimal to talk with people in the physical world face-to-face |
| 07:06 | <MikeSmith> | which I sadly do much less than I should |
| 07:07 | <MikeSmith> | given that I spend whole stretches of time working at home and only stepping out to buy drink |
| 07:08 | <MikeSmith> | coworking thing would be nice but it's just obnoxiously expensive in tokyo |
| 07:08 | <Hixie> | i find things seem to make a lot of sense when i'm talking to someone face to face, and then i go home and think about it some more and realise everything we discussed is based on flawed assumptions and the whole thing was a waste of time :-P |
| 07:08 | <MikeSmith> | heh |
| 07:08 | <MikeSmith> | yeah, that too |
| 07:08 | <MikeSmith> | but it's more entertaining at least |
| 07:35 | <matjas> | how should characters outside the BMP be escaped in CSS? reading http://www.w3.org/TR/CSS21/syndata.html#characters makes it seem that U+1D306 could be written as `\1D306 ` or `\01D306`, but some browsers support escapes like `\D834\DF06 ` instead. Test case: http://jsfiddle.net/mathias/Hqk9T/ |
| 07:36 | <matjas> | WebKit doesn’t support the regular syntax, e.g. "\1D306 " or "\01D306", but does support "\d834\df06 ". Opera supports both. Firefox only supports the regular syntax. |
| 07:37 | <matjas> | I’m not sure who’s right here. Does the CSS spec mention surrogate pairs anywhere? |
| 07:39 | <Hixie> | anyone got a chrome dev build handy? |
| 07:39 | <Hixie> | i want to know if http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1299 still crashes recent chromes (change the 10 to a 100 like it says in the comment) |
| 07:41 | <Hixie> | looks like in shipping chrome if you create more than 96 <audio> elements, the tab crashes |
| 07:43 | <matjas> | Hixie: doesn’t crash in Chrome 18.0.1003.1 canary on OS X |
| 07:43 | <Hixie> | ok good |
| 07:43 | <Hixie> | thanks |
| 08:16 | <matjas> | re: CSS escape sequences, I just filed https://bugs.webkit.org/show_bug.cgi?id=76152 and https://bugzilla.mozilla.org/show_bug.cgi?id=717529 — let me know if I’m missing something |
| 09:07 | <hsivonen> | does anyone happen to recall if IE supports loading document.open()ed docs from cache? |
| 09:08 | <hsivonen> | WebKit doesn't |
| 12:13 | <Velmont> | AryehGregor: An Opera devel has some patches for selection test suite. Where to send? Directly to you? Bug? public-html-testsuite? |
| 12:19 | <Ms2ger> | public-webapps-testsuite, I guess |
| 12:39 | <wilhelm> | jgraham: Fun. The bizarre insert-random-characters-bug has returned. |
| 12:39 | <jgraham> | wilhelm: Have you considered getting an exorcist? |
| 12:40 | <wilhelm> | There is one in the church I was attempted raised in, actually. I could talk to him. |
| 12:41 | <jgraham> | Strange thing for a ghost to do though, repeatedly pressing keys. I thought they were more into pottery. |
| 12:42 | <wilhelm> | Indeed. |
| 15:06 | <AryehGregor> | So now I have patches submitted in Mercurial format and have to figure out how to import them to git. |
| 15:07 | <AryehGregor> | hg-git fails with a mysterious error message, of course. |
| 15:10 | <jgraham> | (why do you need them in git format?) |
| 15:10 | <AryehGregor> | Because I maintain the repo in git, so I need to import the patches to git so I have them in my source repo. |
| 15:10 | <AryehGregor> | I convert the repo to hg for publication. |
| 15:10 | <AryehGregor> | Unfortunately, hg-git is extremely flaky, as it turns out. |
| 15:11 | <AryehGregor> | Clone works okay; push and pull tend to fail in my experience. |
| 15:11 | <AryehGregor> | So I got around pull not working by just re-cloning the git repo every time before pushing. |
| 15:11 | <AryehGregor> | But I guess I have to manually reformat the hg patch in lieu of getting hg push working. |
| 15:12 | <AryehGregor> | hg help git lists commands gclear, gexport, gimport, and git-cleanup, but they themselves have no help text. SIGH. |
| 15:14 | <AryehGregor> | Hmm. |
| 15:15 | AryehGregor | is very confused |
| 15:20 | <AryehGregor> | Okay, I finally managed to get it to work by a series of very confusing and annoying steps. |
| 15:20 | <AryehGregor> | I think I'll manually convert the patch to git format next time. |
| 15:20 | <AryehGregor> | And maybe update the spec saying that the preferred manner to submit updates is a pull request on github. |
| 15:21 | Philip` | wonders why git can't just import "patch"-compatible patches directly |
| 15:21 | <AryehGregor> | It can, but they don't have the authorship info, right? |
| 15:22 | Philip` | generally doesn't care about authorship info |
| 15:23 | AryehGregor | does when someone has taken the effort to submit a well-formed patch |
| 15:40 | <AryehGregor> | Hixie, so what I really want is one selection for every document that's returned by the .document property of some window, I think. Is that the same as "in a browsing context"? |
| 15:40 | AryehGregor | wants to clearly define this, because it seems browsers don't all agree |
| 15:44 | <AryehGregor> | Hmm. |
| 15:44 | <AryehGregor> | Maybe I should just say all Documents have a Selection. |
| 15:44 | <AryehGregor> | That seems to be how IE behaves. |
| 15:44 | <AryehGregor> | It's certainly simplest. |
| 15:45 | AryehGregor | wonders what happens if you addRange() a Range whose endpoints don't lie in the document . . . |
| 15:47 | <AryehGregor> | So should document.implementation.createHTMLDocument("").getSelection() return a) null (WebKit), b) the same Selection as document.getSelection() (Gecko/Opera), c) a different Selection (IE)? |
| 15:47 | <AryehGregor> | (c) is simplest . . . |
| 15:49 | <gsnedders> | Does that not imply that iframes can have a different selection to the parent? |
| 15:50 | <AryehGregor> | gsnedders, iframes definitely have a different selection from the parent, according to everyone AFAIK. |
| 15:50 | AryehGregor | double-check |
| 15:50 | AryehGregor | double-checks |
| 15:51 | <AryehGregor> | gsnedders, http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cp%3Eabc%3C%2Fp%3E%0D%0A%3Ciframe%3E%3C%2Fiframe%3E%0D%0A%3Cscript%3E%0D%0AgetSelection().selectAllChildren(document.querySelector(%22p%22))%3B%0D%0Aw(getSelection())%3B%0D%0Aw(document.querySelector(%22iframe%22).contentWindow.getSelection())%3B%0D%0A%3C%2Fscript%3E |
| 15:51 | <AryehGregor> | The question is whether documents that aren't the active document of any window also have their own selection. |
| 15:53 | <AryehGregor> | Previously I decided no, now I've changed my mind. |
| 15:53 | <AryehGregor> | Hmm. |
| 15:53 | <AryehGregor> | I think I'll ask the list. |
| 16:09 | <gsnedders> | Yay for Google+ browser sniffing. :\ |
| 17:11 | <dglazkov> | good morning, Whatwg! |
| 17:45 | <sneethling> | hey all |
| 17:45 | <sneethling> | question regarding input type=date |
| 17:46 | <sneethling> | is there a way to specify the locale it should use or does it simply use the current locale? If so, does it choose the server locale or the user locale? |
| 17:50 | <Ms2ger> | user |
| 18:23 | <Ms2ger> | Go, WebKit |
| 18:24 | <Ms2ger> | (https://bugs.webkit.org/show_bug.cgi?id=76121) |
| 18:55 | <Ms2ger> | jgraham, did you have time to look at disabling output for testharness.js? |
| 20:01 | <AryehGregor> | Does anyone know why 2D transforms allow 2x3 matrices, but 3D transforms allow 4x4 instead of 3x4? What meaning does something like matrix3d(1,0,0,0, 0,1,0,0, 0,0,0,1, 0,0,1,0) have? It will map (0,0,z,1) to (0,0,1,z), which is (0,0,1/z,1)? That's not affine . . . |
| 20:05 | <hober> | AryehGregor: linalg makes my head hurt, but smfr or cmarrin in #webkit could answer... |
| 20:05 | <jamesr> | the bottom row is to map the homogeneous coordinate, not the z |
| 20:05 | <jamesr> | i can't remember why it's important... |
| 20:05 | <AryehGregor> | jamesr, I know, but isn't the homogeneous coordinate always supposed to be 1? I thought that was the point of homogeneous coordinates. |
| 20:06 | <AryehGregor> | What's the geometric significance of a homogeneous coordinate that's not 1? |
| 20:06 | <jamesr> | http://en.wikipedia.org/wiki/Transformation_matrix#Perspective_projection |
| 20:07 | <AryehGregor> | Oh, so basically you divide out by the last coordinate, just like in mathematics. And that's not affine, but it produces some other useful visual effect. |
| 20:08 | <jamesr> | it's all smoke and mirrors, but it ends up being very useful |
| 20:08 | <AryehGregor> | Okay, I'll take it on faith that it makes sense graphically. |
| 20:09 | <AryehGregor> | I assume that any point whose last component is 0 is just assumed not to exist graphically (the "point at infinity"). |
| 20:11 | <jamesr> | i think it's done this way by convention, not out of pure mathematical necessity, but that does sound like the usual interpretation |
| 20:12 | <Philip`> | AryehGregor: Points at infinity are useful for e.g. representing the position of the sun in lighting equations |
| 20:13 | <AryehGregor> | Philip`, don't you want the sun to be in a particular direction? The point at infinity isn't in any particular direction in the formalism I'm familiar with. |
| 20:13 | <Ms2ger> | How many dimensions? |
| 20:14 | <AryehGregor> | Any number. |
| 20:14 | <Philip`> | AryehGregor: If the vector is (x,y,z,0) then it's in the same direction as (x,y,z,1) but infinitely far away |
| 20:14 | <AryehGregor> | (I learned all about homogeneous coordinate systems in an undergrad number theory research program) |
| 20:14 | <AryehGregor> | Philip`, oh, so you don't treat all vectors with last coordinate 0 as equivalent? Interesting. |
| 20:14 | <Ms2ger> | In projective geometry, we had lines at infinity as well |
| 20:14 | <AryehGregor> | Oh, right. |
| 20:15 | <AryehGregor> | I guess it's really just an n-2-manifold at infinity or something like that. |
| 20:15 | <AryehGregor> | I also got this paper out of the undergrad research: http://www.sciencedirect.com/science/article/pii/S0022314X11001788 |
| 20:15 | <Ms2ger> | Sounds right |
| 20:15 | <AryehGregor> | But it's cool that there are applications to my job too. |
| 20:16 | <Ms2ger> | Or n-1 |
| 20:20 | <AryehGregor> | Oh, yeah, that makes more sense. A hyperplane at infinity. |
| 20:20 | <AryehGregor> | So it's a point at infinity for the homogeneous line, which is really S^1. |
| 20:20 | <AryehGregor> | Or something. |
| 20:21 | <AryehGregor> | I guess I was thinking P^n was the one-point compactification of R^n or something? Well, whatever. |
| 20:31 | AryehGregor | discovers that testing that matrix entries are correct to within +-1.5 means that an entry of 1 instead of 0 or vice versa passes |
| 20:31 | <AryehGregor> | Who'd have thought? |
| 23:13 | <jamesr> | TabAtkins: yo, what's the status of http://dev.w3.org/csswg/css3-box/ compared to http://www.w3.org/TR/CSS2/visuren.html / http://www.w3.org/TR/CSS2/visudet.html for defining basic block / inline layout modes? |
| 23:15 | <jamesr> | does css3 box superscede 2.1 visres/visdet, or is it incomplete? |
| 23:50 | <ojan> | TabAtkins: qq...flex-align: stretch on parent, flex-item-align: auto on child. getcomputedstyle of flex-item-align on the child should return stretch or auto? |
| 23:50 | <ojan> | TabAtkins: and what spec covers this? |
| 23:51 | <ojan> | computed style continue to confound me |
| 23:51 | <ojan> | i guess this is covered by cssom |
| 23:54 | <jamesr> | ojan: http://dev.w3.org/csswg/cssom/#resolved-values |
| 23:54 | <jamesr> | since flex-align-item isn't in the magic list for getComputedStyle, it returns the computed value of that property |
| 23:56 | <ojan> | jamesr: yeah...i've read that before...but i don't see anything defining what "resolved value" and "computed value" actually are |
| 23:56 | <jamesr> | resolved value = computed value, except for that list of properties in which case resolved value = something else |
| 23:56 | <jamesr> | computed value = this guy: http://dev.w3.org/csswg/css3-cascade/#computed |
| 23:57 | <jamesr> | so basically after matching selectors and doing inheritance |
| 23:58 | <ojan> | jamesr++ |
| 23:58 | <ojan> | thx |
| 23:58 | <jamesr> | i think it'd be auto in your case unless there's something that modifies it in some way due to inheritance |
| 23:59 | <ojan> | jamesr: yeah...it won't be auto. flex-align modifies it due to inheritance |