02:38
<heycam>
did the Location interface have stringification requirements that were removed? or am i misremembering?
02:39
<heycam>
i thought i remembered the interface having [Stringifies=href] on it
04:00
<shepazu>
does the case-insensitivity extend to attribute values, or just element and attribute names?
04:01
<heycam>
shepazu, which case-insensitivity?
04:03
<heycam>
i don't think in general the case of attribute values is changed: http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#attribute-value-%28double-quoted%29-state
04:04
<heycam>
when parsing
04:11
<shepazu>
ok, thanks, heycam
05:54
<othermaciej>
Hixie is a brave man
06:24
<hober>
indeed. I can't imagine going near the codec discussion
06:38
<shepazu>
Hixie: is heycam correct that attribute values are not parsed in a case-insensitive manner?
06:40
<shepazu>
(to unwind that a bit: my understanding is that attribute values, unlike element and attribute names, are left as the case they are encountered, and not matched to different values in the DOM)
06:52
<Hixie>
shepazu: correct -- iirc the only processing the parser does on attribute values (other than finding their delimiters) is &entity; expansion
06:57
<shepazu>
Hixie: ok, good... I was worried for a moment about path commands (relative and absolute)
06:59
<shepazu>
now, what implications does that have on things like <text text-anchor="miDDle"/> ? will that be "error-corrected", somehow permitted as is, or fail?
07:13
<heycam>
shepazu, afaict, that's purely within our (svg's) purview
07:17
<shepazu>
heycam: then that opens up a question we should be addressing, no?
07:17
<heycam>
yeah i think it would be good to talk about that
07:18
<Hixie>
shepazu: same as in XML.
07:18
<heycam>
especially for consistency with style="text-anchor: miDDle", which is allowed i think
07:18
<shepazu>
heycam: the one place it really matters, IIRC, is path@d
07:19
<shepazu>
oh, and anyplace where an @id is referenced
07:19
<Hixie>
personally I would encourage you to keep as much as possible case-sensitive as you can
07:19
<Hixie>
but that if you do have anything case-insensitive, you make it work the same in XML as in text/html
07:20
<shepazu>
Hixie: we would certainly do the latter
07:20
<Hixie>
but that's just me :-)
07:20
<shepazu>
you've argued for case-insensitivity for element and attribute names... this seems a bit inconsistent
10:29
<takkaria>
gsnedders: have you written content-type sniffing tests for abarth's draft? someone said you might have done
10:31
<gsnedders>
takkaria: I started playing around with ome but never got anywhere really
10:31
<gsnedders>
takkaria: gimme a min
10:32
<gsnedders>
takkaria: Ah, http://gsnedders.html5.org/tests/content-type.php is what I have
10:33
<gsnedders>
heh. Opera 10b thinks i'm on a slow network.
10:35
<takkaria>
gsnedders: hm useful, is the source for that around anywhere?
10:35
<gsnedders>
no
10:35
<takkaria>
ok :)
10:35
gsnedders
can grab it if you want
10:36
<takkaria>
no, don't worry, it was more out of curiosity than anything
10:37
<gsnedders>
takkaria: http://gsnedders.html5.org/tests/ct.tgz
10:38
<takkaria>
gsnedders: ta
11:27
<virtuelv>
Hixie: do you have anything written on Channel messaging yet? or are there requirements available anywhere?
11:34
<Hixie>
virtuelv: Channel messaging?
11:47
<gsnedders>
takkaria: yt?
11:47
<takkaria>
gsnedders: yup
11:48
<gsnedders>
takkaria: Had lunch?
11:48
<takkaria>
nope :)
11:48
<gsnedders>
takkaria: Want to go and get lunch? :)
11:48
gsnedders
uh, forgot to bring lunch
11:49
<takkaria>
ah
11:49
<takkaria>
I have curry to reheat
11:49
<gsnedders>
(and everyone around here has had lunch)
12:56
<virtuelv>
Hixie: yes
12:56
<virtuelv>
8.1 Event definitions
12:56
<virtuelv>
Messages in server-sent events, Web sockets, cross-document messaging, and channel messaging use the message event.
12:57
<virtuelv>
the link to channel messaging points to a description not yet written
13:02
<virtuelv>
I'm having a hard time understanding where this fits into anything
13:07
<virtuelv>
and is there any particular reason step six in the description uses Window instead of WindowProxy?
13:07
<virtuelv>
«Queue a task to dispatch the event created in the previous step at the Window object on which the method was invoked. The task source for this task is the posted message task source.»
13:08
<virtuelv>
ah, nevermind that last one
13:14
<gsnedders>
What's a good hex editor?
13:14
<jgraham>
gsnedders: I don't know what a good one is but ghexedit or something is a hex editor
13:15
<jgraham>
Oh it's called ghex
13:17
<Philip`>
I used to like Biew
13:30
jgraham
notes that not having ready access to windows makes testing jas harder
13:30
<jgraham>
*JAWS
13:31
Philip`
wonders how hard it can be to get access to Windows
13:32
<takkaria>
jgraham: I think there should be some vmware images around on the network somewhere, if that's any use
13:32
<Philip`>
It's not like you're trying to get access to a SunOS 4 machine or something
13:34
<jgraham>
takkaria: Well I have windowws 7 in Virtualbox but I don't think it likes my audio. Plus I don't really want to spend my weekends in the office
13:35
<virtuelv>
gsnedders: you could also have a look at Bless as a hex editor
13:35
<virtuelv>
(and since we're on such tools, presumably in a Gnome environment - Meld is by far the nicest diff viewer)
13:36
<takkaria>
you say 'meld', I say 'less' :)
13:40
<Philip`>
gsnedders: If the file you want to hex edit is actually executable code, then http://www.hex-rays.com/idapro/ is great
13:40
<gsnedders>
Philip`: It is not
13:40
<Philip`>
gsnedders: Then you should find some executable code and use IDA on that
13:44
Philip`
wishes Google hadn't decided to start looking all ugly with a wide margin around the search results page
16:30
<Philip`>
http://everything2.com/title/W3C - "There is not a single browser on the market that complys with all W3C standards, not even Microsoft Internet Explorer."
16:30
<Philip`>
(Thu Nov 02 2000 at 2:39:22)
16:30
<Philip`>
The first part of that sentence is still true, but I can't imagine anyone nowadays saying the second part
16:31
<Philip`>
I guess things haven't really changed that much, they've just shuffled around a bit
16:38
<tantek>
It may not be possible for any single implementation of anything to comply with all W3C standards - have you seen how many there are? http://www.w3.org/TR/
17:04
<takkaria>
things which suck: when your internet connection will only show you your broadband company's activation page and it's all in swedish
17:07
<Philip`>
All you need to do is activate the internet connection and then put the page through Google Translate to see what it said
17:35
<Midler>
takkaria: U want help with swedish?
17:35
<Midler>
Or is it way to late?
17:35
<takkaria>
I've sorted it now
17:36
<Midler>
In that way, Sweden is just like poland and other countries
17:36
<Midler>
U have ascii characters that u've never used...almost
17:36
<Midler>
=)
17:36
<takkaria>
my swedish is picking up rapidly
17:36
<takkaria>
but that's what happens when you sit and translate things word-by-word
17:37
<Midler>
:D
17:37
<Midler>
why not using google translate or something similar?
17:40
<takkaria>
I have time on my hands, and a swedish-english dictionary
17:41
<takkaria>
if I'm in sweden for twelve weeks, I might as well use all the opportunities I can to learn the language
17:41
<Midler>
oh, thats a long time
17:42
<takkaria>
I'm going to have a go at reading the paper in a bit
17:43
<Midler>
I guess comics might be an easy way to start. Seing their expressions makes it easier to understand what they say...i guess
18:00
<remysharp>
is there anyone here that knows a bit about the offline apps/manifest?
18:01
<remysharp>
I've got this url in Firefox (from what I can tell, RC3 == final release): http://html5demos.com/offlineapp - allow it to store and refresh I get a content encoding issue -
18:02
<remysharp>
does this look like a bug to do with FF or is this something I've done wrong?
18:02
<remysharp>
the manifest is being served under the right mime type (or should be - I did change it)
18:02
remysharp
works fine in Safari 4 + iPhone version.
18:26
<beowulf>
remysharp: i have the same problem
18:26
<remysharp>
cool - then it's not just me
18:26
<remysharp>
or - you've got same issue on my url or your own test?
18:27
<beowulf>
remysharp: when i try a hello world with a manifest I get a Content Encoding Error with ff3.5
18:27
<beowulf>
remysharp: http://www.carisenda.com/sandbox/storage/
18:27
<remysharp>
aye - cool - then it's a bug
18:28
<remysharp>
great. damn, that's some badass functionality and it doesn't work in FF :-(
18:28
<beowulf>
well, it would appear to be worse than doesn't work, it prevents you from using it
18:28
<remysharp>
Is that you then, beowulf == Stephen ? \o. hi if so!
18:29
<beowulf>
unless we've both done something stupid :)
18:29
<beowulf>
remysharp: hi! :)
18:31
<remysharp>
:)
18:31
<beowulf>
remysharp: it seems to work ok in ff 3.05
18:31
<beowulf>
3.0.5
18:32
<remysharp>
then it's definitely a regression bug.
18:32
<remysharp>
perfect, that's what I needed.
18:32
<remysharp>
thanks!
18:33
beowulf
checks his logs...
18:34
<beowulf>
ah, it asked me if it was ok to store stuff but doesn't look at the manifest
18:35
<remysharp>
my server logs do when I'm checking it
18:35
<remysharp>
or do you mean FF3.0.5?
18:35
<beowulf>
yeah, i'm trying 3.05 (which confused me by asking if it was ok to store stuff offline)
18:40
<ttepasse>
Is it just me or is whatwg.org down?
18:41
<ezyang>
http://downforeveryoneorjustme.com/whatwg.org
18:41
<ttepasse>
thanks.
18:43
<Midler>
ezyang: Now i just need to remember that DNS name...
18:43
<Midler>
:P
18:43
<beowulf>
remysharp: does your server use mod_deflate?
18:43
<remysharp>
nope
18:43
<remysharp>
but it does use gzip
18:43
<ezyang>
http://lmgtfy.com/
18:44
<ezyang>
;-)
18:44
<Midler>
hahaha
18:44
<Midler>
well, my brain only has place for a byte
18:44
<Midler>
oops to low
18:45
<Midler>
16bytes
18:45
<Midler>
i dont count .com
19:56
<Philip`>
How boring, Yahoo fixed their iso-2022-kr XSS vulnerability quickly :-(
19:56
Philip`
wonders how many other sites are vulnerable to the same issue
19:56
<ezyang>
linky?
19:57
<Philip`>
ezyang: http://search.yahoo.com/search?p=%EC%84%9E%E8%81%BD%E9%9D%91%E5%AD%83%EC%88%8D%E8%A8%8A%E6%98%B1%E7%A9%BF%E5%88%B7%E5%A5%84%E5%89%94%EA%B5%89%EB%82%91%EC%84%9E%E6%98%8C%E4%BE%84%E5%BE%9E%EC%97%86&ei=UTF-8&eo=iso-2022-kr
19:57
<Philip`>
in Chrome
19:57
<Philip`>
before today
19:57
<ezyang>
holy crap long url
19:57
<Philip`>
Safer than pasting Unicode into IRC :-)
19:58
<ezyang>
my irc client doesn't actually make it easy for me to open that :-(
19:58
<ezyang>
(the url marches off the side of my terminal)
19:58
<Philip`>
http://krijnhoetmer.nl/irc-logs/whatwg/20090630#l-406
19:58
<ezyang>
sweet
19:59
<ezyang>
So, what does the text "do"?
20:00
<Philip`>
$ echo '섞聽靑孃숍訊昱穿刷奄剔굉낑섞昌侄從없' | iconv -t iso-2022-kr
20:00
<Philip`>
)C</title><script>alert(123)</script>x
20:00
<ezyang>
whoa :-P
20:01
<ezyang>
Ok, so this only affects sites that let you set the page's output encoding
20:01
<Philip`>
so Yahoo encodes the text into that string of bytes, and Chrome doesn't support iso-2022-kr so it decodes as iso-8859-1 (or utf-8 or something) and executes the script
20:01
<Philip`>
(All other browsers do support iso-2022-kr, as far as I can tell)
20:01
<ezyang>
That's... a chrome bug
20:02
<ezyang>
Also, I still don't understand why search engines let you twiddle the output encoding
20:02
<Philip`>
It's not a bug for a browser to not support a relatively obscure encoding
20:03
<Philip`>
Even if it did, you might be able to find aliases that are supported by a server but not by the browser
20:03
<Philip`>
Twiddling the output encoding does seem like a pretty stupid idea
20:03
<Philip`>
but Google and Yahoo both allow it, so I guess they've got reasons
20:04
<Philip`>
Ultraseek allows it too, and didn't even forbid EBCDIC when I last tried it, so you can XSS it that way too
20:04
<ezyang>
So, the only good reason is accessibility
20:05
<ezyang>
For those pauvre browsers that don't support UTF-8
20:05
<ezyang>
Search engines like Google have a weird notion of what accessible needs to be
20:06
<ezyang>
I wonder what UTF-8 penetration is these days
20:07
<Philip`>
It doesn't help those browsers unless the users intentionally follow a link with the query parameter that sets the output encoding
20:07
<ezyang>
I thought that Google would do tricky user agent detection?
20:12
<Philip`>
If it was doing that, why would it put the output encoding in the query parameter instead of doing the tricky user agent detection when serving the results page?
20:56
<gsnedders>
jgraham: Can't get a external HD shipped to SE from amazon.co.uk :P
21:32
<eighty4>
gsnedders: buy one from dustinhome, I bet they're way better then crapy amazon
21:40
<gsnedders>
eighty4: And several hundred kronor more expensive
21:41
<eighty4>
so :)
21:44
<eighty4>
not several
21:44
<gsnedders>
several with what I need to buy :)
21:45
<eighty4>
ok
21:45
<eighty4>
I'm thinking of buying this: http://www.amazon.co.uk/Samsung-Spin-Point-Internal-Drive/dp/B000XZ5JBW/ref=sr_1_19?ie=UTF8&s=electronics&qid=1246394761&sr=1-19 from dusinhome
21:45
<gsnedders>
Amazon isn't the place to buy internal HDs :)
21:46
<eighty4>
costs 865 sek from amazon and 895 from dusin
21:50
<Dashiva>
gsnedders: Why not buy it, you know, from somewhere in Sweden?
21:50
<gsnedders>
Dashiva: It costs more?
21:50
<gsnedders>
Dashiva: Also, the Swedish websites are, uh, in Swedish.
21:53
<Dashiva>
I'm pretty sure Samsung Spin Point is the same :P
21:54
<gsnedders>
Dashiva: But actually buying stuff is more challenging :P
21:54
<Dashiva>
And surely you aren't the person to back away from a good challenge
21:56
<Dashiva>
Why must people insist on using obviously flawed analogies?
21:58
<gsnedders>
Dashiva: Where's the fun in coherent analogies?
21:58
<Dashiva>
I was thinking maybe the goal of an analogy was communication, not fun :)
21:59
<gsnedders>
Dashiva: You misguided bugger
22:23
<Sirisian>
I'm confused. whatwg is linked from the W3C site, but what is it?
22:23
<Sirisian>
Just a group of developers?
22:23
<Hixie>
http://wiki.whatwg.org/wiki/FAQ#What_is_the_WHATWG.3F
22:23
<Sirisian>
oh nvm
22:23
<Sirisian>
yeah just saw it
22:23
<Sirisian>
it's also on the main site. I didn't scroll down
22:29
<Sirisian>
Something I noticed. I've been looking through the Editor's Draft for the HTML 5 stuff. It has a section for the context menus and it says "When an element's context menu is requested [...] the UA must fire a simple event called contextmenu that bubbles and is cancelable at the element for which the menu was requested." I was going to post in the forum with a long post, but why isn't this extended to mouse up and mouse down
22:29
<Sirisian>
to set a standard for preventDefault for like middle click events by the browser. (Making a standard that says it should be "cancelable").
22:30
<Hixie>
there is another spec to handle the actual user interaction events
22:30
<Hixie>
like mousedown and mouseup
22:30
<Hixie>
separate from html5
22:30
<Hixie>
(it applies to things like svg also)
22:30
<Hixie>
that spec hasn't been written yet though
22:30
<Sirisian>
You mean for how browsers should react and work with them?
22:30
<Hixie>
yeah
22:30
<Sirisian>
oh is there a link? I must have missed it
22:30
<Hixie>
and how they should be cancelable and stuff
22:30
<Hixie>
it doesn't exist yet
22:32
<Sirisian>
oh that's why. Well that's good to here. Right now FF and IE have methods to allow a web developer to fully take over the mouse. I've been looking forward for a while when that idea is passed onto other browsers. Thanks for the response.
22:32
<Sirisian>
hear*
23:18
<roc>
I find it unfathomable that in 2009 people still argue that HTML5 should have draconian parsing
23:23
<othermaciej>
wait, what? who?
23:50
<roc>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=6777#c13