02:27
<quiznilo>
how long do you guess IETC will upgrade 6455 to a draft standard?
02:27
<quiznilo>
*will it take
02:34
<quiznilo>
nevermind, I see here there are no more 'draft standards' by rfc 6410
06:00
<MikeSmith>
if I want to cache binary objects on the client side for use in a Web app, what should I use?
06:00
<MikeSmith>
indexed DB?
06:25
<AryehGregor>
MikeSmith, how many, and how big?
06:29
<MikeSmith>
AryehGregor: a few dozen and around 500K or so I guess
06:29
<AryehGregor>
Around 500K each?
06:29
<MikeSmith>
up to 500K I gues
06:29
<AryehGregor>
Then IDB is probably your only real choice, I guess.
06:31
<MikeSmith>
OK
06:31
<MikeSmith>
I know there's the File System API
06:32
<MikeSmith>
but it is seeming like that might not get implemented across browsrs
08:00
<jgraham>
hsivonen: So, my feeling is that window.stop should probably abort the parser, but still go through The End in the normal way. Doesn't match Gecko or WebKit but almost matches Opera (and hopefully no one really relies on this behaviour)
08:32
<zcorpan>
should the html spec be changed to not suggest vendor prefixes?
08:39
<niloy>
does anyone know how to access a spdy enabled website, I tried the URL "spdy://google.co.in" but chrome just did a search on it
08:40
<annevk>
zcorpan: they have not been that harmful for HTML thus far, mostly due to them not being used
08:41
<annevk>
zcorpan: actually the one instance where they have (fullscreen) it has been harmful...
08:41
<jgraham>
annevk: "yes" was the shorter version of that answer
08:42
<annevk>
jgraham: hindsight
08:49
<doublec>
niloy: just accessing it normally with http://... should use spdy if the website supports it
08:57
<niloy>
doublec, thanks
09:08
<annevk>
for http://[12::]:12 Gecko omits [ and ] for .host
09:08
<annevk>
for .hostname sorry
09:08
<annevk>
no other browser does that
09:10
<annevk>
all these parsers are buggy as hell
10:54
<hsivonen>
Hixie: I'm here now
11:13
<annevk>
are there any URLs that do not have an authority component but do have a path component?
11:27
<zcorpan>
about:blank?
11:29
<zcorpan>
javascript:...
11:42
<annevk>
zcorpan: afaik those don't have a path component because e.g. javascript:"/../" does not get normalized
11:47
<AryehGregor>
annevk, file://?
11:48
<annevk>
AryehGregor: doesn't that have an implied authority component of localhost in most browsers?
11:48
<annevk>
(hence the triple slash)
11:49
<AryehGregor>
Um, I dunno. Syntactically it has no authority component.
11:49
<annevk>
yes it does
11:49
<AryehGregor>
Hmm.
11:50
<AryehGregor>
RFC 3986 doesn't allow an empty authority component, AFAICT.
11:51
<annevk>
yes
11:51
<annevk>
host -> reg-name -> *( unreserved / pct-encoded / sub-delims ) -> allows nothing
11:52
<AryehGregor>
Okay, I guess.
11:52
<annevk>
not that RFC 3986 is followed much
11:53
<annevk>
I'm sketching a new parser
11:58
<matjas>
Ms2ger, jgraham: would you mind giving some feedback on http://mathias.html5.org/tests/html/named-character-references/w3c/? trying to use the test harness correctly
11:58
<matjas>
is is ok like that, or would you prefer separate test() calls for each test?
11:59
<Ms2ger>
Mm.. Maybe
11:59
<Ms2ger>
With one test, you stop at the first failure
12:00
<Ms2ger>
So I guess if there are still browsers that fail anything there, a test() per entity would be better
12:00
<annevk>
put sketch for URL parser online http://wiki.whatwg.org/wiki/URL
12:01
<annevk>
[tm]: ^^
12:05
<zcorpan>
annevk: some browsers expose .pathname for those urls, though
12:06
<annevk>
zcorpan: yeah, even query
12:06
<annevk>
zcorpan: not sure that makes a lot of sense, but we could reuse .pathname to store the data I guess
12:07
<zcorpan>
matjas: one test() per test is better. there's a convenience function called generate_tests() or some such, though i haven't used it myself
12:07
<matjas>
zcorpan: the source said generate_tests() might be removed in the future, so i got scared
12:08
<zcorpan>
matjas: heh, ok. the other thing still applies though :-)
12:08
<matjas>
Ms2ger, zcorpan: thanks, on it!
12:09
<zcorpan>
general guideline: failing one test() shouldn't change the total number of tests
12:10
<jgraham>
matjas: generate_tests is pretty safe I would say
12:10
<Ms2ger>
matjas, generate_tests is deprecated-ish :)
12:10
<jgraham>
Heh
12:10
<matjas>
12:10
<matjas>
what just happened
12:10
<jgraham>
Well it can be problematic
12:10
<Ms2ger>
Unless jgraham removes that comment
12:10
<jgraham>
If you have an exception in the code that is generating the tests, you suddenly get 0 tests
12:11
<matjas>
so, like this? http://mathias.html5.org/tests/html/named-character-references/w3c/ what’s with the metadata warning?
12:11
<zcorpan>
jgraham: but that'd apply for a custom-written test generator equally, right?
12:11
<jgraham>
zcorpan: Yes
12:11
<jgraham>
matjas: plinss being over0enthusiastic
12:12
<matjas>
jgraham: ok, i’ll just ignore it then. do the tests look ok otherwise?
12:13
<zcorpan>
matjas: looks ok from a quick look
12:14
<jgraham>
I don't see anything wrong at a glance.
12:15
<matjas>
ok, thanks for taking a look
12:16
<[tm]>
xslt2 implemented in Javascript, friends
12:17
<[tm]>
performantly
12:17
<zcorpan>
now implement javascript in xslt2 in javascript
12:17
<[tm]>
hahaha
12:18
<zcorpan>
performantly
12:18
<[tm]>
this odds a sure sign of the apocolypse
12:19
<[tm]>
i would like to see the code
12:19
<[tm]>
wait no i really wouldn't
12:20
<zcorpan>
hey icaaq
12:20
<[tm]>
next we implement visual basic in js
12:20
<icaaq>
zcorpan: hey!
12:20
<Ms2ger>
[tm], naaah
12:20
<Ms2ger>
Visual Basic in js-on-xslt-on-js
12:20
<zcorpan>
icaaq: how's things?
12:21
<icaaq>
zcorpan: great! tgif. long time since I was in this channel :) you?
12:22
<zcorpan>
icaaq: good good. i'm still here regularly :-)
12:24
<icaaq>
zcorpan: :) I heard chaals is leaving opera
12:24
<zcorpan>
yeah
12:27
<[tm]>
they are hiring Assa Dotzler to replace chaals
12:28
<Ms2ger>
Asa*
12:28
<Ms2ger>
And my condolences
12:28
<icaaq>
really?
12:28
<zcorpan>
icaaq: don't pay attention to these guys :-P
12:29
<Ms2ger>
Friday night for [tm]
12:29
<Ms2ger>
And, well, [tm]
12:29
<annevk>
also, Ms2ger
12:29
<annevk>
known troll
12:29
<icaaq>
haha
12:29
<Ms2ger>
Yw
12:30
<[tm]>
every night is Friday night for Mikey
12:30
<annevk>
Lachy: http://dbaron.org/css/intrinsic/
12:30
<Lachy>
annevk, thanks. I found that already
12:31
<annevk>
I was going to update the wiki, but it eh is down?
12:31
<Lachy>
but stig wants me to pick something related to docxs instead of layout
12:31
<Lachy>
wiki works for me
12:32
zcorpan
wonders what Lachy and annevk are discussing
12:32
<Lachy>
zcorpan, I'm being assigned to more spec work. I need to find a spec that needs working on.
12:33
<zcorpan>
ah. nice!
12:33
<zcorpan>
going into css land?
12:33
<Lachy>
current options now include DOM Events and WebECMAScript stuff documented in the wiki.
12:33
<Ms2ger>
Apparently not
12:34
<zcorpan>
ok
12:35
<annevk>
Lachy: could talk to olak as well
12:35
<annevk>
Lachy: sitecompat usually means a ton of spec issues somewhere
12:36
annevk
updated http://wiki.whatwg.org/wiki/Specs_todo#CSS
12:50
<matjas>
can i submit this test somewhere? http://mathias.html5.org/tests/javascript/identifiers/
12:51
<Ms2ger>
To ECMA? No
12:52
<Ms2ger>
They don't accept tests unless you work for a member company
12:55
<matjas>
and the W3C only accepts HTML/CSS tests, right?
12:56
<Ms2ger>
I guess
12:56
<annevk>
matjas: well, they typically accept tests for specs
12:56
<Ms2ger>
I'd take them ;)
12:56
<annevk>
matjas: but I think if you talk to plh / mike you can get a place for JavaScript tests
12:57
<matjas>
the above test tests for ECMAScript 5.1 + Unicode 6.1 + http://wiki.whatwg.org/wiki/Web_ECMAScript#Identifiers
12:57
<matjas>
so it’s more of a compatibility tests than anything else
12:58
<matjas>
as the ES spec allows engines to use Unicode 3 (and doesn’t require the Web ECMAScript exception for identifiers)
12:59
matjas
looks at [tm]
13:10
<[tm]>
matjas: we shall make a place for js tests
13:11
<[tm]>
but out will have to
13:11
<[tm]>
will have to wait till really next week
13:12
<[tm]>
because the maker OS deep i
13:12
<[tm]>
is deep into his 7th chu-hai
13:16
<jgraham>
matjas: Tack för testerna :)
13:16
<matjas>
jgraham: du är välkommen! more useful ones coming
13:17
<matjas>
(blame Google Translate if that came out wrong)
13:17
<icaaq>
what's up with the swedish? :)
13:18
<matjas>
[tm]: deal
13:19
<hasather>
matjas: "Var så god" would've been correct :) "Välkommen" is "welcome" as in the greeting.
13:19
<jgraham>
icaaq: Varför inte Svenska? (actually I don't really know. Maybe I just feel bad that I know so little Swedish after living here for so long)
13:19
<icaaq>
du bor i sverige? var :)
13:20
matjas
sets lang=en
13:21
<jgraham>
I Linköping
13:22
<jgraham>
matjas: Sorry, I didn't mean to waken all the Swedes ;)
13:22
<icaaq>
ahh ok. Jag bor i Mora. *back to english*
13:23
<icaaq>
:)
13:29
<[tm]>
jag förstår inte
13:42
<Ms2ger>
[tm], also webgl, please :)
13:44
<[tm]>
hai
13:47
<annevk>
so yeah, I'm no longer the editor of CSSOM View
13:47
<annevk>
but it seems nobody else is either
13:47
<annevk>
http://dvcs.w3.org/hg/csswg/log/1693d8c293a4/cssom-view/Overview.src.html
13:48
<annevk>
:/
13:50
<[tm]>
annevk: hahaha
13:50
<[tm]>
real leadership going on over there
13:51
<[tm]>
great example to all of us of how to run things
13:52
<[tm]>
crazy plus crazy
13:53
<[tm]>
taking out to
13:53
<[tm]>
taking it to a whole nothet
13:53
<[tm]>
nother level
14:04
<[tm]>
ah Julian is such a treasure
14:06
<[tm]>
annevk: the answer is that if he wants that kind of spec, he can write one
14:06
<[tm]>
but he's not ever going to do that of course
14:07
<annevk>
I doubt that is feasible too
14:07
<[tm]>
far easier to just forever play guardian of purity from the sidelines
14:07
<annevk>
it's kind of hard to say conclusively of course, but I already like my pseudo-code a lot better than RFC 3986 to answer questions about URLs...
14:07
<[tm]>
yup
14:20
<[tm]>
Christ
14:20
<[tm]>
annevk: this is a really special neurosis
14:21
<[tm]>
that thread
14:21
<[tm]>
dude really needs to find another hobby
14:48
<zcorpan>
annevk: how silly of you not to consider what webkit will do tomorrow
14:48
<annevk>
heh
14:48
<annevk>
my idea was to write down what webkit will do tomorrow
14:51
<annevk>
the "patch the existing spec" tactic has fallen out of favor with me for specs that are not actively maintained
14:52
<annevk>
HTML, Media Queries, Cookies, MIME, DOM, Encoding all showed that to be much harder than starting with a fresh perspective
14:57
<annevk>
matjas: what the hell is MySQL three-byte UTF-8
14:58
<matjas>
annevk: MySQL’s `utf8` collation doesn’t support astral symbols
14:58
<annevk>
matjas: not natively or not at all?
14:59
<matjas>
annevk: not at all
14:59
<annevk>
I wonder how I manage to store those characters in blog posts then...
14:59
<gavinc>
BMP characters only :(
14:59
<matjas>
annevk: wat
15:00
<matjas>
annevk: to get the 𝌆 symbol in http://mathiasbynens.be/notes/javascript-encoding I had to use str_replace()
15:00
<matjas>
annevk: got an example of such a blog post?
15:00
<annevk>
hmm maybe I don't
15:04
<matjas>
annevk: i believe wordpress has a workaround, you might have written something similar
17:53
jwalden
is bemused to see that some readonly attributes in WebIDL can have setter functions associated with them in the ECMAScript bindings
18:10
<aklein>
smaug____: yt? got a question about an oddity with Firefox MutationObservers and the style attribute...
18:12
<smaug____>
aklein: yes?
18:15
<aklein>
smaug____: it appears that changing non-existent style properties triggers a MutationRecord creation. here's the test that passes in WebKit but fails in Firefox: http://trac.webkit.org/browser/trunk/LayoutTests/fast/mutation/observe-attributes.html#L741
18:17
<aklein>
not a big deal, seems to me, but it's the one test that failed when I ran Aurora through our tests just now
18:18
<smaug____>
aklein: so the problem isn't about mutationobserver
18:18
<smaug____>
just about whether attr should change in that case...
18:18
<smaug____>
"Mutating the CSSStyleDeclaration object must create a style attribute on the element (if there isn't one already) and then change its value to be a value representing the serialized form of the CSSStyleDeclaration object."
18:19
<aklein>
true, it's a lower layer than MutationObserver
18:19
<smaug____>
aklein: so, per spec there should be a mutation
18:21
<aklein>
smaug____: thanks for the spec citation, I think this then buckets into the other set of failures (general behavioral differences). I very much doubt this would be a problem in practice.
18:21
<smaug____>
yeah
18:27
<smaug____>
aklein: in your test, perhaps you shouldn't use shouldBe
18:27
<smaug____>
but something like todo
18:27
<smaug____>
if you have todo
18:27
<smaug____>
so that it clearly indicates what the result should be
18:36
<matjas>
I bet Hixie is gonna ♥ JSON even more once he sees https://www.w3.org/Bugs/Public/show_bug.cgi?id=17490#c2
18:45
<hober>
matjas: yeah, i rejected a patch to json.el that added octal & hex literals and the patch author was incredulous
19:26
<Ms2ger>
So, if you pass 4 arguments to drawImage, what should happen?
20:46
<Hixie>
oh are you effing kidding me
20:47
<Hixie>
how can json not support hex literals
20:47
<Hixie>
especially when it supports hex character escapes!
20:47
<Hixie>
wtf
20:48
<Hixie>
matjas: why the str()? does python not convert integers to strings when concatenated with strings?
20:49
<SimonSapin>
Hixie: no, to avoid surprises like '1'+1 being '11' … or maybe 2
20:50
<Hixie>
so it's a dynamically typed language that doesn't do dynamic conversion? awesome.
20:50
<Hixie>
all the disadvantages of strong typing with none of its advantages
20:51
<SimonSapin>
yes, strong+dynamic
20:51
<SimonSapin>
strong/weak is not the same as dynamic/static
20:51
<Hixie>
yes, i am aware :-)
20:53
<SimonSapin>
right. Don’t feed the troll ;)
21:04
<zewt>
trying to get webgl folks to do the right thing can try one's endurance
23:12
<aklein>
Hixie: the HTML parser doesn't reference DOM4's mutation methods when constructing the tree, does it? from the text it seems to know how to create a tree all on its own.
23:49
<aklein>
Hixie: perhaps next week. in the meantime, http://lists.w3.org/Archives/Public/public-webapps/2012AprJun/1198.html