| 00:02 | <TabAtkins> | The Wii download screen is mario jumping into a block for every % the download progresses. |
| 00:03 | <annevk> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=9224 -- it seems highly unlikely implementations will follow through |
| 00:03 | <JonathanNeal> | I can't predict how Opera will handle video, the closest approximation would be video { height: 150px; width: 300px; } video[src] { height: auto; width: auto; } |
| 00:05 | <Hixie> | video { height: auto; width: auto; } would be accurate, since the spec says <video>'s intrinsic width defaults to 300x150 iirc |
| 00:07 | <JonathanNeal> | oh, does the spec also say <audio> should default to 300x25? |
| 00:07 | <gsnedders> | http://auriethepixie.deviantart.com/art/The-Secret-Life-of-Trees-158431899 |
| 00:07 | <gsnedders> | The lighting of that is awesome. |
| 00:08 | <JonathanNeal> | Opera reports inline for audio, but shouldn't it be inline-block? |
| 00:10 | <annevk> | I guess we just make it a replaced element |
| 00:10 | <nessy> | spec says nothing about audio default width/height |
| 00:10 | <annevk> | seems to be it should do the same as <img> / <video> ... |
| 00:11 | <annevk> | pretty sure it is inline for <img> |
| 00:11 | <JonathanNeal> | Yea. |
| 00:11 | <JonathanNeal> | I always thought that was a grandfathered inline. |
| 00:12 | <JonathanNeal> | I'll follow your guys advice on this one audio { display: none; } audio[controls] { display: inline; } video { display: inline; } |
| 00:13 | <annevk> | "The audio element, when it is exposing a user interface, is expected to be treated as a replaced element about one line high, as wide as is necessary to expose the user agent's user interface features. When an audio element is not exposing a user interface, the user agent is expected to hide it, irrespective of CSS rules." |
| 00:14 | <annevk> | so in effect you should not be able to display it even if you use display:block |
| 00:14 | <annevk> | that seems kind of contrary to <head> and <title> and all... |
| 00:16 | <JonathanNeal> | yes, it does. |
| 00:16 | <JonathanNeal> | and contrary to <img> |
| 00:17 | <JonathanNeal> | If you expect to express itself visually, it should be much more like <img> and <video> |
| 00:18 | <JonathanNeal> | but thanks for the heads up, annevk :) |
| 00:25 | <Hixie> | iirc people demanded that it be like that |
| 00:25 | <Hixie> | i forget why |
| 00:26 | <nessy> | maybe because it is assumed that most use of audio is as background music to the site? |
| 00:26 | <JonathanNeal> | Interactive Illinois Report Card? |
| 00:26 | roc | wonders what Mark Pilgrim has to say about http://blog.chromium.org/2010/03/bringing-improved-support-for-adobe.html |
| 00:26 | <JonathanNeal> | :-P |
| 00:27 | <JonathanNeal> | nessy, that makes sense for a control-less <audio> but once controls are introduced it seems very off. |
| 00:28 | <nessy> | I'm not saying it's the right decision :) |
| 00:28 | <nessy> | might be worth another discussion on the list, actually |
| 00:29 | <JonathanNeal> | "the list" |
| 00:29 | <Hixie> | JonathanNeal: wait, what? why does it make no sense when you have controls? |
| 00:30 | <Hixie> | i thought that was the one case that DID make sense |
| 00:30 | <annevk> | roc, uhuh |
| 00:31 | <annevk> | roc, in general putting much effort into a better plugin API sort of seems like a waste of time |
| 00:31 | <roc> | yeah |
| 00:31 | <roc> | honestly I don't think anyone at Mozilla is excited about Pepper |
| 00:31 | <roc> | we're not blocking it |
| 00:31 | <roc> | in fact this is the first sign I've seen that Adobe is interested in it |
| 00:32 | <Hixie> | JonathanNeal: btw if you want to look into why that rule is there, it was added in r4083 on 2009-10-05 |
| 00:32 | <JonathanNeal> | Okay, I'm very ready to be wrong about everything all the time, but it just seemed off. |
| 00:32 | <annevk> | oh, I have seen some other signs of people working for Adobe pushing for a better plugin API |
| 00:33 | <JonathanNeal> | I guess you don't want to be able to hide <audio controls> right? |
| 00:33 | <JonathanNeal> | Irrespective of CSS rules. |
| 00:33 | <annevk> | I managed to convince my manager I had better things to do then work on that |
| 00:34 | <JonathanNeal> | I think my issue was that I had never thought about "irrespective of CSS rules". If that is the case, why would head get display: none in the rendering guide section? |
| 00:34 | <annevk> | I phrased that poorly, we discussed it briefly; it didn't require much convincing either way... |
| 00:34 | <JonathanNeal> | Does that make sense what I just asked, Hixie? |
| 00:34 | <JonathanNeal> | (also, I do not know how to access the log for r4083 on 2009-10-05 |
| 00:35 | <annevk> | sicking, what you're suggesting will give us much of the same problems we faced with deploying HTML5 forms in the future whenever we do something like it again... |
| 00:35 | <annevk> | sicking, just allowing everything is not a good strategy; especially for extensibility it is harmful |
| 00:36 | <annevk> | and from all things conformance requirements help with that is probably the one I care about the strongest |
| 00:38 | <JonathanNeal> | I would love to compile a list of "irrespective" and "suggested" css rules. |
| 00:39 | <annevk> | JonathanNeal, you can display the head element using CSS |
| 00:39 | <annevk> | JonathanNeal, try <style>head, style { display:block } </style> in some cool browser |
| 00:40 | <annevk> | if you then put contenteditable on the <style> element you can even edit the CSS rules on the fly |
| 00:40 | <JonathanNeal> | I hope not. If you can, it is an "irrespective" css rule, which would be contrary to #display-types |
| 00:40 | <annevk> | no it's not |
| 00:41 | <othermaciej> | roc: I think on some level the motivation for better plugin API is Chrome OS |
| 00:41 | <othermaciej> | roc: not that I have any specific information from Google about that |
| 00:41 | <annevk> | i really love data:text/html, <style contenteditable>head, style { display:block } </style> |
| 00:41 | <annevk> | it's deliciously weird and useless, but still a lot of fun |
| 00:41 | <othermaciej> | annevk: you just made my head explode |
| 00:42 | <annevk> | maybe I should blog about that tomorrow for mass head explosion effects around the globe |
| 00:43 | <annevk> | (or at least the tiny part of the population that tracks it :) ) |
| 00:43 | <JonathanNeal> | hahaha |
| 00:44 | <JonathanNeal> | Just did <style contenteditable style="display: block; white-space: pre;" type="text/css"> in Chrome |
| 00:44 | <annevk> | via sicking: http://search.twitter.com/search?q=%23shaverfacts |
| 00:44 | <JonathanNeal> | You need the white-space to prevent the <p> getting in there from Opera. |
| 00:50 | <JonathanNeal> | In IE <head> is irrespectively display: none; |
| 00:50 | <JonathanNeal> | In Chrome and Opera it is malable. |
| 00:50 | <JonathanNeal> | (and Firefox) |
| 00:51 | <JonathanNeal> | Is it meant to be irrespectively hidden, or is it recommended to be hidden? |
| 00:59 | <Hixie> | JonathanNeal: not sure i understand the questions you were asking earlier (Sorry was afk) |
| 01:44 | <boblet> | Hixie: I have a question about the hr element and 3.2.5.2 paragraphs. Apart from the p element what other elements can hr be between? flow minus heading and sectioning elements? |
| 01:45 | <Hixie> | ? |
| 01:49 | <boblet> | “The hr element represents a paragraph-level thematic break”, with paragraph linking to 3.2.5.2 |
| 01:50 | <boblet> | but I’d like to confirm if that means hr only between p elements or eg between lists too |
| 01:53 | <boblet> | hr is used a lot as a section break in Japan, mainly because it translated well on mobile browsers. Am I right in thinking this would be incorrect now? |
| 01:55 | <Hixie> | 3.2.5.2 in which version of the spec? :-) |
| 01:56 | <Hixie> | oh 3.2.5.3 Paragraphs in the complete.html spec, ok |
| 01:56 | <Hixie> | it's just paragraph-level |
| 01:56 | <boblet> | Hixie: curses! transcribing error; 3.2.5.3 (http://www.whatwg.org/specs/web-apps/current-work/multipage/content-models.html#paragraphs) |
| 01:56 | <Hixie> | lists are paragraph-level |
| 01:56 | <Hixie> | pretty much everything is paragraph-level :-) |
| 01:57 | <Hixie> | the spec is pretty explicit when something isn't allowed or if there are restrictions |
| 01:57 | <boblet> | Hixie: ohrly? hmm… sectioning and heading content too? |
| 01:57 | <Hixie> | what it means is that <hr> can be used anywhere <p> can, basically |
| 01:58 | <Hixie> | though the real way to work out where you can put <hr> is to look at the content models |
| 01:58 | <Hixie> | not try to decode mostly informative text :-) |
| 01:58 | <boblet> | aha! well, I completely misread #paragraphs then |
| 01:58 | <Hixie> | the word "represents" is the key word that means "this element means..." e.g. for rendering or analysis purposes |
| 01:58 | <Hixie> | so "paragraph-level" here just means that <hr> can be treated like a paragraph for the purposes of rendering |
| 01:59 | Hixie | tries to work out why his browser isn't showing him what the server is returning |
| 01:59 | <boblet> | so hr between section elements or after h1 would be fine then? |
| 02:02 | <Hixie> | after <h1> would be pretty silly but sure |
| 02:02 | <Hixie> | between section elements is pretty much redundant |
| 02:03 | <boblet> | …because sections convey the semantics & style can be applied via borders or background images right? |
| 02:04 | <Hixie> | yeah |
| 02:04 | <Hixie> | <hr> basically says "end of one section, start of another" |
| 02:04 | <Hixie> | which is the same semantic as </section> <section> |
| 02:05 | <Hixie> | (though in the case of <hr> the sections are generally more thematic rather than explicit -- e.g. in a novel think <section> being like a chapter and <hr> being like a bigger space between paragraphs) |
| 02:08 | <boblet> | hixie: I misread this as basically hr only between paragraphs of text (eg theme change in novel), which is a lot more restrictive than you’ve explained |
| 02:09 | <Hixie> | ah, ok |
| 02:09 | <boblet> | I’ve submitted a comment about this (#9371). Thanks for your explanation! |
| 02:09 | <Hixie> | np, thanks for the bug! |
| 02:09 | <boblet> | I better go update my html5doctor article :) |
| 02:16 | <gsnedders> | WTH, the KLM website is certain that to fly from Montreal to Amsterdam, where there are directly flights, I should go by Detroit |
| 02:19 | <Hixie> | gsnedders: looks like anolis has a bug with astral plane characters |
| 02:19 | <gsnedders> | Hixie: How so? |
| 02:19 | gsnedders | notes it's probably not Anolis itself, but rather a parser issue |
| 02:19 | <Hixie> | <span>𝓏</span> in the input turns into <span>��</span> in the output |
| 02:20 | Hixie | discovers some of the named character references are in fact combining characters |
| 02:21 | <gsnedders> | Gah. |
| 02:21 | <Hixie> | i'd really rather the output was just ascii personally |
| 02:21 | <Hixie> | i thought that's what i'd set it to |
| 02:21 | <Hixie> | haven't checked if my config is wrong yet |
| 02:22 | <gsnedders> | It should be, if you set it to be |
| 02:23 | <Hixie> | oh it is ascii |
| 02:23 | <Hixie> | i was briefly confused because 
 turned into a real newline, but of course that's still ascii... |
| 02:24 | <Hixie> | so yeah, looks like there's an astral plane problem |
| 02:25 | <gsnedders> | Hixie: html5lib serializer bug |
| 02:26 | <gsnedders> | jgraham: ping |
| 02:27 | <Hixie> | i was about to say there was also a validator bug, since the validator didn't catch it, but of course i validate my post-processed pre-anolis source file, not the post-anolis file |
| 02:28 | <gsnedders> | Hixie: This on PMS I presume? |
| 02:28 | <Hixie> | yeah |
| 02:30 | <gsnedders> | https://code.google.com/p/html5lib/issues/detail?id=143 |
| 02:31 | <gsnedders> | I guess if the spec depends on this it's somewhat urgent? :) |
| 02:32 | gsnedders | is too tired to want to look at this now |
| 02:33 | <Hixie> | i'm adding the glyphs to the big table at the bottom |
| 02:33 | <Hixie> | it'll make the output invalid, but i don't mind it being invalid for a while |
| 02:33 | <Hixie> | i doubt anyone will notice the broken glyphs for a few days |
| 02:33 | <Hixie> | (other than people who are reading this -- hi logs readers!) |
| 02:34 | gsnedders | thinks for a minute, and concludes that combining chars are always legal |
| 02:34 | <gsnedders> | Even at the start of a string, I think |
| 02:35 | <Hixie> | the combining chars thing is a separate problem that i'll need to talk to the mathml guys about |
| 02:36 | <gsnedders> | I'll try and fix the html5lib bug tomorrow |
| 02:36 | <gsnedders> | Well, later today |
| 02:36 | <Hixie> | does python have a "string S is one of strings A,B,C" construct? |
| 02:36 | <Hixie> | as in "s in ['aaa','bbb','ccc']" or some such? |
| 02:36 | <gsnedders> | Yeah |
| 02:36 | <gsnedders> | Exactly that syntax :) |
| 02:36 | <Hixie> | cool |
| 02:37 | <gsnedders> | Except if it's a fixed set, you probably want to use a frozenset |
| 02:37 | <Hixie> | and does it have ternary-? syntax? as in "foo ? true : false" ? |
| 02:37 | <gsnedders> | (as then it's just a hashtable lookup) |
| 02:37 | <gsnedders> | Hixie: no |
| 02:37 | <Hixie> | k |
| 02:37 | <Hixie> | frozenset? |
| 02:37 | <gsnedders> | a set which is immutable |
| 02:37 | <Hixie> | i just want a set literal |
| 02:37 | <gsnedders> | set([1,2,3]) |
| 02:37 | <Hixie> | k |
| 02:37 | <gsnedders> | that gives a mutable set |
| 02:38 | <gsnedders> | s/set/frozenset/ gives an immutable one |
| 02:38 | <Hixie> | what does "[1,2,3]" give? |
| 02:38 | <gsnedders> | A list |
| 02:38 | <Hixie> | is frozenset() a method call, ora unary operator that gets optimised away? |
| 02:39 | <gsnedders> | It's a method call |
| 02:39 | <Hixie> | so probably not what i want here, since i'm in a loop |
| 02:39 | <Hixie> | there's no set literal syntax? |
| 02:39 | <gsnedders> | If it's a constant set, just have it declared outside the loop |
| 02:40 | <gsnedders> | No |
| 02:40 | <Hixie> | pah |
| 02:40 | <Hixie> | silly language |
| 02:40 | <Hixie> | i'll just use a list, it's short |
| 02:40 | gsnedders | notes there are plenty of things that are silly in Python |
| 02:40 | <Dashiva> | If you want dictionary lookup, can't you just use a dict instead of an array? |
| 02:40 | <gsnedders> | Dashiva: Then they need a value |
| 02:41 | <Hixie> | here's what i want: |
| 02:41 | <gsnedders> | Dashiva: If you want O(1) checking something is in a set, you don't need a value for every key |
| 02:41 | <franksalim> | a ternary operator was added to python in 2.5, btw |
| 02:41 | <Dashiva> | gsnedders: But you can't make a set literal |
| 02:41 | <Hixie> | (in pseudocode): variable = foo in ['aaa', 'bbb', 'ccc'] ? 'xxx' : 'yyy' |
| 02:41 | <Hixie> | and i'm in a loop |
| 02:42 | <gsnedders> | franksalim: Oh, yeah, duh |
| 02:42 | <Hixie> | so i'd rather not do anything expensive like constructing something or allocating memory in every loop |
| 02:42 | <Dashiva> | And you can't declare the lookup list outside the loop? |
| 02:42 | <Hixie> | well i'd rather not since then to read the code you have to look in two places to work out what this does |
| 02:43 | <gsnedders> | variable = 'xxx' if foo in ['aaa', 'bbb', 'ccc] else 'yyy' |
| 02:43 | <Hixie> | sweet jesus that's worse than perl |
| 02:43 | <Hixie> | i'll just use a simple if/else :-P |
| 02:43 | <franksalim> | it's two lines |
| 02:43 | <gsnedders> | Hey, at least it's not PHP. |
| 02:44 | <Dashiva> | You can optimize for the common case by using assignment + only if! |
| 02:44 | <Hixie> | i really miss pascal. it was so clear compared to all these modern languages |
| 02:45 | gsnedders | notes there are still pascal compilers in existance |
| 02:45 | <Dashiva> | Javascript isn't so different from pascal, is it? |
| 02:45 | <Hixie> | Dashiva: javascript doesn't have a sane class system (sorry smalltalk people :-P) |
| 02:45 | <Hixie> | and it's not strongly typed |
| 02:45 | <Hixie> | gsnedders: i know, i'm writing my game in it |
| 02:46 | <Dashiva> | I will never understand how people can like dynamic strong typing |
| 02:46 | <Dashiva> | Worst of both worlds! |
| 02:47 | <franksalim> | sigh |
| 02:48 | <Hixie> | dynamic strong typing? |
| 02:48 | <Hixie> | pascal uses static strong typing |
| 02:48 | <Hixie> | as in, checked by the compiler |
| 02:48 | <Hixie> | well, provided you don't use Variants, but those are wacko and i recommend staying well clear unless you're doing weird COM stuff |
| 02:50 | <roc> | Pascal doesn't get used because important stuff like variable-length arrays didn't make it into the language |
| 02:51 | <Hixie> | any modern pascal has variable-length arrays |
| 02:51 | <Hixie> | but yeah, the original pascal was a pretty limited language |
| 02:51 | <roc> | if you want static type joy in a good language, use ML or Haskell |
| 02:51 | <Dashiva> | My bad, it's been too long since I used pascal, I misremembered it being staticly typed |
| 02:51 | <Hixie> | (pascal actually had quite a wide following on dos and windows, since borland made possibly the best RAD environments for those platforms) |
| 02:52 | <roc> | I was a big Turbo Pascal fan back in the day |
| 02:52 | <roc> | but C won |
| 02:52 | <Hixie> | (unfortunately borland's stuff was never truly cross-platform and they died when microsoft hired when of their people to do .NET and when microsoft made Visual Studio all one needed to work on windows) |
| 02:52 | <Hixie> | yeah |
| 02:53 | <ment> | Go is also static typed |
| 02:54 | <Hixie> | go is pretty nice in many ways, though they made some syntax decisions that are a bit odd |
| 02:54 | <Hixie> | still a very new language though |
| 02:55 | <Dashiva> | I always wanted to try D, but I can never get anything done unless I have a project to work on |
| 02:56 | gsnedders | feels the same with C/C++ |
| 02:56 | <Hixie> | sanity check |
| 02:56 | <Hixie> | combining characters |
| 02:56 | <Hixie> | they come after teh character they combine with right? |
| 02:56 | <gsnedders> | Yes |
| 02:56 | <Dashiva> | Yes |
| 02:56 | <Hixie> | k |
| 02:56 | <Hixie> | i'm not going insane |
| 02:56 | <Hixie> | just checking |
| 02:56 | <Hixie> | or at least, that's not a symptom of my going insane |
| 02:58 | <Dashiva> | Going insane would require you to be sane to begin with |
| 02:58 | <gsnedders> | Hixie: I think you went insane a long time ago |
| 03:00 | <Dashiva> | Speaking of insane, was the last last week post the last? |
| 03:11 | <Hixie> | re my last checkin: if anyone can work out why (a) the columns no longer work in firefox or (b) why the :hover effect doesn't seem to work anywhere, please let me know |
| 03:11 | <Hixie> | (afk food) |
| 03:13 | <Dashiva> | Let's try validating |
| 03:14 | <Dashiva> | Just complaints about NFC... |
| 03:15 | <Dashiva> | Script error, though |
| 03:27 | <Dashiva> | The problem seems to be that the position: relative on the td is ignored. Probably because of the column stuff, since Opera doesn't support that and gets the positioning right. |
| 03:36 | <Dashiva> | Maybe this? "The effect of 'position:relative' on table-row-group, table-header-group, table-footer-group, table-row, table-column-group, table-column, table-cell, and table-caption elements is undefined." |
| 05:40 | <Hixie> | Dashiva: hm, thanks |
| 05:41 | <Hixie> | Dashiva: i figured they'd have defined it by now :-) |
| 05:52 | <JonathanNeal> | hello |
| 06:00 | <JonathanNeal> | Hixie, you around? |
| 06:10 | <Hixie> | JonathanNeal: vaguely |
| 06:22 | <JonathanNeal> | heh |
| 06:22 | <JonathanNeal> | So Hixie, why is it that <head> can be display: block; but <audio> shouldn't be overwritable? |
| 06:47 | <boblet> | JonathanNeal: thanks for iecss.com — great work! |
| 06:47 | <JonathanNeal> | boblet, thanks, and thank the guys on here for encouragement and advice. |
| 06:48 | <JonathanNeal> | Especially TabAtkins who wrote a php file I used to figure out how to write php files (which is how that table is rendered) |
| 08:25 | <hsivonen> | Well, this is interesting. And it's not the April Fools Day, yet. http://blog.chromium.org/2010/03/bringing-improved-support-for-adobe.html |
| 08:38 | <othermaciej> | hsivonen: I wonder if Mark Pilgrim will decry this move |
| 08:42 | <Hixie> | JonathanNeal: i don't recall, but the rule was added in r4083 on 2009-10-05, if you want to do some archeology to find out |
| 08:42 | <Hixie> | othermaciej: that's what roc said! |
| 08:44 | <othermaciej> | Hixie: yeah, I'm stealing his line |
| 08:44 | <Hixie> | ah :-) |
| 08:45 | <hsivonen> | othermaciej: that was my first thought, too |
| 08:45 | <Dashiva> | Man, what are we going to do when othermaciej goes to jail for theft |
| 08:46 | <othermaciej> | Dashiva: there will still be two co-chairs left |
| 08:46 | <othermaciej> | good thing we have double redundancy |
| 08:47 | <othermaciej> | hsivonen: I wonder if the bundle will come with the "Evil and Lazy" t-shirt |
| 08:47 | <Dashiva> | I'm not convinced you're all the same :P |
| 08:47 | <Dashiva> | It's more like a balance of power |
| 08:47 | <hsivonen> | othermaciej: :-) |
| 08:48 | <JonathanNeal> | where do i do said archeology? |
| 08:48 | <JonathanNeal> | Hixie |
| 08:49 | <roc> | my lines are available under an open source licence |
| 08:51 | <Dashiva> | JonathanNeal: Searching whatwg and public-html mail archives, and the irc channel logs if they go that far back |
| 08:51 | <othermaciej> | roc: viral or permissive? |
| 08:52 | hsivonen | wonders if Pepper surfaces can be GPU-hosted textures |
| 08:53 | <JonathanNeal> | Searched the mail archives but didn't see anything offering a detailed explanation. |
| 08:53 | <roc> | viral |
| 08:54 | <JonathanNeal> | Just instructional making-it-work-this-way talk |
| 08:54 | <roc> | in case anyone plans to make heaps of money out of books based on #whatwg logs |
| 08:54 | <roc> | hsivonen: I assume so, otherwise it'd be worse than what we already have (or will soon have) |
| 08:55 | <hsivonen> | roc: but GPL allows the making of heaps of money |
| 08:55 | <roc> | sure, in theory it does |
| 08:56 | <roc> | it destroys all the obvious business models and then says "good luck!" |
| 08:56 | <Dashiva> | It doesn't destroy the business model of relying on people thinking it's reasonable to pay for stuff |
| 08:56 | <roc> | note: I am a fan of the GPL for many things |
| 08:56 | <hsivonen> | or the business model of putting a Google search field on the product |
| 08:57 | <JonathanNeal> | Maybe there never was a reason to make <head> display blockable, but <audio> none or inline iirrespective of css |
| 08:57 | <JonathanNeal> | it should be consistant. |
| 08:57 | <Dashiva> | Making head display:block has been possible a long time, so I wouldn't be surprised if at least one site depends on it :) |
| 08:58 | <roc> | hsivonen: yeah, we're lucky |
| 09:00 | <roc> | the GPL makes it hard to make monopoly profits, which is what I think of as "heaps of money" :-). That's a good thing. |
| 09:00 | <JonathanNeal> | I understand, no controls on audio, then no audio, but having no control over the audio control's width or height or display seems very non <video / image / canvas / object> -ish |
| 09:06 | <JonathanNeal> | *then no audio = then no display |
| 09:25 | Philip` | suggests "s in ('aaa', 'bbb', 'ccc')" |
| 09:25 | <Philip`> | since rounded brackets are prettier, and tuples are theoretically faster than lists |
| 09:26 | <Philip`> | (If you've only got n=3 then I don't think it's really worth worrying about a O(n) list/tuple search vs O(1) set search) |
| 09:30 | <workmad3> | Philip`: are you sure that went into the right chan? :) |
| 09:30 | <Hixie> | yeah he was responding to something i said a while bacm |
| 09:30 | <Hixie> | back |
| 09:30 | <Hixie> | thanks Philip` |
| 09:31 | <workmad3> | ah, I thought I'd accidentally joined #python again :) |
| 09:41 | <jgraham> | gsnedders: Unless I am misreading, it is actually a bug in the python stdlib |
| 09:49 | <jgraham> | (and I tend to agree with Philip`; for n == 3 you probably don't care about O(1) vs O(N) but about constant factors |
| 09:50 | <jgraham> | ) |
| 09:50 | <Philip`> | (If you're using Python you probably don't care about constant factors either) |
| 09:51 | <jgraham> | (Well you might if you were constrained to use Python and this was a particularly slow part of the code. But that seems unlikely in the extreme) |
| 09:51 | <Dashiva> | If it was a critical section of the code, I don't think moving the list out of the loop would be unacceptable |
| 09:52 | <jgraham> | Moving the list out of the loop seems like it should be acceptable in either case. If the loop body is so long that that seriously impacts readbility, you probably have two problems |
| 09:52 | <Dashiva> | He didn't say he was using regular expressions |
| 09:53 | <Philip`> | If it was a critical section of the code, rewriting that whole part of the program so it doesn't have to deal with strings in a tight loop may not be unacceptable |
| 10:04 | <zcorpan> | ukai: hi |
| 10:05 | <ukai> | zcorpan: hi |
| 10:06 | <zcorpan> | ukai: do you know if the webkit websocket tests need pywebsocket as an apache module or standalone server, or doesn't matter? |
| 10:06 | <ukai> | standalone |
| 10:06 | <zcorpan> | ok, thanks |
| 10:07 | <ukai> | WebKitTools/Scripts/run-webkit-tests launches pywebsocket's standalone.py in openWebSocketServerIfNeeded() |
| 10:08 | <zcorpan> | ukai: i'm thinking of running them with a remote server instead of a local server |
| 10:08 | <zcorpan> | ukai: the remote server might well be running all the time |
| 10:09 | <ukai> | zcorpan: what do you mean the remote server in this context? |
| 10:10 | <zcorpan> | ukai: on a different machine than the one running the test |
| 10:10 | <zcorpan> | ukai: the webkit tests seem to use 127.0.0.1 |
| 10:11 | <ukai> | zcorpan: yes, but you need to rewrite all *.js in layout test code |
| 10:11 | <zcorpan> | ukai: yeah |
| 10:12 | <ukai> | zcorpan: so, what is the motivation? |
| 10:12 | <zcorpan> | ukai: trying to get it running in opera's regression testing system |
| 10:13 | <ukai> | ukai: ah. is there any problem with running on 127.0.0.1? |
| 10:14 | <ukai> | zcropan: ah. is there any problem with running on 127.0.0.1? |
| 10:15 | <zcorpan> | ukai: not sure, but it seemed like it would be simpler to install a websocket server on one machine than installing it on all test slaves |
| 10:18 | <ukai> | zcorpan: hmm, i'm not sure, but i believe we need to start pywebsocket server each time to run layout tests. some *_wsh.py need to be initialized before running test. |
| 10:22 | <zcorpan> | ukai: so different tests start the server with different configurations? |
| 10:26 | <ukai> | zcropan: no, it runs with same configurations. |
| 10:26 | zcorpan | -> lunch, bbiab |
| 10:27 | <ukai> | zcorpan: so, it might work with remote server that keep running. |
| 10:34 | <hsivonen> | is any browser knows to sniff charset-unlabeled JavaScript files for encoding patterns using a detector? |
| 10:34 | <hsivonen> | s/knows/known/ |
| 10:47 | <jgraham> | hsivonen: I have a feeling that Opera did/does (and that it is considered a bug; don't recall if it got fixed for 10.50) |
| 12:09 | <zcorpan> | ukai: ok, thanks |
| 12:25 | hsivonen | concludes that most style sheets and scripts have to be ASCII if Opera has been get away with its external resource behavior with document.open()ed docs |
| 12:26 | <hsivonen> | when the behavior differs from IE, Gecko and WebKit |
| 12:36 | <zcorpan> | hsivonen: what's different? |
| 12:38 | <hsivonen> | zcorpan: the other's decode the external resources as UTF-8. Opera decodes as Windows-1252. |
| 12:38 | <hsivonen> | *others |
| 12:39 | <hsivonen> | scary how I've started to put ' randomly in front of trailing s |
| 13:00 | <zcorpan> | hsivonen: maybe there aren't that many document.open()ed docs that use external style sheets and scripts |
| 13:00 | <zcorpan> | hsivonen: could you file a bug about that? |
| 14:55 | <gsnedders> | jgraham: That was actually my blind guess |
| 15:43 | <divya> | @TabAtkins you mentioned (w.r.t to my query on legend elm's content model) < "Content model" describes what the element is allowed to *contain*. |
| 15:43 | <divya> | "Categories" describe what the element *is*. (And then "Context"...describes where it can be used.) > Am I right in assuming HTML4 did not explicitly declare category/context and only HTML5 does so? |
| 15:44 | <TabAtkins> | divya: I'd have to check HTML4 to be sure. Been a while since I've read it. |
| 15:44 | <Dashiva> | It declared context, surely |
| 15:44 | <Dashiva> | Isn't that what the DTD is all about? |
| 15:45 | <divya> | @Dashiva e.g. http://www.w3.org/TR/REC-html40/struct/global.html#h-7.5.1 for Body element there is explicit content model associated (%block; | SCRIPT)+ +(INS | DEL) |
| 15:46 | <divya> | @Dashiva Nothing explicit about context. |
| 15:46 | <Philip`> | The "%block" definitions are sort of about context |
| 15:47 | <Philip`> | but in general the child->parent relationships aren't explicit, you have to derive them from te parent->child data in the DTD |
| 15:47 | <divya> | @Philip` %block defines what can go WITHIN body element and not what content model the element ITSELF is. |
| 15:48 | <Philip`> | divya: But some other part of the DTD says that <div> etc *is* a %block, i.e. saying it can be used in all contexts that accept blocks |
| 15:49 | <divya> | @Philip` is there a ref here: http://www.w3.org/TR/REC-html40/sgml/dtd.html |
| 15:50 | <Philip`> | <!ENTITY % block "P | %heading; | %list; | %preformatted; | DL | DIV | NOSCRIPT | BLOCKQUOTE | FORM | HR | TABLE | FIELDSET | ADDRESS"> |
| 15:51 | <Philip`> | That's kind of the explicit declaration of category |
| 15:52 | <divya> | @Philip` so any element in that list has a category of Block? |
| 15:53 | <Philip`> | Yes, and can be used in any element whose content model includes "%block;" |
| 15:54 | <divya> | @Philip` ah k. but contexts are not specified though I think. They are assumed. E.g. ins and del. |
| 15:55 | <jgraham> | divya: (this is not twitter, no need for the @ sign) |
| 15:56 | <divya> | jgraham: sorry :) too used to it. |
| 15:56 | <jgraham> | np :) |
| 15:59 | gsnedders | apparently just bought some "Nytol 0-A-Night" if you believe the recepit |
| 16:00 | <Philip`> | That doesn't seem very good product design, if you're expected to take none a night |
| 16:00 | <Philip`> | They'll never get any repeat customers |
| 16:00 | <jgraham> | Maybe you take it during the day |
| 16:00 | <Dashiva> | That doesn't stop homeopathy |
| 16:01 | <Philip`> | Dashiva: That's not how homeopathy works - you have to keep buying their expensive water regularly |
| 16:01 | <Philip`> | so they've got the right idea |
| 16:02 | <Dashiva> | You could just use half, and then refill yourself |
| 16:02 | <gsnedders> | Hopefully I can finally get off a 3am-10am sleep pattern :\ |
| 16:02 | <gsnedders> | Being ill and ending up with that sort of sleep pattern scuks |
| 16:02 | <gsnedders> | *sucks |
| 16:02 | <Dashiva> | At least you have the excuse of being ill |
| 16:03 | <gsnedders> | It was so sad, I missed going to Oslo. |
| 16:03 | <Philip`> | gsnedders: You could move to a timezone several hours westwards, and then your pattern would be fine |
| 16:04 | <gsnedders> | I tried to do that, but the clocks changed too. |
| 16:04 | <gsnedders> | So I'm still in the same timezone, despite moving west. |
| 16:04 | <Philip`> | Go west more |
| 16:04 | <gsnedders> | Philip`: Get me a plane ticket. |
| 16:04 | <Dashiva> | Take up residence on a hot-air balloon |
| 16:05 | <Philip`> | Build an underground city |
| 16:07 | <Dashiva> | I don't have a problem being awake at night, it's just that other people insist on only being available during working hours |
| 16:10 | <Philip`> | It's strange that everyone has roughly the same working hours |
| 16:10 | <Philip`> | It means you're busy working while everyone else is available, and as soon as you stop working and have some free time everybody else stops working and becomes unavailable |
| 16:11 | <Philip`> | It'd be more efficient if everybody was assigned a random number between 0 and 24 at birth, which is the offset applied to all their working hours |
| 16:12 | <Philip`> | because then a large fraction of people would be available at all times, and you'd never have to worry that all the shops are about to shut |
| 16:16 | <Dashiva> | Well, it's sort of annoying if your business hours don't overlap with people you're supposed to interact with |
| 16:16 | <Dashiva> | (E.g. anyone who's ever been on a team with more than six hours time zone difference) |
| 16:16 | <jgraham> | In Sweden the shops almost all shut for much of the weekend |
| 16:17 | <jgraham> | typical opening hours might be 11-15 on Sat. only |
| 16:18 | <jgraham> | There are literally would-be shoppers listlessly wandering the steets looking for places to spend money |
| 16:18 | <jgraham> | This pretty much disproves economics |
| 16:19 | <Dashiva> | There are social arguments for having non-business hours |
| 16:19 | <Dashiva> | So the economists might just be repressed and unable to optimize |
| 16:19 | <daedb> | All the swedish weekend money gets spent in bars and clubs instead ^^ |
| 16:59 | <TabAtkins> | othermaciej: I volunteer to write counter-proposals to all of Shelley's recent issues, except maybe the one for Issue 92 (that sounds like it might actually be reasonable; I need to review it more closely). |
| 16:59 | <TabAtkins> | I can do it by next Friday. |
| 16:59 | <gsnedders> | Is it just me or is Spotify really bad at the moment? |
| 17:05 | <JonathanNeal> | ahoy! |
| 17:06 | <JonathanNeal> | I've checked the public html mailing list, and I still haven't found an explanation as to why <audio> and <audio controls> are to be styled the way they are irrespective of CSS. |
| 17:07 | <JonathanNeal> | Would anyone currently on have enlightenment regarding this? |
| 17:10 | <jgraham> | http://www.publications.parliament.uk/pa/ld200910/ldhansrd/text/100303-0001.htm |
| 17:10 | <TabAtkins> | I suppose that <audio> is hidden no matter what because otherwise it's really annoying. I've had problems caused by me adding a border to all body>*, and getting <style> and <script> showing up. |
| 17:11 | <jgraham> | (that link is unlikely to be enlightening on almost any topic) |
| 17:11 | <TabAtkins> | <audio controls>, not as sure, but I suspect it's so that users can depend on their browser always exposing the same-looking controls. |
| 17:11 | <gsnedders> | jgraham: Lovely. |
| 17:12 | <TabAtkins> | Essentially, I think <audio controls> is supposed to act like a replaced element. Styling the contents of replaced elements is officially out-of-scope for CSS. |
| 17:12 | <jgraham> | gsnedders: The punchline is a thing of beauty |
| 17:15 | <JonathanNeal> | I'm looking for documentation on "replaced elements" |
| 17:16 | <JonathanNeal> | Hmmm, reading now. |
| 17:20 | <JonathanNeal> | I see why I never noticed this before, "4 The elements of HTML" seems not to be the section that would mention these things. |
| 17:21 | <JonathanNeal> | Where would I learn that <audio> is supposed to act like a replaced element? Where would I learn the intrinsic height/width/ etc values it would have? |
| 18:15 | <JonathanNeal> | Found some @ http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#replaced-elements |
| 18:15 | <hober> | TabAtkins: you can skip ISSUE-95; I've got that one. |
| 18:52 | <JonathanNeal> | How, in CSS, might you represent that an img had successfully loaded a src? |
| 18:56 | <JonathanNeal> | though it doesn't exist, I feel like img[complete] would be correct. "image.complete Returns true if the image has been downloaded, decoded, and found to be valid; otherwise, returns false." |
| 19:11 | <JonathanNeal> | Actually, in theory a pseudo class would better represent this, like :complete |
| 19:22 | <annevk> | JonathanNeal, yeah |
| 19:23 | <annevk> | some browsers have something like that pseudo class I believe |
| 19:23 | <annevk> | maybe at some point we'll add something like that |
| 19:23 | <JonathanNeal> | I'm in la la hypothetical land, I know. I just find it interesting and possibly useful as a reference guide. |
| 19:28 | <JonathanNeal> | how do you think sectioning-content would be defined? :section ? |
| 19:31 | <gsnedders> | Wow. Transport is Scotland is really diabolical today. |
| 19:33 | <JonathanNeal> | annevk, http://pastie.org/897474 |
| 19:34 | <annevk> | JonathanNeal, replaced elements don't need inline-block |
| 19:34 | <annevk> | also, controls is a content attribute so you can just do audio[controls] |
| 19:35 | <annevk> | also, audio is by default not displayed |
| 19:35 | <JonathanNeal> | They don't need it, no. This is a reference guide to what you may be seeing. Like, the presentational requirements and recommendations as represented in css. |
| 19:35 | <JonathanNeal> | annevk, by default it is not displayed? But, in browsers where it is not supported I thought it works like a container? |
| 19:35 | <annevk> | and for :section I assume we'd use :section(<section-level>) |
| 19:35 | <annevk> | if it's not supported it's display:inline |
| 19:35 | <JonathanNeal> | Got it! |
| 19:36 | <annevk> | because then all properties are just their initial value |
| 19:36 | <JonathanNeal> | So, if it had to be represented, there would be audio {} and audio:supported {} ? |
| 19:37 | <annevk> | no |
| 19:37 | <annevk> | if it's not supported there's no representation |
| 19:37 | <annevk> | isn't that obvious? |
| 19:38 | <JonathanNeal> | It is obvious. It wasn't obvious to me because I was realising that unsupported doesn't equate unstyled in some browsers, including IE with the createElement hack. |
| 19:40 | <JonathanNeal> | So, I was just being hypocrtical, willing to talk in theory, while then over-factoring in reality. |
| 19:41 | <JonathanNeal> | annevk, you're saying that :section() would possibly work like this? :section h1 == :section(1) h1, :section :section h1 == :section(2) h1 ? |
| 19:43 | <annevk> | maybe |
| 19:43 | <annevk> | I think we probably need :heading(<heading-level>) too |
| 19:44 | <annevk> | or maybe we do not need either and authors can just style them based on knowing their structure |
| 19:44 | <annevk> | would probably be more efficient since :section/:heading would need some fairly complicated data structures if everything is to remain performant |
| 19:48 | <JonathanNeal> | annevk, funny you say that, I just changed my examples to have :heading |
| 19:49 | <JonathanNeal> | I didn't use :heading in the way you described though, since hgroup is also a heading. |
| 20:02 | <Hixie> | so of the change proposals shelley wrote that i've looked at so far, all the bugs for which she asked for changes that i rejected, she's now proposed removing the whole section instead, and for all the bugs where she asked me to remove the whole section, she's proposed changes instead |
| 20:03 | <gsnedders> | So, basically, it appears I got rejected for Edinburgh on grounds that they have a lot of good applicants, and having slightly worse grades than some on grounds of illness is sufficant reason to get rejected. |
| 20:04 | <Hixie> | d'oh |
| 20:06 | <Hixie> | (oh ok that was only the first few. the rest are more in line with what she originally proposed) |
| 21:05 | <othermaciej> | good afternoon |
| 21:05 | <TabAtkins> | Yo, Maciej. |
| 21:06 | <othermaciej> | TabAtkins: looks like you may get your wish to write some counter-proposals |
| 21:06 | <TabAtkins> | Indeed. |
| 21:07 | <TabAtkins> | I volunteer to write one for all of them except maybe 92 (I need to check it out more first) and maybe 95 (hober wrote one, I haven't reviewed it yet). |
| 21:08 | <othermaciej> | well we need to review the incoming proposals and have a discussion period first |
| 21:08 | <othermaciej> | but your general willingness is noted |
| 21:16 | <Hixie> | anyone remember why we droped <dialog>? |
| 21:16 | <daedb> | Because people bitched about it? |
| 21:16 | <Hixie> | people are bitching about wanting it back too :-) |
| 21:16 | daedb | still likes <dialog> |
| 21:16 | <Hixie> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=9270 |
| 21:16 | <Hixie> | me too |
| 21:16 | <Hixie> | but we can't just keep flipflopping |
| 21:17 | <JonathanNeal> | So object's as plugins (eg flash) are to be treated as replaced elements, right? |
| 21:18 | <Hixie> | dialog was dropped on 2009-09-15 |
| 21:18 | <annevk> | Hixie, confusing with ARIA, contradicts HTML4 suggested markup |
| 21:18 | <annevk> | Hixie, doesn't hit 80/20 |
| 21:20 | <TabAtkins> | Hixie: Dropped <dialog> because it was insufficient for the common use-cases. |
| 21:21 | <TabAtkins> | It worked for the simplest type of conversations, but not if there were any other information, like timestamps (common use-case for marking chat logs), stage directions (common use-case for marking scripts). |
| 21:21 | <Hixie> | based on my archeology, it looks like daedb has the real reason |
| 21:21 | <TabAtkins> | Well, yes, people bitched because of that. |
| 21:21 | <TabAtkins> | There may have been unjustified bitching as well. |
| 21:22 | <daedb> | Couldn't timestamps be handled by just allowing <time> as a child of <dialog>? Stage directions could be <p> children or whatever :p |
| 21:23 | <TabAtkins> | Sure, of course we could have. |
| 21:23 | <Hixie> | timestamps actually were handled |
| 21:24 | <Hixie> | and i had added an element to the parser to do stage directions eventually |
| 21:24 | <Hixie> | (had to get it parsing right first) |
| 21:28 | <jgraham> | I recall dialog being dropped because some people were unhappy with the design and no one could really make a convincing case for the benefits |
| 21:30 | <annevk> | pretty sure my reasons were part of it too |
| 21:33 | <othermaciej> | I think the bottom line is that people had their own preferred solutions, and many were not convinced that <dialog> was overall an improvement, so they didn't want it recommended as *the* way to do things |
| 21:45 | <annevk> | btw, according to w3.markmail.org Ian still has the most messages in public-html |
| 21:46 | <JonathanNeal> | TabAtkins, watcha think of this css? http://pastie.org/897474 |
| 21:47 | <annevk> | and RB is out of the top 10, but still high |
| 21:47 | annevk | is almost out of the top 10 |
| 21:48 | <annevk> | W3C-wide it is Ian, Julian, Bjoern, Anne, Dan... |
| 21:48 | <annevk> | with Ian having a solid lead |
| 21:48 | <annevk> | (not counting the bugzilla bot) |
| 21:49 | <annevk> | and public-html is on route of overtaking www-style |
| 21:49 | <annevk> | o_O |
| 21:49 | <annevk> | www-style has 35000 and public-html 28000 |
| 21:49 | <TabAtkins> | JonathanNeal: This an attempt at a default UA stylesheet? |
| 21:50 | <JonathanNeal> | Yes, a "HTML5 Presentational Requirements and Recommendations Guide" following what I could gather from the spec, and making up a few hypothetical css selectors in the process. |
| 21:53 | <TabAtkins> | JonathanNeal: Why are you assuming a :section(n), but not a :heading(n)? |
| 21:53 | <JonathanNeal> | There's never an instance I could find where :heading(n) was more appropriate (where heading = h1 h2 h3 h4 h5 h6 hgroup) |
| 21:54 | <JonathanNeal> | Unless I misunderstand how :heading would work. |
| 21:55 | <TabAtkins> | Well, it would make a lot of your rules more simple. Rather than ":section(1) h2, :section(2) h1" you can just have :heading(2). |
| 21:55 | <JonathanNeal> | eg: <section><hgroup><h1>foo</h1></hgroup></section> would be what :heading(n)? |
| 21:58 | <JonathanNeal> | My concern is that I wouldn't want hgroup to multiply the effect of :heading |
| 22:11 | <jgraham> | JonathanNeal: Presumably it would match all headings at level n, which iirc is defined in the spec |
| 22:12 | <jgraham> | (well the spec defines the outline and the heading for each section) |
| 22:13 | <jgraham> | (but I don't recall what it says about <hgroup> with multiple heading descendants; I think it might only talk about the first since that is the one that forms the actual heading) |
| 22:21 | <JonathanNeal> | jgraham, I understand -- i just wouldn't want 3em to become 9em |
| 22:23 | <jgraham> | JonathanNeal: I don't understand why you think that might happen |
| 22:23 | <JonathanNeal> | say <hgroup> and <h1> are both :heading's |
| 22:24 | <JonathanNeal> | :heading(1) { font-size: 3em; } <hgroup><h1> becomes 3em 3em |
| 22:24 | <jgraham> | I wouldn't expect <hgroup> to match :heading |
| 22:26 | <annevk> | it would make sense |
| 22:27 | <jgraham> | for it to match :heading? |
| 22:27 | <jgraham> | Not really, it's not a heading |
| 22:27 | <othermaciej> | annevk: yeah, I see that Ian is barely beating me on public-html still |
| 22:27 | <JonathanNeal> | it's heading content |
| 22:27 | <jgraham> | othermaciej: The winners on that list are the ones at the bottom :) |
| 22:28 | <othermaciej> | jgraham: the real winners are the ones not on the list |
| 22:28 | <jgraham> | That's like being at the bottom but turned up to 11 |
| 22:31 | <JonathanNeal> | "The hgroup element represents the heading of a section." |
| 22:35 | <jgraham> | JonathanNeal: Unless I am misreading it (which is most likely) the draft is buggy |
| 22:36 | <jgraham> | OK, it is not quite buggy jsut confusing |
| 22:39 | <JonathanNeal> | Do you think it could make sense that :heading matches <hgroup> |
| 22:39 | <hober> | no |
| 22:40 | <jgraham> | JonathanNeal: Yes. I retract my earlier position. It should match hgroup but not a <hx> element in <hgroup> |
| 22:41 | <JonathanNeal> | jgraham, really you think it shouldn't match the hn's "These elements represent headings for their sections." |
| 22:43 | <jgraham> | JonathanNeal: Only if they are in <hgroup> |
| 22:43 | <jgraham> | Otherwise you have the problem you stated before |
| 22:45 | <JonathanNeal> | Eep, I dunno that sounds more confusing to me because then your single :heading pseudo-selector changes randomly. |
| 22:46 | <JonathanNeal> | randomly == perceivably unexpectedly |
| 22:48 | <jgraham> | Well you can't have it both ways :) What makes most sense? |
| 22:48 | <jgraham> | The outline algorithm actually says (contrary to what I previously stated) that it is the hgroup, not the <hx> inside, that forms the heading |
| 22:50 | <jgraham> | Which is why you might expect it to match :heading |
| 22:51 | <JonathanNeal> | Interesting, so you interpret the spec to say that h1-6 loses :heading status when inside <hgroup> |
| 22:53 | <jgraham> | JonathanNeal: See http://www.whatwg.org/specs/web-apps/current-work/multipage/sections.html#outlines |
| 22:54 | <JonathanNeal> | I just realized the spec doesn't say if body should have margin or how much it should have. |
| 22:54 | <jgraham> | In particular "When entering a heading content element" |
| 22:54 | <jgraham> | "If the current section has no heading..." |
| 22:54 | JonathanNeal | is reading that now |
| 22:55 | <jgraham> | But if you are in <hgroup> the current section already has a heading (the hgroup) so the <hx> elements are not seperately assigned as headings |
| 22:56 | <jgraham> | One could argue they should match :heading nonetheless |
| 22:56 | <jgraham> | But it doesn't work if you consider :heading to be a simple result of the elements assigned as headings by that algorithm |
| 22:57 | <JonathanNeal> | I don't suppose anyone in this chan has power to say one way or another? |
| 22:58 | <jgraham> | Well Hixie could confirm that we have read HTML5 correctly. No one can really say about :heading yet though since it doesn't actually exist afaik |
| 22:59 | <JonathanNeal> | Right, if he can confirm which way to read the spec, then I'll update immediately. |
| 22:59 | <JonathanNeal> | In the meantime, :section seemed safe (and matched the way they had presented it where :section was x) |
| 23:00 | <JonathanNeal> | "When a Document is in quirks mode, margins on HTML elements at the top or bottom of body, td, or th elements are expected to be collapsed to zero." but I don't see a margin being defined for body. |
| 23:01 | <JonathanNeal> | "then a default value of 8px is expected" oh there it is. |
| 23:07 | <JonathanNeal> | Now to figure out what the font size is supposed to be. |