07:05
<matjas>
“Like all of the WHATWG specs, it initially looks like the aftermath of a cluster bomb in a scrabble factory, but once you’ve read it for the 5th time and wiped the blood from your eyes, it’s actually pretty interesting” — http://www.html5rocks.com/en/tutorials/speed/script-loading/
09:31
<zcorpan>
so webkit/blink allow insertRule() for cross-origin style sheets but gecko/presto don't
09:33
<zcorpan>
ie10 allows it as well
10:03
<Ms2ger>
MikeSmit1, ping
10:05
<MikeSmit1>
yup
10:06
<Ms2ger>
Something fun about the list archives
10:06
<Ms2ger>
I hit [ respond to this message ] on http://lists.w3.org/Archives/Public/public-webappsec-testsuite/2013Jun/0001.html
10:07
MikeSmith
looks
10:07
<jgraham>
MikeSmith: (btw, while you are here, were you going to send Rebecca a list of tests for TestTWF Tokyo?)
10:07
<jgraham>
*test areas
10:07
<Ms2ger>
But the In reply to link at http://lists.w3.org/Archives/Public/public-test-infra/2013AprJun/0095.html managed to end up broken
10:07
<Ms2ger>
It links to http://www.w3.org/mid/%3C370C9BEB4DD6154FA963E2F79ADC6F2E27AEAD7F⊙Dcec%3E
10:07
<Ms2ger>
Where it should be http://www.w3.org/mid/370C9BEB4DD6154FA963E2F79ADC6F2E27AEAD7F⊙Dcec (without the <>)
10:08
<MikeSmith>
jgraham: didn't yet but if you have time, can you?
10:08
<MikeSmith>
I'm on my mobile
10:08
<jgraham>
MikeSmith: I guess I can try and do that. I don't remember what we decided though
10:08
<Ms2ger>
Logs are your friend :)
10:09
<jgraham>
Indeed :)
10:09
<Ms2ger>
And hallvors had some stuff too in #webapps
10:09
<MikeSmith>
w didn't decide -- just bounced a few ideas
10:09
<MikeSmith>
hallvors mentioned xhr tests
10:10
<Ms2ger>
(Somehow I suspect none of them will want to review the webappsec PRs)
10:11
<MikeSmith>
Ms2ger: no clue on that archive weirdness. I'll alert the systems team
10:11
<hallvors>
I sent Rebecca an E-mail and she replied.
10:11
<Ms2ger>
MikeSmith, ta
10:12
<Ms2ger>
I asked her about box-sizing on IRC; she wanted to hear what fantasai thought, but fantasai didn't want to think
10:12
<odinho>
WAT!
10:12
<hallvors>
They have a list of stuff to focus on (some of it especially related to Japanese issues, which seems like a good idea)
10:12
<odinho>
I wanted help with CORS tests :P -- And EventSource-fix for mozilla.
10:12
<Ms2ger>
hallvors, marquee? ;)
10:12
<jgraham>
hallvors: You already sent an XHR list?
10:13
<jgraham>
https://etherpad.mozilla.org/VkkW1BU54C
10:13
<jgraham>
Did we decide form elements?
10:13
<hallvors>
I just point out the list I already put in a comment on https://github.com/w3c/web-platform-tests/pull/128
10:18
<jgraham>
I feel like suggesting applet is too cruel
10:25
<jgraham>
hallvors: Do you have any recollection of whether the script scheduling tests tested XML?
10:25
<jgraham>
http://www.whatwg.org/specs/web-apps/current-work/multipage/the-xhtml-syntax.html#parsing-xhtml-documents "When an XML parser"
10:27
<Ms2ger>
You were talking cruel?
10:27
jgraham
is sad that those are still under old-tests
10:27
<jgraham>
Well it's less bad than the HTML case :)
10:30
<jgraham>
The amount of stuff still under old-tests/submissions/ is sadness
10:30
<Ms2ger>
Getting Aryeh's reflection tests merged would be nice too...
10:33
<Ms2ger>
Lovely: https://www.w3.org/Bugs/Public/show_bug.cgi?id=22291
10:34
jgraham
is glad that bz doesn't want to be a supervillan
10:35
<jgraham>
Also, getting people to address issues on existing pull requests/reviews is a good idea
10:40
<Ms2ger>
Is there any way to get a diff that doesn't cut off all the text in https://github.com/dontcallmedom/web-platform-tests/commit/1179ab58108d5fb7e0eb2ef55fe90b2401be9cb3 ?
10:44
<odinho>
12:37 < jgraham> Also, getting people to address issues on existing pull requests/reviews is a good idea <<--- this what I wanted with the CORS tests
11:11
<jgraham>
Hmm, there are some nicy easy looking reviews for anyone that knows CSP
11:15
<jgraham>
I wonder if we have a policy about UA-specific cruft in tests
11:16
<jgraham>
abarth has submitted CSP tests with X-WebKit headers in
11:18
<Ms2ger>
I believe the policy is "no"
11:18
<jgraham>
So do I
11:18
<jgraham>
Not sure we have it written down anywhere though
11:19
<jochen__>
that might just an oversight
11:19
<jochen__>
given that blink wants to go prefix-less nowadays
11:30
<hsivonen>
Are there now X-WebKit headers???
11:30
<hsivonen>
or just did you just mean WebKit-specific X-Foo?
11:32
<jgraham>
Well
11:33
<jgraham>
The test has Content-Security-Policy, X-Content-Security-Policy and X-WebKit-Content-Security-Policy
11:33
<jgraham>
Oh, sorry
11:33
<hsivonen>
sadness
11:33
<jgraham>
X-WebKit-CSP
12:16
<hallvors>
jgraham: sorry, didn't see question "Do you have any recollection of whether the script scheduling tests tested XML?" before
12:16
<hallvors>
but I'm pretty sure the answer is "no, they don't"
12:37
<hsivonen>
what should my schedule expactiations be for the CSS WG processing http://lists.w3.org/Archives/Public/public-css-testsuite/2012Dec/0000.html and publishing a revised version of the CSS 2.1 test suite?
12:37
<hsivonen>
*expectations
12:37
<hsivonen>
it's harmful to have outdated test suites out there
12:41
<Ms2ger>
TabAtkins, ^
12:42
<hsivonen>
some more UTF-16 uselessness: https://bugzilla.mozilla.org/show_bug.cgi?id=879753
13:52
<GPHemsley>
Safari is the most frustrating browser when it comes to unsupported MIME types...
13:53
<GPHemsley>
it just downloads everything it doesn't recognize automatically
13:53
<GPHemsley>
so now I have a whole downloads folder of Safari cruft
13:53
<GPHemsley>
all without my consent
13:55
<zewt>
asking for user consent seems to be a dying idea
13:55
<zewt>
remember when things would actually ask permission before updating? yeah i can hardly remember that myself
14:01
<GPHemsley>
oh cynicism
14:16
<zewt>
skype nagged me to update (to a version I knew had an unusable UI) over and over, then it just updated without asking
14:18
<GPHemsley>
ah
14:18
<GPHemsley>
but I'm more concerned about Safari downloading something that
14:18
<GPHemsley>
's potentially unsafe
16:06
<GPHemsley>
Hixie_: Oh, one reason that the context change is more than editorial is because it should refer directly to mimesniff
16:07
<GPHemsley>
http://mimesniff.spec.whatwg.org/#context-specific-sniffing
16:07
<GPHemsley>
(I'm fleshing these out)
16:12
<Hixie_>
what difference will it require in implementations?
16:13
<GPHemsley>
assuming the implementations are not already interoperable, it will require them to become so
16:13
<GPHemsley>
in terms of when/what/how to sniff in each context
16:13
<GPHemsley>
because the contexts do not all have the same rules
16:13
<GPHemsley>
(or won't, when they're actually written)
16:13
<Hixie_>
doesn't HTML already explicitly say which rules to use when?
16:16
<GPHemsley>
only for certain contexts
16:16
<GPHemsley>
not for others
16:16
<GPHemsley>
but I still concede this is still ~90% editorial
16:17
<Hixie_>
which doesn't it specify it for?
16:19
<GPHemsley>
I think browsing, image, and audio/video specified
16:19
<GPHemsley>
but the rest aren't, AFAIK
16:19
<Hixie_>
do any of the rest do any sniffing at all?
16:20
<GPHemsley>
perhaps not yet, but they should—these are edge cases that perhaps people haven't considered yet
16:20
<Hixie_>
i'm pretty sure i've considered them :-)
16:21
<GPHemsley>
there are a number of issues on file with Mozilla about how to sniff in the style context that I'm tackling now
16:22
<Hixie_>
style context?
16:22
<GPHemsley>
see, you have no definition ;)
16:23
<GPHemsley>
the "style context" is whenever a stylesheet is loaded
16:23
<GPHemsley>
in HTML, that's mostly <link rel=stylesheet>
16:23
<Hixie_>
that shouldn't sniff at all
16:23
<GPHemsley>
in CSS, it's @import
16:23
<GPHemsley>
what happens if the resource is served without a Content-Type header?
16:23
<Hixie_>
treat as text/css
16:24
<GPHemsley>
that has security implications
16:24
<GPHemsley>
if an untagged HTML file is treated as text/css, unexpected things can happen
16:24
<Hixie_>
like what?
16:25
<Hixie_>
you can't read the DOM unless it's same origin
16:25
<GPHemsley>
https://bugzilla.mozilla.org/show_bug.cgi?id=560388
16:25
<GPHemsley>
https://bugzilla.mozilla.org/show_bug.cgi?id=562377
16:26
<Hixie_>
that first one seems to apply to all "contexts", it's just about distinguishing bogus from missing
16:27
<Hixie_>
that second one seems to just say that i was wrong about the no-type case
16:27
<Hixie_>
it should be treated as bogus
16:28
<GPHemsley>
sorry, was trying to have two conversations at once
16:29
<GPHemsley>
apparently that's hard
16:29
<GPHemsley>
anyway
16:29
<GPHemsley>
the bottom line with 562377 was that the style context can't use the same sniffing rules as the browsing context
16:30
<GPHemsley>
err
16:30
<GPHemsley>
560388
16:30
<Hixie_>
seems like the conclusion is that it shouldn't sniff at all
16:30
<Hixie_>
it definitely shouldn't use the same sniffing rules as the browsing context, nobody is suggesting it should, as far as i can tell
16:30
<Hixie_>
certainly not the spec, right?
16:31
<GPHemsley>
well, that's the part I'm trying to clarify
16:31
<GPHemsley>
with these contexts
16:31
<GPHemsley>
but no, I don't think HTML says it should
16:31
<GPHemsley>
mimesniff used to seem that way, though
16:31
<GPHemsley>
which is the part I'm working on now
16:35
<Hixie_>
i think if mimesniff just provides algorithms, and doesn't invoke any of them itself, it's quite sufficient
16:39
<GPHemsley>
Hixie_: I'm not sure what you mean; what is it currently invoking (or what do you think I want it to be invoking)?
16:39
<Hixie_>
i'm not saying it is or will
16:39
<Hixie_>
was just being complete in my description
16:39
<GPHemsley>
ah
16:39
<GPHemsley>
well, yeah, then I agree
16:40
<Hixie_>
this context stuff seems like a very convoluted way of doing it :-)
16:40
<GPHemsley>
really? convoluted how? I thought it makes things clearer
16:40
<GPHemsley>
(and annevk seemed to like it)
16:41
<Hixie_>
"run sniffing algorithm #3 for resource /x/, to obtain sniffed type /y/" seems nice and trivial and doesn't involve having to define contexts
16:41
<GPHemsley>
if you always fetch through the lens of a context, you never have confusion as to which set of rules to follow
16:41
<GPHemsley>
I mean, it's just a name
16:41
<Hixie_>
if you always just call the sniffing algorithm directly there's no confusion either :-)
16:42
<Hixie_>
what is confused today?
16:42
<GPHemsley>
this modularizes the sniffing algorithm so that you don't have one big complex thing
16:42
<Hixie_>
we don't have one big complex thing today either do we?
16:42
<GPHemsley>
like I said, prior to the context stuff I added today, it seemed like the MIME sniffing algorithm should be used everywhere
16:43
<GPHemsley>
which, as written, is/was not appropriate
16:43
<Hixie_>
oh well i'm fine with adding non-normative text that says "don't use this unless you have a legacy compat reason to do so"
16:43
<Hixie_>
i kinda assumed we already had that
16:43
<GPHemsley>
I think a lot of this issue is bogged down by legacy terminology
16:44
<GPHemsley>
in that "sniffing" is a bit of a misnomer for my current vision of this spec
16:44
<Hixie_>
(re "anne likes it" -- anne has a habit of making these same kinds of "simplifications" that require tons of changes to html too :-P)
16:44
<GPHemsley>
my current vision of this spec is to describe how to determine the MIME type of a file—whether through sniffing or otherwise
16:45
<Hixie_>
the mime type of a file without sniffing is trivial
16:45
<Hixie_>
it's the value of the Content-Type header
16:46
<GPHemsley>
it's not always that simple
16:46
<GPHemsley>
but I feel like we could argue this around in circles
16:47
<GPHemsley>
this is mostly about terminological homogenization
16:48
<GPHemsley>
Hixie_: Did you see this? http://lists.w3.org/Archives/Public/public-webappsec/2013Jun/0027.html
16:49
<GPHemsley>
also, the "How to use a context" set of steps on the wiki page?
16:49
<Hixie_>
i don't object to stuff like this in general
16:49
<GPHemsley>
I can't tell if you're objecting to the idea, or just the cost-benefit ratio of adding it to the HTML spec :P
16:50
<Hixie_>
i just object to it while we have higher-priority stuff missing or broken
16:50
<Hixie_>
the latter
16:50
<Hixie_>
i'm not updating the HTML spec for this kind of stuff any time soon (like, coming year, probably)
16:50
<Hixie_>
i've got a ton of higher-priority stuff
16:50
<GPHemsley>
that's fine
16:50
<Hixie_>
same with anne's fetch spec, that's why he has the "legacy" algorithms
16:51
<Hixie_>
or as i call them, "the algorithms" :-P
16:51
<GPHemsley>
heh
16:52
<Hixie_>
btw if you want something more important to do, the editing spec is in dire need of an editor
16:52
<Hixie_>
since aryeh has had to move on to other things
16:52
<GPHemsley>
link?
16:52
<Ms2ger>
https://dvcs.w3.org/hg/editing
16:52
<Hixie_>
...ms2ger beat me to it
16:52
<GPHemsley>
ugh, you want me to work with the W3C? >_>
16:53
<Ms2ger>
It's a CG
16:53
<Hixie_>
it's in a cg, like html
16:53
<Hixie_>
if you want to merge it with the whatwg cg that's fine by me
16:53
<GPHemsley>
I have to admit, I've remained relatively ignorant of the bureaucracy over there
16:54
<Hixie_>
there's basically none for cgs
16:54
<Hixie_>
if you did want to work on that spec and wanted to skip all bureaucracy, you could just take it and make it a whatwg spec the same way as the mime sniffing spec
16:54
<Hixie_>
since it's an open license spec
16:55
<Hixie_>
and the old version is not maintained
16:56
<GPHemsley>
FFR, this was more the type of I link I was interested in: https://dvcs.w3.org/hg/editing/raw-file/tip/editing.html
16:56
<GPHemsley>
;)
16:56
<GPHemsley>
but apparently the styling is broken
16:56
<Hixie_>
oh i thought that's what Ms2ger gave, my bad :-)
16:59
<GPHemsley>
does the whatwg stylesheet prevent cross-site loading?
16:59
<Hixie_>
not in theory, but there might be cross-protocol blocking going on
16:59
<Hixie_>
whatwg is not https:
17:00
<GPHemsley>
ah
17:00
<GPHemsley>
indeed, Aurora is giving me a half shield
17:01
<GPHemsley>
there we go
17:01
<GPHemsley>
"This is an open issue. All issues other than this one are non-normative. "
17:01
<GPHemsley>
somewhat paradoxical
17:04
<GPHemsley>
oh, this would take me ages to understand...
17:04
<Hixie_>
that's how you know it's important :-P
17:04
<GPHemsley>
I've never even *used* HTML editing
17:06
<GPHemsley>
you'd _definitely_ have to pay me to work on this :P
17:06
<Hixie_>
hah
17:09
<Ms2ger>
That's how he sucked in Aryeh, and we know how that ended...
17:12
<GPHemsley>
I mean, I'm generous with my time, but I'm not *this* generous ;)
17:12
<Hixie_>
:-)
17:13
<GPHemsley>
I do have bills that I need paid
20:14
<GPHemsley>
It's always fun to see which sites are parsing Accept-Language incorrectly
20:15
<GPHemsley>
and by "fun" I mean "sad"
20:21
<Hixie_>
isn't it harder to find some parsing it right? :-)
20:32
<GPHemsley>
probably... but there's a clear distinction between parsing it wrong and not parsing it right, and my particular Accept-Language settings bring that to light
22:47
<Hixie_>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=21964 ??
22:49
<miketaylr>
O_o
22:54
<Hixie_>
heycam|away: ping
22:55
<Hixie_>
heycam|away: ping https://www.w3.org/Bugs/Public/show_bug.cgi?id=21946
23:16
<zewt>
stream of consciousness bugs
23:22
<zewt>
what the heck does "font security" even mean? are people encoding passwords as character widths in fonts? heh
23:23
<zewt>
seems like a bizarre pair of words
23:30
<Hixie_>
zewt: i guess you could get Roboto 3.0's font metrics out of Google before Google announced it, or something
23:30
<Hixie_>
(i've no idea if Roboto has a 3.0 or if it's even a Google-made font)
23:37
GPHemsley
wishes Hixie_ would clarify his commit messages a bit
23:39
<Hixie_>
which ones?
23:39
<Hixie_>
most of hte patches seems self-explanatory
23:40
<Hixie_>
r7941's was a mistake (s/Example/Allow/)
23:40
<Hixie_>
but for the editorial ones...
23:40
<Hixie_>
they're fixing editorial problems
23:40
<GPHemsley>
the patches come after the click ;)
23:40
<Hixie_>
what more can i say
23:41
<GPHemsley>
"clarification" and "match reality" are rather non-descript when all you see is the commit message
23:42
<GPHemsley>
just saying
23:42
<Hixie_>
what would you suggest instead?
23:42
<Hixie_>
i'm open to making it better, but i'm not sure how to
23:42
<Hixie_>
short of just putting the diff in the commit message...
23:43
<GPHemsley>
just add a few words about which small part of the humongous spec you're changing ;)
23:43
<GPHemsley>
keep in mind that many people (or just me?) follow along via Twitter
23:43
<GPHemsley>
you gotta click the link to see the patch
23:44
<GPHemsley>
"clarify X" or "make Y match reality" would be a major improvement
23:44
<Hixie_>
i don't know what section it is half the time
23:44
<GPHemsley>
heh
23:44
<Hixie_>
but i'll try to include a vague idea of what topic it might be :-)
23:45
<Hixie_>
at least for the non-editorial or "clarification" ones
23:45
<GPHemsley>
much appreciated
23:47
<GPHemsley>
because otherwise it turns out like this: http://imgur.com/IU2z3bz
23:49
<Hixie_>
yeah, well, given the kind of crap i've been fixing today, that's about right :-)
23:52
<GPHemsley>
:P