01:04
<Hixie>
1/away afk
01:04
<Hixie>
wow that went wrong.
01:05
<WeirdAl>
Hixie: FF12 will have support for XHR timeout 8)
01:06
<WeirdAl>
hm, I guess anne's the one who really needs to know
01:07
<WeirdAl>
I guess I was just thinking of those little browser icons indicating support for features in HTML5
03:11
<zewt>
i sure wish the whole "URL" vs "URI" distinction would go away; almost nobody even knows the difference, heh
03:12
<zewt>
it's particularly silly that we have a function called "createObjectURL" that says "return a URI"
03:12
<zewt>
"UR" + r(and()&1? "I":"L")
05:59
<Hixie>
zewt: that's why i basically only refer to things as URLs.
08:48
<zcorpan>
booo! google code search is no more :-(
09:01
<asmodai>
zcorpan: best replacement so far is koders.com
09:02
<zcorpan>
asmodai: cheers
09:02
<zcorpan>
+ <script>
09:02
<zcorpan>
+ document.write('<link rel=stylesheet href=foo.css><script></script>x');
09:02
<zcorpan>
+ // at this point, whether the DOM contains an "x" or not depends on how quickly the foo.css file could be processed
09:02
<zcorpan>
+ </script>
09:03
<zcorpan>
hsivonen: could we make this deterministic by awaiting a stable state before parsing the "x"?
09:11
<asmodai>
zcorpan: keep in mind that you need to search in lowercase, since camel casing triggers some special search IIRC
09:11
<asmodai>
http://googlecode.blogspot.com/2012/01/making-web-speedier-and-safer-with-spdy.html
09:11
<asmodai>
Seems SPDY is gaining traction.
09:17
<zcorpan>
yeah i noticed
09:17
<zcorpan>
the camelcase thing seems stupid
09:18
<zcorpan>
if you wanted to search for the terms separately, you'd use space between them
09:23
zcorpan
sent feedback to koders
09:33
<hsivonen>
zcorpan: the parsing is deterministic in Gecko in that case
09:33
<hsivonen>
zcorpan: computed style isn't
09:34
<hsivonen>
zcorpan: I'm not particularly keen at throwing engineering effort at making the computed style deterministic in this case
09:34
<hsivonen>
zcorpan: I think make the parsing non-deterministic would be bad, though, which is why it's deterministic in Gecko
09:34
<hsivonen>
zcorpan: possibly in violation of the spec depending on how my spec comments have been addressed
09:50
<zcorpan>
hsivonen: ok. if the comment is correct, then the spec's parsing is non-deterministic
09:55
<hsivonen>
zcorpan: I disapprove of the spec on this point, then
10:03
<zcorpan>
http://cheezburger.com/View/4088691968
11:24
<MikeSmith>
hsivonen: Thanks for the review. I incorporated your review changes and I also add code for conditionally serving up the additional W3C images
11:24
<MikeSmith>
just sent the updated patch to you
11:26
MikeSmith
now steps away for a beer break
11:30
<hsivonen>
MikeSmith: I replied
11:30
<hsivonen>
too early for beer over here, unfortunately
11:31
<MikeSmith>
:)
11:31
hsivonen
has lunch
11:31
<MikeSmith>
thanks
11:31
<louisremi>
Hi, I can't login to w3.org/wiki using my W3C member's login and password, where can I get some help?
11:35
<louisremi>
(or who should I ask?)
11:38
<jgraham>
MikeSmith is a good person to ask but
11:38
<jgraham>
05:27 * MikeSmith now steps away for a beer break
11:39
<louisremi>
MikeSmith: are you there?
11:46
<zcorpan>
does "overlap" imply "not completely nested"?
11:46
<jgraham>
In what context?
11:46
<jgraham>
I assume "yes"
11:47
<zcorpan>
well i'm asking because i wanted to suggest changing the name of http://dev.w3.org/html5/webvtt/#webvtt-file-using-only-nested-cues
11:48
<zcorpan>
to "WebVTT file using no overlapping cues", but that wouldn't be accurate either if nested cues are also "overlapping"
11:49
<jgraham>
I guess it's not clear in that case]
11:49
<jgraham>
I was thinking of markup
12:32
<louisremi>
MikeSmith: are you there? I'm having issues logging in to w3.org/wiki
12:44
<hsivonen>
argh. no single page link from the MathML 3 spec
12:44
<hsivonen>
neither from MathML 2
12:46
<hsivonen>
does MathML have its own <img>-like element?
12:48
<zcorpan>
http://www.w3.org/TR/MathML2/chapter7.html#id.7.1.4.2
12:49
<hsivonen>
zcorpan: thanks
12:50
<hsivonen>
so mglyph takes a font name, not a URL
12:50
<hsivonen>
so MathML has no elements that go and load external resources by URL?
12:50
zcorpan
also finds http://www.w3.org/TR/MathML3/chapter4.html#contm.image
12:50
<hsivonen>
zcorpan: different meaning of "image" :-)
12:51
<zcorpan>
yeah, i suspected
13:06
<MikeSmith>
louisremi: here now
13:13
<MikeSmith>
louisremi: btw I fixed the character borkedness in the user DB
13:21
<david_carlisle>
hsivonen: mathml3 mglyph takes images specified by url
13:23
<david_carlisle>
hsivonen: on image, yeah we explicitly tried to avoid name clash but no one told us that browsers implemented image until we'd had image for a decade (Hixie put us right:-)
13:25
<david_carlisle>
hsivonen: in addition mathml3 annotations can be specified via url rather than inline but you are allowed to (and i assume browsers would) ignore such annotations
13:28
<david_carlisle>
hsivonen: I had a non standard build of the mathml spec as a single page at one time for someone, was it you? it wouldn't be hard to do if it would be useful
13:29
asmodai
wishes more browsers would fix their mathml displaying.
13:29
<asmodai>
Oh well, at least my mathml displays well in ff and better and better in opera.
13:31
<david_carlisle>
amodai: it also should work well in IE with mathplayer (Ive not had an IE without mathplayer since 5.0 which is a while ago now)
13:31
<asmodai>
david_carlisle: I try to go with the least amount of additional dependencies. :)
13:32
<david_carlisle>
asmodai: so do I really, but needing Ie is a commercial rather than technical issue. "IE" is a bunch of related dlls and if you have to get one of them from a different source so be it, it;s not as if any of it is open source or anything
13:33
<david_carlisle>
needing mathplayer, I meant
13:33
<asmodai>
Fair point.
13:34
<david_carlisle>
Still, i wish it just worked without having to give the same lame excuse for a decade:-)
13:34
<Ms2ger>
They have SVG now... Maybe someone should write a script that renders MathML to SVG
13:35
<asmodai>
It's sad, because their math input panel can generate mathml to the clipboard
13:36
<david_carlisle>
There are some, i think, but I'd rather they did the decent thing and implemented it natively (MS has a very nice MathML system in its text rendering software behind the Ms Office suite)
13:37
<asmodai>
yeah, think the math input panel is part of that
13:37
<david_carlisle>
asmodai: Actually, not (It's a big company:-)
13:38
<asmodai>
ah
13:38
<asmodai>
Mmm, was it part o fmy base 7 install then?
13:39
<david_carlisle>
asmodai: The math input panel (unlike office) is part of windows itself. Of course though they appreciate the benefits of open standard formats for moving data from one to the other, or something....
13:40
<Ms2ger>
Yeah, they did embrace mathml, apparently
13:41
<david_carlisle>
Microsoft has been on the Math WG for quite a while (Giorgi from opera also) now it would be nice to have representatives from say CHROME in the WG as well:-)
13:42
<Ms2ger>
Agreed, those damned Chromies ;)
14:14
<hsivonen>
david_carlisle: oh, indeed mglyph in MathML 3 is different. thanks
14:14
<hsivonen>
it's nice that it uses the src attribute in no namespace for loading external content, so I have it covered already
14:15
<hsivonen>
does MathML autoload external content by URL using an attribute other than src?
14:15
<david_carlisle>
hsivonen: no
14:16
<hsivonen>
david_carlisle: thanks
14:16
<hsivonen>
whoa. there's a cd attribute
14:17
<hsivonen>
oh, but it takes an NCName
14:17
<david_carlisle>
hsivonen: yes but like the mathml2 definitionURL attribute there is no implication taht you have to do anything with it
14:18
<david_carlisle>
ther is also href in mathml3 which should act like html's a/@href
14:18
<hsivonen>
david_carlisle: ok
14:18
hsivonen
sees also cdgroup
14:18
<hsivonen>
and altimg
14:18
<david_carlisle>
hsivonen: again it's metadata on where one might find the underlying semantics but a browser neededn't fetch
14:19
<david_carlisle>
altimg ah, you have the spec open (I should) well you only need that if you don't render the mathml natively
14:21
<hsivonen>
david_carlisle: ok.
14:21
<hsivonen>
david_carlisle: I'm concerned about the case where someone puts MathML with external content references in email in order to get the message phone home when viewed
14:22
<david_carlisle>
grepped for anyuri in the relax schema and src href definitionurl altimg cdgroup is the full list
14:22
<hsivonen>
david_carlisle: thanks
14:23
<david_carlisle>
hsivonen: My brain isn't wired like a browser implementor's, could you expand on the concern
14:23
<hsivonen>
david_carlisle: the general expectation is that viewing an email message is a private operation (unlike viewing a Web page)
14:24
<hsivonen>
david_carlisle: so it's expected that when you view a Web page, as a side effect, stuff gets logged on servers
14:24
<hsivonen>
david_carlisle: when viewing an email, it's expected that no one but your computer and your email server find out that you viewed an email sent to you
14:25
<david_carlisle>
hsivonen: so what do you do with html img? (well I know that mail clients typically offer the user an option to fetch or not)
14:25
<hsivonen>
david_carlisle: I'm limiting HTML img only to cid: and data: URLs
14:26
<david_carlisle>
well my expectation (aka hope) would be that mathml mglyph does whatever img does and any others can be blocked by default in a context where these cncerns apply
14:27
<hsivonen>
david_carlisle: ok
14:28
<david_carlisle>
hsivonen: Is this a mozilla thing, or should we be specifying this stuff somewhere?
14:30
<Ms2ger>
Mozilla thing
14:30
<Ms2ger>
But maybe we should
14:33
<zewt>
hsivonen: well, it's up to the UA to handle that (gmail hides images by default until you click a button)
14:33
<hsivonen>
david_carlisle: this is a Mozilla thing
14:33
<hsivonen>
zewt: this is for a more paranoid than usual mode in Thunderbird
14:34
<hsivonen>
david_carlisle: there's no plan to expose this as a Web API, so no spec needed in that sense
14:34
<zewt>
right, a UA concern
15:02
<hsivonen>
FWIW, IE exposes an API of this nature to Web content and, IIRC, Collin Jackson used it as an example of a bad Web platform security feature in a presentation
15:03
<hsivonen>
(because getting interoperable impls. would involve a lot of spec surface that needs to change as the rest of the platform changes)
15:03
<louisremi>
MikeSmith: Thanks for fixing my name. Do you have any idea why I cannot connect to the wiki?
15:03
<Ms2ger>
toStaticHTML?
15:03
<hsivonen>
Ms2ger: yes
15:04
Ms2ger
thought it was a bad idea at the time
15:05
<hsivonen>
ok. now I need to write test cases for this stuff
15:05
<Ms2ger>
Eh
15:15
<MikeSmith>
louisremi: you mean you can't connect to it at all?
15:16
<MikeSmith>
..or do you mean you can't edit?
15:16
<louisremi>
MikeSmith: Yes, I'm using my W3C member login and password and it always fails
15:16
<MikeSmith>
what URL?
15:16
<louisremi>
MikeSmith: http://www.w3.org/wiki/index.php?title=Special:UserLogin&returnto=Special:UserLogin
15:17
<louisremi>
MikeSmith: oh wait, I just used the password that you originally sent me and it works [facepalm].
15:17
<MikeSmith>
:)
15:18
<MikeSmith>
yeah, that main wiki is not tied to the ldap user db
15:18
<louisremi>
MikeSmith: I thought updating my password would update it across all W3C sites
15:18
<MikeSmith>
it's a separate passoword
15:18
<louisremi>
MikeSmith: alright
15:18
<MikeSmith>
yeah, confusing
18:26
<Ms2ger>
Hixie, so you're saying the test is correct?
18:26
<Hixie>
i didn't check the test itself
18:27
<Hixie>
hence not closing the bug
18:27
<Hixie>
well that was bad timing
18:27
<Ms2ger>
Philip`, you check https://www.w3.org/Bugs/Public/show_bug.cgi?id=14421 please
18:45
<AryehGregor>
Does IE not support <base href=""> with relative URLs?
18:47
<AryehGregor>
document.querySelector("base").href = document.querySelector("base").href; fixes it.
19:04
<Hixie>
is there an appropriate wg for web api features relating to signing content before form submission?
19:07
<Ms2ger>
The wg that's going to spec window.crypto?
19:09
<Hixie>
yeah, that would do
19:09
<Hixie>
which is that one? :-)
19:10
<Hixie>
public-web-security⊙wo will do
19:10
<Ms2ger>
Sounds good
19:30
<AryehGregor>
Oh . . . except in some browsers (but not others!), document.querySelector("base").href = document.querySelector("base").href; resolves document.querySelector("base").href according to itself.
19:30
<AryehGregor>
Which really it should.
19:30
<AryehGregor>
Per spec.
19:30
<AryehGregor>
But it seems IE and Gecko don't.
19:31
<AryehGregor>
I should catch that in my reflection tests, but I didn't test setting URLs with <base> active.
19:31
<Ms2ger>
I think I remember some bug about that for Gecko...
19:31
AryehGregor
works around some more
19:33
<Ms2ger>
AryehGregor, https://bugzilla.mozilla.org/show_bug.cgi?id=600809
19:33
<GPHemsley>
When it comes to a datetime value, is there an explicit, complete list of valid characters somewhere?
19:34
<Ms2ger>
GPHemsley, how so?
19:34
<GPHemsley>
Ms2ger: What do you mean?
19:34
<Ms2ger>
Well, what do *you* mean? :)
19:34
<GPHemsley>
I'm looking for a complete list of valid characters in a datetime value
19:35
<AryehGregor>
Ms2ger, that bug's fix is not per spec.
19:35
<AryehGregor>
I think.
19:35
<Ms2ger>
It might not be
19:35
<Ms2ger>
The spec may or may not have said at the time what it says now
19:35
<Ms2ger>
GPHemsley, http://www.whatwg.org/specs/web-apps/current-work/multipage/common-microsyntaxes.html#valid-date-string-with-optional-time
19:36
<AryehGregor>
Ms2ger, https://www.w3.org/Bugs/Public/show_bug.cgi?id=15752
19:37
<Ms2ger>
Thanks
19:37
<GPHemsley>
Ms2ger: I'm not looking for a parser definition, just a list of valid characters. Otherwise, I'd have to go through each of the ~15 algorithms to make sure I didn't miss something.
19:37
<Ms2ger>
GPHemsley, click the links, they don't point to algorithms
19:38
<GPHemsley>
they point to disparate descriptions of each type of datetime value
19:39
<Ms2ger>
... Yes?
19:39
<GPHemsley>
That's not what I'm looking for
19:40
<GPHemsley>
There is a finite set of characters that can appear in a valid datetime value.
19:40
<GPHemsley>
I'm looking for a complete iteration of the characters in that set.
19:41
<Ms2ger>
It's the union of the characters mentioned in those descriptions
19:41
<GPHemsley>
but even those descriptions do not list their complete set at once; they only list them as they come up in the description/algorithm
19:42
<Ms2ger>
Yes
19:43
<GPHemsley>
nevermind
19:43
<GPHemsley>
what about the maximum length of a datetime value?
19:44
<Philip`>
Unlimited
19:45
<Philip`>
since you can have dates like 999999-01-01
19:45
<GPHemsley>
ok
19:47
Philip`
doesn't see any permitted characters other than [0-9-+TZ :.]
19:47
Philip`
provides no guarantee of correctness, though
19:49
<GPHemsley>
Philip`: Well, there are the duration characters, too. But thanks.
19:50
<Philip`>
Durations aren't datetimes, I think
19:53
<GPHemsley>
they are according to the HTML definition of "datetime value"
21:02
<asmodai>
jgraham / wilhelm : managed to snatch a screenshot :)
21:23
<jgraham>
asmodai: Oh, brilliant
21:23
<jgraham>
Let me know if there is a problem getting it onto the bug
21:28
<asmodai>
jgraham: emailed to the email address from my gmail, no bounce, so ought to be ok.
21:28
<asmodai>
Along with accompanying explanation
22:18
<zewt>
haha
22:18
<zewt>
i google for firefox, and the first link it gives me is Firefox 3.6
22:19
<zewt>
which goes to a page with no "no, really, this isn't what you want" notices
22:43
<TabAtkins>
zewt: Huh. My first link is mozilla.org/en-US/firefox/, which lets me download v9
23:32
<zewt>
TabAtkins: the top part is just the top-level "Mozilla Firefox Web Browser", which my eyes just automatically skip past (since download links 99.9% of the time show up on google as a subsection link); then the first one I actually read is 3.6, then to the right of that "Download" (which actually goes to the snapshot builds)
23:32
<TabAtkins>
Ah yes, indeed.
23:47
<ojan>
sigh, this stuff is so confusing http://lists.w3.org/Archives/Public/public-script-coord/2012JanMar/0104.html
23:48
<ojan>
it's not clear to me if the end result here addresses whether variable declarations shadow named properties
23:51
<TabAtkins>
OJAN
23:52
<TabAtkins>
OJAN HELP ME WITH INTRINSIC SIZING OF VERTICAL MULTI-LINE FLEXBOXES
23:52
<ojan>
TABATKINS
23:52
<ojan>
lol
23:52
<ojan>
vertical text is so crazy complicated
23:52
<TabAtkins>
Well, I just mean flex-flow:column wrap; flexboxes.
23:53
<ojan>
yeah
23:53
<TabAtkins>
It acts almost exactly like a vertical text flow, though. :/
23:53
<ojan>
right
23:53
<TabAtkins>
I may wait to finish this until Paris when I have Fantasai in the room with me.
23:54
<ojan>
TabAtkins: yeah...we haven't implemented wrap at all yet
23:54
<TabAtkins>
Yeah, single-line is easy because you dont linebreak at all.
23:55
<TabAtkins>
Hey, how do you do shrinkwrap on a single-line row flexbox? Do the normal calculations (pretend it's a block, etc.) then sum the items widths?
23:58
<ojan>
TabAtkins: we basically pretend it's an inline-block
23:58
<TabAtkins>
Okay, that's what I was assuming.
23:58
<ojan>
TabAtkins: we pretend each item is an inline-block
23:59
<ojan>
TabAtkins: you don't shrinkwrap the flexbox itself, right?
23:59
TabAtkins
wonders how far he can get with just handwaving that it acts "like text" and specifying linebreaking beahvior.
23:59
<TabAtkins>
If the flexbox is floated, yeah you do.
23:59
<ojan>
TabAtkins: oh right...we probably get that wrong :)