00:27
<TabAtkins>
Hixie: Is this better (for :read-only) http://dev.w3.org/csswg/selectors4/#rw-pseudos
01:04
<Hixie>
TabAtkins: so all elements match one or the other?
01:04
<Hixie>
TabAtkins: what should type=radio match? it's neither user-alterable, nor read-only...
01:05
<Hixie>
TabAtkins: (also if you could make user-alterable some sort of obvious hook, e.g. give it an ID and make it italics or bold or something (<dfn>), that'd be ideal)
01:07
<othermaciej>
Hixie: I'm thinking about implementing srcset and I have a question about the spec
01:07
<Hixie>
othermaciej: shoot
01:07
<othermaciej>
(since I actually have free time to code this weekend)
01:08
<othermaciej>
ok, if an image has both src and srcset set, and src appears first textually
01:08
<Hixie>
(the order doesn't matter)
01:08
<othermaciej>
if the parser treats this as the src attribute being added and then the srcset attribute being added
01:08
<Hixie>
(attribute order never matters)
01:08
<othermaciej>
then the obvious consequence is that the load for the src resource will be initiated first
01:09
<Hixie>
the HTML parser is defined as adding all the attributes simultaneously
01:09
<othermaciej>
is that wrong?
01:09
<othermaciej>
ok
01:10
<Hixie>
hence the "or both" in: "A user agent that obtains images immediately must synchronously update the image data of an img element whenever that element is created with a src attribute, a srcset attribute, or both."
01:16
<othermaciej>
thanks
01:19
<othermaciej>
Hixie: I think WebKit probably has quite a few bugs in cases where that matters
01:19
<othermaciej>
Hixie: hmm, on second thought, maybe not
01:19
<othermaciej>
sorry for the noise
01:20
<othermaciej>
it looks like we set them all, then process one at a time, but each attribute when processed checks if the relevant others have been set
02:39
<othermaciej>
man, if building wasn't so slow I'd be done implementing srcset by now (at least the subset handling resolution descriptors only)
03:02
<Hixie>
othermaciej: yeah, i've actually seen cases of browsers having bugs due to the attribute order mattering in practice
03:02
<Hixie>
othermaciej: surprisingly few, all things considered
03:02
<othermaciej>
I think I misinterpreted our code (though I certainly wouldn't promise we are bug-free in this area)
03:10
<TabAtkins>
Hixie: Yeah, it's either/or. This was based off the HTML definition. ^_^
03:11
<TabAtkins>
type=radio I'd consider user-alterable. You can alter the checkedness, which is *like* it's content.
03:32
<cabanier1>
hixie: ping
03:33
<cabanier1>
I saw that you replace 'any' with a union in the Canvas spec
03:35
<cabanier1>
I think that that will change the behavior
03:38
<Hixie>
TabAtkins: html's was based on css' old one, that's why i asked fantasai to change it :-) there's some bug about it
03:38
<Hixie>
cabanier1: sounds plausible
03:38
<Hixie>
cabanier1: which change in particular?
03:39
<cabanier1>
[e] (0) Clean up some IDLs that use 'any' to specify the exact type instead, and clean up some surrounding markup
03:40
<cabanier1>
Fixing https://www.w3.org/Bugs/Public/show_bug.cgi?id=17952
03:40
<cabanier1>
Affected topics: Canvas, DOM APIs, Microdata
03:40
<cabanier1>
git-svn-id: http://svn.whatwg.org/webapps@7260 340c8d12-0b0e-0410-8428-c7bf67bfef74
03:40
<Hixie>
what behaviour would it change?
03:41
<cabanier1>
if you pass it something that can't be converted to a string, you will get a type error.
03:41
<cabanier1>
while it should ignore it instead
03:41
<Hixie>
what can't get converted to string?
03:42
<cabanier1>
the argument you pass to strokestyle/fillstyle
03:42
<Hixie>
do you have a concrete example of some code that would change behaviour?
03:42
<Hixie>
sorry i'm having trouble understanding
03:44
<cabanier1>
let's say you pass in an object that implements its own 'tostring' method that returns a csscolor
03:44
<cabanier1>
with your change, you can pass in that object and it will be accepted
03:45
<Hixie>
like context.fillStyle = { toString: function () { return 'blue' } }; ?
03:45
<cabanier1>
yes
03:45
<Hixie>
yeah, that would have been ignored before but is now treated properly, right?
03:45
<cabanier1>
the wording doesn't allow that
03:46
<cabanier1>
that's true
03:46
<cabanier1>
I guess the prose should change
03:46
<cabanier1>
and we should try if it is allowed
03:46
<Hixie>
what prose? it all looks right to me...
03:47
<cabanier1>
"The style can be either a string containing a CSS color, or a http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#canvasgradient or http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#canvaspattern object"
03:48
<Hixie>
oh the non-normative text
03:48
<Hixie>
yeah i guess the next sentence is wrong, huh
03:48
<Hixie>
oh, no
03:48
<Hixie>
it's still true
03:49
<Hixie>
because strings that don't parse as css colors are still ignored
03:49
<cabanier1>
I'm still trying to understand the WedIDL stuff...
03:49
<cabanier1>
yes. That's true
03:49
<cabanier1>
so, you can set an object. But when you look at the value, you would get a string back
03:49
<Hixie>
i'm assuming "you" in those two sentences refer to different people
03:50
<cabanier1>
correct
03:50
<Hixie>
if you mean that this: context.fillStyle = { toString: function () { return 'blue' } }; ...would result in the function defined in the canvas prose being handed a string, that's my understanding.
03:50
<Hixie>
(well, setter, not function)
03:52
<cabanier1>
as long as people don't expect this to work: var mycolor="…"; context.fillStyle = { toString: function () { return mycolor; } }; ...
03:52
<cabanier1>
canvaspattern and canvasgradient allow that
03:52
<Hixie>
why would that not work?
03:52
<cabanier1>
"When set to a http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#canvaspattern or http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#canvasgradient object, the assignment is http://www.whatwg.org/specs/web-apps/current-work/multipage/infrastructure.html#live, meaning that changes made to the object after the assignment do affect subsequent stroking or filling of sh
03:52
<Hixie>
oh you mean, dynamically
03:52
<Hixie>
yeah
03:53
<Hixie>
it'll work, it'll just set the fillStyle to whatever mycolor's value is at setting time
03:53
<cabanier1>
and it won't be live
03:53
<cabanier1>
yes
03:54
<cabanier1>
do you think that needs to be written down somehow?
03:55
<Hixie>
well it's unambiguously required by the webidl spec
03:55
<Hixie>
and it seems like a really esoteric detail to put in non-normative prose
03:56
<cabanier1>
I agree
03:57
<cabanier1>
I tried "ctx.fillStyle={ toString: function () { return 'blue' } };"
03:57
<cabanier1>
and the fillstyle is black
04:02
<cabanier1>
same on FireFox, Safari and Chrome...
04:04
<cabanier1>
If you have a moment, I think "If x0 = x1 and y0 = y1, then the linear gradient must paint nothing." and "If x0 = x1 and y0 = y1 and r0 = r1, then the radial gradient must paint nothing." in the spec should change.
04:04
<cabanier1>
The gradient should just become a line or a filled circle...
04:05
<cabanier1>
same with color stops at the same offsets
04:08
<Hixie>
file a bug :-)
04:08
<Hixie>
(or mail the list, either way it ends up on my radar)
04:10
<Hixie>
(whatwg.org/newbug and whatwg⊙wo respectively)
05:52
<cabanier1>
OK
05:52
<cabanier1>
I will file a bug on both issues :-P
08:07
<Calthropstu>
hi
08:09
<Calthropstu>
I have a question... if I use say html4 or xhtml doc type statements will I still be able to use html5 tags?
08:11
<deane>
Calthropstu: Hi. Yes you can AFAIK, but, it wont validate.
08:13
<deane>
Calthropstu: Just use <!DOCTYPE html> Why would you want to use old doctypes?
08:13
<Calthropstu>
Im dont
08:13
<Calthropstu>
*I dont
08:14
<matjas>
annevk (assuming you read the logs): could you perhaps chime in here? https://bugs.webkit.org/show_bug.cgi?id=17352
08:15
<Calthropstu>
I'm being considered for a position and when I looked at their work I noticed their website is using a mishmash of xhtml1.0 strict, xhtml 1.1 strict, and html4 transitional
08:15
<Calthropstu>
and the biggest requirment is html 5
08:16
<deane>
Calthropstu: OIC. Sorry to jump to the wrong conclusion.
08:16
<Ms2ger>
Calthropstu, so, doctypes basically don't do anything
08:16
<Calthropstu>
they do
08:16
<Ms2ger>
Calthropstu, except for standards/quirks mode
08:16
<Ms2ger>
And some more modes in IE
08:17
<Ms2ger>
To answer your question, yes, you can use new HTML features regardless of doctype
08:17
<Calthropstu>
well I do know if you dont have 1 at all the whole thing goes nuts
08:18
<Ms2ger>
Right
08:18
<Ms2ger>
Only in IE, though; other browsers will still support those features
08:18
<Calthropstu>
not exactly... because it also starts treating css wrong
08:19
<deane>
A good resource -> http://hsivonen.iki.fi/doctype/
08:19
<Calthropstu>
Ive had that issue as recently as may
08:19
<Ms2ger>
Calthropstu, I'd advise to just search&replace their doctypes, it's not going to break anything
08:20
<Ms2ger>
Unless they rely on quirks-mode differences
08:20
<Calthropstu>
well I would if I already had the position :-)
08:20
<Ms2ger>
Any of the doctypes here: http://www.whatwg.org/html/#obsolete-permitted-doctype-string are fine too, fwiw
08:21
<Calthropstu>
ok
08:48
<deane>
Calthropstu: Good luck with the job.
08:59
<Calthropstu>
thx
09:00
<Calthropstu>
my first resume submittal did not specify html5 so I have to resubmit it on tuesday
09:01
<Calthropstu>
I just figured "html" would cover ALL forms of html and CSS would cover all forms of CSS but they specifically demand CSS3 and HTML5
09:01
<deane>
Yeah, I guess then it's best to be specific.
11:12
<jesusruiz>
hi, group
18:11
<gsnedders>
Anyone got tests for meta pre-parsing?
18:12
<Ms2ger>
Not me
18:12
<gsnedders>
Looks like html5lib has none. And it's totally broken in Py3.
18:12
<gsnedders>
hsivonen: I think I saw some from you?
18:12
<gsnedders>
hsivonen: (meta pre-parsing tests)
19:44
<zewt>
... the synchronous nature of javascript is part of the web platform's *use* of javascript, not javascript/ecmascript per se, right?
19:45
<zewt>
(nothing in the language seems to care one way or the other, but I'm not familiar with the spec itself)
20:24
<gsnedders>
zewt: asynchronous you mean?
20:24
<gsnedders>
Then no, it has nothing to do with JS. It's just how all the host object APIs for the DOM have developed with callbacks.
20:25
<gsnedders>
Probably in large part because most JS impls (i.e., all but Opera's) block everything, so synchronous APIs would be Rather Bad.
20:30
<zewt>
gsnedders: synchronous in the sense of not seeing asynchronous side-effects
20:30
<zewt>
(but yes, we're talking about the same thing)
20:30
<gsnedders>
Ah, right.
20:30
<gsnedders>
That's just a side-effect of impls, really.
20:31
<gsnedders>
The JS spec doesn't address multiple threads of impl at all.
20:33
<zewt>
yeah, that's what I've always assumed intuitively, was just looking for a sanity check before it came up in a debate :)
20:34
<zewt>
is anyone else driven mad by chrome's broken text searching, heh
20:35
<zewt>
i type "foo" and instead of finding the next "foo" from where I am now, it goes to the first "f" in the document before I can type the rest
20:38
<zewt>
particularly annoying in the html spec when it makes me lose my place
21:08
<smaug____>
is rick waldron ever here?
21:25
<zewt>
it's pretty (something) that anyone would seriously suggest this atrocity as a solution to anything: https://github.com/kriskowal/q/blob/master/examples/all.js