| 00:00 | <Rik`> | funny for a contest but useless |
| 00:09 | <AryehGregor> | Dashiva, it's only within 1% if you don't account for the JS required to decode it. |
| 02:28 | <shepazu> | Hixie: reading http://www.w3.org/TR/html5/webappapis.html#event-loops … would events from different devices, such as a mouse and a keyboard, go into the same or different task queues (for the same task loop)? |
| 02:28 | <shepazu> | I was a little confused by your example |
| 02:29 | <abarth> | MikeSmith: enabling fragment parsing is now in the commit pipeline |
| 02:29 | <shepazu> | is it intended that different implementations may organize their task queues differently? |
| 02:29 | <Hixie> | shepazu: looking... |
| 02:29 | <Hixie> | shepazu: (as an aside, you really don't want to be using the TR/ version, it's woefully out of date) |
| 02:30 | <Hixie> | looking at http://www.whatwg.org/specs/web-apps/current-work/complete/webappapis.html#event-loops now |
| 02:30 | <Hixie> | shepazu: they'd likely all go into the user interaction task source |
| 02:31 | <Hixie> | shepazu: the rule of thumb is that you'd use the same task source for two tasks if it matters in what order those two tasks are processed |
| 02:31 | <Hixie> | shepazu: that is in fact the reason we have different task sources -- so that we don't define that, e.g., a packet coming from the network before a keystroke must be processed in that order |
| 02:32 | shepazu | is adding a section in DOM3Events about task loops and how they relate to the model used by D3E |
| 02:32 | <shepazu> | I see |
| 02:32 | <shepazu> | it's essentially defining a state machine, right? |
| 02:33 | <Hixie> | how so? |
| 02:33 | <Hixie> | it's an event loop |
| 02:34 | <Hixie> | basic building block of software for the past 20 or 30 years :-) |
| 02:37 | <shepazu> | maybe I've got that relationship reversed :) |
| 02:37 | <Hixie> | which relationship? |
| 02:37 | <Hixie> | i'm confused |
| 02:43 | <shepazu> | nm |
| 02:43 | shepazu | goes back to editing D3E |
| 03:18 | <MikeSmith> | abarth|afk: sweet -- thanks for the heads-up |
| 04:23 | <MikeSmith> | heh |
| 04:23 | <MikeSmith> | http://hg.diveintohtml5.org/hgweb.cgi/rev/7a0f186d7e178b6e87117ebf007dc4fe6d07349a |
| 04:23 | <MikeSmith> | "more MIME types for fonts. please stop emailing me about this. browsers don't care and neither do I." |
| 04:33 | <aho> | browsers also don't care about file extensions |
| 04:34 | <aho> | had to rename some of the fonts to .txt since bloody iis will pretend a file doesnt exist if it doesn't know its mime type |
| 05:02 | <MikeSmith> | http://lists.w3.org/Archives/Public/public-geolocation/2010Aug/0012.html |
| 05:02 | <MikeSmith> | "I expect more people would want to know the direction of gravity" |
| 05:03 | <MikeSmith> | dino |
| 05:03 | <MikeSmith> | I thought he was kidding at first |
| 05:03 | <MikeSmith> | I understand what he means in context |
| 05:03 | <MikeSmith> | but that sentence on its own is nice too |
| 05:11 | <MikeSmith> | abarth|afk: I predict that your statement "Claiming something is obsolete or deprecated without giving authors better tools for solving their problems is waste of time." is going to end up being quoted quite a lot. |
| 06:37 | <micheil> | Hixie: with websockets, what happens if key3 is > 8 bytes long? |
| 06:37 | <Hixie> | how can it be? |
| 06:38 | <micheil> | for instance, a malformed connection |
| 06:38 | <micheil> | eg, extra 0x0d 0x0a bytes |
| 06:38 | <Hixie> | those wouldn't be part of the key |
| 06:38 | <micheil> | so something like: '^n:ds[4U\r\n\r\n\r\n\r\n\r\n' |
| 06:38 | <micheil> | was sent |
| 06:39 | <Hixie> | the 8 characters after the CR LF CR LF are key3 |
| 06:39 | <Hixie> | everything after that is frames |
| 06:39 | <micheil> | as the last part of the message, should the server only "read" 8 bytes for key3? |
| 06:39 | <Hixie> | the server should do what the spec says, which is indeed to just read 8 bytes |
| 06:39 | <Hixie> | read the spec :-) |
| 06:39 | <micheil> | k |
| 06:39 | <Hixie> | (i'm assuming it hasn't changed since i last touched it -- i don't edit that spec anymore so i could be wrong) |
| 06:39 | <micheil> | I wasn't sure from reading the spec. |
| 06:39 | <micheil> | good point. |
| 06:39 | <Hixie> | really? |
| 06:39 | <Hixie> | i thought it was pretty clear |
| 06:40 | <Hixie> | doesn't it just say "read 8 bytes" |
| 06:40 | <Hixie> | hm i guess it doesn't actually say how you get them |
| 06:40 | <Hixie> | it just says "The eight random bytes sent after the first 0x0D 0x0A 0x0D 0x0A sequence in the client's handshake" |
| 06:40 | <Hixie> | but that's still pretty unambiguous |
| 06:41 | <micheil> | it just mentions that the key3 should be 8 random bytes, not what to do if it isn't. |
| 06:42 | <micheil> | (for what ever reason, should the 8 random bytes get separated from the headers, I've got stuff to stitch the two back together, is that actually to spec?) |
| 06:42 | <micheil> | or should we just assume that the client is invalid and reject the connection? |
| 06:42 | <Hixie> | it says that key3 is "The eight random bytes sent after the first 0x0D 0x0A 0x0D 0x0A sequence in the client's handshake" |
| 06:43 | <Hixie> | the only way there could not be a key3, therefore, is if the client disconnects |
| 06:43 | <Hixie> | in which case it doesn't much matter what you do |
| 06:44 | <Hixie> | if the client disconnects, then you just have to keep reading til you get an 0x0D 0x0A 0x0D 0x0A sequence, and then the next 8 bytes are key3 |
| 06:44 | <micheil> | hmm.. |
| 06:44 | <Hixie> | doesn't matter if there's a frame after that or not |
| 06:44 | <Hixie> | er |
| 06:44 | <Hixie> | s/if the client disconnects/if the client doesn't disconnect/ |
| 06:44 | <micheil> | that's the slightly annoying thing about the websocket protocol; it assumes that you read() from the socket, not that you get data events |
| 06:46 | <micheil> | hmm.. I think I'll just reject the connection then, if key3 is missing / not correct length. |
| 06:47 | <abarth> | MikeSmith: hopefully in a good way... |
| 06:47 | <Hixie> | micheil: i still don't understand what you mean by "not correct length" |
| 06:48 | <Hixie> | micheil: if the server disconnects, though, then obviously you can also just reject the connection |
| 06:48 | <micheil> | Hixie: in my case, under node.js, we don't have a read(), we have to wait until we get a data event |
| 06:48 | <Hixie> | micheil: how is that different |
| 06:48 | <micheil> | if you connect to a websocket server using something like telnet, and pass correct headers, but send \r\n after the key3, then key3 ends up being 10 bytes instead of 8 bytes |
| 06:49 | <Hixie> | no |
| 06:49 | <Hixie> | it's just 8 bytes |
| 06:49 | <micheil> | we are left to guess what key3 is. |
| 06:49 | <Hixie> | the \r\n are the start of first frame |
| 06:49 | <Hixie> | after the key |
| 06:50 | <Hixie> | imagine a client sending the handshake, \r\n\r\n, 8 bytes, and then 0xFF [length] [data] 0xFF [length] [data] |
| 06:50 | <Hixie> | all at once |
| 06:50 | <MikeSmith> | abarth: yeah |
| 06:50 | <Hixie> | you should treat that as a handshake followed by two frames |
| 06:50 | <Hixie> | (at least unless the spec is changed) |
| 06:51 | <micheil> | Hixie: fair point. just trying to work out how to represent that in code for node.js |
| 06:51 | <Hixie> | MikeSmith: thanks, i threw your mail onto my pile |
| 06:52 | <Hixie> | MikeSmith: i'll think about how to do it... maybe a data- attribute? |
| 06:52 | <Hixie> | MikeSmith: though then i guess i'll have to strip it for the w3.org version |
| 06:52 | <MikeSmith> | data- attribute sounds fine to me |
| 06:52 | <Hixie> | k |
| 06:52 | <Hixie> | anyway, afk for now |
| 06:53 | <MikeSmith> | and if you do go that route, please leave it in the w3.org version too.. I'll deal with teh consequences |
| 06:58 | <boblet> | argh, CSS3 animation plus transition on same element is a world of hurt |
| 08:20 | <hsivonen> | interesting. Google search puts <link rel=prefetch> inside an <ol> |
| 08:21 | <hsivonen> | sure looks like a use case for <a rel=prefetch> |
| 08:21 | <hsivonen> | since the same link follows as an <a href> |
| 08:21 | <Peter`> | <a rel=prefetch> is allowed now afaik |
| 08:21 | <hsivonen> | Peter`: but is it supported by Gecko, etc.? (AFAIK, no.) |
| 08:22 | <Peter`> | The WebKit change should be trivial |
| 08:22 | <Peter`> | for <a> and <area> support anyway |
| 08:22 | <Peter`> | if the <link> version already works, outside of <head> as well, it makes sense for Google to be using it I guess |
| 08:23 | <Peter`> | even though it's not exactly clean |
| 08:23 | <hsivonen> | I'm not saying it doesn't make sense for them |
| 08:25 | <Peter`> | I didn't interpreted it as that |
| 08:28 | <hsivonen> | <iframe name=wgjf style=display:none src="" onload="google.j.l()" onerror="google.j.e()"> |
| 08:28 | <hsivonen> | that's also in Google's search results |
| 08:28 | <hsivonen> | relying an about:blank's onload can't be good... |
| 08:31 | <annevk5> | why not? |
| 08:31 | <annevk5> | it ought to work |
| 08:32 | <abarth> | it's pretty fast too |
| 08:32 | <abarth> | synchronous even |
| 08:32 | <hsivonen> | relying on synchronous stuff is suspicious when a page is breaking... |
| 08:33 | <annevk5> | the onerror is a bit dubious though |
| 08:33 | <hsivonen> | I wonder why this CSS is brought in as a script as opposed to straight CSS file: http://www.google.com.au/extern_chrome/46e7c64d3eb6ebfb.js |
| 08:33 | <annevk5> | that should never fire, though maybe they load more in it later? |
| 08:34 | <annevk5> | "The requested URL /extern_chrome/46e7c64d3eb6ebfb.js was not found on this server." |
| 08:34 | <annevk5> | what are you researching btw? |
| 08:34 | <annevk5> | fun parser things? |
| 08:36 | <hsivonen> | I'm researching why google.com.au serves single-pixel GIFs as data: URLs when UA substring "Firefox" exists *and* html5.enable=true but serves the right JPEGs as data: URLs when the UA substring "Firefox" exists and html5.enable=false |
| 08:37 | <hsivonen> | (aside: when even Google sniffs for "Firefox" rather than "Gecko/", the "Gecko is Gecko" thing should be recognized as a lost cause.) |
| 08:37 | <Hixie> | ooh, hsivonen has returned! |
| 08:43 | <hsivonen> | it appears the purpose of this over-complex setup is to both put the images and the HTML in the same resource *and* to order them in such a way that the text loads before the image data is loaded |
| 08:43 | hsivonen | wonders if this optimization is really necessary and really an optimization |
| 08:43 | <hsivonen> | SPDY could help here... |
| 08:44 | hsivonen | worders why Google puts single-letter comments all over the place |
| 08:45 | <hsivonen> | *wonders |
| 08:45 | <dbaron> | Hmmm... I can't quite yet mark Mozilla bug 915 invalid based on HTML5, since a tiny piece of the presentational attributes on col/colgroup crept in to HTML5. |
| 08:45 | <Hixie> | what part! |
| 08:45 | <Hixie> | tell me where so i can nuke it! |
| 08:45 | <dbaron> | Hixie, see my most recent message to public-html |
| 08:45 | Hixie | gets out the gattling run |
| 08:45 | <annevk5> | gun? |
| 08:45 | <dbaron> | Hixie, http://lists.w3.org/Archives/Public/public-html/2010Aug/0265.html |
| 08:46 | <Hixie> | gatling gun even |
| 08:46 | <Hixie> | who knew that only had one t |
| 08:46 | <hsivonen> | and there's no useless use of <em> and <cite> where they could save a few bytes by using <b> |
| 08:47 | <annevk5> | s/no//? |
| 08:47 | <hsivonen> | annevk5: oops. yes |
| 08:47 | <hsivonen> | or s/no/now/, rather |
| 08:48 | <Hixie> | dbaron: fixed |
| 08:48 | <annevk5> | hsivonen, fun stuff (in a way) |
| 08:50 | <annevk5> | Hixie, you changed thead into head |
| 08:50 | <annevk5> | Hixie, for the first set of rules |
| 08:50 | <Hixie> | oops |
| 08:51 | <hsivonen> | and yes, they do end up loading something else into the iframe |
| 08:52 | <hsivonen> | it would be nice if Google had a non-minified debug mode... |
| 09:01 | <TabAtkins_> | hsivonen: There's been a feature request for that for a while now. |
| 09:06 | <hsivonen> | Is there someone at Google who I could CC in the Mozilla bug to find out what exactly they are trying to do? |
| 09:06 | <Hixie> | you can mail me (ianh⊙gc) if you want me to stick someone on the bug |
| 09:07 | <hsivonen> | Hixie: thanks |
| 09:07 | <Hixie> | please mention in the e-mail what general area we're talking about |
| 09:07 | <Hixie> | since, you know, it's a big company and all |
| 09:11 | <othermaciej> | good evening |
| 09:11 | <abarth> | othermaciej: evening |
| 09:27 | <hsivonen> | Hixie: I sent email with my @mozilla.com address in the From field |
| 09:28 | <Hixie> | thanks |
| 09:29 | <Hixie> | nn |
| 09:31 | <annevk5> | g'n |
| 09:36 | <Ms2ger> | Hixie, "... then always assume the normative part is wrong." Really? |
| 09:41 | <roc> | TabAtkins: do you remember off the top of your head what was the result of that interminable www-style thread about what a 10px blur means in box-shadow? |
| 09:42 | <ruby_on_tails> | hello |
| 09:42 | <ruby_on_tails> | why don't I see a shadow here http://software.hixie.ch/utilities/js/canvas/?c.clearRect(0%2C%200%2C%20640%2C%20480)%3B%0Ac.save()%3B%0Ac.fillStyle%20%3D%20c.createPattern(img3%2C%20%27repeat%27)%3B%0Ac.strokeStyle%20%3D%20%27black%27%3B%0Ac.lineWidth%20%3D%208%3B%0Ac.lineCap%20%3D%20%27round%27%3B%0Ac.beginPath()%3B%0Ac.moveTo(300%2C%20200)%3B%0Ac.arc(300%2C%20200%2C%20100%2C%20-Math.PI%2F4%2C%20Math.PI%2F4%2C%20true)%3B%0Ac.closePath()%3B%0Ac.fill() |
| 09:44 | <ruby_on_tails> | anyone alive :-/ |
| 09:45 | <espadrine> | perhaps the intent was to c.stroke() ? |
| 09:45 | <espadrine> | since you set c.strokeStyle... |
| 09:45 | <ruby_on_tails> | so I can't have a shadow with a stroke ? |
| 09:46 | <ruby_on_tails> | I removed stroke, still no shadow :( |
| 09:47 | <TabAtkins_> | roc: Yeah. |
| 09:47 | <TabAtkins_> | roc: The blur should approximate a gaussian with a stdev of half the specified length. |
| 09:48 | <TabAtkins_> | roc: So, for that case, blur with a stdev of 5px. |
| 09:48 | <roc> | hmmmm ok |
| 09:48 | <TabAtkins_> | That gives you 97% transparency or so at the specified length. |
| 09:49 | <micheil> | anyone know if any browsers actually support websockets over TLS? |
| 09:49 | <micheil> | (using wss:// instead of ws://) |
| 09:50 | <ruby_on_tails> | anyone can tell why I can't see a shadow ? |
| 09:50 | <espadrine> | ruby_on_tails: you get a shadow with c.shadowOffsetX and c.shadowColor set |
| 09:51 | <espadrine> | ruby_on_tails: read the spec: http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#the-canvas-element |
| 09:51 | <ruby_on_tails> | I am on that page |
| 09:52 | <ruby_on_tails> | but I applied those on the link I pasted yet I don't see the shadow |
| 09:52 | <ruby_on_tails> | is it visible on your screen ? |
| 09:52 | <othermaciej> | TabAtkins_: I am hella curious about your recent twitter posts about "sunk costs" and all that |
| 09:52 | <othermaciej> | TabAtkins_: was there some kind of fight over something? |
| 09:52 | <TabAtkins_> | othermaciej: Minor. |
| 09:53 | othermaciej | enjoys watching drama that he doesn't have to be involved in |
| 09:53 | <espadrine> | ruby_on_tails: add the lines c.shadowOffsetX = 5 and c.shadowColor = 'blue' |
| 09:53 | TabAtkins_ | deleted a few tweets in that vein. |
| 09:53 | <othermaciej> | TabAtkins_: was it about the test suite or something? |
| 09:53 | <TabAtkins_> | Yeah. |
| 09:53 | <othermaciej> | are you at liberty to state what the actual point of controversy was? |
| 09:54 | <othermaciej> | I was trying to guess from the twitter record (from you and others) and I couldn't even guess wtf people were talking about |
| 09:54 | <ruby_on_tails> | espadrine: http://tinyurl.com/3yqwuaw |
| 09:55 | <TabAtkins_> | I dunno. I'll take it private just in case. |
| 09:56 | <espadrine> | ruby_on_tails: if you don't have a shadow here, use a browser that implements canvas shadows |
| 09:56 | <ruby_on_tails> | I have been drawing on canvas since 2 days using this browser, shadows can be missing ? |
| 09:56 | <ruby_on_tails> | I am on FF 3.0.8 |
| 09:57 | <ruby_on_tails> | oh well I see the shadow on Shiretoko |
| 09:57 | <ruby_on_tails> | damn |
| 09:57 | <roc> | FF 3.0.8 is somewhat ... old |
| 09:57 | <espadrine> | ruby_on_tails: yes, they can. I don't know if FF 3 has it |
| 09:58 | <ruby_on_tails> | hmm |
| 10:00 | <espadrine> | roc: tell that to my website designer, she is still using FF 2.0 |
| 10:00 | <TabAtkins_> | lolwut |
| 10:04 | <ruby_on_tails> | oh well, now the shadow is getting applied to all stuff, why is that happening ? |
| 10:05 | <ruby_on_tails> | is shadow applied to everything as a rule ? |
| 10:05 | <TabAtkins_> | "All drawing operations are affected by the four global shadow attributes." |
| 10:06 | <hsivonen> | ruby_on_tails: it's a really bad idea to continue to use a Web browser that's no longer getting security patches |
| 10:06 | <ruby_on_tails> | TabAtkins_: missed that, sorry |
| 10:06 | <ruby_on_tails> | set it to transparent |
| 10:06 | <hsivonen> | oops, I meant Firefox 2.0 but that was mentioned by espadrine instead |
| 10:06 | <ruby_on_tails> | hsivonen: am too lazy :P my lappy is in a bad condition latesty wise |
| 10:07 | <hsivonen> | well, actually, Firefox 3.0 has been EOLed, too |
| 10:08 | <ruby_on_tails> | EOLed means ? |
| 10:08 | <hsivonen> | end-of-life'ed |
| 10:09 | <hsivonen> | no longer maintained |
| 10:09 | <ruby_on_tails> | ok |
| 10:10 | <ruby_on_tails> | I actually don't use the latest FF because it doesn't have anti-aliased text |
| 10:11 | <hsivonen> | ruby_on_tails: huh? which platform? |
| 10:12 | <ruby_on_tails> | ubuntu |
| 10:12 | <ruby_on_tails> | my FF3.0.8 has it |
| 10:12 | <hsivonen> | that's really odd. |
| 10:13 | <hsivonen> | Ever since at least dapper, every version of Firefox I've seen on Ubuntu has respected the Gnome anti-aliasing prefs |
| 10:13 | <hsivonen> | Works for me on Lucid |
| 10:13 | <ruby_on_tails> | I am on 9.04 |
| 10:14 | <ruby_on_tails> | the only way to make hover effects like tooltips are possible with divs on above the canvas? |
| 10:15 | <hsivonen> | ruby_on_tails: anti-aliased text worked for me on 9.04 |
| 10:15 | <ruby_on_tails> | my laptop is broken maybe lol |
| 10:15 | <ruby_on_tails> | god knows |
| 10:15 | <hsivonen> | ruby_on_tails: how did you obtain Firefox? |
| 10:15 | <ruby_on_tails> | I need a complete reformat |
| 10:15 | <ruby_on_tails> | I don't remember |
| 10:15 | <ruby_on_tails> | I have FF 3.0.8 and Shiretoko(from synaptec) |
| 10:18 | <hsivonen> | hmm. a-a text in Shiretoko from backports worked for me, too |
| 10:18 | <hsivonen> | I wonder if there's a pref that allows you to override the Gnome AA prefs in Gecko... |
| 10:21 | <ruby_on_tails> | hmm |
| 11:30 | <MikeSmith> | I thought the spec explicitly disallowed the xml declaration in text/html documents |
| 11:30 | <MikeSmith> | but I can't find any statement of that constraint in the spec |
| 11:34 | <hsivonen> | MikeSmith: it parses as a bogus comment and bogus comments are parse errors |
| 11:34 | <hsivonen> | MikeSmith: that might not count as "explicit", though |
| 11:34 | <MikeSmith> | hsivonen: ok |
| 11:34 | <MikeSmith> | yeah |
| 11:35 | <MikeSmith> | It seems like that it needs to be explicitly defined as a document-conformance error as well |
| 11:35 | <hsivonen> | all parse errors are |
| 11:36 | <hsivonen> | and I believe there's nothing in the "writing" section permitting it |
| 11:36 | <hsivonen> | what's not permitted is prohibited |
| 11:36 | <hsivonen> | (unlike in RSS...) |
| 11:37 | <MikeSmith> | ok |
| 11:38 | <MikeSmith> | I still think it would be better to explicitly disallow it |
| 11:38 | <MikeSmith> | for the sake of clarity |
| 11:38 | <hsivonen> | it probably makes sense to have a note that calls it out |
| 11:38 | <MikeSmith> | yeah |
| 12:02 | <hsivonen> | I wonder how many Objection Polls I missed while I was on vacation... I've noticed 2 so far. |
| 12:10 | <Lachy> | hsivonen, what was the last poll you did see before you went on vacation? |
| 12:12 | <hsivonen> | Lachy: versioning and ascii-ref |
| 12:16 | <jgraham> | <math><mi><object><span></span></object></mi><mi> |
| 12:16 | <jgraham> | The second <mi> ends up in the HTML namespace |
| 12:16 | <jgraham> | spec bug? |
| 12:16 | <jgraham> | Or bug in my head? |
| 12:19 | <hsivonen> | I'm positively surprised by http://lists.w3.org/Archives/Public/public-html/2010Aug/0116.html |
| 12:28 | <tantek> | hsivonen - whoa - I missed that too, similarly positively/pleasantly surprised. |
| 12:34 | <Lachy> | what have we had since those two polls? Have I missed some too? |
| 12:36 | <hsivonen> | Lachy: aria section title + characterization and removing srcdoc |
| 12:36 | <hsivonen> | (and indeed, there haven't been more) |
| 12:40 | <wirepair> | silly question, how would one remove a nameless cookie? |
| 12:40 | <wirepair> | like document.cookie = '=asdf'; |
| 12:42 | <wirepair> | er. how would one remove the nameless cookie from javascript ;) |
| 12:46 | <wirepair> | nevermind got it |
| 12:55 | <david_carlisle> | jgraham: "The second <mi> ends up in the HTML namespace," watching with interest... sort of related to bug 9887 |
| 12:57 | <david_carlisle> | also who registered the irc name "davidc" ? nickserv told me to rename myself, what you need is namespaces to avoid such clashes. |
| 12:58 | <annevk5> | namespaces on IRC? haha |
| 13:00 | <Ms2ger> | /nick http://www.w3.org/1998/Math/MathML/davidc? |
| 13:01 | <Lachy> | obviously, we need a prefix mechanism to make that more usable |
| 13:01 | <Peter`> | david_carlisle: he registered almost 10 years ago |
| 13:02 | <jgraham> | david_carlisle: The problem is, I think that when you have a scoping element inside a <mn> then an end tag causes you to switch back to the secondary insertion mode i.e. out of the mode in which you insert things in a non-HTML namespace |
| 13:02 | <jgraham> | or <mi> |
| 13:02 | <jgraham> | even |
| 13:02 | <jgraham> | it looks like a simple bug in the spec |
| 13:03 | <jgraham> | hsivonen: Any opinion? |
| 13:03 | <hsivonen> | jgraham: if the second <mi> is not in the MathML namespace, I'd consider it a bug somewher |
| 13:03 | <hsivonen> | e |
| 13:03 | <jgraham> | (I say "simple bug" but I'm not quite sure what the right fix is of if it is simple) |
| 13:04 | <jgraham> | (possibly it shouldn't worry about scoping elements when deciding to switch or not) |
| 13:05 | <jgraham> | (maybe that plays badly with <foreignContent> though) |
| 13:06 | <hsivonen> | foreignObject definitely needs to be scoping |
| 13:07 | <hsivonen> | (and everyone keeps calling it foreignContent) |
| 13:07 | <hsivonen> | maybe I misunderstood your remark about scopingness here |
| 13:08 | <jgraham> | Argh |
| 13:08 | <jgraham> | I always do that :) |
| 13:12 | <jgraham> | hsivonen: The point is that in the "Any other end tag" part of In Foreign Content, you switch to the SIM if there is no non-HTML element in scope. But there can be a non-HTML element on the stack of open elements even if it is not in scope |
| 13:12 | <jgraham> | In which case you switch too early |
| 13:12 | <hsivonen> | jgraham: oh |
| 13:13 | <jgraham> | I think |
| 13:13 | <hsivonen> | this context-dependent stuff is harder than it first seemed |
| 13:13 | <jgraham> | At least that is consistent with what gecko and html5lib do |
| 13:13 | <hsivonen> | still beats making the author use xmlns |
| 13:13 | <jgraham> | Yeah, but it is disappointingly complex |
| 13:14 | <hsivonen> | is it just me or isn't it rather out of orderish to start discussing longdesc use cases right *after* the poll on the topic? |
| 13:16 | <jgraham> | AFAICT the main conclusion from that is that there exists a contingent who believe that Process is important and must be strictly followed whilst they percieve it is helping their cause, but who ignore it once a decision goes against them |
| 13:17 | <hsivonen> | jgraham: shocking |
| 14:43 | <cyberix_> | Is browser support for XHR2 being tested/tacked somewhere? |
| 14:44 | <annevk5> | not really |
| 14:45 | <annevk5> | I plan to write XHR2 tests at some point, but haven't done it so far |
| 14:45 | <cyberix_> | I found out that you could use the file writer API to generate blobs from raw data |
| 14:45 | <annevk5> | CSSOM has somewhat higher priority and XHR1 tests are still not properly published (but they are at least written) |
| 14:46 | <cyberix_> | so I guess it should then be possible to post raw data with XHR2 |
| 14:46 | <annevk5> | cyberix_, well, that feature hasn't even made it into the draft yet |
| 14:46 | <annevk5> | cyberix_, oh, posting Blob objects is in there, yes |
| 14:47 | <annevk5> | responseBlob not yet, but I think we pretty much agreed on how it ought to work |
| 14:47 | <cyberix_> | Well, that we already have working |
| 14:49 | <cyberix_> | we just read it as a string, and do (charcode & 0xff) |
| 14:49 | <annevk5> | cyberix_, cross-origin support is in WebKit/Gecko |
| 14:49 | <cyberix_> | I have tried it out |
| 14:49 | <annevk5> | cyberix_, I think posting Blob might be too |
| 14:50 | <annevk5> | there's too much going on to really keep track, even being the editor |
| 14:50 | <cyberix_> | not blaming you |
| 14:51 | <annevk5> | no worries, I don't feel blamed :) |
| 14:51 | <cyberix_> | just scared of writing this without seeing a working example first :-) |
| 14:56 | <boblet> | any chrome ppl around? I’ve found a fun CSS3 transform+animation issue; works when in head, doesn’t when linked |
| 14:56 | <annevk5> | are you sure the linked style sheet applies? |
| 14:57 | <boblet> | annevk5: well, the anim fires in both situations, and I can see styles applied in inspector, but animated object ends up in different place when linked |
| 14:58 | <boblet> | not sure if I’m just doing something wrong or not |
| 14:58 | <boblet> | also only happens in new tab, not on reload |
| 14:59 | <cyberix_> | annevk5: Should there not be one "HTML5: the Blob" document that would explain the datatype? |
| 15:00 | <cyberix_> | at the moment it is defined in, what, ten places? |
| 15:00 | <boblet> | annevk5: will set up example to show you… |
| 15:02 | <annevk5> | cyberix_, I'm not very happy with the whole Blob/File mess to be honest |
| 15:04 | <annevk5> | cyberix_, but I don't want to make it my problem |
| 15:07 | <cyberix_> | I see |
| 15:08 | <boblet> | annevk5: this is as intended (animation and transition CSS in head, other CSS in linked stylesheet): http://east.webdirections.org/wde/2010/ |
| 15:08 | <boblet> | annevk5: this is the same code moved from head to stylesheet: http://east.webdirections.org/wde/2010/test.html |
| 15:09 | <boblet> | in Chrome the animated arrows bg-img jumps after animating when in linked styles, but not in Safari |
| 15:10 | <boblet> | this new-fangled CSS3 stuff is confounding me |
| 15:14 | <cheeser> | it's awesome :) |
| 15:15 | <Rik`> | boblet: why are you putting some transitions if you use animations ? |
| 15:16 | <boblet> | Rik`: better browser support for transitions, but the transition on page load can only be done via an animation or JS |
| 15:17 | <hsivonen> | annevk5: is HTML parsing in responseXML still a possibility for XHR2? |
| 15:17 | <boblet> | cheeser: was that re: my links? heh, thanks :) apart from the broken bit (and occasional flashing after transition in Safari, mostly sorted) |
| 15:18 | <annevk5> | hsivonen, it's still part of XHR2 |
| 15:18 | <annevk5> | hsivonen, I tried getting it in XHR1 but that was controversial |
| 15:19 | <hsivonen> | annevk5: OK |
| 15:21 | jgraham | wishes responseXML didn't exist |
| 15:22 | <jgraham> | Also I wish the live DOM viewer showed namespaces |
| 15:22 | <jgraham> | and a pony |
| 15:22 | <jgraham> | with a single horn from its forehead |
| 15:22 | <Rik`> | boblet: you're only using -webkit-transition so, browser support is quite the same as animations |
| 15:22 | <annevk5> | responseXML is nice |
| 15:23 | <jgraham> | annevk5: Why? being able to parse XML from a string is much more flexible |
| 15:23 | <jgraham> | and it is hardly any more code |
| 15:23 | <boblet> | Rik`: once it actually works… :p |
| 15:23 | <jgraham> | or any slower |
| 15:25 | <annevk5> | yeah, and strings are not needed either if we have binary |
| 15:26 | <annevk5> | but okay... |
| 15:27 | <jgraham> | Not really a reasonable argument |
| 15:27 | <jgraham> | I mean we don't have responseJSON or responseSexp or whatever |
| 15:27 | <annevk5> | Sexp? |
| 15:28 | <jgraham> | s-expression |
| 15:28 | <annevk5> | we will get responseBlob |
| 15:28 | <annevk5> | responseJSON has been proposed at some point |
| 15:28 | <jgraham> | It just seems like API bloat to me |
| 15:28 | <espadrine> | Isn't there some JSONRequest adoption in browsers? |
| 15:28 | <jgraham> | making priviledged accessors for some formats |
| 15:29 | <espadrine> | I know XHR doesn't like that |
| 15:29 | <annevk5> | XHR doesn't like it? |
| 15:29 | <espadrine> | meaning the spec says it doesn't want JSONRequest |
| 15:30 | <annevk5> | that's CORS |
| 15:30 | <espadrine> | yep |
| 15:30 | <annevk5> | but yeah, it's not going into browsers |
| 16:00 | <boblet> | nn |
| 16:09 | <Fyrd> | Does anyone here know if being able to load a data URI as a standalone file in a browser window is part of a standard? |
| 16:11 | <Fyrd> | Currently all browsers but IE support this...I'd like IE9 to support it too, but they probably won't if there isn't a spec that says they should. |
| 16:11 | <cyberix_> | BlobBuilder seems to only work for text. Is there a way of turning a data url into a blob? |
| 16:12 | <Peter`> | Fyrd: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-August/028011.html |
| 16:12 | <Peter`> | Currently data: URLs actually don't even validate |
| 16:14 | <Fyrd> | Ah, thanks Peter` . So this information is currently lacking, I see... |
| 16:17 | <Fyrd> | Too bad no one's responded to that email yet. |
| 16:19 | <TabAtkins_> | Fyrd: IE doesn't allow it as a "security issue". Not sure of the details. |
| 16:20 | <annevk5> | oh, hybi realizes Shelby Moore is a troll |
| 16:21 | <annevk5> | that's about time |
| 16:21 | <annevk5> | I've ignored most emails from the past week or so |
| 16:22 | <Fyrd> | TabAtkins_: Yeah, that's what I suspected, though they never did explain the reason in the bug I submitted... |
| 16:22 | <Rik`> | is there a definitive answer in one spec about what "'onload' in window" should return ? |
| 16:22 | <annevk5> | Peter`, data: URLs should validate... |
| 16:22 | <annevk5> | Peter`, why do you keep insisting they don't? |
| 16:22 | <Ms2ger> | Rik`, I expect WebIDL would say yes |
| 16:23 | <Rik`> | Ms2ger: so it is a bug in Firefox if it doesn't ? |
| 16:23 | <Ms2ger> | It doesn't? Could you file a bug? |
| 16:24 | <Ms2ger> | (And cc me) |
| 16:24 | <Rik`> | Ms2ger: I'm just doing some background research first |
| 16:24 | <Rik`> | 'onhashchange' in window returns true |
| 16:24 | <annevk5> | Opera also says false |
| 16:24 | <Rik`> | but 'onpopstate' in window returns false |
| 16:24 | <Rik`> | it's very inconsistent |
| 16:37 | <Rik`> | ok, IDL is too strange for me to absorb right now, so i'll file a bug :) |
| 17:41 | <Rik`> | Ms2ger: created https://bugzilla.mozilla.org/show_bug.cgi?id=590198 |
| 17:43 | GPHemsley | wonders why none of tantek's time/date proposals have gotten any response |
| 17:43 | <tantek> | GPHemsley - quite a bit of follow-up on the wiki |
| 17:43 | <GPHemsley> | positive? because so far I like all of your ideas :) |
| 17:43 | <tantek> | (which is where my emails requested people follow-up, rather than getting lost in email arhives :) ) |
| 17:43 | <GPHemsley> | (I'm currently catching up on my WHATWG e-mail) |
| 17:43 | <tantek> | mostly positive yes |
| 17:45 | <GPHemsley> | oh, hmm... actually, I'm not sure I agree with your AM/PM proposal |
| 17:46 | <tantek> | http://wiki.whatwg.org/wiki/Time_element |
| 17:46 | <tantek> | what are your concerns about am/pm? |
| 17:46 | <tantek> | we've found it quite useful in microformats |
| 17:47 | <tantek> | (already have multiple interoperable implementations) |
| 17:47 | <tantek> | (and it helps reduce duplicate data) |
| 17:47 | <tantek> | parsing it is trivial |
| 17:47 | <tantek> | and it really helps content stay slimmer |
| 17:47 | <tantek> | (markup in particular) |
| 17:48 | <GPHemsley> | I just don't see the DRY value |
| 17:49 | <Cesarino> | Hi all, HTML4 and so on works in a html5 page? |
| 17:50 | <GPHemsley> | tantek: Well, alright, I concede that the proposal addresses my immediate concerns... but I personally don't like it, still |
| 17:51 | <Ms2ger> | Rik`, thanks |
| 17:51 | <tantek_> | what don't you like about it? |
| 17:52 | <GPHemsley> | tantek_: It just doesn't feel right. IDK, I have no evidence to back up my opinion :P |
| 18:03 | AryehGregor | gives kudos to mpilgrim on http://wearehugh.com/public/2010/08/html5-video-accessibility/ and http://wearehugh.com/public/2010/08/html5-web-workers/, both awesome |
| 18:30 | <Ms2ger> | Rik`, fwiw, we've got a test for `"onhashchange" in window` |
| 18:36 | <Hixie> | Ms2ger: ah crap, i means the non-normative part |
| 18:37 | <Ms2ger> | Hixie, thought so |
| 19:57 | <annevk5> | latest Shelby Moore: "I knew that, I just wanted you to admit it. \I have known for a while what sort of systems you are building." |
| 19:59 | <abarth|adoptPtr> | i didn't realize he's the axiomatic proof guy |
| 20:00 | <annevk5> | I hope my links will make people realize they are wasting their time and have the chairs put a ban on him |
| 20:00 | <annevk5> | but maybe I'm optimistic |
| 20:08 | <annevk5> | oh, "Last email to the list." |
| 20:09 | <annevk5> | curious to see how long that lasts |
| 20:11 | <Ms2ger> | So is this Shelby or Shelly? |
| 20:12 | <Hixie> | they are distinct people |
| 20:17 | <gavin> | what list is this? |
| 20:17 | <annevk5> | hybi |
| 20:17 | <annevk5> | it's the same Shelby Moore from www-style 2002-2005 (fwiw) |
| 20:19 | <Hixie> | more like 2002,2005 iirc :-) |
| 20:21 | <ojan> | annevk5: are there any cases where returning "auto" from getComputedStyle is correct? |
| 20:22 | <annevk5> | yeah |
| 20:22 | dbaron | looks around for scrollback |
| 20:22 | <annevk5> | ojan, e.g. for 'top' |
| 20:22 | <annevk5> | iirc |
| 20:23 | <ojan> | annevk5: i'm looking at https://bugs.webkit.org/show_bug.cgi?id=43559. there's a case where WebKit returns 0px and Gecko returns auto |
| 20:23 | <dbaron> | also for other things that are display:none or inside something display:none, I think |
| 20:23 | <annevk5> | ojan, the reason for that is because I aligned with WebKit; Gecko returns the used value there |
| 20:24 | <annevk5> | ojan, oh, that's the reverse |
| 20:24 | annevk5 | looks |
| 20:24 | <ojan> | it's getting the computed width/height on an uncached image |
| 20:24 | <annevk5> | dbaron, yeah, or where the property is in the special list and does not apply to the element |
| 20:24 | <ojan> | shift+reload http://tinyurl.com/24qmgck to see the difference |
| 20:25 | <annevk5> | ojan, I guess that is because in Gecko the image is treated as inline element so 'width' does not apply and therefore returns the computed value rather than the used value |
| 20:26 | <ojan> | oh, interesting. |
| 20:26 | <annevk5> | ojan, pretty sucky that pages can depend on how browsers generate boxes for replaced elements while loading :/ |
| 20:27 | <ojan> | this sounds like a scary thing to change (for webkit or gecko) :( |
| 20:27 | <annevk5> | sounds like an annoying thing to define too |
| 20:27 | <annevk5> | i guess HTML5 should define that |
| 20:28 | annevk5 | checks the rendering chapter |
| 20:30 | <annevk5> | http://www.whatwg.org/specs/web-apps/current-work/complete/rendering.html#images-0 has the WebKit behavior |
| 20:30 | <ojan> | annevk5: i'm missing something. according to http://www.w3.org/TR/css3-values/#specified0, the computed value should never be "auto" no? |
| 20:31 | <ojan> | oh, nm |
| 20:31 | <ojan> | i misread |
| 20:31 | <annevk5> | http://www.w3.org/TR/CSS21/visudet.html#the-height-property "the percentage or 'auto' (see prose under <percentage>) or the absolute length; 'auto' if the property does not apply" |
| 20:32 | <annevk5> | (and it does not apply with the Gecko <img> model; assuming I guessed all this correctly) |
| 20:34 | <crash\> | there is a Bug for that |
| 20:34 | <crash\> | https://bugzilla.mozilla.org/show_bug.cgi?id=321919 |
| 20:34 | <crash\> | pretty much the same problem |
| 20:35 | <annevk5> | crash\, with the same conclusions |
| 20:35 | <annevk5> | :) |
| 20:36 | <crash\> | what does getPropertyValue("width") for <input type=hidden> return? |
| 20:36 | <crash\> | should be auto too, right? |
| 20:37 | <annevk5> | yes |
| 20:39 | <crash\> | maybe I should create a bug and test case for that to clarify that |
| 20:39 | <Hixie> | Ms2ger: any idea what you meant on http://www.w3.org/Bugs/Public/show_bug.cgi?id=9642 ? |
| 20:40 | <annevk5> | make a test suite for getComputedStyle, that'd be neat :) |
| 20:42 | <crash\> | I can't see where the specs say. what behaviour is right |
| 20:43 | <annevk5> | http://dev.w3.org/csswg/cssom/#dom-window-getcomputedstyle defines it |
| 20:43 | <annevk5> | if anything is unclear please email www-style |
| 20:43 | <Ms2ger> | Hixie, that pre eats the first linefeed, if that is relevant to the section I filed it on |
| 20:43 | <annevk5> | Ms2ger, not if you are in PLAINTEXT mode |
| 20:43 | <annevk5> | (I think) |
| 20:44 | <Hixie> | i really didn't understand the bug at all |
| 20:45 | Ms2ger | checks |
| 20:46 | <annevk5> | oh actually, it just emits it as character token |
| 20:46 | <annevk5> | Hixie, if you're plain text file starts with a newline the parser would eat it |
| 20:46 | <annevk5> | s/you're/your/ geez |
| 20:46 | <Hixie> | good to know |
| 20:47 | <Hixie> | i'll fix it, no need to file a bug |
| 20:47 | <Hixie> | (since i'm between bugs right now) |
| 20:47 | <crash\> | "If the property applies to the element or pseudo-element and the resolved value of the 'display' property is not none, the resolved value is the used value." for width/height |
| 20:47 | <annevk5> | (that is the bug from Ms2ger :) ) |
| 20:47 | <crash\> | but a not loaded image is not of type display:none; |
| 20:48 | <annevk5> | crash\, no, it's of type inline in Gecko |
| 20:48 | <Hixie> | annevk5: no it wouldn't |
| 20:48 | <crash\> | yes, so it should be the computed value |
| 20:48 | <crash\> | so I guess 0px would be right |
| 20:48 | <annevk5> | no, auto |
| 20:48 | <annevk5> | we just discussed this in this channel |
| 20:48 | <Hixie> | however, document.open('text/plain');document.write('\n'); would, so i'll fix that instead |
| 20:49 | <annevk5> | I even quoted http://www.w3.org/TR/CSS21/visudet.html#the-height-property |
| 20:49 | <crash\> | ok, since it's inline |
| 20:49 | <annevk5> | Hixie, ah yeah, that is the bug |
| 20:49 | <Ms2ger> | Hixie, only because I filed the text/plain one earlier :) |
| 20:49 | <annevk5> | Hixie, didn't mean text/plain loading |
| 20:49 | <Hixie> | Ms2ger: :-D |
| 20:50 | <crash\> | but when it was replaced it can have a width/height |
| 20:50 | <crash\> | wired |
| 20:52 | <annevk5> | I never said this was easy ;p |
| 20:52 | <Hixie> | man some of these bugs are remarkably terse |
| 20:52 | Ms2ger | pleads guilty |
| 20:53 | <Hixie> | not all yours! |
| 20:54 | <crash\> | ist still dont think "auto" is right ;) |
| 20:54 | <crash\> | why it can't be 0px and when it can't be replaced and then alt="" is used be "auto"? |
| 20:55 | <Ms2ger> | Be happy I didn't file >300 bugs ;) |
| 20:55 | <annevk5> | hmm, one email and now I get private spam from Shelby Moore |
| 20:55 | <annevk5> | crash\, sure, that is what HTML5 defines right now |
| 20:55 | <Hixie> | i love bugs like http://www.w3.org/Bugs/Public/show_bug.cgi?id=9633 |
| 20:55 | <annevk5> | crash\, but apparently that breaks sites |
| 20:55 | <Hixie> | i have no idea WHAT should read that way... |
| 20:55 | <annevk5> | crash\, again see above for the WebKit bug report |
| 20:56 | <Ms2ger> | Hixie, add more ids |
| 20:57 | <Ms2ger> | When trying to use types other than image/png, authors can check if the image was really returned in the requested format by checking to see if the returned string starts with one the exact strings |
| 20:57 | <crash\> | ok, that confident me |
| 20:58 | <Hixie> | Ms2ger: yeah, i worked it out already, don't worry :-) |
| 20:58 | <annevk5> | anyways, nn |
| 20:59 | <Ms2ger> | Hixie, if you can work out HTML parsing, terse bug reports shouldn't be much of a challenge |
| 21:07 | <Hixie> | how about http://www.w3.org/Bugs/Public/show_bug.cgi?id=9664 |
| 21:07 | <Hixie> | does that make any sense to anyone? |
| 21:29 | <jcranmer> | Hixie: possible interpretations |
| 21:29 | <jcranmer> | 1. The bug reporter wants more identification of the meanings of a, b, etc. |
| 21:29 | <jcranmer> | 2. The bug reporter wants an explanation of transformation matrices |
| 21:31 | <jcranmer> | I'm not quite sure what he intends in the last clause of the disjunction |
| 21:31 | <jcranmer> | but the most clue I can get out of it is that the user lacks knowledge of transformations |
| 21:59 | <Hixie> | jcranmer: yeah that's as much as i could get out of it |
| 22:00 | <jcranmer> | an informative reference to 2D transformations may be helpful |
| 22:01 | <Hixie> | should i link to http://lmgtfy.com/?q=2d+transformations ? |
| 22:02 | <jcranmer> | too unstable a site |
| 22:02 | <jcranmer> | I said "may be" |
| 22:04 | <Hixie> | :-) |
| 22:27 | <sicking> | Hixie: you are aware that the spec currently does mandate filtering <option>s with empty values only in <datalist> and not in <select>, right? |
| 22:29 | <sicking> | sorry, i don't think we explicitly mentioned this in the original bug filing |
| 22:34 | <Hixie> | sicking: sure but <select> still needs live updating |
| 22:37 | <Hixie> | sicking: i don't see why it would be harder to decide not to show an option than it would be to show it with different text |
| 22:57 | <sicking> | Hixie: updating the UI is a wholly different set of code than the code that implements the .options list |
| 22:58 | <sicking> | Hixie: also, changing the value in the UI is very different from shrinking the length of the UI, or the length of the .options list |
| 22:58 | <Hixie> | wait, what's .options got to do with anything |
| 22:58 | <Hixie> | i thought we were talking about the autocompletion ui |
| 22:59 | <sicking> | Hixie: the .options list is what the spec says should should be affected by .value |
| 23:00 | <sicking> | Hixie: oh, hmm.. i guess it doesn't |
| 23:01 | <Hixie> | you mean datalist.options ? |
| 23:01 | <sicking> | yeah |
| 23:01 | <Hixie> | that's just a list of <option> elements |
| 23:01 | <Hixie> | no magic |
| 23:01 | <sicking> | ok, cool |
| 23:01 | <sicking> | we probably won't filter the UI either, for the same performance reasons |
| 23:02 | <sicking> | but since you can't test that, i care less what the spec says :) |
| 23:02 | <Hixie> | :-) |
| 23:02 | <Hixie> | that's mostly a quality of implementation issue |
| 23:02 | <sicking> | Hixie: for implementing lists like that, you generally want to know the number of options |
| 23:02 | <sicking> | as to calculate how big you make the scrollbar |
| 23:02 | <sicking> | so it's actually a lot more work to filter out empty-string ones, than to just show them as empty |
| 23:03 | <Hixie> | fair enough |
| 23:03 | <Hixie> | not sure you'd ever want a scrollbar for this stuff anyway |
| 23:03 | <Hixie> | but that's another story |
| 23:03 | <sicking> | true |
| 23:05 | <sicking> | Hixie: it's still a strange difference between <select> and <datalist>, <select> doesn't mention filtering out empty-value <option>s |
| 23:12 | <Hixie> | well it doesn't make much sense to show the ones whose value is empty |
| 23:12 | <Hixie> | since the behaviour would be to autocomplete nothing |
| 23:12 | <Hixie> | but it makes sense with <select> |
| 23:12 | <Hixie> | since the empty string is a valid value |
| 23:27 | <Hixie> | does css have a feature yet so that you can use the first inline box in a block as a marker, and have everything else indented as far as needed to get past that box on subsequent lines? |
| 23:38 | <tantek> | Hixie, as a marker like a list-item marker? or do you mean something else by marker? |
| 23:47 | <Hixie> | tantek: yeah like a list item marker |
| 23:48 | <tantek> | I don't know how you specify what the first inline box in a block would be. There's the first line box. And the first letter. But assuming you mean something different. |
| 23:48 | <tantek> | (since I have no AXIOMATIC PROOF ;) ) |
| 23:49 | <Hixie> | i mean something like <span> <input type=radio> foo bar baz </span> -- i want the foo bar baz to flow to multiple lines but all aligned so that subsequent lines start immediately below the "foo" |
| 23:50 | <tantek> | ah, you're talking about (essentially) display:marker |
| 23:50 | <Hixie> | yes, essentially |
| 23:50 | <Hixie> | but i don't mind if it's actually that |
| 23:50 | <Hixie> | just wandering if such a feature exists yet |
| 23:50 | <tantek> | I don't think that has made any progress since you and I touched that draft |
| 23:50 | <Hixie> | k |
| 23:50 | <tantek> | I can check tomorrow |
| 23:50 | <tantek> | at the meeting |
| 23:50 | <tantek> | (CSSWG @ Oslo) |