| 00:12 | <benschwarz> | Hixie: !ping |
| 02:34 | <rektide_> | are cross site <script type="text/javascript" src="..."></script>'s no longer allowed? |
| 02:34 | <rektide_> | i'm not seeing my dev channel chrome attempting the request |
| 02:35 | <rektide_> | i thought jsonp was a still used workaround? is it still possible? if so, how? or is there something amiss with my chrome not attempting the above <script> request |
| 02:37 | <rektide_> | sorry error elsewhere in my html; my mistake |
| 10:03 | <annevk> | The RTC-Web effort (Real Time Collaboration on the World Wide Web) ... http://rtc-web.alvestrand.com/ |
| 10:03 | <annevk> | The Real Time Communication on the Web (RTC-Web) ... http://rtc-web.alvestrand.com/home |
| 10:04 | <annevk> | what is it? |
| 10:21 | <hsivonen> | annevk: "What is it?" in what sense? |
| 10:24 | <annevk> | Communication or Collaboration |
| 10:24 | <hsivonen> | annevk: oh |
| 10:24 | <jgraham> | Hmm, I wonder if smaug wrote a testharness.js-to-mochitest adapter. Someone has to do that eventually |
| 10:25 | <hsivonen> | annevk: the sites fail to provide a quick elevator pitch like "Chatroulette without Flash" |
| 10:25 | <hsivonen> | (Or G+ hangouts without Flash, to give a less creepy usecase) |
| 10:27 | <espadrine`> | annevk: I guess it is more about communication, collaboration seems like a particular use case rather than the whole thing |
| 10:27 | <hsivonen> | BTW, what's the business case for Adobe for having videoconferencing enablement in Flash Player? Is it only about making Flash Player more sticky so that they can leverage it for other business or does Adobe monetize videoconferencing directly? |
| 10:28 | <hsivonen> | espadrine`: seems to me to be more about realtime audio and videa than about either communication or collaboration on a general level |
| 10:28 | <Ms2ger> | jgraham, I did :) |
| 10:29 | <hsivonen> | Ms2ger++ |
| 10:29 | <Ms2ger> | Now I just need someone to review it |
| 10:29 | <espadrine`> | hsivonen: oh well. |
| 10:29 | <jgraham> | Ms2ger: You rock |
| 10:29 | <Ms2ger> | Thanks :) |
| 10:30 | <hsivonen> | annevk: maybe the C is like X in XBL :-) |
| 10:30 | <espadrine`> | "realtime audio/video" is not enterprisy enough |
| 10:30 | Ms2ger | wonders if he could land it with r=jgraham... |
| 10:31 | espadrine` | agrees with jgraham |
| 10:31 | <annevk> | Ms2ger, you should probably tell smaug |
| 10:31 | <annevk> | Ms2ger, before he rewrites all those tests |
| 10:31 | <Ms2ger> | CCd him |
| 10:35 | <annevk> | Ms2ger, btw |
| 10:36 | <annevk> | Ms2ger, <ol type=TEST> ol[type=test] { background:lime } will style the <ol> |
| 10:36 | <annevk> | Ms2ger, in every browser |
| 10:36 | <Ms2ger> | Yeah |
| 10:36 | <annevk> | which is wrong per HTML at the moment |
| 10:36 | <Ms2ger> | Is it really? |
| 10:37 | <annevk> | "The exception to the list above is the type attribute on ol elements, which must be treated as case-sensitive." |
| 10:37 | <Ms2ger> | Where's that? |
| 10:38 | <annevk> | first subsection of "Matching HTML elements using selectors" |
| 10:39 | <Ms2ger> | Huh |
| 10:55 | <hsivonen> | espadrine`: what kind of enterpriseyness do you perceive there to be about WebRTC and what enterpriseyness do you think it should have? |
| 11:20 | <annevk> | it's still window.URL right? |
| 11:20 | <annevk> | http://html5-mediasource-api.googlecode.com/svn/trunk/draft-spec/mediasource-draft-spec.html has some confusing URI.createObjectURL() in it... |
| 11:22 | <Ms2ger> | Yes |
| 11:25 | <annevk> | I wonder if I should bother filing a bug on Mozilla about attribute values and selectors |
| 11:25 | <annevk> | Or just file a bug on HTML5 to remove the exception for <ol type> |
| 11:26 | <annevk> | Guess I'm going to do the former. The behavior is magical and ugly... |
| 11:33 | <annevk> | http://blog.whatwg.org/weekly-rtc-download |
| 11:51 | <AnselmBradford> | For the area element, what is meant by a "dead area on an image map" (in reference to one form the element can take)? |
| 11:51 | <Ms2ger> | One that doesn't link anywhere? |
| 11:52 | <Ms2ger> | Like <area coords=0,0,100,100 href=foo><area coords=25,25,75,75> |
| 11:54 | <AnselmBradford> | Ms2ger: gotcha thanks... so corresponds to this text in the spec "If the area element has no href attribute, then the area represented by the element cannot be selected, and the alt attribute must be omitted." |
| 11:54 | <Ms2ger> | Sounds right |
| 11:55 | <AnselmBradford> | What's a use case for an area without an href? |
| 11:58 | <Ms2ger> | Making parts of an image unclickable, I suppose... This was long before my time |
| 11:59 | <annevk> | AnselmBradford, you can add a title attribute for instance |
| 12:00 | <AnselmBradford> | annevk: ah, I see... so on hover the title shows, but it wouldn't be a hyperlink to anywhere? |
| 12:00 | <annevk> | not sure how <area> is layered but I can imagine you have a large clickable layer and then some parts of that should not be clickable maybe |
| 12:00 | <annevk> | AnselmBradford, right |
| 12:02 | <smaug____> | annevk: would it be possible to have some text in your test that they are Public Domain (or whatever license they use) |
| 12:08 | <annevk> | I rather not put that stuff in each test |
| 12:09 | <annevk> | last time I contributed tests to Mozilla I just said so in a bug report |
| 12:11 | jgraham | wonders why these tests can't go via W3C |
| 12:11 | <jgraham> | Which should solve Mozilla's licensing issue and make them avaliable for everyone |
| 12:12 | <annevk> | I think I was waiting for comments before putting them in even more repositories |
| 12:12 | <annevk> | Potentially also for the W3C to agree on testharness.js or something else |
| 12:14 | <jgraham> | Seems like it might be time to move them then :) |
| 12:16 | <annevk> | Still waiting for comments |
| 12:18 | <hsivonen> | the enclosure thread shows why you need to bikeshed early and forcefully |
| 12:18 | <hsivonen> | who'd have thought that Dave Winer's pick for an RSS element could threaten to leak into HTML rel values one day |
| 12:23 | <jgraham> | annevk: My comment is you should move them to some W3C venue before I comment further :) |
| 12:23 | <smaug____> | heycam|away: what should happen if some method takes an dictionary, but the object which implements the dictionary has wrong types. Are those properties automatically converted to right type or something? |
| 12:24 | <Ms2ger> | smaug____, yes, afaict |
| 12:25 | <Ms2ger> | But maybe not if that method is overloaded |
| 12:28 | <annevk> | jgraham, not falling for that one :p |
| 12:52 | <annevk> | Extra arguments added to methods are ignored right? |
| 12:53 | <annevk> | looks like it |
| 12:53 | <annevk> | am I glad entityReferenceExpansion is the last argument |
| 12:54 | <annevk> | we can just remove it without hassle |
| 13:24 | <hsivonen> | I wonder what would happen if an Editorial Assistant WONTFIXed a bug Hixie filed. |
| 13:24 | <annevk> | try it! |
| 13:25 | <annevk> | although maybe wait for a bit, so I can have some popcorn with that |
| 13:27 | <hsivonen> | Another thing I've been wondering about today: Should I write email to Bob Leif expressing that using XSD 1.1 isn't a use case. |
| 13:29 | <annevk> | I am hoping the chair will handle it |
| 13:29 | <hsivonen> | yeah, maybe I should wait for Norm to deal |
| 13:42 | <annevk> | I wonder why the DOM specifications are so verbose |
| 13:42 | <annevk> | Well, the pre-DOM Core stuff |
| 13:43 | <annevk> | Looking at DOM Traversal it is mostly complicated language for a few simple concepts |
| 13:48 | <hsivonen> | annevk: I thought the old DOM specs weren't verbose enough |
| 13:48 | <hsivonen> | annevk: maybe they spend words on the wrong things, though |
| 13:49 | <annevk> | well yeah, they are not exactly clear |
| 13:49 | <annevk> | they explain models by examples rather than well, defining them |
| 13:59 | <annevk> | smaug____, why can't withCredentials be a simple attribute on EventSource just as it is with XHR? |
| 14:00 | <annevk> | if we need withCredentials at all |
| 14:14 | <smaug____> | annevk: EvenSource opens the connections when ctor is called |
| 14:16 | <annevk> | no |
| 14:16 | <annevk> | it queues a task to do that |
| 14:17 | <smaug____> | ah, that is true |
| 14:17 | smaug____ | hates the API inconsistency in XHR vs EventSource |
| 14:17 | <smaug____> | so, yeah, perhaps withCredentials attribute could work |
| 14:18 | <smaug____> | that would give some consistency |
| 14:20 | <smaug____> | annevk: I haven't seen any good reason to not have withCredentials |
| 14:20 | <annevk> | still think it wasn't the best idea to add it to XHR |
| 14:22 | <smaug____> | but to do what? |
| 14:22 | <smaug____> | never send credentials? |
| 14:26 | <annevk> | always send credentials |
| 14:28 | <smaug____> | why |
| 14:29 | <annevk> | because that is how cross-origin requests worked already |
| 16:23 | <annevk> | Hmm, Traversal depends on mutation quite a bit |
| 16:49 | <TabAtkins> | annevk: Hm, I guess that bit about <ol=type> needs to be reworded, then, so that it specifically references the known type values but lets everything else be case-insensitive. |
| 16:59 | <smaug____> | Interesting. Per spec calling command.click() ends up firing click event twice |
| 17:02 | <annevk> | TabAtkins, that would be even worse |
| 17:02 | <annevk> | TabAtkins, also does not match impls afaik |
| 17:05 | <TabAtkins> | annevk: Man, really? so <ol type="A"> is matched by ol[type='a']? |
| 17:09 | <annevk> | TabAtkins, always has been |
| 17:09 | <annevk> | TabAtkins, the change I am hoping we can make is to treat attribute values as case-sensitive, always |
| 17:09 | <annevk> | TabAtkins, I think that would be better than having a magic list of attributes |
| 17:11 | <TabAtkins> | annevk: I'm wondering how impls actually apply list styling, then. I assumed it was done via ua stylesheet. |
| 17:11 | <annevk> | UA style sheet has to be special anyway because of XHTML |
| 17:11 | <annevk> | e.g. input[type=button] would not work well in XHTML |
| 17:15 | <annevk> | smaug____, prolly a bug |
| 17:48 | <annevk> | TabAtkins, and it seems browsers have bugs because of their UA style sheet |
| 17:48 | <annevk> | e.g. compare data:text/xml,<html xmlns="http://www.w3.org/1999/xhtml"><input type="BUTTON"/><input type="button"/></html> between Gecko / WebKit / Opera |
| 17:50 | <TabAtkins> | Yay! |
| 18:09 | <rillian> | hey all |
| 18:09 | <rillian> | I'm having trouble understanding http://www.w3.org/TR/html5/webappapis.html#provide-a-stable-state |
| 18:10 | <rillian> | I understand the idea is to wait for any (asynchronously running?) synchronous sections to complete |
| 18:10 | <rillian> | so the attributes available to javascript are coherent |
| 18:10 | <rillian> | but isn't that at odds with "A synchronous section never mutates the DOM ... or has any other side-effects."? |
| 18:11 | <TabAtkins> | Why do you think it's at odds? |
| 18:12 | <rillian> | I suppose because firebug shows global js objects in the DOM tab |
| 18:12 | <rillian> | so I don't see how you could update object state without mutating the DOM |
| 18:18 | <annevk> | I recommend a) looking at a version of the draft that is not multipage so you can see where the terms are referenced from and b) not using TR/ |
| 18:21 | <annevk> | rillian, fwiw, mutates the DOM means mutating the Document tree |
| 18:21 | <annevk> | rillian, it is not about changing IDL attributes |
| 18:22 | <rillian> | so "DOM" there is different from the view in javascript's namespace? |
| 18:24 | <annevk> | not sure what you mean by that |
| 18:24 | <rillian> | annevk: sorry, by not using 'TR' do you mean http://www.whatwg.org/specs/web-apps/current-work/#provide-a-stable-state ? |
| 18:24 | <annevk> | for instance |
| 18:25 | <annevk> | TR/ is always out of date |
| 18:25 | <rillian> | 'k thanks |
| 18:26 | <annevk> | DOM mutation is generally understood to mean things like removeChild() |
| 18:26 | <rillian> | so things that affect the document object from javascript's point of view? |
| 18:26 | <annevk> | things that fire mutation events |
| 18:27 | <annevk> | you could ask for that to be made clearer though |
| 18:27 | <rillian> | sorry, by "DOM" I pretty much meant "What firebug shows in its DOM tab" |
| 18:27 | <rillian> | I'm trying to understand what's spec, what's current implementation, and so on, with the event loop |
| 18:28 | <annevk> | as currently phrased you could think setting e.g. <link>.disabled would be such a thing |
| 18:28 | <rillian> | *nod* |
| 18:29 | <rillian> | I think the spec would also be improved with some less-normative language about what stable-state was for |
| 18:29 | <rillian> | right now it feels a bit circular |
| 19:03 | <Xeon06_> | How legal is it to create our own attributes on existing elements in HTML5? |
| 19:03 | <TabAtkins> | You can use the data-* attributes for whatever you want that's private to the page. Otherwise, it's invalid. |
| 19:04 | <AryehGregor> | Xeon06_, you can make up your own attributes as long as they start with "data-". |
| 19:04 | <Xeon06_> | Okay, thanks guys1 |
| 19:04 | <Xeon06_> | !* |
| 19:06 | <Xeon06_> | Wow, this is what jQuery's .data does! |
| 19:07 | <TabAtkins> | Yes. |
| 19:19 | <Hixie> | hsivonen: if you wontfix a bug i filed i won't notice. but i would be interested in hearing about it if you ever find a bug i filed be something you might consider wontfixing :-) |
| 19:19 | <Hixie> | hsivonen: especially since they're generally editorial things :-) |
| 19:24 | <janv_> | Hixie: hi, could you take a look at https://bugzilla.mozilla.org/show_bug.cgi?id=617528#c109 ? |
| 19:27 | <janv_> | Hixie: thanks |
| 19:30 | <Hixie> | just got their in my bugmail :-) |
| 19:31 | <janv_> | :) |
| 19:31 | <janv_> | this is the last issue |
| 19:31 | <Hixie> | er, there |
| 19:31 | <Hixie> | not their |
| 19:31 | <janv_> | I see |
| 19:32 | <janv_> | I hope we land the patch soon |
| 21:14 | <annevk> | Hixie, why does ev = new EV("url"); ev.withCredentials = true; not work? |
| 21:14 | <annevk> | new EV; queues a task |
| 21:15 | <annevk> | when the task queue is spinned and that task gets done you get the value from withCredentials and all should be fine |
| 21:48 | <Hixie> | annevk: workers |
| 21:49 | <annevk> | EventSource is defined differently for workers? |
| 21:50 | <annevk> | how does onopen get registered on time in workers? |
| 21:51 | <Hixie> | workers don't spin the event loop often |
| 21:51 | <Hixie> | or, they don't have to spin the event loop often, rather |
| 21:51 | <Hixie> | since their scripts can run forever |
| 21:52 | <annevk> | either way you need to wait for the next task |
| 21:52 | <annevk> | and new ev(); ev.wC is a single operation |
| 21:52 | <Hixie> | the EventSource constructor doesn't wait for the event loop to start the connection |
| 21:57 | <annevk> | oh I see |
| 21:57 | <annevk> | it doesn't queue a task at all |
| 21:57 | <annevk> | it does some magic |
| 21:58 | <annevk> | should XHR use similar magic? |
| 21:58 | <annevk> | hmm |
| 21:58 | <Hixie> | not magic, just async |
| 21:59 | <annevk> | i thought queue was our async |
| 21:59 | <annevk> | not return and do both set of things |
| 22:00 | <Hixie> | there's all kinds of different models depending on what the needs are |
| 22:01 | <annevk> | i had no idea |
| 22:01 | <Hixie> | working on a background thread like this, queuing work on an event loop, using "stable states", hooking in other parts of the event loop, etc |
| 22:05 | <annevk> | guess XHR could use this too |
| 22:05 | <annevk> | hmm |
| 22:10 | <Hixie> | if you use 'fetch' then you already use this |
| 22:10 | <Hixie> | since 'fetch' only makes sense on a background thread, not in a task |
| 22:10 | <Hixie> | (unless you're doign sync xhr, of course) |
| 22:10 | <annevk> | i think it queues a task to fetch atm |
| 22:11 | <annevk> | bit sad that nobody actually reviews the spec in detail |
| 22:11 | <annevk> | or maybe next to nobody understands those parts |
| 22:11 | <jamesr> | the implications can be very subtle |
| 22:11 | <annevk> | yeah, way hard |
| 22:12 | <annevk> | anyway, got to go |
| 22:13 | <rillian> | at least I'm not the only one confused on these points :/ |
| 23:08 | <AryehGregor> | Hey, someone else actually was using the Live DOM Viewer upload/download feature at the same time as me. That's a first. |
| 23:11 | <TabAtkins> | Did you collide? |
| 23:11 | <AryehGregor> | Yes, that's how I knew. |
| 23:17 | <AryehGregor> | Can anyone explain to me why the www-html list still exists? |
| 23:18 | <TabAtkins> | Looks like just the older version of public-html, from before we switched list naming standards. Dunno why it's still open. |
| 23:19 | <Philip`> | I think the switch to public-html was primarily to ensure subscribers agreed to the patent policy |
| 23:19 | <Philip`> | (since those already on www-html hadn't) |
| 23:20 | <Hixie> | AryehGregor: i assume you collided with my window.find() thing, sorry :-) |
| 23:21 | <Hixie> | also, window.find() is incredibly buggy |
| 23:21 | <Hixie> | well, not incredibly |
| 23:21 | <Hixie> | but it has six boolean arguments and of those, only one works the same in gecko and webkit |
| 23:23 | <jamesr> | i only use the upload/download feature when i am trying to get a permalink and forget which link it is to do that |
| 23:23 | <jamesr> | which is pretty much every time i try to get a permalink |
| 23:23 | <Philip`> | Isn't it the link that says "permalink"? |
| 23:24 | <Hixie> | i think he means the short permalink |
| 23:24 | <Hixie> | which is "save" |
| 23:24 | <jamesr> | right, that one |
| 23:24 | <Philip`> | Ah |
| 23:24 | <AryehGregor> | I use it constantly when testing among multiple browsers. |
| 23:24 | <Hixie> | yeah me too |
| 23:24 | <AryehGregor> | Particularly since one of them is on a different computer. |
| 23:24 | <Hixie> | indeed |
| 23:25 | <AryehGregor> | Although it seems like in IE9 and IE10, "download" only works once until you restart the browser. |
| 23:25 | <AryehGregor> | At least on my machine. |
| 23:25 | <Hixie> | actually my first implementation of the upload/download thing was for a clipboard on my portal page |
| 23:25 | <Hixie> | IE's wacked |
| 23:25 | <AryehGregor> | If you hit it a second time, it just re-downloads the same thing it did last time. No idea why. |
| 23:25 | <AryehGregor> | Also, the permalinks don't work in IE or Chrome because they're XSS. :( |
| 23:25 | <Hixie> | it wants different cache flags, i expect |
| 23:25 | <Hixie> | the whole tool is a giant XSS :-P |
| 23:26 | <jamesr> | Hixie: send the XSS header then! |
| 23:26 | <Hixie> | the XSS header is lame |
| 23:26 | <Hixie> | and unspecified |
| 23:27 | <Hixie> | i may one day switch to base64-encoding the payload |
| 23:27 | <Hixie> | but i'm in no rush |
| 23:28 | <Hixie> | so for window.find() |
| 23:28 | <Hixie> | should i only specify what's interoperable, or should i specify the superset and just call every implementation buggy? |
| 23:28 | <AryehGregor> | Specify the subset that's either interoperable or useful. |
| 23:28 | <Hixie> | not clear any of it is useful |
| 23:29 | <AryehGregor> | Then just the part that's interoperable, I'd say. |
| 23:29 | <Hixie> | and no subset it is 100% interoperable |
| 23:29 | <Hixie> | of it |
| 23:29 | <TabAtkins> | Your work is done, then. |
| 23:29 | <AryehGregor> | Well, then make something up. :) |
| 23:29 | <AryehGregor> | I mean, no subset of execCommand() is 100% interoperable, that's for sure . . . |
| 23:29 | <TabAtkins> | Also, this is a horrible function and I hate everyone involved in the creation of it. |
| 23:30 | <AryehGregor> | Wow, it only returns a boolean? I'd have thought it would at least return the location of the first occurrence. |
| 23:31 | <AryehGregor> | That seems profoundly useless. |
| 23:31 | <TabAtkins> | Yeah, it seems completely useless. |
| 23:31 | <Hixie> | it selects the match |
| 23:31 | <AryehGregor> | Oh, I see. |
| 23:31 | <AryehGregor> | That's . . . also amazingly useless. |
| 23:31 | <TabAtkins> | Ah, MDN doesn't mention that. |
| 23:32 | <TabAtkins> | Can you query for what parts of the document are selected? |
| 23:32 | <AryehGregor> | getSelection()? |
| 23:32 | <TabAtkins> | kk, wasn't sure. |
| 23:32 | <jamesr> | is it used? |
| 23:32 | <TabAtkins> | So I guess that'll return a disconnected selection if there are multiple matches. |
| 23:32 | <AryehGregor> | If you couldn't, I wouldn't have had much to do for the last six months. :) |
| 23:32 | <AryehGregor> | I imagine it would just select the first match. |
| 23:33 | <AryehGregor> | Or the first one after the current selection. |
| 23:33 | <AryehGregor> | Like Ctrl-F. |
| 23:33 | <AryehGregor> | Although that doesn't actually select anything. |
| 23:33 | <TabAtkins> | Well, my ctrl+f selects all the matches, and then scrolls to the first. |
| 23:33 | <AryehGregor> | It doesn't select them, it highlights them, no? |
| 23:33 | <Hixie> | window.find() selects the first match after the cursor position |
| 23:35 | <AryehGregor> | Apparently someone named Tim Powell used window.find() for bookmarklets as of 2001: https://bugzilla.mozilla.org/show_bug.cgi?id=9550#c14 |
| 23:36 | <Hixie> | the only reason i'm even looking at this is i got some feedback from oepra saying they're getting feedback from developers asking for it |
| 23:37 | <Hixie> | in Gecko, window.find() pops up a dialog |
| 23:37 | <Hixie> | but Ctrl+F pops up an inline find bar |
| 23:38 | <AryehGregor> | This all sounds amazingly horrible even by web standards standards. |
| 23:38 | <Hixie> | these two find UIs in gecko have different features... |
| 23:38 | <Hixie> | (neither is a subset of the other) |
| 23:39 | <Hixie> | wow |
| 23:40 | <Hixie> | i got gecko's to fail with NS_ERROR_INVALID_POINTER |
| 23:41 | <AryehGregor> | This sounds like a feature everyone has forgotten about for ten years. |
| 23:41 | <smaug____> | window.find() is find, ctrl+f is typeaheadfind |
| 23:41 | <smaug____> | who uses window.find() anymore? |
| 23:41 | <Hixie> | i don't understand why when i make it pop up the dialog, it comes prepopulated with "cat" as teh search string |
| 23:41 | <Hixie> | smaug____: can we drop support for it? that would make my life way easier. |
| 23:41 | <TabAtkins> | Presuambly that's the last thing you searched for, Hixie? |
| 23:41 | <Hixie> | TabAtkins: no |
| 23:42 | <smaug____> | Hixie: dropping window.find(). Yes please. |
| 23:42 | <TabAtkins> | It seemed likely. |
| 23:42 | <Hixie> | TabAtkins: i searched for any number of other things, but every time the dialog comes back, it says "cat" |
| 23:42 | <Hixie> | smaug____: any idea who makes the call on this? |
| 23:42 | <smaug____> | I just wonder if it is possible to drop support for that |
| 23:43 | <TabAtkins> | Hixie: Mine doesn't do that, so shrug. |
| 23:43 | <Hixie> | TabAtkins: oh i'm sure it's something i did |
| 23:44 | <Hixie> | wow my gecko is seriously messed up now |
| 23:44 | <Hixie> | every time i call it with a non-empty first argument, i get this pointer exception |
| 23:44 | <jamesr> | sounds like you need to restart |
| 23:44 | <Hixie> | i did! |
| 23:45 | <Hixie> | oh, i know what's wrong |
| 23:45 | <Hixie> | i'm calling it on a frame i've made display:none |
| 23:47 | <Hixie> | webkit has another crazy bug, if yo call it before onload it just returns false regardless |
| 23:48 | <smaug____> | so you can't search on long loading page? |
| 23:48 | <Hixie> | not in webkit |
| 23:48 | <Hixie> | you can in gecko |
| 23:48 | <Hixie> | (dunno exactly when it starts being possible in webkit) |
| 23:49 | <Hixie> | (it might work progressively, just not immediately) |
| 23:53 | <jarek> | Hi |
| 23:53 | <jarek> | is there anything wrong in using custom (not defined in specification) tags in HTML5 documents? |
| 23:53 | <jarek> | I mean I know they don't validate |
| 23:54 | <jarek> | but besides that, are there any other practical drawbacks? |
| 23:54 | <Hixie> | yes |
| 23:54 | <Hixie> | if we start using them as elements in future revisions of HTML, your pages will break |
| 23:54 | <jarek> | I see, I can live with that :) |
| 23:54 | <Hixie> | also, today nobody will know what your pages mean, since they won't be following the standard |
| 23:55 | <Hixie> | also, there's a high likelihood you will accidentally run into weird behaviour, e.g. <image> parses weirdly even though it's not valid, as does <plaintext> and <xmp>, etc. |
| 23:55 | <Hixie> | there's probably a better way to do what you want to do |
| 23:55 | <jarek> | I was thinking about using custom XML namespaces, but most of the the new stuff (e.g. contentEditable) does not work with them |
| 23:57 | <jarek> | HTML5 is still mostly a document format, I wish there were more XUL-like tags |
| 23:57 | <Hixie> | like what? |
| 23:57 | <Hixie> | smaug____: filed https://bugzilla.mozilla.org/show_bug.cgi?id=672395 |
| 23:58 | <jarek> | like <box>, there is not sematics in HTML5 for describing many widgets |
| 23:58 | <jarek> | s/not/no |
| 23:58 | <Hixie> | <box> is presentational. use CSS for that. |
| 23:58 | <Hixie> | what widgets are you missing? |
| 23:59 | <jarek> | dialogs, menubars, toolbars, tabs... |
| 23:59 | <jarek> | I mean the usual stuff that is used in desktop apps |
| 23:59 | <Hixie> | we're looking at dialogs -- if you have any feedback on what you need there, please e-mail the list |
| 23:59 | <Hixie> | menubars and toolbars we have - <menu> |
| 23:59 | <Hixie> | tabs is a presentational thing - just a way to display multiple sections |