00:24
<webben>
hsivonen: Of course - by nature, any usage of DE is theoretical until actually implemented, no?
01:18
<vbulant>
caues
08:32
<hsivonen_>
webben: well, maybe we shouldn't cater for theoretical extension ahead of time in abstract architectures and should instead cross bridges as we get to them
08:33
<annevk>
I don't think that is still a maybe
08:52
<hsivonen>
looks like bluegriffon tries to be polyglot
08:52
<hsivonen>
a bit too much, even: there's an XML decl
08:59
<annevk>
jgraham_, the DOMImplementation-createDocument.html test gives very weird results
08:59
<hsivonen>
I wonder if the Super Friends have opinions on http://www.w3.org/Bugs/Public/show_bug.cgi?id=10589
08:59
<annevk>
jgraham_, "assert_throws:
08:59
<annevk>
function() { document.implementation.createDocument(namespaceURI, qualifiedName, doctype) } threw with code INVALID_CHARACTER_ERR (5) expected INVALID_CHARACTER_ERR (5)"
09:14
<annevk>
does the end of an SRT timestamp really look like "SS:,FFF"?
09:14
<annevk>
or "SS:.FFF"
09:16
<annevk>
looks like a bug in the spec
09:17
<annevk>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=10607
09:26
<annevk>
jgraham_, more interesting is that in Chrome this problem does not occur
09:34
<annevk>
oh yeah, the WebKit way of omitting DOM arguments is not very great
09:35
<annevk>
e.g. by createDocument it will give you a root element of "undefined"
10:04
<annevk>
does IE have isSupported on Node?
10:04
<annevk>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cscript%3E%20w(%22isSupported%22%20in%20document)%20%3C%2Fscript%3E
10:04
<annevk>
true / false?
10:05
<hsivonen>
annevk: true in IE9 PP4
10:06
<hsivonen>
or at least I think I have PP4
10:06
hsivonen
wishes the PPs had nicer version numbers
10:06
<annevk>
bah
10:06
<annevk>
but thanks
10:06
<hsivonen>
annevk: were you hoping to zap that feature?
10:06
<annevk>
yes
10:07
<annevk>
it's no different from document.implementation.hasFeature as far as I can tell
10:07
<annevk>
which makes it a fairly stupid addition of DOM3Core
10:09
<annevk>
there's lots of features I'd really like to zap
10:10
<annevk>
the draft has a whole list at the end already of features marked historical that have not been implemented everywhere and are fairly close to useless
10:10
<othermaciej>
you should still zap it if it is unused
10:11
<annevk>
agreed, but that is hard to determine
10:11
<annevk>
but it seems somewhat unlikely that people use this one
10:12
<hsivonen>
kinda sad that XOM that's supposed to be a DOM replacement that doesn't suck has indexed child access instead of firstChild/nextSibling links
10:12
<hsivonen>
now that we've learned on the browser side that indexed children are worse
10:13
hsivonen
keeps a cache of table index in the XOM tree builder to make foster parenting have better performance characteristics
10:14
<annevk>
codesearch yields one somewhat relevant result, but that script adding support for the method via prototyping, and it doesn't actually use it...
10:15
<hsivonen>
didn't the IE team say they prioritized DOM additions based on use in the wild?
10:15
<annevk>
maybe they also implemented additions that were easy?
10:16
<annevk>
no idea really
10:16
<annevk>
hopefully they come to TPAC
10:16
<annevk>
cause mailing list communication...
10:16
<hsivonen>
fwiw, your test doesn't work in IE8
10:17
<hsivonen>
fails before logging anything
10:17
<hsivonen>
s/anything/true or false/
10:18
<annevk>
weird
10:18
<annevk>
oh, maybe that is their "security feature"
10:18
<hsivonen>
annevk: two errors:
10:18
<hsivonen>
Function expected
10:19
<hsivonen>
Object doesn't support this property or method
10:19
<annevk>
MSDN doesn't list it
10:20
<annevk>
I guess that is sufficient
10:20
<othermaciej>
hsivonen: what is worst of all is supporting both indexed and previous/next
10:21
<othermaciej>
(though I suppose indexed is almost inevitably poor in the face of remove/insert)
10:21
<hsivonen>
XOM: http://pastebin.mozilla.org/788007
10:22
<hsivonen>
DOM: http://pastebin.mozilla.org/788008
10:22
<othermaciej>
you have to reverse-engineer previous/next, eh?
10:23
<hsivonen>
well, I have to emulate them by first searching the node's index in the parent
10:23
<othermaciej>
I kinda wish the DOM didn't have indexed access at all (which it does via childNodes)
10:23
<othermaciej>
then the fundamental accessors and mutators could all be O(1) in all cases
10:24
<hsivonen>
also, JavaScript FTW for not having to do all this casting
10:29
<annevk>
that code looks terrible
10:48
<hsivonen>
well. this isn't nice. I've updated GWT at some point and now the GWT Single Script Linker refuses to link the JS version of the parser
10:50
<hsivonen>
boo. and the GWT development mode plug-in would require downgrading to Firefox 3.5
10:57
<hsivonen>
I don't like it at all that GWT wants to default to generating browser-specific alternative scripts instead of generating a single-script with built-in feature sniffing
10:59
hsivonen
leaves the GWT build broken and code untested and moves on with making Firefox better
11:37
<annevk>
I wonder why Google News thinks that Koran burning has anything to do with either Science or Technology
11:37
<annevk>
is that a US thing?
11:39
<othermaciej>
no, Google News is confused
11:39
<annevk>
j/k ;)
11:40
<hsivonen>
jgraham_: have you figured out if the Ubuntu jumpiness since August is a reported and known problem?
12:28
<webben>
hsivonen: Maybe. I wasn't arguing for DE. I was providing hober with examples of how it might be used.
13:49
<annevk>
so per http://www.w3.org/TR/DOM-Level-3-Core/core.html DOMTimeStamp should have been Date in ECMAScript bindings
13:49
<annevk>
but that never got implemented that way :/
13:50
<annevk>
in any event, it seems that should move to Web IDL just like DOMString
13:50
<annevk>
should just name those string and timestamp really, but I guess people do not want a largescale s&r
13:55
<MikeSmith>
about http://www.w3.org/TR/css3-text/#white-space-rules
13:56
<MikeSmith>
"A tab (U+0009) is rendered as a horizontal shift that lines up the start edge of the next glyph with the next tab stop. Tab stops occur at points that are multiples of 8 times the width of a space (U+0020) rendered in the block's font from the block's starting content edge."
13:56
<annevk>
hey mike
13:56
<MikeSmith>
do any browsers actually do that
13:56
<MikeSmith>
annevk: hey
13:56
<annevk>
filing some Web IDL bugs I wonder if they should maybe be emailed to public-script-coord as well
13:56
<MikeSmith>
yes
13:56
<MikeSmith>
would be a good idea
13:57
<annevk>
i think for white-space:pre tabs might be rendered, yes
13:57
<annevk>
that is why some browsers have this proprietary (prefixed) tab-size property
13:57
<annevk>
e.g. -o-tab-size
13:57
<MikeSmith>
oh yeah
13:57
<MikeSmith>
that works
13:58
<MikeSmith>
but white-space-collapse does not
13:58
<MikeSmith>
that's what I get for reading specs
13:59
<annevk>
oh, white-space-collapse is some new stuff that nobody does
13:59
<MikeSmith>
I see
13:59
<annevk>
kind of wondering why that is being added now
14:01
<MikeSmith>
heh, ES5 spec has the typo "hyphon"
14:01
<MikeSmith>
cute
14:02
<MikeSmith>
jacobolus: hey, I used some page of yours the other day that I seem to remember finding extremely useful
14:02
<MikeSmith>
but at the moment I cant remember what page it was
14:03
<MikeSmith>
ah
14:03
<MikeSmith>
jacobolus: ah, the pages you have about customizing Cocoa config
14:03
<jacobolus>
MikeSmith: hey, what's up?
14:04
<jacobolus>
MikeSmith: this one? http://www.hcs.harvard.edu/~jrus/site/cocoa-text.html
14:04
<MikeSmith>
yeah, and this: http://www.hcs.harvard.edu/~jrus/Site/Cocoa%20Text%20System.html
14:04
<jacobolus>
MikeSmith: that's a hard link to the same file :)
14:04
<MikeSmith>
ah
14:05
<MikeSmith>
anyway, very useful info
14:06
<jacobolus>
thx
14:06
<jacobolus>
let me know if anything is confusing, &c.
14:07
<MikeSmith>
I think I got everything figured out from that now
14:07
<MikeSmith>
re-bound my ^y key to cycle through kill ring
14:07
<jacobolus>
MikeSmith: btw, these might be useful too http://www.hcs.harvard.edu/~jrus/site/selectors.html and http://www.hcs.harvard.edu/~jrus/site/system-bindings.html
14:08
<MikeSmith>
thanks!
14:08
<MikeSmith>
handn't seen those yet
14:09
<MikeSmith>
I did also add a ^W binding to deleteWordBackward already
14:10
<MikeSmith>
<manusporny> Ha! The vim editor gets HTML5+RDFa and Microdata support: http://ht.ly/2CXH5 /via @cwaring #rdfa #microdata
14:10
<MikeSmith>
that's pretty cool
14:11
<MikeSmith>
adds autocomplete of attribute names
14:11
<henrikbjorn>
when i launch vim it shows some errors that is removed but the editor finishing loading, is there some place where does get logged ?
14:11
<jacobolus>
I set it up so that move up down left right is ⌃i ⌃k ⌃l ⌃j and forward/backward by a word is ⌃o ⌃u and to beginning/end of line is ⌃h ⌃;
14:11
<MikeSmith>
oh, http://github.com/othree/html5.vim says it has ARIA support also
14:12
<hsivonen>
how odd. a tree builder bugfix I made and that works in Java seems to have regressed a tokenizer behavior on the C++ side
14:12
<MikeSmith>
jacobolus: oh, those sound good.. Ive been using the default control-option combos, which are a PITA to hit, especially on my Japanese laptop keyboard
14:13
<jacobolus>
yeah. I don't like moving my hand down to the arrow keys
14:13
<jacobolus>
and the "emacsish" ⌃a ⌃e &c. are not intuitive
14:14
<MikeSmith>
I guess I'm used to those from shell
14:15
<MikeSmith>
but in shell I use esc combos for word movements
14:16
<MikeSmith>
e.g, esc than b, esc than f
14:17
<MikeSmith>
henrikbjorn: :messages maybe?
14:17
<MikeSmith>
hmm, no
14:18
<henrikbjorn>
nope nothing is happening
14:18
<MikeSmith>
yeah, that was just guess
14:18
<henrikbjorn>
okay :)
14:19
<MikeSmith>
hmm, docs says that should work
14:19
<MikeSmith>
that and :echo errmsg
14:19
<henrikbjorn>
nope nothing :S
14:20
<MikeSmith>
darn
14:25
<MikeSmith>
kennyluck: do you know who this is? http://blog.othree.net/
14:25
<kennyluck>
MikeSmith: A Taiwanese folk I know. Why?
14:30
<MikeSmith>
kennyluck: dude has developed some HTML5-related software
14:31
<MikeSmith>
http://github.com/othree/html5.vim
14:32
<kennyluck>
MikeSmith: Oh, he's a Web Developer. :)
14:33
<MikeSmith>
kennyluck: hey, btw, about wu wei, you use 無為 in Taiwanese, right? (just like the Japanese word)
14:33
<kennyluck>
This is something I don't know why. But most of Taiwanese programmers prefer vim over emacs.
14:33
<kennyluck>
Yes. (I actullay don't know this term appears in Japan)
14:33
<annevk>
MikeSmith, othree is on this channel
14:33
<MikeSmith>
instead of 无为 I mean
14:33
<MikeSmith>
oh
14:34
<MikeSmith>
othree: thanks for vim stuff
14:34
<kennyluck>
Do we have a emacs html5 already BTW? :)
14:35
<kennyluck>
mode, I mean.
14:35
<kennyluck>
http://github.com/hober/html5-el
14:35
<kennyluck>
found it. Cool.
14:36
<hsivonen>
isn't that for XHTML5, though?
14:36
<MikeSmith>
yeah
14:36
<MikeSmith>
uses an XML parser
14:36
<MikeSmith>
but it still is great
14:36
<kennyluck>
Huh. I see. It does have nxml-mode dependency.
14:37
<MikeSmith>
we should make an nxml-mode for vim
14:37
<hsivonen>
we should make an incremental HTML5 parser for emacs
14:37
<MikeSmith>
yeah
14:38
<MikeSmith>
could maybe use James's nxml code as a starting point
14:38
<MikeSmith>
his parser code
14:38
<MikeSmith>
just have to learn emacs lisp
14:39
<MikeSmith>
your IQ will skyrocket if you learn lisp, I heard
15:06
<hsivonen>
Has Chrome 6 gone out of beta yet?
15:06
<hsivonen>
and Chrome 7 has already branched?
15:08
<annevk>
didn't they want to release every six weeks?
15:08
<annevk>
actually, I think Chrome 6 might be out
15:08
<hsivonen>
oh
15:08
<annevk>
http://en.wikipedia.org/wiki/Google_Chrome#Release_history
15:08
<hsivonen>
I was wondering if they have two releases that have branched but haven't gone final
15:10
<hsivonen>
*sigh* at the "Standards" section of the wikipedia article
15:11
<annevk>
i was just reading that
15:11
<evanchooly>
1
15:12
<hsivonen>
the "Standards" section for Firefox has this wonderful bit: "Firefox also implements[74] a proprietary protocol[92] from Google called "safebrowsing" (used to exchange data related with "phishing and malware protection"), which is not an open standard."
15:13
<annevk>
so they call out safebrowsing but not XUL, XBL, XPCOM, etc?
15:22
<Rik`_>
annevk: btw, @list is now filtering the propositions based on input
15:23
<annevk>
Ms2ger, hey, welcome back
15:23
<Ms2ger>
Thanks :)
15:24
<annevk>
I made a bunch of changes while you were gone, maybe you can review them
15:25
<annevk>
I was also wondering if we should file bugs on browser vendors to implement some of the changes to see if such a thing gains any traction
15:25
<hsivonen>
Ms2ger: http://lists.w3.org/Archives/Public/public-webapps/2010JulSep/0703.html
15:25
<Ms2ger>
I'll look at them and reply to your email after I skimmed through the rest of my inbox
15:26
<annevk>
e.g. removing xmlVersion, isSupported, etc.
15:26
<annevk>
Ms2ger, great
15:26
<annevk>
Ms2ger, hsivonen, regarding DOM Range, Hixie also suggested moving Selection there
15:26
annevk
goes off to do some shopping
15:26
<Ms2ger>
annevk, I saw that too :)
15:27
<Ms2ger>
And hsivonen, I was watching that bug
15:37
<hsivonen>
Ms2ger: cool
15:44
<Ms2ger>
Oh, fun, Shelby Moore in my bugmail
15:45
<erlehmann>
HTTP needs a status code for "not here yet"
15:45
<erlehmann>
(I am on a slow connection and have 333MB of PDFs to upload.)
15:49
<hsivonen>
Ms2ger: which bug tracker?
15:49
<Ms2ger>
Mozilla
15:50
<Ms2ger>
https://bugzilla.mozilla.org/show_bug.cgi?id=156388
18:44
<AryehGregor>
http://www.aregooglesbouncingballshtml5.com/
18:46
<aho>
balls!
18:54
<erlehmann>
balls are touching!
18:55
<aho>
:v
19:22
<AryehGregor>
Why does this display with no shadow in a Firefox nightly? data:image/svg+xml,<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd"><svg xmlns="http://www.w3.org/2000/svg"><defs><filter id="Gaussian_Blur"><feGaussianBlur in="SourceGraphic" stdDeviation="5"/></filter></defs><rect x="20" y="20" width="100" height="100" style="fill:black;filter:url(#Gaussian_Blur)"/></svg>
19:25
Ms2ger
was wondering the same thing
19:26
<AryehGregor>
Same in 3.6.
19:29
<aho>
saving this as svg, however, works: <svg xmlns="http://www.w3.org/2000/svg"><defs><filter id="Gaussian_Blur"><feGaussianBlur in="SourceGraphic" stdDeviation="5"/></filter></defs><rect x="20" y="20" width="100" height="100" style="fill:black;filter:url(#Gaussian_Blur)"/></svg>
19:30
<aho>
same thing prefixed with "data:image/svg+xml," does not work
19:30
<aho>
odd
19:31
<aho>
also doesnt work with ff4b6
19:32
<aho>
works with chrome, parsing error with opera *shrug*
19:35
<annevk>
Ms2ger, not sure the File API exceptions will stay
19:35
<annevk>
Ms2ger, that's why I did not add them
19:36
<Ms2ger>
Sure, but I'd like the list to stay in sync with HTML5
19:37
<annevk>
hmm, the dependency really should be the other way around, but ok, it'll be fixed eventually
19:39
<Ms2ger>
Yes, but until then we shouldn't have conflicting requirements
19:43
<hsivonen>
aho: no fragment id in data URLs
19:44
<aho>
ah
23:19
<hober>
hsivonen: re: emacs comment earlier, I've started work on one
23:19
<hober>
(an html parser in elisp that impls the html5 algorithm)
23:20
<hober>
it's nowhere near usable; I'll let you know when it gets to the point where it's worth looking at
23:39
<jgraham>
hober: Awesome
23:39
<jgraham>
hsivonen: No
23:40
<jgraham>
anne: (for the logs) the test cndition is just wrong. I will check in a fix at some point
23:40
<jgraham>
(i.e. the way that assert_throws determines pass/failiure in that case)