| 00:02 | <astearns> | odd that 'shown' is more common than 'web' or 'link' |
| 00:03 | <dgorbik> | Hixie: hello. There is a WebVTT requirment "The 'display' property on WebVTT Ruby Objects must be set to 'ruby'." But for just ruby elements in HTML we have those styles: "text-indent: 0;". Should they be different? Or probably display: ruby is not implemented yet in WebKit? |
| 00:03 | <Hixie> | TabAtkins: do i have to say anything for :user-error in the HTML spec? |
| 00:04 | <Hixie> | dgorbik: text-indent doesn't matter in webvtt's case because it doesn't apply and thus can't be set to anything but zero |
| 00:04 | <Hixie> | dgorbik: the problem in HTML is that text-indent inherits, and so anything that's block-like in a phrasing context has to have its text-indent overridden back to 0 so it doesn't get its ancestor's indent |
| 00:05 | <dgorbik> | Hixie: then why don't we set display: ruby for HTML? |
| 00:07 | <Hixie> | we do |
| 00:07 | <Hixie> | search for "ruby {" |
| 00:07 | <Hixie> | the first match is where it's set to display:ruby |
| 00:08 | <Hixie> | wait, where did you find it set text-indent to zero in the spec? |
| 00:08 | <dgorbik> | Hixie: I don't see that in html.css |
| 00:08 | <Hixie> | it probably should, but i can't find that line... |
| 00:08 | <Hixie> | what's html.css? |
| 00:08 | <dgorbik> | No, I've found that in html.css |
| 00:08 | <Hixie> | oh |
| 00:08 | <Hixie> | i thought you were asking about the spec |
| 00:08 | <dgorbik> | these are default css styles in webkit for elements |
| 00:08 | <Hixie> | i'm not a webkit developer, so i've no idea |
| 00:08 | <dgorbik> | I am now trying to understand whether we have to set display: ruby for html elements as well. |
| 00:09 | <dgorbik> | Hixie: so the only thing I need — display: ruby should be set for HTML elements as well, not just WebVTT ruby elements? |
| 00:10 | <Hixie> | according to the spec, yes, but i have no idea how webkit implements ruby |
| 00:10 | <Hixie> | if it implements ruby |
| 00:10 | <Hixie> | maybe webkit has special optimisations to do it differently internally |
| 00:10 | <dgorbik> | Hixie: okay, thanks, I will try to find that |
| 00:10 | <Hixie> | the spec just says what the results have to be like, not what the implementation strategy has to be |
| 00:10 | <gsnedders> | Hixie: Why do you not know everything? |
| 00:29 | <esprehn_> | dgorbik: I don't think there's a ruby display type |
| 00:30 | <dgorbik> | esprehn_: is it being emulated somehow or just not implemented? |
| 00:30 | <esprehn_> | dgorbik: oh I see, css3-ruby specifies specific display types |
| 00:31 | <esprehn_> | dgorbik: webkit has a hack in RenderObject::createObject that checks the qname and looks at display and produces either an inline ruby or a block ruby |
| 00:31 | <esprehn_> | css3 ruby |
| 00:32 | <esprehn_> | err, css3 ruby's display types seem broken since it doesn't support block and inline modes |
| 00:32 | <esprehn_> | this spec is busted |
| 00:32 | <zewt> | it uses ... display types? o_O |
| 00:36 | <esprehn_> | zewt: yeah, but doesn't define an inline-ruby or block-ruby or I dunno |
| 00:36 | <esprehn_> | I don't think webkit could implement the spec without breaking existing ruby content |
| 00:38 | <zewt> | sounds like the wrong axis, though i havn't looked at that spec |
| 00:39 | <Hixie> | empty numeric control, you hit the step up button. what value does it take? |
| 00:39 | <Hixie> | same question, step down button |
| 00:39 | <dgorbik> | esprehn: do we render ruby the way specs require aside the fact we don't set some properties? |
| 00:39 | <zewt> | unless there's some use case for a particular behavior it's probably arbitrary |
| 00:40 | <zewt> | by default i'd probably just pick 0 (or 1 or an arbitrary value in range if there's a range limit) |
| 00:40 | <Hixie> | possible answers: minimum, maximum, default step base, step base, zero... |
| 00:41 | <zewt> | min and max don't seem correct, at least |
| 00:41 | <Hixie> | i was thinking just snap to the minimum |
| 00:41 | <Hixie> | let's see what real ones do... |
| 00:41 | <zewt> | what if the range is -9999999999 to 99999999999? |
| 00:42 | <Hixie> | what if indeed |
| 00:42 | <zewt> | zero or one at least don't go crazy if the range is huge (or if there isn't any range) |
| 00:43 | <dgorbik> | esprehn, Hixie: by the way (I was about to ask this at one point). Specs require that 'background-' style rules that are applied to WebVTT objects should be applied to the root of those objects instead (it's called webvtt cue box). So here is a question — do you have an idea of how to implement that? As far as I understand we create a RenderStyle for the parent first and then for the children, so we will have to recalc styles after w |
| 00:43 | <zewt> | after we$ |
| 00:43 | <Hixie> | chrome default to zero then takes a step in the given direction, snapping to min/max if it's out of range, as far as i can tell |
| 00:44 | <Hixie> | dgorbik: not all webvtt objects, just the root, right? |
| 00:44 | <Hixie> | dgorbik: just apply the background to the cue's wrapping box rather than the root box |
| 00:44 | <dgorbik> | esprehn: or do some crazy preprocessing... Like manually walking over the rules and splitting them so that 'background-' rules would apply to parent nodes. And only then create renderers |
| 00:44 | <Hixie> | ("just") |
| 00:44 | <Hixie> | opera seems to default to minimum if you go up and maximum if you go down |
| 00:45 | <dgorbik> | Hixie: I thought that's for all objects... If any object (even deeply nested one) has this property, it should be applied to the cue box which could be several levels above. |
| 00:46 | <dgorbik> | no? |
| 00:46 | <Hixie> | no, that would make no sense :-) |
| 00:47 | <Hixie> | how would you decide which one was the background? |
| 00:47 | <dgorbik> | Hixie: that what I was thinking too |
| 00:48 | <dgorbik> | Hixie: hm, is looks like it's now clear about this. I don't remember it was in the previous revision. |
| 05:27 | <bambu> | I have index.html, and I copy it to another file called about.html. I alter a paragraph containing the content, and I have a two page site! |
| 05:28 | <bambu> | However, to me it seems redundant to have basically two files that are the same apart from a small paragraph, lots of wasted space. |
| 05:28 | <bambu> | Is this just how its done? |
| 05:28 | <bambu> | Any alternatives? |
| 05:46 | <bambu> | Anyone care to read: http://pastie.org/private/7jdn5nelu1zgg6m69pg2dg |
| 06:21 | <zewt> | i'll never understand pages with image viewers that hide the image until it finishes loading |
| 06:22 | <zewt> | as i sit there waiting for several seconds on a spinner or whatever annoyed because they're jumping hoops to disable incremental loading that browsers have spend tons of time making work |
| 06:23 | <zewt> | sometimes i can only wonder if peoples' brains have an on/off switch (stuck in the more unfortunate position) |
| 10:09 | <ashemedai> | So this whole new Opera ICE using Webkit. I was amazed. |
| 10:09 | <ashemedai> | (Or Ice rather.) |
| 14:15 | <MikeSmith> | so who's the guy in the Opera Ice video? |
| 14:15 | <MikeSmith> | I don't recognize him |
| 15:01 | <espadrine> | MikeSmith: I thought that was Opera's CEO? |
| 15:35 | <Ms2ger> | Opera has a CEO? |
| 16:01 | <MikeSmith> | Ms2ger: yeah Opera has a CEO but they're planning to replace him with WebKit |
| 16:01 | <Ms2ger> | I see |
| 16:01 | <MikeSmith> | espadrine: that's no Opera's CEO in that video |
| 16:01 | <Ms2ger> | Insufficient pointy hair? |
| 16:02 | <MikeSmith> | heh |
| 16:02 | <MikeSmith> | well, for one thing, he wouldn't be caught dead wearing the shirt that guy is wearing |
| 16:03 | <MikeSmith> | there are a lot of varieties of bad fashion sense in the world, but there's nothing quite like the northern European variety |
| 16:04 | <MikeSmith> | especially Scandinavian |
| 16:05 | <MikeSmith> | I mean what message is dude sending with that shirt |
| 16:05 | <Ms2ger> | And that coming from MikeSmith :) |
| 16:06 | <MikeSmith> | well, I would make that shirt look good |
| 16:06 | <MikeSmith> | it's all about how you wear it |
| 16:08 | <Ms2ger> | Adding a cowboy hat, maybe? |
| 16:10 | <MikeSmith> | either that or one of my pimp canes |
| 16:40 | <Ms2ger> | "Leading the Web to its fool potential" |
| 16:47 | <divya> | :))) |
| 16:47 | <divya> | thats what i was wondering MikeSmith |
| 16:48 | <divya> | that dude looks so not like lars |
| 17:54 | <MikeSmith> | divya: indeed |
| 17:54 | <MikeSmith> | reminds me of somebody, but definitely not Lars |
| 17:55 | <MikeSmith> | https://mobile.twitter.com/kripken/status/292685421282988033 |
| 18:23 | <hsivonen> | has annevk's URL spec been implemented in Java yet? |
| 19:20 | <zewt> | gar i need to write a gm script to somehow stop href from being changed during onmousedown, to kill those obnoxious fake-link things google and other sites do |
| 19:34 | <zewt> | http://pastebin.com/vcjZ7nuh bad enough that google was doing it, now other sites are starting to copy it |
| 19:50 | <zewt> | oh i left it killing all onmousedowns, that won't exactly scale heh |
| 21:33 | <Moo-_-> | zewt: aah... so that's how they do it |