01:08
<Hixie>
http://lists.w3.org/Archives/Member/member-i18n-core/2008Aug/0008.html is fantastic
01:10
<Hixie>
woot, html5 being used on actual google pages
01:10
<Hixie>
e.g. http://www.google.com/privacy.html
01:36
<jruderman>
what does that page use from html5? the short doctype and <meta charset>?
01:37
<jruderman>
innerHTML? ;)
01:38
<Hixie>
the short doctype and the intent that it comply to html5
01:38
<Hixie>
as opposed to the usual google policy of "hey it works in ie"
01:38
<jruderman>
oh, wow, the use of <h1> and alt text at the top is beautiful
01:45
<jacobolus>
also, it cuts out the <head>, <body> tags, all unnecessary closing tags, etc.
01:45
<Hixie>
indeed
01:45
<jacobolus>
what's the <div></div> about?
01:46
<Hixie>
probably to hang style off
01:46
<Hixie>
but i haven't checked
01:47
<billmason>
Looks like it just clears the floats from the lists of specific privacy policies above it.
01:49
<jacobolus>
ah, it does a { clear: left; }
01:57
<Hixie>
jesus, ITS is over-engineered for this.
02:25
<Hixie>
i really should figure something out as far as an event queue is concerned
02:25
<Hixie>
maybe i need to define the main event loop
05:26
<roc>
anyone know if the @font-face font linking in Safari 3.1 imposes a same-origin restriction on linked fonts, or can you link to any font on the Web from any page?
05:37
<bdash>
roc: it doesn't
05:38
<roc>
yeah, I just tested it myself, thanks
05:38
<bdash>
roc: there's a FIXME in the code that suggests that the policy is not set in stone though
05:38
<roc>
interesting
05:39
<roc>
I think we might want to impose a same-origin policy
05:59
<hsivonen>
roc: wouldn't same-origin policy prevent new business models?
05:59
<roc>
huh?
05:59
<hsivonen>
roc: like a foundry hosting a font and metering access
06:00
<roc>
we could support Access Controls for font file loads if the server with the font file wants to be more permissive
06:09
<hsivonen>
takkaria: thanks
06:10
<hsivonen>
Hixie: I disagree. I think if your content and style are properly separated, your style can be written ahead of time enumerating the constructs you want to use.
06:10
<hsivonen>
Hixie: where a 'table with headings' is a construct
06:11
<hsivonen>
Hixie: not table with two columns and the second column takes percentages
08:12
<hsivonen>
hmm. GWT 1.5 RC2 is out
08:12
<hsivonen>
still using *old* versions of Gecko and WebKit
12:18
hsivonen
needs apache gzip negotiation for dummies
12:26
<zcorpan>
http://simon.html5.org/specs/web-dom-core
12:26
<jcranmer>
bah, how do I send an
12:27
<jcranmer>
= in post data properly?
12:27
<zcorpan>
=3D ?
12:36
<gDashiva>
zcorpan: the entire spec is one huge issues block? :)
12:40
<zcorpan>
gDashiva: yeah
12:40
<zcorpan>
gDashiva: before i had notes all over the place, at least now they're all in 1 place :)
12:41
<gDashiva>
It does serve as a kind of pseudo-toc as well
12:54
<Lachy>
I have to buy myself a Drobo from the USA. It's 3 times more expensive to get the first gen (USB only, no firewire and no drives) in Norway
13:22
<zcorpan>
hsivonen: got the "undef error - Insecure dependency in exec while running with -T switch at /usr/share/perl5/Mail/Mailer/sendmail.pm line 22." again
13:26
<hsivonen>
zcorpan: :-( I don't know what to do about it.
13:26
<hsivonen>
zcorpan: thanks
13:28
<hsivonen>
JohnResig: you blogged about parsing HTML in JS earlier. I got GWT to compile the Validator.nu HTML Parser into JS: http://livedom.validator.nu/
13:28
<JohnResig>
hsivonen: ah, neat
13:32
<zcorpan>
hsivonen: cool
13:45
<Lachy>
hsivonen, could that be merged with Hixie's live dom viewer and made to show both the HTML5 parser and the browser DOMs side by side?
13:46
<hsivonen>
Lachy: that should be technically possible
13:46
<hsivonen>
Lachy: I don't promise to do it right now
13:54
zcorpan
notes that <p><b><b><b><p>x<p>x<p>x is parsed differently in webkit and html5
13:56
<zcorpan>
webkit ends up with 10 elements, html5 with 16
13:58
<zcorpan>
with <p><b><b><b><b><b><b><b><b><b><p>x<p>x<p>x<p>x<p>x<p>x<p>x<p>x<p>x it's 28 vs 100
14:14
zcorpan
sees window.parseHtmlDocument(textarea.value, iframe.contentWindow.document, afterParse, null);
14:14
<zcorpan>
hsivonen: is there any documentation on how to use the js parser?
14:14
<hsivonen>
zcorpan: no
14:14
<hsivonen>
but I can document it here
14:14
<zcorpan>
cool
14:14
<hsivonen>
there is one tested entry point
14:14
<hsivonen>
(the above)
14:14
<hsivonen>
and one untested entry point
14:15
<hsivonen>
oops
14:15
<hsivonen>
correction
14:15
<hsivonen>
there's only one entry point
14:15
<hsivonen>
the first argument is the HTML5 source as a string
14:16
<hsivonen>
the second argument is a DOM document object, whose children will be removed and replaced by the result of the parse
14:16
<hsivonen>
the fourth argument is a function object that will be called when the parse finishes
14:16
<hsivonen>
s/fourth/third/
14:16
hsivonen
can't count
14:17
<hsivonen>
the fourth argument is supposed to be an error reporting callback, but it's not implemented
14:17
<hsivonen>
so the fourth argument does nothing
14:18
<hsivonen>
so currently, the JS parser doesn't support fragments
14:18
<hsivonen>
and it can't attach to an existing tree without destroying it
14:18
<hsivonen>
I want to add those features in the future
14:19
<zcorpan>
was that what you had in mind as the second entry point?
14:19
<hsivonen>
the fragment entry point
14:19
<zcorpan>
ok
14:19
<zcorpan>
thanks
14:19
<hsivonen>
but its design sucked, so I removed it
14:19
<hsivonen>
note that parseHtmlDocument returns before the parse is complete
14:20
<zcorpan>
hence the callback?
14:20
<hsivonen>
to avoid hanging the browser, the parser pumps data through itself using timeouts
14:20
<hsivonen>
yes, hence the callback
14:22
<hsivonen>
oh, and calling parseHtmlDocument overwrites the write and writeln methods on the document object that is passed to it
14:22
<hsivonen>
and replaces them with its own implementations
14:23
<zcorpan>
hsivonen: this is all pretty neat work
14:23
<hsivonen>
thanks
14:26
<zcorpan>
hsivonen: does the parser support innerHTML?
14:28
<hsivonen>
zcorpan: the Java code support the fragment cases, but the JS version doesn't provide entry to that code and doesn't overwrite innerHTML setters
14:29
<zcorpan>
hsivonen: ok
14:46
<mpilgrim>
lachy or hsivonen or someone with blog.whatwg.org admin rights: can you create a category for "This Week in HTML 5"? (or give me admin rights so i can do it myself)
14:47
<Lachy>
mpilgrim, sure
14:47
<Lachy>
what do you want me to call it?
14:47
<mpilgrim>
no preference
14:48
<zcorpan>
html5-weekly
14:48
<zcorpan>
?
14:48
<gDashiva>
as the html5 turns
14:48
<mpilgrim>
as long as it gets its own feed
14:49
<mpilgrim>
my boss's boss's boss wants to follow it
14:49
<mpilgrim>
seriously
14:49
<Lachy>
I was thinking weekly-review
14:49
<mpilgrim>
weekly-review works for me
14:51
<Lachy>
done and I added the first post too it already
14:51
<mpilgrim>
thanks
14:51
<mpilgrim>
so, nominations for this week's episode?
14:51
<Lachy>
http://blog.whatwg.org/category/weekly-review/feed works
14:51
<mpilgrim>
hooray
14:52
<mpilgrim>
i've already written up the bit about the navigator object
14:53
<Lachy>
outerHTML was apparently added.
14:53
<gDashiva>
huzzah
14:54
<Lachy>
xml:lang and lang are now allowed in both HTML and XHTML
14:55
<hsivonen>
Lachy: except xml:lang in HTML and xml:lang in XHTML are totally different
14:55
<Lachy>
hsivonen, yeah, I know
14:56
<hsivonen>
Almost every XMLism adopted in XHTML 1.0 causes trouble when mapping HTML5 to XML infoset
14:57
<Lachy>
were there backwards compatibility issues with doing magic namespace processing for xml:lang in HTML documents?
14:58
<jacobolus>
hsivonen: the livedom thing is neat
14:59
hsivonen
goes file browser bugs exposed by it
15:00
<jacobolus>
hsivonen: that could presumably be used to hack really old browsers to make proper doms, no?
15:01
<jacobolus>
(not that there's necessarily any particular use for that)
15:01
<hsivonen>
jacobolus: yeah, I want to make a library that sniffs for HTML5 parsing and if not present, document.writes <plaintext>, attaches the JS parser to the DOM and parses the plaintext content
15:02
<hsivonen>
jacobolus: I think such a JS library would be useful for bootstrapping SVG and MathML in text/html once new browsers support HTML5 parsing so that only old browsers would take a perf hit
15:03
<jacobolus>
ah, good point
15:03
<jacobolus>
also might help for support of the various structure tags
15:07
<jacobolus>
hsivonen: might also be nice to show a box with conformance problems in that page, if it's possible
15:08
<hsivonen>
jacobolus: I could add parser-level error reporting, but porting the entire conformance checker to JS is a bit too much :-)
15:09
<jacobolus>
hsivonen: or maybe just a button that says "check conformance" and sends the text out to your conformance checker, putting the problems found back in a box
15:09
<hsivonen>
GWT is pretty smart about removing dead code, though, so I'm pretty sure the code size would go up notably if the error reporting code didn't count as dead
15:10
<hsivonen>
jacobolus: perhaps when a browser ships with Access-Control support
15:33
<hsivonen>
has anyone figured out yet if the new <a> content models have uglier consequences than the <ins> content model?
15:43
hsivonen
finds it amusing that Hixie changed the ins/del content models in the same check-in
15:43
<gDashiva>
Isn't the new <a> content model just what the de-facto content model was all along?
15:44
<hsivonen>
gDashiva: I mean conformance consequences
16:23
<kangax>
is it possible to work with canvas in rhino?
16:38
<mpilgrim>
http://blog.whatwg.org/this-week-in-html5-episode-2
16:50
mpilgrim
still can't believe he's getting paid for this
16:52
<MikeSmith>
mpilgrim: I think those writeups a massively good. It would so great if you can manage to keep doing them
16:52
<MikeSmith>
Hixie's productively might make pretty hard to keep up
16:52
<MikeSmith>
productivity
16:54
MikeSmith
wishes he knew what rhino is
16:54
<Dashiva>
MikeSmith: The javascript engine?
16:55
<MikeSmith>
Dashiva: for which browser?
16:56
<Dashiva>
None that I know of. It's a standalone thing. In java, even.
16:56
<MikeSmith>
ah
16:56
<zcorpan>
hsivonen: http://livedom.validator.nu/?%3C!DOCTYPE%20html%3E%0D%0A%3Csvg%3E%3Cscript%3Ea%3Cfoo%3Eb%3C%2Fscript%3E
16:57
<zcorpan>
gives different result than http://parsetree.validator.nu/?parser=html5&content=%3C%21DOCTYPE+html%3E%0D%0A%3Csvg%3E%3Cscript%3Ea%3Cfoo%3Eb%3C%2Fscript%3E&submit=Print+Tree
16:57
<zcorpan>
intended?
21:38
<zcorpan>
http://www.brandonfrohs.com/2008/08/whatwg-they-know-html-dont-they/
21:46
<Hixie>
if mark's boss' boss' boss wants to follow the "This week in html5" stuff, that means my boss' boss does too
21:46
<Hixie>
good to know!
21:46
<Hixie>
guess i'd better stop slacking off
21:47
<Hixie>
btw next time mark asks for someone to configure the whatwg blog for him or whatever
21:47
<Hixie>
just make him an admin
21:47
<Hixie>
:-P
22:03
<Dashiva>
We're the html-wg, not the pretty-sites-wg!
22:03
Lachy
commented too
22:03
<Lachy>
on the blog about the desing
22:03
<Lachy>
*design
22:04
<Hixie>
originally when i read that blog post i thought he was going to compliment us on not wasting time with frivalous things
22:04
<Dashiva>
Although I must agree default phpbb is ugly like little else
22:04
<Hixie>
:-)
22:05
<Lachy>
the default wordpress is quite ugly too. But I don't have design skills to redesign it.
22:05
<jacobolus1>
Dashiva: and not just ugly on the outside
22:05
<Dashiva>
jacobolus: Don't remind me, please :)
22:05
<Lachy>
photoshop and other image editors are way too complicated for me to produce anything even remotely good looking
22:06
<jacobolus>
Lachy: what good looking are you trying to produce?
22:06
<jacobolus>
there are various simple image editors …
22:07
<Lachy>
jacobolus, put it this way: the design of http://lachy.id.au/ is about the limit of my design skills. I know how ugly it is, but regardless of the image editor, I don't have design skills.
22:08
<jacobolus>
I like the color scheme, but the design is very busy :)
22:08
<Lachy>
but if someone were to give me a photoshop mockup, I could cut out the images and code it in a matter of days
22:09
<Lachy>
initally, I liked the colour scheme too. I used a colour scheme tool to pick it. But I'm quite tired of it, since I've had the same design for 4 years now
22:13
<Hixie>
ok i really think i'm going to have to define some sort of event loop
22:14
<Hixie>
that will make things much simpler to specify in all the places where we have "and then at the next available opportunity..." stuff