| 00:45 | <annevk> | oh, didn't realize the HTTPOnly guys are planning to fix the mess that is cookies |
| 00:45 | <annevk> | http://www.ietf.org/mail-archive/web/http-state/current/msg00000.html |
| 00:46 | <annevk> | I guess that's good news, although I hope they won't make them much more complicated along the way... |
| 08:55 | <hsivonen> | The releases of IE8 (compat view button) and Firefox 3.1 (<video>) are near... I wonder if I should start promoting the HTML5 doctype on my doctype page even though the HTML5 parsing algorithm hasn't shipped in a browser release... |
| 09:18 | <zcorpan___> | Hixie: could you apply the toc script to the multipage version too? also it would look better with .toc input { vertical-align: middle; } |
| 09:22 | <zcorpan> | although a fully expanded toc is useful for searching |
| 09:22 | <zcorpan> | maybe you should have an expand all button? |
| 09:22 | <Philip`> | Hixie: It should be displayed fully expanded when printing |
| 09:24 | <zcorpan> | (or a checkbox "use collapsable toc" which can remember the setting with a cookie) |
| 09:30 | <Philip`> | Or don't collapse the TOC at all, but add a separate table above it that just lists the major section headings and then links to the relevant parts of the TOC |
| 09:37 | <zcorpan> | yeah |
| 09:42 | <Lachy> | Hixie, running the toc script onload causes an annoying delay because the spec is so large. If you add the script just after the TOC and let it execute immediately, the delay would be less noticable |
| 09:55 | <jgraham> | Oops. I seem to have acidentially joined the distributed extensibility discussion |
| 11:14 | <annevk> | with a bit more publicitiy someone could make a fulltime job out of following HTML5 tweets on twitter |
| 11:14 | <hsivonen> | anyone who'd want to digest the latest batch of TAG minutes into a Last Month in the TAG blog post? |
| 11:15 | <hsivonen> | annevk: and correcting tem |
| 11:15 | <hsivonen> | them |
| 11:15 | <annevk> | we have lastweekinthetag now? |
| 11:15 | <annevk> | that'd be an awesome blog |
| 11:16 | <hsivonen> | perhaps someone should make one |
| 11:17 | <hsivonen> | there are interesting bits like: "<jar> When I mentioned this metadata discovery issue to a library scientist, they said: Why on earth would anyone ask the publisher? They're not qualified to provide this kind of information." |
| 11:18 | <annevk> | that's sort of a blow for RDFa |
| 11:18 | <annevk> | maybe we should have the blog on IRC instead :) |
| 11:20 | <svl> | Think Mr. Last Week would be so kind as to then do all the hard work of pushing it out in blog format? |
| 11:20 | <annevk> | one can hope |
| 11:21 | <annevk> | hmm http://orderedlist.com/articles/structural-tags-in-html5 is a lot more markup than I typically need for my blog homepage |
| 11:22 | <annevk> | s/my/a/ |
| 11:22 | <hsivonen> | annevk: "TBL: Calling it Access Control" is misleading. It's about privacy." |
| 11:22 | hsivonen | agrees with TimBL that Access Control is misleadingly named. |
| 11:22 | annevk | agrees as well |
| 11:22 | <annevk> | it was thought to be too late to change |
| 11:23 | <annevk> | semantics are more like share-with-origin than access-control-allow-origin |
| 11:23 | <hsivonen> | Cross-Origin Data Sharing? |
| 11:24 | <annevk> | well, we can change the spec title, we can change the spec shortname, but we can't change the header names anymore I think |
| 11:24 | <hsivonen> | I agree that changing headers at this point would be harmful. |
| 11:29 | <annevk> | I suppose we could rename it, dunno |
| 11:29 | <annevk> | though s/Data/Resource prolly as it's also for font files, images, etc. |
| 11:30 | <annevk> | not sure if it's worth the trouble |
| 11:31 | Philip` | has never found the name confusing, since he just treats it as an opaque identifier |
| 11:32 | <jgraham> | Philip` is unusual |
| 11:33 | Philip` | takes that as a compliment |
| 11:39 | <jgraham> | annevk: That blog post goes a bit overboard on the use of <header> and <section> but basically looks OK. More to the point every single comment is positive which is encouraging. |
| 11:40 | <Philip`> | It would be useful to know where the people writing these articles get their information from |
| 11:41 | <Philip`> | e.g. do they read the actual spec? |
| 11:41 | <Philip`> | and do they follow the mailing lists? |
| 11:41 | <Philip`> | or are they mostly echoing other already-summarised documents? |
| 11:44 | <Philip`> | "Waiting for official support for HTML5 structural tags from search engines… any news from this field?" - we need to spread rumours that Google favours sites using HTML5 markup, and then all the SEO experts will rush to adopt it |
| 11:48 | <zcorpan> | Philip`: will they replace <body> with <body><article> and </body> with </article></body>? |
| 11:48 | <zcorpan> | i wonder if someone's going to comment about firefox 2 |
| 11:50 | <Philip`> | zcorpan: No, Google has an advanced secret algorithm developed by teams of PhDs and augmented by billions of pages of training information, that will detect when you're using HTML5 elements stupidly and will penalise you for it |
| 11:50 | <Philip`> | according to my anonymous insider source |
| 11:52 | <Philip`> | You may not believe me given the total lack of evidence, but surely you can't take the risk of losing your number one search results to your competitors by ignoring me |
| 11:59 | <jgraham> | Hmm. If people are going to implement live preview on their comment forms it would be nice if it corresponded to what actually got posted. |
| 11:59 | <zcorpan> | it seems the createElement hack doesn't work as expressions (and ie8 seems to ignore 'behavior' in ie8 mode) |
| 12:00 | <jgraham> | It would also be nice if they documented whatever magic syntax it is they are using for formatting |
| 12:00 | <annevk> | I thought expressions were a killed feature in super standards mode |
| 12:01 | <jgraham> | Anyway I replied to the article :) |
| 12:13 | <annevk> | how is http://www.w3.org/TR/SVGMobile12/ a REC? didn't it depend on DOM Level 3 Events? |
| 12:14 | <annevk> | I guess it doesn't |
| 12:15 | <Philip`> | Are you attempting to apply logic to the status of the spec? |
| 12:16 | <annevk> | Philip`, yes |
| 12:17 | <Philip`> | No good will ever come of such foolishness |
| 12:18 | <annevk> | did you just call me a fool? |
| 12:18 | <annevk> | :p |
| 12:18 | <Philip`> | Certainly, coming in here and attempting to use logic! |
| 12:20 | <Lachy> | Philip`, FYI, I just filed the bug about the textbox on notebookreview.com that you mentioned yesterday |
| 12:21 | <Philip`> | Lachy: Okay, thanks |
| 12:22 | <zcorpan> | Hixie: the way the toc is set up in http://www.w3.org/TR/SVGMobile12/ seems more useful than html5 |
| 12:24 | Philip` | notes that there was a suggestion many months ago for each multipage page to include the subset of the TOC corresponding to that page |
| 12:24 | <Philip`> | so it seems like it could be a good idea |
| 12:24 | <Philip`> | (but non-trivial to implement) |
| 12:25 | <annevk> | why non-trivial? can't you just keep a copy of the toc around and copy the relevant part in? |
| 12:26 | <annevk> | i do think it's sort of useful to have the complete toc as well, maybe list both on the first page? |
| 12:26 | <Philip`> | What is "the relevant part"? |
| 12:26 | <annevk> | the part that matches first heading of the page to the last heading of the page |
| 12:27 | <Philip`> | particularly when the page starts at a subsubsection, so it needs to work out to indent the first TOC entry by the right number of levels |
| 12:28 | <Philip`> | It would probably still be a reasonable thing to do; it just needs a little bit of thought, which is why I haven't done it yet :-) |
| 12:31 | <Philip`> | ((I guess the only non-triviality is that it can't just extract a subtree (or consecutive sequence of subtrees) from the TOC's DOM, since it's not always split along tree lines)) |
| 12:34 | <takkaria> | I find it very useful as it is now |
| 12:34 | <takkaria> | I know that if I want to find a treebuilder step, I can hit cmd-F and type "insertion mode" |
| 12:35 | <Philip`> | Has anyone written a mail client that detects if your message contains a string "(have|is|see) attached" and warns you if you're sending it without an attachment? |
| 12:36 | <gavin> | there's http://nic-nac-project.de/~kaosmos/index-en.html#ar |
| 12:36 | <takkaria> | Philip`: there's a thunderbird extension which does it, I 'm pretty ure |
| 12:37 | <gavin> | (which is a thunderbird extension that does it) |
| 12:37 | <Philip`> | Ah, neat :-) |
| 12:45 | <Philip`> | http://orderedlist.com/articles/getting-started-with-html5 - oh, sounds like it's the start of a series |
| 12:46 | <annevk> | it is |
| 12:46 | zcorpan | wonders where jgraham's comment went |
| 12:48 | <jgraham> | zcorpan: It's on the next one in the series http://orderedlist.com/articles/structural-tags-in-html5 |
| 12:48 | <Lachy> | Hixie, please remove the TOC script. I can't find anything now! |
| 12:49 | <hallvors> | (Pegasus mail has had the "notify about missing attachment" feature for ages) |
| 12:49 | <zcorpan> | jgraham: ah |
| 12:50 | <jgraham> | Hmm. the TOC script seems to make the spec take ages to load in Firefox too which I guess is only fair since it already takes ages to load in Opera |
| 12:52 | <Philip`> | Takes about a second to toggle the TOC sections in Firefox too |
| 12:52 | <jgraham> | Oh it seems to ork OK in epiphany-webkit |
| 12:55 | <annevk> | TOC script? |
| 12:56 | <jgraham> | annevk: Load the WHATWG spec |
| 12:56 | annevk | tries |
| 12:57 | <Philip`> | The non-multipage one in particular |
| 12:57 | <annevk> | christ |
| 12:57 | Philip` | notices that orderedlist.com uses <cite> around comment authors' names |
| 13:07 | <hdh> | http://www.whatwg.org/specs/web-apps/current-work/multipage/status-documentation.html is sent as text/plain |
| 13:07 | <annevk> | Hixie, ^^ |
| 13:10 | zcorpan | notes that the slider at http://orderedlist.com/contact/ has differently sized steps (1,2,3,4,5,7.5,10,20,30,40,50) |
| 13:11 | <annevk> | i guess we could let step="" accept some kind of function at some point? |
| 13:12 | <zcorpan> | how would that work? |
| 13:12 | <annevk> | though I'm not sure why budget should follow such a function |
| 13:13 | <annevk> | step="1.5x" or something? dunno |
| 13:57 | <Philip`> | annevk: Sounds like that probably falls outside the 80% |
| 13:58 | <Philip`> | and anyway you can just have your script / server interpret the linear slider output as the input to a non-linear function that gives the real value |
| 14:00 | Lachy | adds toc.js to his AdBlocker. Problem solved :-) |
| 14:04 | Philip` | added *.com.com to his block list maybe a year ago, but recently discovered that some legitimate sites use com.com to host all their static images and CSS |
| 14:12 | <zcorpan> | Philip`: the slider under discussion is linear but with some steps removed |
| 14:14 | <Philip`> | Oh, okay |
| 14:14 | Philip` | didn't notice since the slider doesn't work for him in Opera |
| 14:24 | <Philip`> | http://www.antplc.com/ant_galio_browser.asp?menu=153 - "Remote Event Support (HTML 5)" |
| 14:52 | <wakaba> | Hixie: i didn't read irc last week, sorry |
| 14:54 | <wakaba> | re missing module, it is included in cvs version of tarball |
| 14:54 | <wakaba> | i added link to that tarball in the webpage. thank you! |
| 14:56 | <MikeSmith> | http://www.palm.com/assets/js/dist/mojo.js is interesting |
| 14:56 | <MikeSmith> | I'm wondering what other *.js filenames we could fish for there |
| 15:01 | <takkaria> | Apple make use of non-linear sliders in a few places in OS X |
| 15:27 | <MikeSmith> | so it seems some admin dumbass at Wikipedia is poised to delete the entry for JohnResig because he's an "NN person" |
| 15:27 | <MikeSmith> | http://twitter.com/jeresig/status/1113245384 |
| 15:27 | <MikeSmith> | wtf does "Prodding after speedy per request" |
| 15:30 | <Lachy> | The article is missing references that would support his notability. Merely being the creator of JQuery isn't enough and based on the content of the article, it's not really surprising it was up for deletion |
| 15:30 | <Lachy> | surely there must be some news articles from reputable sources that talk about JQuery and mention JohnResig that could be cited. |
| 15:30 | <MikeSmith> | jesus |
| 15:31 | <MikeSmith> | with all the other crap that makes up the bulk of Wikipedia, I'd think they'd find lots of other things to spend their time on |
| 15:33 | <jgraham> | http://twitter.com/diveintomark/statuses/917843268 |
| 15:34 | <MikeSmith> | heh |
| 15:34 | <jgraham> | Once you realise that wikipedia makes far more sense |
| 15:35 | <MikeSmith> | Mark is spot-on as usual |
| 15:37 | <Philip`> | If some people didn't spend time marking single-line bios for deletion, what would stop people adding a billion pages about everyone on the planet and resulting in a huge burden to anyone who does anything that's O(n) in the size of the wiki? |
| 15:39 | <Lachy> | I'm still waiting for someone to write a wikipedia article about me |
| 15:39 | <MikeSmith> | Lachy: I'll write one about you |
| 15:40 | <Lachy> | MikeSmith, if you think I'm notable enough, and have sources to back it up, go for it |
| 15:40 | <Philip`> | I'll nominate you for deletion |
| 15:40 | <Lachy> | thanks Philip` |
| 15:41 | <Philip`> | No problem |
| 15:41 | <MikeSmith> | I don't care about notable or sources... I'll just make it all up, along with the citations |
| 15:41 | <MikeSmith> | I've done it before |
| 15:42 | <MikeSmith> | that's the great thing about Wikipedia |
| 15:42 | <Philip`> | Cross-post it to Uncyclopedia |
| 15:42 | <Lachy> | ok. I guess you could list the specs I've edited, my interview on WSG or something |
| 15:43 | <Lachy> | I'll be happy if it makes it on deletionpedia |
| 15:43 | <jgraham> | MikeSmith totally needs a wikipedia page as Mike_TM_Smith |
| 15:43 | <Philip`> | Alas, it looks like my page was already deleted |
| 15:43 | <Philip`> | ("Philip Taylor, is the great-great-great-great grandson of U.S. President Taylor, and is a junior from Calvary Chapel Christian School in...") |
| 15:46 | Philip` | gets highly confused, because the online weather monitoring system for his building says there is 0 mm/hr of rain, but his window says there is much more |
| 15:46 | <Philip`> | The internet makes life far too hard |
| 15:48 | <MikeSmith> | I'm working on the Lachy article; so far, I've got the parts about his illegitimate love child, his secret meth lab, panty thievery... wondering what else I need to add |
| 15:49 | <MikeSmith> | Philip`: maybe it only gets updated hourly |
| 15:49 | <MikeSmith> | or fortnightly |
| 15:51 | <MikeSmith> | oh wait, I had confused the article I was writing about Lachy with the one I wrote about myself |
| 15:53 | MikeSmith | ponders the use of "free" in "Wikipedia: The Free Encyclopedia" |
| 15:54 | <MikeSmith> | I think "free" in software/online contexts has pretty much come to mean, "Anything that attracts a lot of guys who are very much like the Comic Book Store Guy in the Simpsons" |
| 15:55 | <Philip`> | MikeSmith: The graph seems to get updated at least every ten minutes, and the textual version at least every minute, so it's not such an easy explanation :-( |
| 15:56 | <MikeSmith> | Philip`: maybe it's just very finicky and expects only a certain quality of rain, regardless of quantity |
| 15:57 | <Philip`> | Perhaps it was calibrated to work with unicorn tears, not bog standard rain |
| 15:59 | jgraham | assumes that the rain monitoring system was merely designed by the same people who designed the climate control system at the CMS and therefore has opposite notions of "hot", "cold", "wet" and "dry" to those employed by mortal humans |
| 16:00 | <gsnedders> | MikeSmith: It's Comic Book Guy, not Comic Book Store Guy. |
| 16:02 | <MikeSmith> | gsnedders: thanks, you've proven you have the qualities to participate in "free" |
| 16:02 | <gsnedders> | :P |
| 16:02 | gsnedders | types on his laptop running a "non-free" OS/ |
| 16:02 | <MikeSmith> | Philip`: easy way to test that: just get a unicorn and have it watch "Bambi" or "Old Yeller" |
| 16:03 | <Philip`> | One of the options for second-year CS group projects here was to develop a climate control system for the CS building (particularly the room in which all the projects were going to be demonstrated), but they found it difficult because the real building admin people wouldn't let them open any doors or windows because it would disturb the rest of the building |
| 16:03 | <gsnedders> | My main issues with Win7 so far: No decent POSIX shell, and no virtual desktops. |
| 16:04 | <Philip`> | MikeSmith: I find it much quicker to squeeze certain parts of the unicorn's anatomy until it cries |
| 16:04 | <gsnedders> | (in that order of from most to least important) |
| 16:04 | <MikeSmith> | Philip`: I like the way you think |
| 16:04 | <MikeSmith> | gsnedders: Cygwin is "free" |
| 16:05 | <gsnedders> | MikeSmith: Cygwin's shell sucks though. |
| 16:05 | <MikeSmith> | it's bash |
| 16:05 | <Philip`> | It's in a window that's really hard to resize |
| 16:05 | <MikeSmith> | rxvt |
| 16:05 | <gsnedders> | Half of that suckiness is the fact it is workaround the fact that the FS isn't POSIX compliant to start with |
| 16:05 | <rubys> | opera (the site) is down? |
| 16:05 | <gsnedders> | And the other half is that half the folders aren't called what I expect |
| 16:06 | <gsnedders> | Also, ~ isn't my normal user folder |
| 16:06 | <gsnedders> | Also, I can't drag a folder/file into the shell to get the pathname copied there like I can on Terminal.app |
| 16:06 | <MikeSmith> | rubys: seems so.. I can't get to it at least |
| 16:06 | rubys | likes cygwin, makes windows tolerable |
| 16:07 | <Philip`> | gsnedders: If you don't like unexpected folder names, it seems your complaint is really "it's not OS X" rather than "it's not POSIX-compliant" |
| 16:08 | <gsnedders> | Philip`: I mean things like "My Documents" come up as "Documents" |
| 16:08 | jgraham | prefers to avoid windows rather than endure cygwin |
| 16:08 | gsnedders | finds shells on OS X, Solaris, GNU/Linux, FreeBSD, etc. far more bearable than Cygwin |
| 16:08 | <MikeSmith> | gsnedders: I think you can fix that by pouring some water into the vents |
| 16:09 | <Philip`> | gsnedders: Vista (and presumably Win7) doesn't have "My Documents", it just has "Documents" |
| 16:09 | <Philip`> | so that's nothing to do with the shell |
| 16:09 | <gsnedders> | Philip`: It appears in the GUI as My Documents |
| 16:09 | <Philip`> | Hmm |
| 16:09 | <rubys> | cygwin's shell is simply bash, so what people are reacting to is the terminal client? Or am I missing something? |
| 16:09 | <Philip`> | Did they change it from Vista, then? |
| 16:10 | <gsnedders> | rubys: Yeah, it's the terminal client |
| 16:10 | <Philip`> | (Or am I misremembering whether they renamed it on Vista?) |
| 16:10 | <MikeSmith> | so use rxvt or whatever |
| 16:10 | <rubys> | that's easy, use putty |
| 16:10 | <sid0> | it is ~/Documents on both Vista and Win7 |
| 16:11 | <MikeSmith> | gsnedders: please first try my pouring-water-into-the-vents solution and let me know if that fixes things |
| 16:11 | <gsnedders> | rubys: Can that work locally without having a whole virtualized OS? |
| 16:11 | gsnedders | starts up Win7 box |
| 16:11 | <rubys> | it is documents on win7 |
| 16:11 | <Philip`> | sid0: Is it "Documents" in the GUI in both too? |
| 16:12 | <rubys> | gsnedders: yes |
| 16:12 | <sid0> | Philip`: yes |
| 16:12 | <sid0> | I see there's a junction from "My Documents" to "Documents" for compat |
| 16:12 | <Philip`> | Ah |
| 16:13 | <Philip`> | gsnedders: Were you complaining about the shell in Win7, or am I getting entirely confused? |
| 16:13 | <rubys> | create a shortcut to <path>/putty.exe -cygterm - |
| 16:13 | <gsnedders> | I only have My Documents, no "Documents" |
| 16:13 | <gsnedders> | I have "Documents" as a library, but not as a folder |
| 16:13 | <rubys> | http://intertwingly.net/blog/2008/02/13/Making-Myself-at-Home-temporarily-on-XP#c1202931750 |
| 16:14 | <sid0> | gsnedders: oh, sorry, I moved my documents folder elsewhere |
| 16:14 | <sid0> | though it really doesn't matter how it appears in the gui, does it, in the shell it's still Documents |
| 16:15 | <gsnedders> | It's confusing it isn't consistent |
| 16:15 | <gsnedders> | rubys: thx |
| 16:16 | <Philip`> | gsnedders: It'd be more confusing if the file path depended on the user's currently-selected language, and it'd also be more confusing if the GUI displayed "[My] Documents" regardless of the user's currently-selected language |
| 16:17 | <sid0> | yes, I believe that's the reason |
| 16:17 | sid0 | is liking the libraries though |
| 16:18 | <Philip`> | Libraries? |
| 16:19 | <sid0> | er, those virtual views of multiple collated folders |
| 16:19 | <sid0> | even over the network |
| 16:22 | <Philip`> | Ah |
| 16:23 | gsnedders | doesn't care for them |
| 16:26 | <Philip`> | Argh, I keep forgetting FF2 doesn't have contenteditable |
| 16:26 | Philip` | wonders what's the easiest way to hack in equivalent functionality for rubbish browsers |
| 16:26 | <gsnedders> | rubys: -cygterm - just causes "unknown option". Meh. |
| 16:27 | Philip` | supposes he should just use <input>, since he doesn't care about formatting or anything |
| 16:27 | <gsnedders> | http://code.google.com/p/puttycyg/ is needed, obviously |
| 16:35 | <Philip`> | Oh, the weather graph says it rained |
| 16:36 | <Philip`> | Looks like it averages over ~15 minutes, which introduces some lag |
| 16:54 | <hsivonen> | let's try deploying all the recent Validator.nu check-ins finally. |
| 17:10 | <hsivonen> | MikeSmith: your patches should now be live on validator.nu and html5.validator.nu. Thanks! |
| 17:17 | <hsivonen> | What's the condition for not using anti-aliasing for text in IE8 on Windows 7? do some older MS-supplied fonts include bitmap glyphs and if you hit one of those sizes exactly, the legacy bitmaps are used |
| 17:17 | <hsivonen> | it sure is ugly |
| 17:22 | <hsivonen> | http://www.zeldman.com/2009/01/12/jeffrey-zeldmans-web-standards-advisor/ |
| 17:23 | <hsivonen> | I wonder what conformance criteria the hCard and hCalendar validators are based on |
| 17:36 | <gsnedders> | Anyone know anything about this BlogML? |
| 17:47 | <gsnedders> | Hixie: Are things like content-type sniffing feedback moved off the issues list? |
| 17:51 | <gsnedders> | Oh, there they are |
| 17:52 | <gsnedders> | <http://lists.w3.org/Archives/Public/public-html/2007Aug/0671.html>; doesn't appear, yet I can't find any reply to it |
| 17:54 | gsnedders | finds <http://lists.w3.org/Archives/Public/public-html/2008Feb/0410.html>; |
| 17:54 | gsnedders | wonders why on earth that isn't in his email client |
| 18:01 | <Dashiva> | Google news says "try google archives for news older than 30 days" and the newest news I can find there are from 2004 :) |
| 18:28 | <Philip`> | hsivonen: JZWSA sounds somewhat like JSLint, i.e. mostly a waste of time trying to satisfy pointlessly pedantic warnings |
| 18:29 | <Philip`> | ("Gosh, you used this class value only once, so you could have made it an ID instead! That's a really useful thing for you to fix") |
| 18:30 | <Philip`> | "I ran Jeffrey Zeldman’s Web Standards Advisor on the four-year-old markup of this site’s current blog layout, and discovered embarrassing mistakes that [nobody has ever noticed or complained about because they're totally irrelevant in reality]" |
| 18:30 | Philip` | is probably being slightly unfair :-) |
| 19:45 | <Hixie> | zcorpan: the multipage version is generated by Philip` |
| 19:45 | <Hixie> | Philip`: is it not displayed fuly-expanded when printing? |
| 19:46 | <Hixie> | Lachy: i can't easily put the script beyond the toc without also changing the w3c version |
| 19:51 | <Hixie> | wakaba: np |
| 20:12 | <hsivonen> | Hixie: why doesn't the content property of CSS work on actual elements the same way it works on pseudos? |
| 20:13 | <hsivonen> | Hixie: http://www.w3.org/TR/css3-content/#inserting suggests it should |
| 20:14 | <Hixie> | because that's still just a WD |
| 20:15 | <hsivonen> | ok |
| 20:16 | <Philip`> | Hixie: It wasn't fully expanded when I tried print preview in FF2 |
| 20:16 | <Philip`> | (or it might have been FF3, I'm not quite sure) |
| 20:16 | <Hixie> | oh, print preview is a different matter |
| 20:16 | <Hixie> | i thought you meant the pdf |
| 20:17 | <hsivonen> | how wrong is it to use PUA dingbats with @font-face? |
| 20:17 | <Hixie> | so i guess what we should do is have G generate two table of contents |
| 20:17 | <Hixie> | since the collapsing isn't popular |
| 20:23 | <Hixie> | ok how's that like |
| 20:23 | <Hixie> | new toc mechanism |
| 20:24 | <Hixie> | temporary, until G can do it statically |
| 20:26 | Hixie | gives up waiting for shelly to reply |
| 20:38 | <gsnedders> | Hixie: G == me? |
| 20:39 | <gsnedders> | Hixie: And what sort of TOC do you want now? |
| 20:39 | <Hixie> | see the spec :-) |
| 20:40 | <gsnedders> | Hixie: Ah, so one with just the top level? |
| 20:40 | <gsnedders> | (Actually, the second level) |
| 20:40 | <Hixie> | yeah |
| 20:41 | <gsnedders> | (As the top level is just "HTML 5") |
| 20:42 | <gsnedders> | Well, send email (<mailto:geoffers⊙gc> before you ask) |
| 20:44 | <hsivonen> | oh cool. Gecko already has an html5lib tree builder test harness for mochitests |
| 20:52 | <Lachy> | Hixie, it doesn't matter. I just blocked the script entirely because it made the spec less usable |
| 21:01 | <Hixie> | makes it more usable now :-) |
| 21:36 | <Lachy> | Hixie, yeah, that is much more useful. I've unblocked the script now |
| 21:37 | <Lachy> | but it would be useful if that could instead be generated in the spec gen process |
| 21:37 | <Lachy> | with an anolis plugin |
| 21:37 | <Hixie> | there are a number of things i'd like anolis to do :-) |
| 21:37 | gsnedders | knows |
| 21:37 | gsnedders | has several emails :) |
| 21:38 | <gsnedders> | Hixie: As I said, send an email about this |
| 21:38 | <Hixie> | i'd rather you did the others first :-D |
| 21:38 | <hsivonen> | It's a shame that rounded corners aren't anti-aliased in WebKit |
| 21:38 | <gsnedders> | Hixie: :D |
| 21:38 | <hsivonen> | (neither in Safari nor in Chrome) |
| 21:38 | <gsnedders> | Hixie: It's probably easier to just dump it in my email box at least |
| 21:38 | <gsnedders> | Hixie: I probably won't do stuff in order anyway :P |
| 21:39 | <Hixie> | that's why i don't want to put it in your box :-) |
| 21:39 | <hsivonen> | also, rounding em-based widths to pixels in WebKit is different for the vertical and horizontal lines and for curves |
| 21:39 | <gsnedders> | Hixie: Stuff for Lachy and annevk before you, anyway :P |
| 21:39 | <Lachy> | gsnedders, what stuff from me is in your queue? |
| 21:39 | <gsnedders> | Lachy: biblio |
| 21:39 | <Lachy> | oh yeah. |
| 21:40 | <Hixie> | i found biblio to be undesireable because the references churn so fast that the maintainer of the generator script gets tired of updating references |
| 21:41 | <Hixie> | and so the editor ends up having to do it manually anyway |
| 21:41 | <Hixie> | so if you do biblio stuff, try to work around that problem somehow |
| 21:41 | <gsnedders> | Hixie: Well, it should be possible to automatically create a database of RFCs and W3C TRs easily |
| 21:41 | <Hixie> | e.g. fetch data off a wiki or something |
| 21:41 | <Hixie> | if only RFCs and TRs were all that was referenced... |
| 21:41 | <gsnedders> | Well, in HTML 5 they are a fair proportion |
| 21:41 | <Hixie> | (and are you going to regenerate the database every time you run the script? because that would be a lot of work) |
| 21:42 | <Hixie> | it doesn't really matter what the proportion is, all that matters is whether there is a critical mass of cases that change often enough to be annoying |
| 21:42 | <gsnedders> | Hixie: No, just rebuild it per release, and include the script to re-gen it |
| 21:43 | <Hixie> | well it's certainly an improvement over doing it manually |
| 21:43 | <Hixie> | but i doubt it'll be enough |
| 21:43 | <gsnedders> | Hixie: Also, custom databases can be added as well, so just add extra refs needed there |
| 21:44 | <Hixie> | can they be added in the source file? that would side-step my issue a bit |
| 21:44 | <Hixie> | if they could be added in the source file in the output format |
| 21:44 | <Hixie> | and all the script did was add the missing ones and reorder them |
| 21:44 | <Hixie> | that would be ok |
| 21:45 | <gsnedders> | Then just look within <!--begin-references--> … <!--end-references--> (or whatever the string is) for them? |
| 21:47 | <gsnedders> | And just expect it to be a dl? |
| 21:47 | <Hixie> | or just look for the <dl> with id=refernces |
| 21:47 | <Hixie> | or with class="references" |
| 21:47 | <Hixie> | or something like that |
| 21:48 | <Hixie> | also there needs to be a way to put some references in one <dl> and some in another, and a way to annotated the references (e.g. I put "Informative" next to some) |
| 21:48 | <Hixie> | anyway |
| 21:48 | <gsnedders> | Yeah, I know |
| 21:48 | <Hixie> | all these complications are why I gave up with automating this |
| 21:49 | <gsnedders> | Ah, the css3 spec-gen uses <!--normative--> and <!--informative--> |
| 21:50 | <gsnedders> | (At least, that's what the docs claim) |
| 21:51 | gsnedders | has limited faith in the docs, though |
| 21:51 | <Hixie> | bbiab |
| 21:54 | <gsnedders> | Lachy: Does selectors use css3 spec-gen biblio? |
| 21:54 | gsnedders | looks at source and sees the give-away comments |
| 21:54 | <gsnedders> | (i..e, yes) |
| 21:56 | <Lachy> | gsnedders, did you mean selectors api? |
| 21:56 | <gsnedders> | Lachy: yeah |
| 21:56 | <gsnedders> | I have already answered my own question, though |
| 21:56 | <Lachy> | yeah. That's why I want it supported in anolis, cause it's preventing me from switching |
| 21:56 | <Lachy> | though I've already switched for the html5 authoring guide |
| 21:56 | <gsnedders> | Yeah, that's what I was guessing you used it there |
| 21:56 | <gsnedders> | Yeah, I saw. |
| 21:57 | <gsnedders> | (insofar as the numbering is ISO 2145 compliant) |
| 22:07 | Lachy | wants a plugin that outputs non-ISO-2145 compliant numbering |
| 22:07 | gsnedders | notes that won't be easy |
| 22:17 | <annevk> | gsnedders, base it on a wiki that you scrape while running the script and fall back to a backup you have from last time you loaded the script if the network request fails? |
| 22:18 | <gsnedders> | annevk: I like Hixie's suggestion, having thought about it |
| 22:19 | <annevk> | that doesn't encourage people to add them to a place for others to reuse |
| 22:20 | <gsnedders> | indeed it doesn't, but it does leave it wide-open to name conflicts |
| 22:21 | <annevk> | the community will sort that out |
| 22:23 | <gsnedders> | RDF is really fucking confusing. |
| 22:23 | <hsivonen> | gsnedders: how so? |
| 22:24 | <gsnedders> | The primer at least has left me more confuse than before |
| 22:24 | <gsnedders> | Where is rdf:about actually defined? |
| 22:31 | <webben> | gsnedders: Hmm. Good question! |
| 22:33 | <webben> | gsnedders: 2.2.1 http://www.w3.org/TR/1999/REC-rdf-syntax-19990222/ |
| 22:33 | <webben> | I guess. |
| 22:33 | <webben> | "When the about attribute is specified with Description, the statements in the Description refer to the resource whose identifier is determined from the about." and the following |
| 22:35 | <annevk> | the newest version of that document refers to the RDF primer and therefore it does not seem to be normative |
| 22:35 | <annevk> | or at least the newest version isn't |
| 23:11 | <Hixie> | heycam: yt? |
| 23:12 | <heycam> | hi Hixie |
| 23:12 | <Hixie> | hey |
| 23:12 | <Hixie> | question |
| 23:12 | <Hixie> | how do i make an interface that has an anonymous named getter that returns values for strings that aren't properties of the object? |
| 23:12 | <Hixie> | [NameGetter] seems to be defined only in terms of "supported named properties" |
| 23:13 | <heycam> | hmm, you can't currently. what objects need this? |
| 23:13 | <Hixie> | HTMLDocument |
| 23:13 | <Hixie> | http://www.w3.org/mid/op.tzqg9y2hidj3kv⊙hpon |
| 23:13 | <Hixie> | i'm trying to emulate Safari's behavior |
| 23:14 | <Hixie> | <img name=x><script> assert(document.x is a [HTMLImageDocument]); assert(not 'x' in document); </script> |
| 23:14 | <Hixie> | er |
| 23:14 | <Hixie> | HTMLImageElement |
| 23:14 | <heycam> | i see |
| 23:15 | <Hixie> | anonymous name-getter, i should say, not "anonymous named getter" which is an oxymoron |
| 23:16 | <heycam> | :) |
| 23:16 | <heycam> | i'll make a note of it |
| 23:16 | <Hixie> | k |
| 23:18 | heycam | wonders if there's a bot here that will tell me a link to this point in the irc logs |
| 23:18 | <annevk> | nope |
| 23:19 | <Hixie> | does "The name of the corresponding named property is N and will have the DontEnum attribute." mean that you can't do "for (i in foo) { }" to get the names out? |
| 23:19 | <heycam> | correct. do you need to have things returned by [[Get]] that will return false when doing (prop in object) but will be enumerated? |
| 23:20 | <heycam> | since that'd be a violation of ecma-262, i think |
| 23:20 | <Hixie> | i think we might need that for document.all |
| 23:20 | <heycam> | hmm |
| 23:20 | <Hixie> | but i haven't looked at that yet |
| 23:21 | <Hixie> | i was wondering more about 'window' |
| 23:21 | <Hixie> | where the subframes get exposed to enumeration iirc |
| 23:21 | <heycam> | oh actually no it wouldn't necessarily be a violation, you can just define [[HasProperty]] appropriately |
| 23:21 | <heycam> | not that web idl allows you to override [[HasProperty]] at all at the moment |
| 23:22 | <Hixie> | hm, no, i'm wrong |
| 23:22 | <Hixie> | about window enumeration |
| 23:22 | <Hixie> | i'm sure i've seen properties exposed before |
| 23:22 | <Hixie> | where would that be |
| 23:24 | <Hixie> | can't find anywhere |
| 23:24 | <Hixie> | oh well |
| 23:24 | <annevk> | I thought document.all required hacks in ECMAScript parsers |
| 23:24 | <Hixie> | d.all is messed up |
| 23:24 | <Hixie> | i don't know what we'll do for it |
| 23:25 | <annevk> | but I don't know the exact details since nobody wrote them down afaik |
| 23:25 | <Hixie> | i wonder if the JS committee would deal with it for us |
| 23:26 | <annevk> | I've been trying to get them to deal with <!-- without much success so far |
| 23:26 | <annevk> | though I haven't paid close attention to see if they made changes either, to be perfectly honest |
| 23:27 | <Hixie> | <!-- is easy to deal with |
| 23:27 | <Hixie> | and they won't do it? |
| 23:27 | <Hixie> | sheesh |
| 23:27 | <Hixie> | heycam: something else that would be useful (regarding the 'any'/DOMObject issue) would be union types for objects |
| 23:28 | <Hixie> | heycam: e.g. sometimes i have to return an HTMLElement or an HTMLCollection |
| 23:28 | <heycam> | mm |
| 23:29 | <Hixie> | oh i see |
| 23:29 | <Hixie> | IE does enumerate |
| 23:29 | <Hixie> | btu safari does not |
| 23:29 | <Hixie> | maybe my request above is not necessary |
| 23:29 | Hixie | investigates further |
| 23:29 | <Hixie> | yeah |
| 23:29 | <heycam> | such a feature wouldn't be suitable for certain language bindings |
| 23:29 | <Hixie> | i was wrong about 'x' not in document |
| 23:29 | <heycam> | but maybe in those languages it would just map to Object or whatever? |
| 23:30 | <heycam> | Hixie, so i'll remove that feature request? |
| 23:30 | <Hixie> | yeah |
| 23:30 | <Hixie> | sorry abotu that |
| 23:30 | <heycam> | ok |
| 23:34 | <Hixie> | can someone who has IE up confirm that it says "2" "3" in the logs for this?: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E...%3Cimg%20name%3Da%3E%3Cimg%20name%3Da%3E%3Cscript%3Ex%20%3D%20document.a%3Bw(x.length)%3B%3C%2Fscript%3E%3Cimg%20name%3Da%3E%3Cscript%3Ew(x.length)%3B%3C%2Fscript%3E |
| 23:34 | <Hixie> | nevermind |
| 23:34 | Hixie | finds it is even more complicated and launches VMWare |
| 23:37 | <Hixie> | wow, IE returns 2/3 |
| 23:37 | <Hixie> | nice |
| 23:38 | <Hixie> | my life just became 2018451023 times easier |
| 23:39 | jcranmer | sighs |
| 23:39 | <jcranmer> | what is with these eRDF/RDFa/RDF/whatever-the-hell-all-that-stuff-is people? |
| 23:40 | <Hixie> | now what? |
| 23:40 | <Hixie> | i thought we actually made slight progress towards getting some problems listed |
| 23:40 | <annevk> | Firefox actually says 2/2 |
| 23:40 | <Hixie> | yeah firefox is disqualified |
| 23:41 | <jcranmer> | Hixie: from what I read, they're debating over which one to use or something like that |
| 23:41 | <Hixie> | in whatwg? |
| 23:41 | <jcranmer> | yes |
| 23:41 | <Hixie> | yeah i don't know how to direct the conversation towards problems rather than solutions |
| 23:41 | <webben> | jcranmer: That's because Calogero asked them why not eRDF. |
| 23:41 | <Hixie> | i've tried everything i know how to do |
| 23:42 | <Hixie> | i'm just going to delete the e-mails without problems |
| 23:42 | <jcranmer> | webben: well, it doesn't help that one of the people doesn't send emails with References/In-Reply-To |
| 23:43 | <annevk> | I've been reading all those e-mails and I'm not that much wiser, though those of hsivonen usually have good stuff |
| 23:44 | <Hixie> | hey ojan |
| 23:46 | <Hixie> | heycam: do i have to define how to handle named index |
| 23:46 | <Hixie> | er |
| 23:47 | <Hixie> | heycam: do i have to define how to handle name index property access when the name isn't in the list of names, or does webidl define that that returns undefined? |
| 23:47 | <Hixie> | heycam: (or does JS define that) |
| 23:47 | <heycam> | web idl defines that normal object property lookup is done |
| 23:48 | <heycam> | web idl defines a whole [[Get]] to use on the object, which has cases for what to do in the presence of [NamedGetter], etc. |
| 23:48 | <Hixie> | where is that defined? |
| 23:48 | <heycam> | www.w3.org/TR/WebIDL/#Get |
| 23:48 | <heycam> | (or maybe #get?) |
| 23:48 | <Hixie> | i found a [[put]] algorithm, but no [[get]] |
| 23:48 | <Hixie> | #get and #Get return nothing |
| 23:48 | <heycam> | oh right |
| 23:48 | <heycam> | sorry, misremembering |
| 23:49 | <heycam> | since the named/index properties are defined in terms of creating actual properties on the object, the standard [[Get]] is used |
| 23:49 | <Hixie> | k |
| 23:49 | <Hixie> | so i don't have to define that case |
| 23:49 | <Hixie> | ok |
| 23:49 | <heycam> | if overriding [[Get]] to return values for properties that don't exist on the object is needed, then a section defining a [[Get]] will be needed |
| 23:50 | <Hixie> | not needed yet as far as i know |
| 23:54 | <ojan> | hi hixie! |