07:18
<zcorpan>
huh, where does the url snapshot hide part of the title? it's not hidden for me
07:20
<zcorpan>
hahahaha whatwg isn't open enough
07:22
<zcorpan>
oh the hiding was reverted, ok
07:54
<zcorpan>
hsivonen: we could also make innerHTML only be available on html/svg/mathml elements
07:54
<zcorpan>
hsivonen: although mathml just uses Element, doesn't it?
07:56
<hsivonen>
zcorpan: I don't recall what MathML uses
07:57
<zcorpan>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/3211 - Element
07:57
<hsivonen>
oh. fun times. createContextualFragment supports only HTML contexts
07:58
<zcorpan>
could make setting innerHTML throw when context isn't html/svg/mathml. but i think #2 is better
09:57
<annevk>
hsivonen: did you read http://queue.acm.org/detail.cfm?ref=rss&id=2673311
09:58
<annevk>
"The E.U. Parliament seems to have been successfully captured by CA lobbying efforts."
10:25
<hsivonen>
annevk: I haven't read that
10:25
<hsivonen>
thanks
10:28
<hsivonen>
annevk: does that article describe something that's about to make things worse? or just a failure of the EU to make things better?
10:31
<MikeSmith>
hard to imagine what could get worse there
10:34
<MikeSmith>
I suspect It's more a thing of Just the CAs managing to get the eu to avoid Doing any useful regulation of the CAs, or hold them accountable for anything
10:35
<zcorpan>
maybe it will be illegal not to use SHA-1
10:47
<annevk>
hsivonen: latter
11:47
<annevk>
https://twitter.com/DanHarper7/status/514822464673951744 o_O
11:51
<SimonSapin>
annevk: Natasha asks if you're coming to TPAC
12:20
<hsivonen>
annevk: hooray. HZ changes landed. I hope they everything is green this time and they stick.
12:20
<annevk>
teehee
12:20
<annevk>
hsivonen: do you know if they are in Chrome final yet?
12:21
<hsivonen>
annevk: I don't
12:21
<hsivonen>
annevk: but isn't 4 months shorter than the time it takes to get to release for Chrome?
12:22
<annevk>
hsivonen: yeah I guess it is, the weeks vs months thing threw me off
12:28
<annevk>
:-( Found an attack where you can use CSP and HSTS to discover whether someone has visited a domain
12:41
<darobin>
annevk: details?
12:47
<Ms2ger>
In the cve? ;)
13:05
<annevk>
darobin: public-webappsec
13:12
<darobin>
annevk: oh, duh, ta
13:12
<annevk>
the number of problematic things with redirects...
13:42
<beverloo>
annevk, reading the thread now
13:42
<annevk>
beverloo: notifications? yay
13:42
<beverloo>
yup
13:42
<annevk>
beverloo: also https://wiki.whatwg.org/wiki/Sharing
13:42
<beverloo>
that's unrelated to notifications, right?
13:43
<beverloo>
I'll read it again, stumbled across the link a few weeks ago
13:43
<beverloo>
mounir will definitely have opinions about it, too
13:45
<annevk>
beverloo: yeah is
14:07
<wanderview>
annevk: I imagine something like JSONx might be nice for gluing modern things to legacy things in the "enterprise"
14:08
<wanderview>
s/nice/useful/g
14:08
<annevk>
wanderview: I'm now curious about your employment history :p
14:08
<wanderview>
annevk: a lot of xml based systems running around out there is all I'm saying
14:09
<wanderview>
whether we like xml or not
14:09
<annevk>
I'm not opposed, I even designed a successor
14:09
<annevk>
And advocated the hell out of XML MIME types, XHTML, DOCTYPEs, and a bunch of other crap that turned out to be mostly irrelevant for the web
14:14
<wanderview>
annevk: the last enterprise-type system I worked on used a mixture of json, xml, and a fixed-width text file format from the 80s :-\
14:14
<annevk>
wanderview: what was the encoding?
14:15
<annevk>
wanderview: and what kind of newlines did it use?
14:15
<wanderview>
ascii I believe? I don't recall the specifics
14:20
<wanderview>
annevk: the text file format was EDI... took me a while to dredge that up... http://en.wikipedia.org/wiki/Electronic_data_interchange
16:23
<sicking>
Domenic: ping
16:23
<sicking>
Domenic: is it possible to .freeze() a Map?
16:25
<SimonSapin>
annevk: http://www.openwebfoundation.org/legal/the-owf-1-0-agreements/owfa-1-0 now redirects to a landing page rather than 1.0 specifically
16:26
<annevk>
SimonSapin: yeah, I've been meaning to ask tantek how to fix that
16:26
<annevk>
sicking: no
16:26
<annevk>
sicking: freeze() only freezes properties
16:26
<annevk>
sicking: it's not a general mechanism
16:26
<sicking>
annevk: can't functions check the frozen-state and honor it?
16:27
<annevk>
sicking: they could, but no built-in object observes it
16:27
<sicking>
annevk: seems like at least getters/setters need that ability
16:27
<sicking>
annevk: bummer
16:27
<sicking>
that seems pretty broken to me :(
16:27
<annevk>
sicking: it's a feature that came out of the Mark Miller ES5 compromise
16:28
<annevk>
sicking: and is disliked by at least half of TC39 which thinks it should never have existed in the first place afaict
16:28
<annevk>
sicking: and therefore nobody pays attention to it
16:28
<sicking>
annevk: freezing in general?
16:28
<annevk>
sicking: Object.freeze()
16:28
<sicking>
sigh
16:28
<annevk>
sicking: it was added for fairly specific Caja use cases
16:28
<sicking>
i guess what we really need is value object
16:28
<annevk>
sicking: but given a rather prominent name
16:29
<annevk>
I haven't seen much progress on those recently, but they do seem neat
16:29
<sicking>
annevk: seems like the ability to use readonly data structures, like arrays and maps, is not Caja specific
16:30
<sicking>
bummer
16:30
<sicking>
oh well, we can always add more DOM APIs to overcome ES shortcomings, that's not new :(
16:30
<annevk>
sicking: the people that don't care about freeze(), including e.g. dherman, think immutable data structures require something else
16:31
<annevk>
sicking: we can also propose new ES features relatively easily these days
16:31
<annevk>
sicking: by making a thought out proposal and then getting someone to bring it up during a TC39 meeting
16:31
<annevk>
sicking: it's still pretty bad compared to other bodies
16:32
<sicking>
annevk: yeah. Though designing readonly data structures is a pretty big undertaking
16:32
<annevk>
sicking: yeah if you want a generic solution
16:32
<annevk>
sicking: dherman and some people have been working on immutable ArrayBuffers I'm told, but I'm not sure how far along that is
16:32
<sicking>
annevk: i don't need it to be more generic than "readonly lists and maps". But that's still pretty generic
16:32
<annevk>
sicking: very little public stuff to follow, not sure why they're so bad at that
16:33
<sicking>
ok
17:03
<Domenic>
Hixie et al: is it completely safe to use < as long as I always follow it with a space?
17:03
<Hixie>
sounds plausible
17:04
<Hixie>
might not be conforming though
17:04
<Domenic>
heh
17:04
<annevk_>
I always escape <, but never >
17:04
<Hixie>
what anne said
17:08
<Domenic>
I'm looking at algorithm steps basically. While k < length is much nicer than while k &lt; length
17:09
<annevk>
Even less English?
17:09
<annevk>
You have prose somewhere that describes it means less than?
17:10
<annevk>
biab
17:14
<Domenic>
um, if http://en.wikipedia.org/wiki/Less-than_sign counts, then yes
17:26
<Sample>
Does the idea or terminology "constrained by implementation" mean that browser vendors have committed features ahead of adherance to spec? I've browsed webkit chromium and mozilla ticket systems and I've seen at times suggestions "we won't implement that because it's not in spec" but other times acceptance and review of patches that seem to cause this "constrained by implementation"
17:27
<Sample>
or is it more commonly that they've implemented adherance to spec in such a specific way that the specification has no leverage to re-define better intentions?
17:27
<Sample>
in this particular case, I was reading a thread about input.files where webkit implemented an approach through patch right away (ability to assign FileList references I believe)
17:32
<Sample_>
Doh.. what was the last thing I wrote in here and did I get any answers?
17:32
<Sample_>
irssi was letting me type without connection (not being delivered to the channel0
17:34
<Sample_>
I guess I'll repeat myself in case nothing came through (apologies!)
17:35
<Sample_>
Does the idea/terminology "constained by implementation" mean/suggest that browser vendors have committed features ahead of adherance to spec? I've browsed webkit chromium and mozilla ticket systems and I've seen at times suggestions "we won't implement this feature because it's not in spec" and a plea is made to the w3c list
17:36
<caitp>
those messages came through
17:36
<Sample_>
however regarding a w3c ticket on input.files being writeable it was said to be "constrained by implementation" linking a bugs.webkit ticket where patches were made in 2012. I'm curious if this was ahead of specification
17:36
<caitp>
[13:27] <Sample> in this particular case, I was reading a thread about input.files where webkit implemented an approach through patch right away (ability to assign FileList references I believe) --- was the last one before you rejoined
17:36
<Sample_>
oh thanks caitp
17:37
<Sample_>
sorry for spamming
18:00
<Hixie>
anyone got IE?
18:00
<Hixie>
what does it do on http://software.hixie.ch/utilities/js/live-dom-viewer/?saved=3177 ?
18:03
<Hixie>
please describe the results on https://www.w3.org/Bugs/Public/show_bug.cgi?id=26804 if you see this later
18:10
<annevk>
grr, I keep getting disconnected
18:10
<annevk>
I need a better wirless setup I guess
18:19
<Sample_>
annevk: I know the feeling, clearly =)
18:26
<annevk>
https://dom.spec.whatwg.org/#error-names-table is sad
18:26
<annevk>
everyone just mints a bunch of names
18:27
<annevk>
nobody asks for names, they just take up generic names like ReadOnlyError and claim it's about transactions
18:31
<Domenic>
Last time someone asked for names you weren't very receptive IIRC
18:48
<Domenic>
Do HTML import documents require a <!DOCTYPE>? I remember us trying to make them never-quirks but don't remember where that landed.
18:51
<Domenic>
Also do HTML import documents need <title>?
19:52
<TabAtkins>
Sample: Both things happen. Sometimes specs lead browsers, sometimes the reverse, sometimes they develop at the same time with close communication between the implementor and the spec writier.
19:53
<TabAtkins>
Domenic: I have no idea why we'd ever want to make them quirks. If they're not already never-quirks, we better have a damned good reason.
19:59
<annevk>
Domenic: that's fair I guess
19:59
<annevk>
Domenic: I've been trying to make this IDL's problem
19:59
<annevk>
Domenic: or get people to just use TypeError
20:00
<Domenic>
Yeah :/
20:00
<annevk>
Domenic: still need to see code that differentiates on errors though
20:01
<annevk>
There was some flipflopping on the quirks stuff and HTML imports
20:01
<annevk>
There was even some flipflopping on the encoding stuff for HTML imports, but I killed that quickly
20:05
<Domenic>
Oh thank goodness, I hadn't even contemplated having to include a <meta charset> everywhere
20:11
<Hixie>
Domenic: they definitely don't need <title>, i think HTML already has the clause you need to avoid those
20:13
<Sample>
TabAtkins: thanks for the response regarding that terminology
20:54
<TabAtkins>
Domenic: Have you been using markdown headings yet?
21:26
<Domenic>
TabAtkins: yes, but Bikeshed yells at me to give them IDs, so it feels only temporary
21:27
<TabAtkins>
Domenic: Hahaha, I was going to ask you about if you were tagging them with IDs.
21:27
<TabAtkins>
I've finally documented them a bit better - you've been able to give them IDs for forever.
21:27
<Domenic>
Also it is weird that ### gives a h4
21:27
<Domenic>
I am used to the number of #s giving the heading level
21:28
<TabAtkins>
Yeah, the choice was between shifting the levels, or just making === headings useless.
21:28
<TabAtkins>
I went with the former - relative levels are still maintained, they just map to a different level of heading.
21:28
<Domenic>
Hmm
21:29
<Domenic>
Well for those reasons I think I'll probably just switch to <hn> tags when I get around to thinking about what my IDs should be
21:29
<TabAtkins>
"Those reasons" include the ID yelling?
21:29
<TabAtkins>
Because, like I said, you can totally provide IDs for your markdown headings.
21:31
<Domenic>
Well, if I have to give them IDs anyway, and I am already uncomfortable with the mismatch between number of #s and heading level, I might as well use <hn id="x">...</hn>, since at that point I am not saving many characters
21:32
<TabAtkins>
Interesting that ### = <h4> makes you uncomfortable.
21:32
<TabAtkins>
It's the third-level heading in your source, after all.
21:33
<Domenic>
Yeah, I think my brain is well-trained to read ### as "h3" in the mental-verbalization sense
21:33
<TabAtkins>
And your brain is simultaneously well-trained to read "h3" as "second-level heading in my source document"?
21:35
<Domenic>
More like ... there is only one h1, there are many h2s, and below h2 comes h3, ...
21:35
<TabAtkins>
Interesting.
21:36
<TabAtkins>
Related: Do you like using # headings in preference to = and - headings?
21:36
<Domenic>
Yes, definitely
21:37
<TabAtkins>
Okay, our preferences definitely don't match there.
21:37
<TabAtkins>
Wonder if I should rejigger things to have # headings map directly to heading levels, but keep = and - mapping to <h2> and <h3>.
21:39
<Domenic>
I mean, that would make me happy, but not a big deal, and I can understand the consistency argument
21:41
<TabAtkins>
I was just wondering, because I just converted Images 3 to use all the Markdown I support so far.
21:42
<TabAtkins>
And I had to fix the underlined-heading ID handling to actually match Markdown Extra, so I wanted to warn you in advance. ^_^
21:42
<Domenic>
Link to source? Sounds like worth a read.
21:42
<TabAtkins>
http://dev.w3.org/csswg/css-images-3/Overview.bs
21:44
<Domenic>
Very interesting
21:46
<TabAtkins>
Now I'm curious what you find interesting. ^_^
21:47
<Domenic>
Heavy use of indentation, the large minimap blocks... the way ' and '' are useful for CSS specs...
21:48
<Domenic>
your list formatting and line-wrapping choices
21:48
<TabAtkins>
Semantic line-wrapping, yo.
21:48
<TabAtkins>
The large text blocks are supported by Bikeshed. ^_^
21:48
<TabAtkins>
<!-- Big Text: foo --> in your doc, then run `bikeshed source`
21:49
<TabAtkins>
I do it because it makes it much easier to navigate the document via Sublime's minimap.
21:49
<Domenic>
Yeah, definitely, never actually used that technique, but was quite aware of it.
22:41
<Hixie>
so what's the relationship between css animations and web animations?
22:41
<Hixie>
and css transitions, for that matter
22:43
<Domenic>
I think they are meant to be related, and implementations try to back them by the same code, but that aspect of the spec is very hand-wavey
22:58
<jonr22>
acc fb off
22:58
<jonr22>
acc fb on
22:59
<jonr22>
acc fb set nick_format %full_name
22:59
<jonr22>
acc fb off
22:59
<jonr22>
acc fb on