06:51
<AryehGregor>
"It almost seems to me that nobody cares over here what the W3C document actually says, as there is that other "more helpful" version. In which case I wonder why it's published at all?"
06:52
<AryehGregor>
. . . I guess it's probably best not to respond to that.
06:53
AryehGregor
cries when he reads stuff like http://lists.w3.org/Archives/Public/www-style/2012Jul/0267.html that adds ever more declarative special cases instead of writing imperatively
06:54
<annevk>
you mean the calc() feature itself?
06:54
<AryehGregor>
I mean basically all of CSS.
06:55
<AryehGregor>
Which currently consists of lots of declarative requirements scattered across a zillion specs with all interactions undefined unless someone specifically points them out.
06:55
<annevk>
I saw a somewhat plausible answer to why adding imperative stuff to CSS would be problematic
06:55
<annevk>
ooh
06:55
<annevk>
you mean spec-wise
06:55
<AryehGregor>
Right.
06:55
<annevk>
yeah dunno about that
06:55
<AryehGregor>
For that matter, a lot of the terminology is undefined too -- specs will refer to things like "the element's border box" when elements can have any number of border boxes from zero and up.
06:56
<annevk>
uhuh
06:56
<annevk>
it's kind of crazy really
06:56
<AryehGregor>
Clearly it's harder to write imperatively than DOM stuff, because JavaScript is imperative and CSS is declarative.
06:56
<AryehGregor>
But still.
06:56
<annevk>
it's hard to write because CSS is hard
06:56
<annevk>
so instead of defining CSS they describe it which is a little less hard
06:56
<annevk>
but then you run into problems soon enough
06:57
<AryehGregor>
I don't think CSS is inherently any harder than, say, the HTML parser, or editing.
06:57
<AryehGregor>
It's a lot harder than DOM Core, yes, but that's not saying much.
06:58
<annevk>
dunno, CSS is big
06:58
<AryehGregor>
Bigger than HTML?
06:58
<AryehGregor>
One person did that. :)
06:58
<annevk>
I think CSS is bigger and definitely more complex
06:59
<annevk>
but I have hopes of someone fixing it still :)
06:59
<AryehGregor>
Really? The WHATWG spec is a lot bigger than all CSS specs combined.
06:59
<AryehGregor>
Although that's partly because it actually specifies all details, obviously.
06:59
<annevk>
all CSS specs?
06:59
<annevk>
including the object model and such?
07:00
<AryehGregor>
I'm pretty sure.
07:00
<AryehGregor>
I think I once counted using wget and wc -l.
07:00
<AryehGregor>
Retrieving all the CSS specs from the Current Work page.
07:00
<AryehGregor>
CSS2.1 is the only one of appreciable size. (Guess what: co-edited by Hixie.)
07:00
<annevk>
CSS 2.1 alone is half of HTML
07:01
<AryehGregor>
Really?
07:01
<annevk>
I reckon if you rewrite that proper style it will be bigger
07:01
<annevk>
the PDF is 487 pages
07:02
<zcorpan>
AryehGregor: annevk: did you see http://krijnhoetmer.nl/irc-logs/whatwg/20120711#l-533 ?
07:02
<annevk>
and CSS 2.1 does not include the object model and such
07:02
<AryehGregor>
zcorpan, no, what does it do?
07:03
<zcorpan>
AryehGregor: checks http://www.w3.org/community/reports/reqs/
07:04
<annevk>
zcorpan: that's kind of neat, maybe get ij to do something with it?
07:07
<zcorpan>
i'm not sure a bookmarklet is the best way to implement it, maybe something that can be run from the command line is better so editors can run it at the same time as anolis
07:08
<zcorpan>
i checked yesterday and it finds some errors in Fullscreen and HTML Editing APIs
07:08
<annevk>
ideally it's some kind of service that just emails you the violations
07:09
<zcorpan>
or files a bug maybe?
07:09
<annevk>
yeah
07:09
<annevk>
would be cool for the markup validator too
07:10
<zcorpan>
hsivonen: ^
07:10
<MikeSmith>
what would be cool for the validator?
07:10
<zcorpan>
to email or file bug when your page doesn't validate anymore
07:11
<MikeSmith>
oh
07:13
<zcorpan>
but i guess that's possible to implement with v.nu's existing api
07:13
<zcorpan>
with a simple script
07:13
<zcorpan>
just file-upload to v.nu after running anolis and if it gives errors, file a bug
07:35
<hsivonen>
zcorpan: you want the validator to file bugs in the W3C bugzilla?
07:36
<zcorpan>
hsivonen: yeah. except i realized that's possible already without more code in v.nu
07:37
<hsivonen>
I see
08:07
<annevk>
MikeSmith: zcorpan: if you want to allow indexing of W3C Bugzilla https://www.w3.org/robots.txt will need to be modified (note https)
08:08
<MikeSmith>
ah yeah
08:38
<hsivonen>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=17693 someone missed the memo on <ins> and <del> (which became synonyms for <u> and <s> since it's really hard to do what Word does)
09:04
<smaug____>
Uh, http://dev.w3.org/2006/webapi/FileAPI/#dfn-autoRevoke :/
09:05
<smaug____>
no arun here
09:32
<zcorpan>
smaug____: file a bug?
09:33
<smaug____>
perhaps I should discuss with arun first
09:33
<smaug____>
though, I could file a bug to remove "stable state" concept from HTML :)
09:35
<zcorpan>
what's wrong with stable state?
09:39
<smaug____>
zcorpan: it is a weird concept
09:39
<smaug____>
it is not really detectable from JS
09:39
<zcorpan>
what do you suggest instead for algorithms that currently use it?
09:39
<zcorpan>
queue a task?
09:40
<smaug____>
that is often a possible choice
09:40
<smaug____>
There is also microtask
09:41
<smaug____>
microtask was added because stable state and end-of-task are so vague
09:41
<zcorpan>
have a link to microtask?
09:42
<smaug____>
http://www.whatwg.org/specs/web-apps/current-work/#perform-a-microtask-checkpoint
09:42
<smaug____>
for some odd reason the spec talks only about mutation observers
09:44
<zcorpan>
thanks
09:55
<annevk>
is that revoking language even correct?
09:56
<annevk>
you only want it to be revoked after you have used the URL somehow
09:56
<annevk>
that file API spec is turning into a mess
09:59
<smaug____>
why should it say anything about using the URL ?
10:03
<annevk>
if you just create it, then wait a while before using it (waiting on some event) it would already be revoked
10:03
<annevk>
that seems kind of pointless
10:05
<zcorpan>
yeah the autorevoke thing only allows the url to be referenced in the same script. and many things that fetch urls are themselves async, like e.g. <img src>
10:15
<zcorpan>
should websocket binary messages allow scripts to read the message in a streaming fashion while it's being received, like with xhr?
10:15
<zcorpan>
right now the message event is only fired once the whole message has been received, afaict
10:17
<annevk>
pretty much nothing in XHR is streaming fwiw
10:17
<annevk>
only responseText at the moment
10:17
<annevk>
maybe also responseType = "text"
10:35
<zcorpan>
annevk: can't you get the partial data as blob or arraybuffer?
10:36
<annevk>
zcorpan: I don't really see how
10:36
<annevk>
we thought of allowing that at some point
10:37
<annevk>
with "I don't really see how" I mean that the current spec doesn't do it
10:37
<annevk>
see http://dvcs.w3.org/hg/xhr/raw-file/tip/Overview.html#the-response-attribute
10:38
<zcorpan>
ah. i thought it wasn't null while LOADING
10:40
<zcorpan>
wouldn't it work to return a blob early and have it act like a file that loads slowly?
10:41
<zcorpan>
or maybe something different makes more sense for streaming use cases
10:49
<annevk>
yeah that would work
10:49
<annevk>
I think implementors had some concerns
10:50
<annevk>
but we can still make that work in the future
10:52
<annevk>
time to see the outside world
11:06
<Ms2ger>
Pff
11:06
<Ms2ger>
The outside world is overrated
11:23
<hsivonen>
hmm. does the new GWT ingest IDL from Gecko and WebKit or really C++?
11:23
<hsivonen>
if it really eats C++, that would be a failure for IDL
11:25
<smaug____>
hsivonen: link?
11:27
<hsivonen>
smaug____: http://www.youtube.com/watch?v=VOf27ez_Hvg&feature=plcp
11:27
<hsivonen>
(not really a time-efficient reference)
13:52
<zcorpan_>
isn't it awesome when people talk in XML tags on mailing lists and suggest to replace HTML+CSS with XML+XSL?
15:02
<matjas>
hsivonen: re: 404 page on <audio> testbed — http://twitter.com/yvg/status/223346356088094721
15:11
<miketaylr>
matjas: pfffffffffft
15:15
<smaug____>
zewt: an event listener can't know if something else runs before the next stable state
15:15
<smaug____>
(just an example)
15:15
<smaug____>
also, the same task may for example dispatch several events
16:33
<hsivonen>
matjas: thanks
16:58
gsnedders
contemplates starting a new cross-browser JS testsuite, again
17:02
<Ms2ger>
gsnedders, at least join us at the W3C :)
17:25
<sedovsek>
Hey, Is there any alternative for document.webkitHidden on iOs and Android devices?
17:25
<Hixie>
what does it do?
17:30
<sedovsek>
https://developers.google.com/chrome/whitepapers/pagevisibility
17:30
<sedovsek>
Here are some examples.
18:13
<matjas>
sedovsek (if you read the logs): http://mths.be/visibility demonstrates how you can fall back to focus/blur
18:16
<bencc>
are svg animations standard in html5?
18:17
<shepazu>
bencc: technically, because SVG animations (SMIL) are part of SVG 1.1, which is the spec that HTML5 references, they are "standard"… but IE doesn't support them, and HTML does strictly require full support for SVG 1.1
18:19
<bencc>
shepazu: it did support it in the past I think
18:19
<shepazu>
bencc: nope
18:19
<bencc>
shepazu: do you know if there are plans to support SMIL in IE10 or later?
18:19
<shepazu>
no plans for IE10
18:20
<bencc>
shepazu: that's sad :(
18:20
<shepazu>
I agree
18:20
<shepazu>
they are focusing on CSS animations, which you should be able to use with SVG
18:20
<shepazu>
I think that's all still being worked out, though
18:21
<bencc>
shepazu: ok
18:21
<bencc>
shepazu: can I do feature detection for beginElement or animate?
18:21
<shepazu>
bencc: you should think about using FakeSMIL
18:22
<shepazu>
it's a script lib shim that handles most of SMIL in SVG for IE
18:22
<bencc>
cool. this? http://fakesmil.blogspot.co.il/
18:22
<Hixie>
wow, not counting the responsive images threads and not counting e-mails about adding new features, i'm down to just about 250 e-mails!
18:22
<shepazu>
yeah
18:22
<Hixie>
(including those two categories, 1200 e-mails.)
18:23
<Hixie>
(also including bugs: 1700 e-mails.)
18:23
<shepazu>
bencc: it's a bit old, but it still works
18:24
<shepazu>
it's not perfect, but it does the job for most things
18:26
<bencc>
shepazu: I've just tried it and I still get the error
18:26
<shepazu>
bencc: what error?
18:26
<bencc>
shepazu: I'm adding the animate element dynamically so I'm not sure if the script cover this use case
18:26
<shepazu>
it should
18:26
<bencc>
Object doesn't support property or method 'beginElement'
18:28
<shepazu>
bencc: yeah, it's not going to add that support if you're doing feature detection
18:28
<shepazu>
bencc: you have a jsfiddle?
18:29
<bencc>
shepazu: not yet http://dpaste.com/769928/
18:29
<bencc>
shepazu: what do you mean? I'm not doing feature detection
18:29
<shepazu>
ok
18:29
<bencc>
shepazu: I just included the script
18:30
<shepazu>
bencc: that snippet doesn't include fakesmil...
18:30
<bencc>
shepazu: I thought it should just work
18:31
<shepazu>
bencc: I can't debug if I don't have an example that should work...
18:31
<bencc>
shepazu: sure, I'll make one
18:38
<bencc>
shepazu: this is without smil http://jsfiddle.net/DgMDV/1/
18:38
<bencc>
shepazu: works in FF but not in IE9
18:38
<bencc>
shepazu: now trying to add smil but I'm not sure how to use it
18:38
<bencc>
I mean fakesmil
18:39
<shepazu>
http://leunen.me/fakesmile/faq.html#dev
18:43
<bencc>
shepazu: I've added the script to the jsfiddle but there must be something more I need to do
18:43
<shepazu>
bencc: let me make a simple example
18:43
<shepazu>
it might take me a few minutes
18:44
<bencc>
thanks!
18:48
<Hixie>
hmm, i wonder if there's a way to make bugzilla search results more computer-readable
18:53
<bencc>
shepazu: looks like there is a more recent version here: http://bazaar.launchpad.net/~smilteam/smil/MAIN/view/head:/smil.user.js
19:00
<shepazu>
bencc: I just realized I can't test it… I'm on a mac, and I don't have a browser that doesn't support SMIL :(
19:00
<bencc>
shepazu: I can
19:00
<bencc>
the example will be useful in any case
19:07
<bencc>
shepazu: I'm trying to add it with <svg ...><script type="text/ecmascript" xlink:href="smil.user.js"/>...</svg> but I'm not sure if it is loaded
19:09
<MikeSmith>
bencc: you serving it with an XML mime type?
19:10
<bencc>
MikeSmith: no
19:11
<bencc>
MikeSmith: I don't know how to use it
19:12
<MikeSmith>
bencc: I think what you're trying to do won't work unless you serve it with an XML mime type
19:12
<MikeSmith>
but I dunno really
19:12
<MikeSmith>
shepazu knows better than me
19:13
<bencc>
ok
19:17
<shepazu>
bencc: my suggestion is to google uses of fakesmil, and find an example that works for you
19:18
<bencc>
shepazu: I tried but can't find how to dynamically include it
20:21
<MikeSmith>
hmm JSON-LD
22:18
<loincloth>
what what
23:06
<zewt>
cool, "new" paypal page broke the login form in some weird way
23:08
<zewt>
haha and they use a position: absolute block hovering on top of the input form instead of @placeholder, and don't bother turning off pointer-events, so when i click the input, it randomly selects the placeholder instead of the input
23:08
<zewt>
it's always amazing when the highest-profile pages on the internet screw up so badly
23:09
<TabAtkins_>
yay!
23:10
<hober>
s/amazing/sadly unsurprising/
23:10
<zewt>
i was trying to be a *little* generous
23:11
<zewt>
okay I guess I wan't :)
23:11
<zewt>
also wasn't
23:13
<zewt>
(somehow having a simple login form break autocompletion twice and then select the word "address" three times makes me feel ungenerous)