00:06
<Hixie>
zcorpan: wfm
00:20
<annevk>
I think it's confusing personally
00:24
<Hixie>
what i had isn't much better
00:33
<annevk>
I guess
02:13
<AryehGregor>
User Agent: Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.1; Trident/4.0;
02:13
<AryehGregor>
Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1) ; InfoPath.2)
02:13
<AryehGregor>
WHAT THE HECK.
02:13
<AryehGregor>
What genius decided to put both MSIE 6.0 and MSIE 8.0 in the IE8 UA string? Ugh.
02:14
<AryehGregor>
At least sometimes . . .
02:15
<Rik`>
AryehGregor: Adobe
02:16
<AryehGregor>
. . .
02:16
<Rik`>
with the PDF reader iirc
02:17
<AryehGregor>
Are you really saying that Adobe's PDF reader is adding "MSIE 6.0" to IE8 User-Agent strings?
02:17
<AryehGregor>
That's pathologically stupid.
02:17
<Rik`>
yes
02:17
<AryehGregor>
:/
02:17
<Rik`>
IE allows softwares to alter the UA
02:17
<Rik`>
and IE9 won't allow that
02:18
<Rik`>
so no more stupidity like this
02:23
<Rik`>
can't find a link about that, but I'm pretty sure it's an adobe product that inserts that in the registry
02:24
<Rik`>
http://blogs.msdn.com/ie/archive/2010/03/23/introducing-ie9-s-user-agent-string.aspx
02:48
<AryehGregor>
Hurrah.
03:38
<cardona507>
does <!doctype html> cause IE versions 6 and below to render in standards mode?
04:20
<theMadness>
Yes.
04:21
<theMadness>
They do trigger the doctype switching, the shortened doctype was picked because it was enough to incite said selection.
04:39
<cardona507>
thanks theMadness
04:45
<theMadness>
Don't mention it.
06:51
<zcorpan>
so url/URL didn't conclude with anything, or did it?
06:52
<Hixie>
no conclusion so far
06:52
<Hixie>
you've both made good points
06:52
<Hixie>
i'm waiting to see what happens with implementations
06:53
<zcorpan>
i don't really like that approach, i think it'll result in half being url and the other half being URL
06:54
<zcorpan>
and we'll have to stick to that for compat
06:55
<zcorpan>
maybe StorageEvent.url is already too late to change anyway?
07:03
<othermaciej>
oh, I didn't notice StorageEvent.url
07:03
<othermaciej>
it would be really nice to make them all consistent
07:05
<othermaciej>
StorageEvent is the only place in WebKit's IDL that has an "url" attribute currently
07:05
<othermaciej>
vs. Document, EventSource and WebSocket with URL
07:06
<zcorpan>
othermaciej: it was called StorageEvent.uri until a few days ago in webkit
07:06
<othermaciej>
zcorpan: if it just got renamed from "uri" then surely it can be renamed to "URL" too
07:07
<zcorpan>
yeah probably
07:08
<othermaciej>
there is another "uri" though:
07:08
<othermaciej>
./svg/SVGPaint.idl: readonly attribute DOMString uri;
07:08
<othermaciej>
several things with all-caps "URI" as the second word also
07:08
<zcorpan>
there's also node.baseURI
07:08
<othermaciej>
no urn or iri
07:09
<Hixie>
zcorpan: well i wanted to make it all URL but jonas said no. and so I changed it to all url and you said no. So I dunno what to do. For now the only browser that's actually consistently executing on their position is Mozilla, and they use url.
07:09
<othermaciej>
has Mozilla implemented WebSocket or EventSource yet?
07:10
<Hixie>
no
07:10
<Hixie>
they have implemented File though
07:10
<othermaciej>
did they change Document.URL?
07:10
<zcorpan>
they're working on websocket
07:10
<Hixie>
no
07:10
<othermaciej>
I'm ok with everything being "URL"
07:10
<Hixie>
convince jonas, not me
07:10
<othermaciej>
I am also ok with everything being "url" except for Document.URL
07:10
<Hixie>
that's what we have in the spec at the moment and what firefox is doing
07:10
<othermaciej>
I think what we'll actually end up with is an inconsistent mish-mash
07:11
<Hixie>
what is implemented in opera and webkit today is an inconsistent mish-mash already
07:11
<Hixie>
personally I don't buy jonas' camelCase argument
07:11
<othermaciej>
in our trunk we seem to have all URL except for StorageEvent
07:11
<Hixie>
because that would suggest .uRL
07:11
<othermaciej>
heh
07:12
<zcorpan>
maybe you should call it that
07:12
<othermaciej>
well, in WebKit code where the name is not forced by a spec, we try to follow the rule that an acronym is capitalized in the same way an initial letter would be in that place in an identifier
07:12
<othermaciej>
so class URL, method .urlFoo, method .fooURL
07:13
<othermaciej>
so I agree with Jonas on an abstract aesthetic level
07:13
<othermaciej>
but I'm also not super enthusiastic about running around renaming things
07:14
<zcorpan>
othermaciej: we'll have to rename some things anyway
07:14
<othermaciej>
true
07:14
<othermaciej>
may as well go with the scheme Jonas prefers then, since there's nothing wrong with it other than the legacy Document.URL not being able to comply
07:15
<othermaciej>
I'll ask webkit-dev if anyone minds renaming EventSource.URL or WebSocket.URL
07:16
<othermaciej>
of course, File.url should really be named something like File.tempURL to avoid the misleading impression that it is the file: URL for the file
07:20
<Hixie>
othermaciej: anything would be better than what we have now (.urn)
07:29
<Hixie>
is there a way to embed a block of text in mediawiki that might contain wiki-like syntax but have it all be escaped? the equivalent of <xmp> in HTML or <![CDATA[ in XML?
07:30
<danbri>
including markup which would show up as real markup in the Web site, you mean?
07:30
<Hixie>
yeah
07:31
<Hixie>
i want to paste in some wiki markup and have it show up like i wrote it, not interpreted
07:34
<Hixie>
alternatively is there some way of escaping everything?
07:34
<Hixie>
like in html you can just escape < and & and then you're done?
07:38
<othermaciej>
Hixie: .urn at least doesn't imply that it's a file: URL
07:39
<othermaciej>
(but of course it's otherwise terrible
07:39
<othermaciej>
)
07:41
<Hixie>
based on looking at all these subtitle formats, i think the original proposal of using SRT really is the least bad choice
07:41
<Hixie>
http://damowmow.com/temp/srtspec is a possible format based on it
07:42
<Hixie>
i've updated the ::cue() idea based on what zcorpan suggested
08:01
<Hixie>
http://forum.doom9.org/showthread.php?p=1396575#post1396575
08:13
<othermaciej>
wait, there's a caption format called "ASS"?
08:14
<lazni>
An Subtitle Standard?
08:14
<lazni>
s/an/a/
08:14
<othermaciej>
Hixie: by the way - Eric Carlson told me the other day that there are some YouTube videos that have fancy captions using some complicated XML format - he said he'd add it to the wiki, not sure if he's had time
08:15
<zcorpan>
Hixie: i think you can escape as in html
08:15
<zcorpan>
Hixie: at least i've used &lt;!-- for <!--
08:34
<tiglionabbit>
any reason why we don't have expanding textareas in the spec yet?
08:40
<hsivonen>
ooh. Hixie writes BNF. Did we enter a parallel universe?
08:44
JonathanNeal
crosses back over from the parallel universe.
08:54
<zcorpan>
hsivonen: where?
09:13
<Hixie>
hsivonen: BNF is a good format for defining what's valid (though that's pretty much all it's good for)
10:30
<daedb>
Hixie: <nowiki> works in Mediawiki.
11:30
<micheil>
Hixie: is there anything on how websocket servers should determine which protocol version that they should use?
11:30
<micheil>
like, a version negotiation
12:34
<theMadness>
Ok, this is a personal matter, but it's somewhat relevant. I just ended a discussion with a guy who supposedly is "in the know" of the w3c circles, and we have to start a project together.
12:35
<theMadness>
Apparently he's of the idea that using html4-like markup with a html5 doctype and some of the new input types is a reckless approach.
12:36
<theMadness>
What's worse is that he's boasting that it's the common opinion of all the people in the business. And that such an approach should be attempted only for personal or experimental sites.
12:36
<theMadness>
I'm about to call bullshit.
12:37
<boblet>
HTML4.5 (HTML5 doctype but no new elements) is perfectly cromulent
12:38
<theMadness>
thanks boblet, I needed the reassurance.
12:38
<boblet>
the dude is thinking of stuff <= IE 8 doesn’t support without JS
12:39
<boblet>
sectioning elements etc
12:39
<theMadness>
Nono, he's actively against stuff like input type="email".
12:39
<theMadness>
or the attribute "required"
12:39
<boblet>
input types are np — browsers treat em as type="text" if they don’t understand
12:40
<theMadness>
I know! But apparently everyone at the w3 (according to him) is secretly divulging the "do not use yet" mantra while apparently supporting early adoption on the public site.
12:41
<theMadness>
I felt like in a Lewis Carroll book. But this is potentially offtopic, so I'll stfu.
12:41
<boblet>
theMadness: all the people I know at W3 are saying use it in private too
12:41
<boblet>
oh noes! I’m not in the inner circle!! :|
12:41
<boblet>
point him at http://diveintohtml5.org/forms.html#type-email for input types
12:42
<theMadness>
Yeah I gave him that link too.
13:19
<Philip`>
theMadness: It's safer to listen to WHATWG mantras, since there isn't a big private/public divide and people can't be sharing secret opinions :-)
13:22
<boblet>
Philip`: you’re just saying that because you’re one of the inner circle members!! aha! I exposed you
13:22
<boblet>
protestations will only strengthen the case against you, my friend
13:22
<boblet>
;-)
13:22
<theMadness>
Philip`, and that would be?
13:23
<Philip`>
Will admitting it also strengthen the case?
13:23
<boblet>
Philip`: you catch on fast. have you played this game before?
13:24
<boblet>
everyone is guilty until proven innocent, including the people proven innocent
13:24
<Dashiva>
Philip`: What about the treehouse?
13:24
<Dashiva>
Seems private to me
13:25
<Philip`>
theMadness: That would probably be that lots of features are fine to use (like the new doctype, and like new input types as long as you test your site in a browser that supports them to make sure you're not using them wrong)
13:27
<theMadness>
I'm an Opera user, so I'm covered there, the features I'm intrested in using are input attributes and states, plus maybe video for iphone|ipad.
13:29
<theMadness>
On a unrelated node, anyone has a plugin to choke people over skype?
13:33
<theMadness>
Possibly with the "omae wa mo, shindeiru" soundclip prior to said choking.
14:52
<theMadness>
<meta charset="utf-8"> works in all browsers or should I still use content-type?
14:59
<Philip`>
It works in all browsers
14:59
<Philip`>
though ideally you should send the charset in the HTTP Content-Type too
14:59
<Dashiva>
theMadness: The choking is too obviousl. Also, "mou"
15:09
<theMadness>
Dashiva, I'm used to macrons with romaji :(
15:09
<theMadness>
I hate transliterations without the pretty lines, but this stupid client won't let me.
15:10
<Dashiva>
Oh, the irony
15:10
<Dashiva>
You transliterate the original Japanese to a form that's actually harder to write
15:10
<theMadness>
Not here, I have a spiffy keyboard layout that lets me write silly stuff. :D
15:11
<theMadness>
I just need to add a unicode <3 to be the perfect teenager girl. And of course the snowman.
15:22
<Dashiva>
Don't need a keyboard layout for that, IME ❤❥❤
15:22
<Dashiva>
Even comes with ☃
15:26
<theMadness>
Fact is, I'm into finger bondage. If it's not worth getting cramps over, it's not worth typing.
15:38
<theMadness>
「お前はもう死んでいる」
15:38
<theMadness>
took me a while.
16:37
<theMadness>
Where should I submit ideas for css pseudos?
16:38
<boogyman>
pseudo classes?
16:38
<theMadness>
Yep.
16:39
<theMadness>
http://www.w3.org/TR/css3-selectors/ oh, Hixie is there too.
16:39
<boogyman>
I am sure that the w3c site has a contact page
17:11
<theMadness>
Sent a mail to www-style. Hopefully it'll reach someone.
17:11
<theMadness>
But just in case: http://paste2.org/p/806357
17:14
<Lachy>
theMadness, www-style is the correct place. It's a very active mailing list
17:14
<theMadness>
I see.
17:15
<Lachy>
theMadness, but your mail hasn't arrived
17:16
<Lachy>
You should have received a confirmation e-mail from the mailing list asking you to agree to the archival of your message
17:16
<theMadness>
I gave it permission on the review page, id=8e869d7cae2ef4aaa880
17:16
<theMadness>
Yes, I agreed with the topmost option.
17:16
<Lachy>
weird then, that it's not showing up here http://lists.w3.org/Archives/Public/www-style/2010May/
17:16
<theMadness>
It's probably being moderated.
17:17
<Philip`>
It sometimes takes a day or two for your first message to a list
17:17
<Philip`>
(where "sometimes" means "always, in my experience")
17:18
<theMadness>
I see. Well, the paste is archived in the IRC log at least, so it's not completely totally absolutely lost.
17:20
<Lachy>
theMadness, similar ideas to yours have been discussed in the CSSWG many times in the past
17:21
<theMadness>
Always rejected?
17:22
<theMadness>
Weird, it's so clear that people has a use for it, and it's not like it's something complex to put in specification.
17:22
<theMadness>
I guess I'm wishing for a level of defensive coding that is too high.
17:24
<Lachy>
The problem is that the level of support for a feature is not always a binary decision.
17:24
<Lachy>
Some implementations may only ship with partial support for a property, or they may claim full support, while having a bug that causes some significant aspect to function wrongly.
17:24
<theMadness>
But for that we have vendor prefixes right?
17:25
<theMadness>
Of course, but what's the alternative?
17:25
<theMadness>
I mean, I know it wouldn't be perfect.
17:25
<theMadness>
But it's not like we have better stuff at hand...
17:27
<theMadness>
I mean, when I declare color:red, I know that the browser can have a bug and interpret red as green and give me green text, but that can't be a valid argument.
17:27
<Rik`>
theMadness: webkit introduced some new media queries with transforms/transitions, etc
17:28
<Rik`>
http://webkit.org/specs/MediaQueriesExtensions.html
17:28
<theMadness>
That's a very limited form of what I was suggesting.
17:29
<theMadness>
I'm not a strong application developer, I'm much better at web stuff, but would it be that hard to expose a series of flags in pseudoclass form?
17:30
<theMadness>
And yes Lachy, I'd accept the browser position on the subject, if the browser is lying or is sloppy, I'll have to make do.
23:21
<TabAtkins__>
Hixie: I think we can usefully define ::disclosure and ::content pseudos for <details> as the default binding, without stepping on future development of the bindings in real XBL.
23:21
<TabAtkins__>
That would solve the most pressing styling issues.
23:22
<TabAtkins__>
With the structure generally being <details><summary><::disclosure/></summary><::content/></details>.
23:23
<TabAtkins__>
Though there are alternate ways a UA could represent <details> that would still work well with ::disclosure and ::content.
23:26
<ment>
pardon my ignorance, but what's "<::disclosure/>" supposed to mean?
23:27
<ment>
(or any tag starting with double-colon)
23:27
<TabAtkins__>
A pretend ::disclosure pseudoelement (from CSS).
23:27
<TabAtkins__>
Self-closed for example brevity.
23:28
<TabAtkins__>
It's nothing to do with namespaces or anything, it's just a convenient way to express the shadow element that gets created when you use a pseudoelement in CSS.
23:28
<Dashiva>
What happens to content before summary?
23:28
<TabAtkins__>
Dashiva: Two <::content> elements?
23:31
<Dashiva>
Makes me wonder how browsers would render a <details> with content both before and after the summary
23:31
<TabAtkins__>
Me too.
23:31
<TabAtkins__>
I presume the <summary> would jump down when you opened the element.
23:32
<TabAtkins__>
Alternately, everything gets reordered in the box tree so that it comes after <summary>?
23:34
<Dashiva>
The element's shadow tree is expected to take the element's first child summary element, if any, and place it in a first 'block' box container, and then take the element's remaining descendants, if any, and place them in a second 'block' box container.
23:34
<TabAtkins__>
Well, the content model explicitly has <summary> coming first, so shrug.
23:35
<TabAtkins__>
Excellent, that's what I thought.