| 02:22 | <JoePeck> | I'm reading about localStorage at http://dev.w3.org/html5/webstorage/#introduction |
| 02:22 | <JoePeck> | It says "The localStorage DOM attribute is used to access a page's local storage area." |
| 02:22 | <JoePeck> | by "page" does it mean "domain" ? |
| 02:23 | <JoePeck> | I played around in WebKit and it seems to mean domain. I think using the word "domain" would be more clear, but I haven't read too much of the specs. Is it common to use the word "page" and mean domain? |
| 02:24 | <JoePeck> | I haven't read too many of the specs. |
| 02:59 | <othermaciej> | JoePeck: it means "page", but all pages on the same domain share the same ultimate back end for storage |
| 03:00 | <JoePeck> | othermaciej: I think the use of "page" in the example is okay, but the line before the example I think is misleading |
| 03:01 | <JoePeck> | albeit I'm still reading through the document, and I haven't gotten down to the technical documentation of localStorage yet =) |
| 03:45 | <JoePeck> | Is there a legend describing the many different styles of content used in the specifications. For instance, what does it mean when something is boldy underlined in this thick blue-gray line? http://tr.im/oK6j?grabup |
| 03:46 | <JoePeck> | wow, I don't see my entire line (using Client Linkinus) so I'll just repeat the last part |
| 03:46 | <JoePeck> | For instance a thick blue-gray underline |
| 03:46 | <JoePeck> | http://tr.im/oK6j?grabup |
| 03:47 | <JoePeck> | The style rule that applies that underline seems to explain, any paragraph or span without a title gets the underline |
| 03:47 | <JoePeck> | so its probably an indicator for something missing an explanation? |
| 03:53 | <JoePeck> | I'm guessing IRC channel is probably busier during normal work hours? =) |
| 07:07 | <Hixie> | http://www.w3.org/mid/21B8BC99-055B-47AF-9515-6CE680B149BA⊙gc |
| 07:08 | <Hixie> | (follows this e-mail from the same thread: http://www.w3.org/mid/AC39CC88-2B2F-41F2-B69E-5DF02A3F0E5E⊙gc ) |
| 07:26 | <othermaciej> | I had to review a poorly written spec today which seemed to be badly attempting to imitate HTML5 style |
| 07:26 | <othermaciej> | it was full of MUSTs with no obvious referent |
| 07:28 | <othermaciej> | and stepwise algorithms like: "(1) If A, do B, then go to step 3. (2) If !A, do nothing, then go to step 3. (3) Terminate this algorithm." |
| 07:29 | <Hixie> | heh |
| 07:29 | <Hixie> | the only algorithms i have that are that bad are the ones that started much more complicated and that i didn't realise got simple enough to rewrite sanely |
| 07:29 | <Hixie> | sadly there are some i'm sure :-( |
| 08:10 | <Hixie> | woo, got the number of XXXs below 200 |
| 08:10 | <Hixie> | which was my goal for the quarter |
| 08:11 | <annevk42> | CSSOM is done in the CSSWG fwiw |
| 08:11 | <annevk42> | as for where to log it, maybe in the CSSWG issue tracker? |
| 08:13 | <Hixie> | how do i add things to that? i looked in bugzilla but didn't find a css component |
| 08:14 | <Hixie> | (and i didn't think sending anything to www-style would be conducive to anything getting done :-) ) |
| 08:14 | <Hixie> | ah, found http://www.w3.org/Style/CSS/Tracker/ |
| 08:14 | <Hixie> | wow, that's awesome |
| 08:14 | <Hixie> | almost every action is overdue |
| 08:16 | <Hixie> | annevk42: ok added, thanks |
| 08:16 | <Hixie> | ok bed time |
| 08:16 | <Hixie> | nn |
| 08:21 | annevk42 | "fixes" due dates for his actions |
| 08:27 | <annevk42> | nn |
| 08:54 | <remysharp> | morning - did I read somewhere that class manipulation (as in class attributes) is covered somewhere in HTML5/web apps api? |
| 08:55 | <annevk42> | search for classList and DOMTokenList |
| 08:56 | <remysharp> | cheer - exactly what I was after. |
| 08:56 | <remysharp> | *cheers |
| 08:59 | <remysharp> | darn - none of the browsers have it yet though :-( |
| 08:59 | <annevk42> | maybe if you write a compelling test suite ;) |
| 09:00 | <remysharp> | would that spur on development for the classList.add functionality then? |
| 09:01 | <annevk42> | having a good test suite is one of the main concerns when implementing new features |
| 09:02 | <annevk42> | I can't promise anything of course, but it will increase the odds :) |
| 09:02 | <remysharp> | :) |
| 09:02 | <remysharp> | Is there any guidelines on test suites? |
| 09:02 | <remysharp> | not that I've got oodles of time on my hands, but if it's something I can help with, there's no reason I shouldn't. |
| 09:04 | <jgraham> | My new mantra for testsuites is: don't make the harness more complex than the tests :) |
| 09:05 | <jgraham> | (but there are also probably useful guidelines somewhere) |
| 09:05 | <Philip`> | I think I tend to violate that mantra :-( |
| 09:05 | <annevk42> | remysharp, we don't have published guidelines, but in general being able to run them automated would be nice and when opening the test it would be nice if it was directly clear whether it passed or failed |
| 09:05 | <annevk42> | remysharp, e.g. http://tc.labs.opera.com/apis/getElementsByClassName/001.htm is clear on whether it passes or fails (it's not integrated into a test harness though atm) |
| 09:06 | <remysharp> | okay, that'll do for a starting point for me. |
| 09:06 | Philip` | has seven hundred lines of Python for his canvas tests |
| 09:06 | <remysharp> | John Resig has an arse load of tests for jQuery and class manipulation, any reason why I shouldn't glean tests from there? |
| 09:07 | <jgraham> | Philip`: Actually having lots of external code is OK. It's when there are 10,000 lines of javascript in the harness for 1 line tests that yo have a problem |
| 09:09 | <remysharp> | annevk2: seeing as you've appeared to self.clone(true) yourself, did you catch my last question on taking a lead from the jQuery tests for addClass et al? |
| 09:09 | <annevk42> | remysharp, can't think of any |
| 09:09 | <remysharp> | cool |
| 09:09 | <Philip`> | jgraham: Oh, good, I've only got ~250 lines of JS in the harness, plus a couple of dozen lines of HTML boilerplate in each test case |
| 09:09 | <annevk42> | remysharp, that's just my other laptop waking up to play music :) |
| 09:09 | <remysharp> | ah, fair enough, if only there was a self.clone function ... |
| 09:09 | remysharp | dreaming on |
| 09:09 | <annevk42> | :) |
| 09:16 | <jgraham> | remysharp: Parts of http://www.w3.org/Style/CSS/Test/guidelines.html are relevant to HTML tests |
| 09:16 | <remysharp> | cool - cheers. |
| 09:18 | <jgraham> | (specifically all of section 2, and much of section 4) |
| 09:38 | <annevk42> | heh, http://www.youtube.com/watch?v=o4MwTvtyrUQ is great |
| 09:43 | <hsivonen> | annevk42: the sad part is that over here a person with that level of conceptual clarity gets to censor DNS entries |
| 09:52 | <annevk42> | I suppose there's only so many smart people that want to be in governing |
| 09:52 | <annevk42> | s/smart/informed/ |
| 09:53 | <jgraham> | Goos substitution :) |
| 09:53 | <jgraham> | Good |
| 09:53 | jgraham | should give up |
| 09:53 | <annevk42> | maybe take a day off to get over that :p |
| 09:56 | <hsivonen> | well, having to look at the kind of stuff that may qualify for censoring doesn't seem like a job one would want to have |
| 10:01 | <hsivonen> | contains-p? how did that name come about? |
| 10:01 | <hsivonen> | hmm. maybe I misunderstood the diff |
| 10:03 | <hsivonen> | it scares me that one day I or Someone Else will have to review the Gecko script execution choreography against HTML5 |
| 10:03 | <hsivonen> | now I'm using the old script execution dance |
| 10:05 | <annevk42> | in theory you can reveal bugs by tests |
| 10:06 | <annevk42> | I guess fixing the bugs will be the interesting bit |
| 10:13 | <Philip`> | Writing the tests will be 'interesting' too |
| 10:14 | <hsivonen> | reviewing existing test cases vs. spec will be 'interesting', too |
| 12:27 | <annevk42> | I wonder why I'm still surprised into what detail the TAG has studied HTML5: http://lists.w3.org/Archives/Public/www-archive/2009Jun/0092.html |
| 12:32 | <myakura> | Is there something which defines XML syntax for XQuery? |
| 12:32 | myakura | found and kinda amazed there exists such: http://www.w3.org/TR/xqueryx/ |
| 12:33 | <annevk42> | it's pretty crazy |
| 12:34 | <annevk42> | especially when you compare the two |
| 12:34 | <myakura> | http://www.w3.org/TR/xqueryx/#Examples |
| 12:35 | <annevk42> | hmm, in Nov 2004 I said "I once read XQueryX and I'm glad that one didn't made it." |
| 12:35 | <annevk42> | apparently it did in Jan 2007! |
| 12:36 | <JohnResig> | oh Jesus, that's some crazy syntax |
| 12:36 | myakura | is surprised once again, thought it was just a W3C Note or such |
| 12:36 | <JohnResig> | it's not often that you see "here's 3 lines of simple code translated into this 150 line pile of mess, impressed?" |
| 12:37 | <annevk42> | gives "tag soup" a whole new meaning |
| 12:39 | <myakura> | 'The [XML Query 1.0 Requirements] states that "The XML Query Language MAY have more than one syntax binding. One query language syntax MUST be convenient for humans to read and write. One query language syntax MUST be expressed in XML in a way that reflects the underlying structure of the query."' |
| 12:43 | <jgraham> | That is a pisstake, right? |
| 12:47 | myakura | tries to finde an impl. report for that spec. |
| 13:31 | <Philip`> | "The ideal situation: you can take any HTML5 document, convert it to some XML-based language designed for the purpose (not necessarily XHTML), convert it back, and get a semantically equivalent HTML5 document." |
| 13:31 | <Philip`> | That seems trivial - just output <html5><![CDATA[<!doctype html><title>foo</title>...]]></html5> |
| 13:31 | <Dashiva> | Brillant! |
| 13:32 | <Dashiva> | You could also print it out and take a picture of it, and store the image as XML |
| 13:33 | <Philip`> | http://thedailywtf.com/Articles/Interesting-Bitmap.aspx |
| 13:33 | <jgraham> | You mean like <pixel><red>1</red><green>1</green><blue>1</blue></pixel> |
| 13:45 | <Philip`> | jgraham: That's not very extensible |
| 13:45 | <Philip`> | You need to be able to cope with multiple bit-depths |
| 13:46 | <Philip`> | and floating-point pixel values too |
| 13:46 | <Philip`> | and that needs to be configurable per pixel, so you can losslessly merge multiple images into one |
| 13:48 | <annevk42> | <!doctype html><title></title>]]> mwaha |
| 13:51 | <Philip`> | annevk42: Obviously any characters like ']' or '&' or U+0000 etc in the input will get converted to '&#nnn;' before you put the <html5><![CDATA[ wrapper around it |
| 13:53 | <annevk42> | oooh :p |
| 14:00 | <jgraham> | Philip`: To be fair I was assuming pure floating point pixels :) |
| 14:01 | <Philip`> | jgraham: That's not going to work if I want a pixel to be precisely a third of maximum brightness |
| 14:02 | <jgraham> | What if I want a pixel to be exactly 1/pi of the maximum brightness? |
| 14:03 | <Philip`> | You would write a schema for your pixel brightnesses, then use its URI as the namespace for your component elements |
| 14:06 | <Philip`> | (The schema would look like <pixel:schema xmlns:pixel="..."><pixel:type id="eightbit"><pixel:max inclusive="1">255</pixel:max><pixel:min inclusive="1">0</pixel:min><pixel:descr>Linear mapping from value onto pre-gamma brightness</pixel:descr></pixel:type></pixel:schema> with an equivalent bit for your 1/pi scale) |
| 14:23 | <Lachy> | http://www.darkreading.com/security/encryption/showArticle.jhtml?articleID=217801293 |
| 14:24 | <Lachy> | the article says they can create a darket using a page written using PHP and an "HTML 5-based browser" |
| 14:24 | <doublec_> | I guessing they don't say which one |
| 14:24 | <Lachy> | I'm curious what feature of HTML5 they've actually used to that, or whether "HTML5" has become a replacement buzzword for Ajax, and they're simply using some new scripting techniques |
| 14:30 | <Philip`> | Maybe they're using local storage |
| 14:32 | <Lachy> | I'm not sure how localstorage would help you connect to a darknet |
| 14:33 | <Philip`> | It would help you store data that you could share through the darknet |
| 14:33 | <Lachy> | maybe, but I'm more curious about the connection technique |
| 14:33 | <Philip`> | except apparently they don't persist any data after you shut down the browser |
| 14:34 | <Lachy> | I expect they're using XHR to connect to the server, and it's the server itself that handles the connection to other distributed servers |
| 14:34 | <Lachy> | which would explain why they said it needed PHP |
| 14:35 | <Philip`> | http://www.forbes.com/2009/06/15/darknet-hewlett-packard-technology-security-darknet.html - "Someone navigates to a Web site that serves up some JavaScript code that runs in the user's browser. That code uses the local storage capacity built into the latest version of browsers like Google ( GOOG - news - people ) Chrome and Internet Explorer." |
| 14:35 | <Lachy> | so it's not entirely anonymous, but it would be difficult to work out which server in the network a given user is connected to, assuming there are enough servers to make a process of elimination improbable |
| 14:36 | <Philip`> | "It's more like a peer-to-peer network where the computer nodes can't talk to each other directly." - so, it's like a client-server network, and not like a peer-to-peer network at all? |
| 14:38 | <Lachy> | yeah, that makes sense |
| 14:40 | <Dashiva> | http://www.facebook.no/ |
| 14:40 | <Dashiva> | Those are some interesting tooltips you get when hovering Facebook, © and 2009 in the lower corner |
| 14:48 | <gsnedders> | Dashiva: Translate? |
| 14:48 | <gsnedders> | Oh, there |
| 14:48 | <gsnedders> | I looked at the Facebook in top left :P |
| 14:49 | <gsnedders> | Dashiva: The "20" and "09" are different too |
| 14:49 | <Philip`> | The title text on the © is different to what's in the source, too |
| 15:04 | <Lachy> | Philip`, what? |
| 15:06 | <Lachy> | oh, the title is being modified by a script for some reason |
| 17:27 | <annevk2> | http://docs.google.com/View?id=dcgd8hk6_2g7c6zzc6 |
| 17:32 | <annevk2> | https://lists.webkit.org/pipermail/webkit-dev/2009-June/thread.html#8066 also suggests the display types are not actually supported |
| 17:32 | <annevk2> | hmm |
| 17:35 | <JoePeck> | is there a guide/legend for the different types of styles/syntax in the html5 specifications? |
| 17:36 | <annevk2> | JoePeck, section 1.8.2. does not help? |
| 17:38 | <JoePeck> | annevk2: For the most part, and thats what I was looking for, but it didn't help me with the 1 syntax I was actually wondering about. |
| 17:38 | <JoePeck> | annevk2: thick blue underlines: http://tr.im/oOv2?grabup |
| 17:39 | <JoePeck> | although the CSS rule makes it pretty clear its for a span without a title... maybe when the spec is completed these should have been eliminated? |
| 17:39 | <annevk2> | ah, thick blue underlines are an unimplemented feature |
| 17:39 | <annevk2> | they are references to a term defined in another specification |
| 17:40 | <annevk2> | they should really be links, but the preprocessor is not smart enough just yet |
| 17:40 | <JoePeck> | annevk2: ahh, I had that feeling. Thanks! =) |
| 17:44 | <Philip`> | annevk2: It would be nice if that document said who wrote it and who to send feedback to |
| 17:44 | <Philip`> | (Not that I want to send feedback myself) |
| 17:45 | <annevk2> | the second link I provided indicates the author |
| 17:46 | <Philip`> | annevk2: Seems like a more general failing of Google Docs that it gives no indication of who's responsible for a document |
| 17:47 | <Philip`> | At least with normal web documents you can look around the domain it's hosted on to find someone who probably knows about it |
| 18:56 | <JoePeck> | http://downforeveryoneorjustme.com/http://lists.whatwg.org/ |
| 18:58 | <annevk2> | weird, nothing on http://twitter.com/dhstatus |
| 19:01 | <Hixie> | wow did rigo really use "think of the children" as an argument |
| 19:05 | <annevk2> | he was right about P3P |
| 19:05 | <annevk2> | and to a certain extent it is being done again with Powder afaict |
| 19:08 | <Hixie> | not really, one of the bugs he quoted had a comment from someone basically saying the same as i did, and maciej replied saying that the same applied for him |
| 19:08 | <Hixie> | and i can assure you the same applies for chrome |
| 21:39 | <expensivelesbian> | hello. Are questions about seeming quirks in the HTML5 validator welcome here? I seem to be getting an odd response using multiple source elems inside a media elem |
| 21:39 | <annevk2> | there's a good chance hsivonen will pick it up |
| 21:40 | <annevk2> | alternatively you could use http://bugzilla.validator.nu/ |
| 21:40 | <expensivelesbian> | thanks |
| 21:49 | <billyjackass> | expensivelesbian: about validating multiple source elems inside a media elem, what's the question? do have your source or a test case online? |
| 21:50 | gsnedders | doesn't know what to say about a billyjackass and an expensivelesbian in #whatwg… There has to be a joke in there. |
| 21:51 | <krijnh> | :P |
| 21:51 | krijnh | adds a new subtitle |
| 21:51 | <billyjackass> | heh |
| 21:51 | <expensivelesbian> | billyjackass, sure, ignoring doctype stuff, the basic meat of the media elem is quite simple |
| 21:51 | <gsnedders> | krijnh: Do I have to reload the page until I find it? |
| 21:51 | <expensivelesbian> | safe to paste here? |
| 21:52 | <krijnh> | gsnedders: no, why do people keep asking that :) |
| 21:52 | <gsnedders> | krijnh: Where then!? |
| 21:52 | <billyjackass> | expensivelesbian: pastebin.com or somewhere is better, probably |
| 21:52 | <expensivelesbian> | k |
| 21:53 | <annevk2> | krijnh, nobody remembers the URL |
| 21:53 | <krijnh> | annevk2: that's a feature |
| 21:53 | <gsnedders> | krijnh: What is the URL!? |
| 21:54 | <krijnh> | I don't know either :$ |
| 21:54 | <expensivelesbian> | billyjackass http://pastie.org/515647 |
| 21:55 | <gsnedders> | krijnh: You must find it and tell us! We demand it of you! |
| 21:55 | <expensivelesbian> | billyjackass running that through the HTML5 validator produces an error, but I'm not exactly sure why, as it seems like it's following the spec |
| 21:55 | <krijnh> | Then I really have to come up with a new subtitle as well |
| 21:55 | <expensivelesbian> | it doesn't like the source elem for some reason |
| 21:58 | <billyjackass> | expensivelesbian: OK, I'll take a look now |
| 22:29 | <billyjackass> | expensivelesbian: sorry, got distracted. looking at the source for the schema that v.nu uses, it seems that it is requiring that if you have multiple source elements there, the first source in document order must have a type attribute |
| 22:30 | <billyjackass> | I don't know if that's per-spec or a schema bug, looking now |
| 22:30 | <annevk2> | hmm, sees like either type or media in that case would make sense |
| 22:31 | <billyjackass> | annevk2: yeah, that's what the schema requires actually |
| 22:31 | <expensivelesbian> | right, phew. That's a webkit quirk, which I raised a ticket about a while back. |
| 22:31 | <annevk2> | HTML5 however does not seem to make any requirements |
| 22:31 | <expensivelesbian> | thanks for your insights into that |
| 22:31 | <billyjackass> | source.attrs = common.attrs & src & ((type & media?) | media) |
| 22:32 | <expensivelesbian> | if I remember, I switched them around to try and coax iPhone <audio> support into life. That's still an ongoing project |
| 22:32 | <annevk2> | it does require src |
| 22:32 | <annevk2> | billyjackass, if you move the ? to after the ) would it work? |
| 22:34 | <billyjackass> | annevk2: yeah, it'd work but would mean something different of course. I'm looking at the spec now to see what it really should be |
| 22:34 | <annevk2> | well, with work I mean matching the spec :) |
| 22:37 | <expensivelesbian> | I tried adding the type attr to the audio elem for various Safari specific encodes, and it breaks it, hence why I removed it. It was this that I raised as a ticket with the Webkit people. But this is slightly off topic for here |
| 22:38 | <billyjackass> | expensivelesbian: well, as far as the spec goes, the current spec does not require it there (as annevk2 points out) |
| 22:38 | <billyjackass> | maybe some previous draft did, but the current one does not seem to |
| 22:38 | <billyjackass> | so it's a bug in v.nu |
| 22:40 | <billyjackass> | annevk2: it seems like this would match the spec too: source.attrs = common.attrs & src & type? & media? |
| 22:41 | <expensivelesbian> | ok, that's interesting to know. I'd assumed the type was a requirement, and webkit was being naughty. However, the general gist of the ticket is still for them to fix, as they are (by their own design, if I remember) supposed to be ignoring this attr. I can't remember the history of this now, but some dude at Apple said "yeah, it's a bug" |
| 22:41 | <billyjackass> | ok |
| 22:42 | <annevk2> | i think it changed because the spec now actually fetches the resource to determine the type as well |
| 22:42 | <billyjackass> | ah |
| 22:42 | <annevk2> | loading media resources changed like lots of times |
| 22:42 | <billyjackass> | yeah |
| 23:27 | <Hixie> | annevk42: where is the credentials flag initialised in xhr2? |