00:00
<Hixie>
nessy: the only change to the current model would be that instead of advancing whenever you're potentially playing, you advance when the controller says to
00:00
<nessy>
interesting - I'd like to see that specced a bit more
00:01
<Hixie>
nessy: or maybe even you just multiply the playback rate by he controller's rate, and the controller sets its rate to zero whenever anyone is stalled (or when the controller is paused)
00:02
<nessy>
I personally actually don't think the playback of the main resource should be stalled because of subresources
00:02
<nessy>
or at least you should have the possibility to not stall
00:02
<nessy>
e.g. maybe I care about seeing sign language, but not enough to want my movie to be stalled when it falls behind a bit
00:04
<zewt>
the whole "seeking to keyframe" thing might be tricky with multiple video tracks, since they won't necessarily keyframe together
00:04
<zewt>
(very unlikely unless specially authored)
00:05
<Hixie>
nessy: well you definitely don't want it to fall behind a bit
00:05
<nessy>
no, just drop it when out of sync
00:06
<Hixie>
nessy: you might want it to just be muted for a bit while the UA buffers, but in most cases if you need to buffer, and you don't, you'll just never have that track again.
00:06
<Hixie>
unless it just skips ahead and starts buffering in the future, and then rejoins later
00:06
<Hixie>
which seems like very confusing behaviour.
00:06
<Hixie>
i can't really imagine wanting my audio track to do that, so i don't see why a sign-language user would want that to happen to theirs.
00:06
<nessy>
I would prefer that - but maybe I am a strange user
00:07
<nessy>
I am concerned about mixed viewing environments
00:07
<nessy>
when a deaf person watches something with non-deaf people and the sign language track keeps stalling
00:08
<Hixie>
AryehGregor: can you give me a tl;dr on http://www.w3.org/Bugs/Public/show_bug.cgi?id=11912 ?
00:08
<Hixie>
nessy: the web is mostly a single-viewing environment.
00:09
<nessy>
hmm, for now maybe
00:09
<Hixie>
nessy: the mixed-viewing environments tend to have reliable video streams.
00:09
<AryehGregor>
Hixie, he wants to improve HTTP auth in a couple of obvious ways, allowing you to use it via forms and allowing the site to log you out.
00:09
<AryehGregor>
I got into a long argument with him about whether HTTP auth is worth salvaging, but it didn't relate to his actual suggestions.
00:09
<Hixie>
nessy: but in any case, if you're watching something with someone who's blind, and the video needs to buffer but we still have the audio, do you want the video to cut out? or do you want it to pause?
00:09
<Hixie>
AryehGregor: k
00:10
<Hixie>
AryehGregor: thanks
00:10
<nessy>
depends on how much is happening on the video
00:11
<nessy>
but yeah, I guess the main use case is that of keeping in sync and stalling when one active track stalls
00:12
<zewt>
multiple people watching a video seems normal enough, but I'd think if the sign track keeps dropping out, the deaf person's going to be even more annoyed (than the non-deaf people would be if it stalled the whole video)
01:48
<nessy>
say, has anyone ever considered introducing an attribute such as @rows onto the <select> element?
01:49
<TabAtkins>
What would it do?
01:49
<nessy>
I am baffled that there is no way to limit the number of options that are displayed in a <select> for scrollbars
01:49
<TabAtkins>
<select multiple size>?
01:50
<AryehGregor>
Or just <select size>.
01:50
<AryehGregor>
E.g., <select size=6>.
01:50
<AryehGregor>
Or whatever.
01:50
<TabAtkins>
Ah, indeed, I didn't realize you could use @size on a plain select.
01:50
<TabAtkins>
Still ugly as sin.
01:51
<nessy>
no, @size gives you the number of options that are displayed and turns the option drop-down off
01:52
<nessy>
http://stackoverflow.com/questions/570642/height-of-an-html-select-box-dropdown
01:52
<nessy>
this part:
01:52
<nessy>
the height needed to show (x) entries (with scrollbars to see remaining)
01:52
<nessy>
     (20 in Firefox & Chrome, 30 in IE6,7,8, 16 for Opera 10, 14 for Opera 11,
01:52
<nessy>
      22 for Safari 4, 18 for Safari 5, 11 in IE5.0,5.5)
01:53
<nessy>
that number of option rows that are displayed is fixed and browser dependent and there is no way to change it
01:53
<TabAtkins>
@size is the correct answer, you just have to convince browser to make it not look like *ass*.
01:53
<nessy>
but @size turns it from a drop-down to a list
01:54
<TabAtkins>
That's not a necessary effect, it's just how browsers traditionally present it.
01:55
<nessy>
but since all browsers do that, it's the de-facto standard
01:55
<TabAtkins>
Considering how horrible it looks, though, I suspect not too many people are actually using it.
01:55
<TabAtkins>
(I could be wrong.)
01:55
<TabAtkins>
This is something we can test.
01:56
<nessy>
so we need a way to convince the browsers to keep displaying it as a drop-down even when @size is given
01:56
<nessy>
I guess
01:57
<nessy>
hmm… Firefox even seems to completely ignore @size (testing here: http://www.w3schools.com/TAGS/tryit.asp?filename=tryhtml_option)
01:57
<zewt>
there's no size there
01:58
<nessy>
@size is specified to be "the number of options to show to the user" which is not equivalent to restricting the drop-down length, actually
01:58
<nessy>
zewt: no, but it's simple to add it manually
01:58
<zewt>
but it doesn't ignore it when I add it
01:58
<nessy>
and see the effect
01:59
<nessy>
if I set it to size=2, then I get all four options, but as a list rather than a drop-down
01:59
<zewt>
i set it to 5 and it shows all 4 with one empty spot
01:59
<TabAtkins>
size=4 is probably the minimum size
01:59
<AryehGregor>
Hixie, in http://www.w3.org/Bugs/Public/show_bug.cgi?id=11912, to be fair, the reporter did actually convince me that there was a not-totally-theoretical use-case in which HTTP auth was inherently more secure than cookie auth, albeit slightly.
01:59
<TabAtkins>
chrome does the same.
01:59
<zewt>
if i set it to 2 i see 2 options
01:59
<nessy>
same in safari
01:59
<zewt>
haha but
01:59
<zewt>
if i set it to 5, i get 5 rows with a disabled scrollbar
02:00
<zewt>
if i set it to 2, i get 2 rows ... and no scrollbar
02:00
<nessy>
ups: I mixed up my chrome and firefox
02:00
<AryehGregor>
But "not totally theoretical" I mean "He said it applies to his application, but it doesn't apply to any app I know about, because you'd have to be incredibly obsessive about separation of privilege for it to be relevant."
02:00
<AryehGregor>
s/But/By/
02:00
<zewt>
so it shows a scrollbar if it doesn't need one, and doesn't show a scrollbar if it does need one
02:00
<nessy>
actually, firefox is the only one that actually does limit the length
02:01
<Hixie>
AryehGregor: what was the use case?
02:01
<zewt>
nessy: same behavior in Opera as FF
02:02
<nessy>
yup, Opera and FF behave the specced way, I guess
02:02
<AryehGregor>
Hixie, the guy had a system where the application didn't actually have access to the database. The only way for it to access the database was using the users' actual username and login.
02:02
<AryehGregor>
Username and password.
02:02
<nessy>
so webkit only changes from drop-down to list display without actually limiting the size
02:02
<zewt>
webkit (testing in chrome) will extend but not scroll (which again makes it bizarre that it shows a disabled scrollbar)
02:02
<TabAtkins>
AryehGregor: We need to not worry about that ever.
02:02
<AryehGregor>
So every application-level user had a corresponding database user, and the database user had only the rights to access stuff that the application user had.
02:02
<AryehGregor>
And so if the browser submits the password every time, the app just uses it for that request, to access the database, then throws it away.
02:02
<nessy>
so, I would say that's a webkit bug
02:03
<nessy>
but it's also a spec bug to not have a possibility to remain in drop-down mode and limit the length
02:03
<AryehGregor>
But if you use cookies, you'd force the application to remember the user's password somehow, so an application server compromise would result in database server compromise too, even for users who aren't actually executing requests at that particular moment.
02:03
<AryehGregor>
I thought it was cool that the guy actually designed such a system that he apparently uses.
02:03
<AryehGregor>
Or at least he works with such a system, dunno if he designed it.
02:04
<AryehGregor>
But, uh, not exactly mainstream practice.
02:05
<zewt>
(AryehGregor: not necessarily useful to that person, but if I really needed that, I'd probably do something along the lines of storing an encrypted password in the cookie with a key on the server, so the server can decrypt it for each request and then throw it away)
02:05
<AryehGregor>
Neat idea.
02:05
<Hixie>
AryehGregor: how did he add new users to the database?
02:06
<AryehGregor>
Hixie, I didn't ask. I'm guessing this is some workplace thing, and the database is getting the user info from LDAP or some other external authentication source.
02:06
<AryehGregor>
So people are probably added by hand by a sysadmin when they first come there, or something.
02:06
<Hixie>
i can't actually imagine a scenario where i could use that kind of technique
02:07
<AryehGregor>
In principle, though, you could have open login.
02:07
<Hixie>
unless his database has per-row ACLs or something
02:07
<AryehGregor>
It's PostgreSQL.
02:07
<AryehGregor>
Which, yes, can have extremely granular access controls.
02:07
<AryehGregor>
I suspect a nontrivial amount of the application logic was implemented as stored procedures or such on the database server.
02:07
<Hixie>
well anyway
02:07
<AryehGregor>
So users without direct access to particular rows might still have access to functions that would let them read or modify them indirectly.
02:08
<Hixie>
you can do that without http auth
02:08
<Hixie>
in fact most systems i write send the password and username along with every request too
02:08
<Hixie>
they just use xhr :-)
02:08
<AryehGregor>
Yeah, but that doesn't work for regular web browsing.
02:08
<Hixie>
"regular"?
02:08
<Hixie>
you mean web 1.0 ?
02:08
<AryehGregor>
:)
02:09
<AryehGregor>
"something that works in lynx"
02:09
<TabAtkins>
I don't think we're designing the web for lynx.
02:09
<AryehGregor>
Really? Darn.
02:09
<Hixie>
lynx needs to get with the programme and implement js already
02:09
<TabAtkins>
AryehGregor: I could be wrong. Ask Hixie.
02:10
<zewt>
the few times in the last decade I've had to load lynx for one reason or another, it still does the "pause to show a message 10 times per page load" thing
02:11
<zewt>
which is pretty awesome
02:11
<AryehGregor>
You mean like "This site is trying to set a cookie. Do you want to accept it?"
02:11
<zewt>
(for some value of awesome)
02:11
<AryehGregor>
Several times?
02:12
<AryehGregor>
Yeah, lynx is cool that way.
02:12
<zewt>
well come on, cookies are a violation of privacy, you need to ask permission for each one
02:12
<AryehGregor>
One time my school computers (almost all RHEL) were practically unusable because NFS was unresponsive, taking like a second to service simple requests.
02:13
<AryehGregor>
So all GUI stuff totally died because everyone's home directory was on NFS and all GNOME stuff checks every file in your home directory a billion times a second or something.
02:13
<TabAtkins>
True story.
02:13
<AryehGregor>
So I did Ctrl-Alt-F1 and browsed the web using lynx, which more or less worked if you were patient.
02:13
<AryehGregor>
It impressed my officemate.
02:13
<AryehGregor>
TabAtkins, have you ever tried straceing a GNOME process's startup?
02:14
<TabAtkins>
AryehGregor: No. I've never run strace in any capacity.
02:14
<zewt>
i havn't used gnome in like a decade and I hope to keep my streak going
02:14
<AryehGregor>
$ strace gnome-calculator 2>&1 | grep '^open(' | wc -l
02:14
<AryehGregor>
338
02:14
<AryehGregor>
That's starting and then immediately closing.
02:14
<AryehGregor>
srsly.
02:15
<AryehGregor>
zewt, what do you use?
02:15
<zewt>
windows :)
02:15
<zewt>
(linux too, but relegated to a shell)
02:16
<AryehGregor>
Speaking of GNOME, I have to think of new ways to harass them into reviewing the one-line patch I submitted well over six months ago.
02:16
<zewt>
generally if i have to harass people to look at patches, I find other projects to contribute to
02:17
<AryehGregor>
I think I'm going to write up step-by-step instructions for how to apply the patch to your desktop, and post it on all the bugs I filed in various bug trackers.
02:17
<AryehGregor>
Then maybe other people will find it works for them and they'll help me harass them.
02:17
<AryehGregor>
zewt, me too, except that in this case it means I have to either endure the bug or re-patch the GNOME clock every time.
02:17
<AryehGregor>
Currently I've chosen to endure it because I can't be bothered to figure out again how to actually replace the clock applet on a running system.
02:18
<AryehGregor>
But if I write up instructions and post them on the bugs, then at least I'll have something to refer to when I next upgrade.
02:26
<AryehGregor>
Hixie, why is <xmp> not conforming? It would be nice to have when marking up sample HTML markup in HTML by hand.
02:27
<AryehGregor>
(If you're generating the HTML by a script, you can of course just escape it.)
02:27
<AryehGregor>
I hate having to type out stuff like &lt;b>Foo&lt;/b>, it's incredibly awkward.
02:27
<Hixie>
same reason <u> isn't conforming :-)
02:28
<TabAtkins>
This is why I'm writing myself a custom Markdown parser that doesn't understand any HTML, so I can just write things in that and then take the generated HTML.
02:28
<Hixie>
(what semantic would it have?)
02:29
<AryehGregor>
Hixie, it's a syntactic construct, not a semantic one. It should be semantically void, like <span>.
02:29
<AryehGregor>
Or <div>, maybe.
02:30
<AryehGregor>
<xmp><b></xmp> and <span>&lt;b&gt;</span> would be semantically equivalent.
02:33
<Hixie>
interesting
02:33
<Hixie>
file a bug, i guess
02:33
<TabAtkins>
I don't think HTML has any syntactic constructs any more.
02:33
<Hixie>
cc henri, i'm interested in what he'd have to say
02:33
<Hixie>
it'd be an xhtml vs html difference
02:34
<Hixie>
what is supposed to happen with a UTF-8 document that contains a U+D800 (surrogate half) character?
02:34
<TabAtkins>
My only fear is people using it as a way of "safely" embedding user-generated text.
02:34
<Hixie>
did we change this recently?
02:37
<zewt>
AryehGregor: would &amp; &lt; &gt; still resolve to &<>? (to encode </xmp> literally)
02:37
<AryehGregor>
zewt, try it and see: data:text/html,<!doctype html><xmp>&lt;&gt;&amp;</xmp>
02:37
<AryehGregor>
(no)
02:37
<AryehGregor>
(so you can't encode </xmp> literally, just like you can't encode </script> literally inside <script>)
02:38
<zewt>
well, </script> has a workaround; would this?
02:38
<TabAtkins>
Sure - use two <xmp> elements.
02:38
<AryehGregor>
Doesn't work well with default style.
02:38
<TabAtkins>
Encoding the first "<x" in one, and "mp>" in the other.
02:39
<AryehGregor>
Which is block, same as pre.
02:39
<zewt>
that's not the same, though--that gives you two elements
02:39
<AryehGregor>
The workaround would be using <pre> instead, and HTML-escaping.
02:39
<TabAtkins>
Sure, I'm not saying it's perfect.
02:39
<zewt>
(granted, "</scr"+"ipt>" isn't exactly the same, either, but in practical terms it is)
02:39
<AryehGregor>
So <xmp> isn't perfect, that's life.
02:39
<jcranmer>
"\u003c/script>" ?
02:42
<AryehGregor>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=12235
02:43
AryehGregor
goes to bed
02:43
<AryehGregor>
I'll try to think of a use-case for <plaintext>. :)
02:51
<zewt>
speaking of </script>, should innerHTML/outerHTML's non-normative text note that the fragment may not represent the element exactly (may not round-trip)?
02:51
<Hixie>
it already does somewhere
02:53
<zewt>
hmm, down in the HTML serialization algo text ... okay i guess
07:12
<annevk>
Hixie, the multipart boundary string probably needs to be prefixed by "boundary="
07:12
<annevk>
Hixie, when you make the MIME type
07:22
<jamesr__>
hey y'all i'm trying to find documentation or specification for the 'scrolling' attribute on i?frame. it seems pretty widely supported but i can't find references to it
07:24
<zcorpan>
Hixie: <xmp> already has the same semantic as <pre>
07:25
<annevk>
jamesr__, http://www.whatwg.org/specs/web-apps/current-work/complete.html#margins-and-padding
07:39
<annevk>
benschwarz, yo
07:39
<annevk>
benschwarz, how about that blog post?
07:39
<annevk>
;)
07:54
<zcorpan>
annevk: one argument against <object> is that it's not as safe as <img> -- it allows scripted content and plugins, so wouldn't allow it for user generated stuff
07:54
<zcorpan>
maybe <iframe sandbox> solves that, dunno
07:55
<zcorpan>
also svg in object is like interactive and eats click events while svg in img is not interactive and doesn't eat click events
07:56
<zcorpan>
so if you want to use an svg image in a link with rich fallback, <object> makes the link not work
07:59
<boblet>
hsivonen: I think I found a validator.nu bug. Using <style scoped> as the first child of <figure> gives an error
08:11
<zcorpan>
Hixie: + &lt;ul onchange="calculate(form)"> - this wouldn't actually work right? would need to be document.forms.pricecalc or so
08:56
<hsivonen>
boblet: thanks. I kinda wish we had a clear direction on <style scoped>: remove from spec ASAP or implement in browsers ASAP
08:56
<boblet>
oh, is it controversial?
08:57
<boblet>
I love it for code examples — everything bundled into one figure
08:57
<zcorpan>
i don't think it's controversial, it's just low priority and probably non-trivial to implement (especially scoped @-rules)
09:00
<boblet>
yeah, that’s true. I’m just using it with an id hook, so implementation isn’t really so important for me
09:00
<zcorpan>
hsivonen: maybe the simplest thing to do at the moment is have an assertion that bans <style scoped> on the basis that it's not implemented anywhere yet so should not be used yet
09:01
<boblet>
zcorpan: is it actually harmful tho? it’s just added to head styles, no?
09:02
<zcorpan>
boblet: it's harmful if sites start depending on styles being applied to the whole document
09:02
<zcorpan>
boblet: because then the feature can't be implemented without breaking those sites
09:02
<zcorpan>
apparently this has already happened with "use strict";
09:04
<boblet>
I was thinking to write about it for HTML5Doctor, and say to use with an id hook until there’s browser support
09:04
<zcorpan>
i'd wait with recommending use of new features until they're implemented
09:05
<boblet>
well, more if you use it, use it like this. I take your point, but I think ppl will find and use it regardless
09:05
<zcorpan>
i'm frankly surprised that we haven't had any problems with <article> and friends
09:05
<Hixie>
zcorpan, annevk: good points; please file bugs!
09:07
<boblet>
hsivonen: regardless of which way you decide to jump, the current error for style scoped is a bit confusing, so a warning or more accurate admonition would help
09:07
<boblet>
zcorpan: thanks
09:08
<boblet>
l8r
09:08
<zcorpan>
seeya
09:12
<annevk>
filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=12237
09:15
<annevk>
I guess I should use www-dom for DOM Core
09:15
<annevk>
or, we should use
09:15
<annevk>
but Ms2ger is away or something
09:28
<annevk>
so in some browsers you can apply CSS properties to <frameset> and such?
09:28
<annevk>
like 'background'?
09:29
<annevk>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/863
09:29
<annevk>
works in Gecko
09:30
<annevk>
I guess that's a bug then
09:31
<zcorpan>
spelling fail - JaëgerMonkey https://mozillademos.org/demos/dashboard/demo.html
09:33
<zcorpan>
section on CORS talks about CSP
09:42
<jgraham__>
zcorpan: ae is cool and umlauts are cool clearly having both is double plus cool. But they really missed a trick not going for æ̈
09:46
<jgraham__>
MikeSmith: BTW did I mention recently how awesome the HTML version of the ES5 spec is?
09:47
<MikeSmith>
no you didn't
09:47
<MikeSmith>
but glad it's useful
09:51
<zcorpan>
äë?
09:53
<annevk>
AryehGregor, you are not subscribed to www-dom
09:54
annevk
was wondering who would miss out on DOM Core discussion
10:00
<jgraham__>
zcorpan: Nah, the coolness of æ is so much greater than ae that the double diaeresis can't comepete
10:02
zcorpan
was just having an unrelated discussion with his mom about how to type æ on windows
10:17
<zcorpan>
lol http://clubajax.org/presentations/the-fight-over-html5/
10:18
<zcorpan>
funny and accurate, love it
10:21
<Philip`>
AryehGregor: If you want a syntactic construct rather than a semantic one, it seems better design to use something like <![CDATA[ that is purely syntactic and doesn't affect the DOM
10:21
<Philip`>
(Need to find a way to make it apply in non-foreign content though, I guess)
10:22
<karlcow>
zcorpan: misleading and inaccurate the clubajax thingie
10:22
<karlcow>
I left a comment here but still in the moderation queue http://clubajax.org/presentation-the-fight-over-html5/comment-page-1/#comment-7463
10:23
<annevk>
there was only one real error I spotted
10:24
<zcorpan>
namely?
10:25
<zcorpan>
it spells W3C as WC3
10:25
<annevk>
trying to make standards patentable
10:25
<zcorpan>
oh right
10:26
<annevk>
well and the logo cost, but that seems like an obvious joke
10:26
<zcorpan>
you sure? :)
10:26
<annevk>
nah, just seems that way
10:36
<deane>
Hixie: I think the spec needs to define what a document is. It needs a definition or a pointer to a definition. Many people get confused between syntax and document.
10:41
<zcorpan>
deane: file a bug
10:44
<deane>
zcorpan: AFAIK the whatwg doesn't have a bug tracking system
10:45
<zcorpan>
deane: html wg does, there's a comment box in the whatwg spec to use it
10:47
<deane>
zcorpan: yeah, I know, but some people prefer not to deal with the w3c so the whatwg should provide some sort of alternative.
10:48
<zcorpan>
ah, ok. the alternative in whatwg land is whatwg⊙wo i guess
10:48
<deane>
zcorpan: yeah, I'll sort something out
10:55
<jgraham>
Not really clear how you have to "deal with w3c" to file a bug
10:55
<jgraham>
I guess there is the possibility that someone will raise your bug to an Issue
10:56
<jgraham>
but you can ignore it if you like
11:06
<MikeSmith>
the W3C needs an official theme song
11:06
<MikeSmith>
WHATWG already has that T.Rex "Whatever Happened to the Teenage Dream" song
11:06
<MikeSmith>
which is a great song
11:06
<MikeSmith>
but we need a similarly appropriate one for the W3C
11:08
<deane>
I don't particularly wish to use the W3C's buzilla since the last time I used it I was attacked by some crazed fan of Hixies
11:12
<jgraham>
Hixie has /groupies/? This spec editor lifestyle is so much mopre rock and roll than I thought
11:13
<MikeSmith>
I'm a fan of Hixie
11:13
jgraham
is somewhat ashamed that his brain has just started reworking "Money for Nothing" to more apropos form
11:13
<MikeSmith>
I'm aheh
11:13
<MikeSmith>
oh
11:13
<MikeSmith>
heh
11:13
<deane>
I'm a fan of Hixie too
11:13
<MikeSmith>
I'm also crazed maybe
11:14
<MikeSmith>
but I don't know if I'm crazed about Hixie
11:14
<MikeSmith>
I'm just a big fan
11:14
<deane>
nah,MikeSmith, you're ok
11:14
<MikeSmith>
like I'm a Charlie Sheen fan too
11:14
<jgraham>
MikeSmith: Have you ever turned up at his hotel room door without wearing any clothes?
11:14
<deane>
maybe you're his biggest fan
11:14
<jgraham>
That is really the acid test for a groupie
11:14
<deane>
maybe you're like a Hixie stalker
11:15
<MikeSmith>
jgraham: I have done that, and worse
11:15
<MikeSmith>
oh *his* hotel room
11:15
<MikeSmith>
well, no
11:15
<MikeSmith>
but other hotel rooms, yeah
11:15
<deane>
MikeSmith: perhaps you have photos of Hixie on your bedroom wall
11:15
<jgraham>
MikeSmith: Phew. The images I was getting there were seriously disturbing
11:16
<MikeSmith>
Hixie and David Cassidy from the Partridge Family
11:16
<MikeSmith>
I also have pictures of Roy Fielding
11:16
<MikeSmith>
waiting to be autographed
11:17
<MikeSmith>
anyway, I need to drop off
11:17
<MikeSmith>
going to watch a movie
11:17
<MikeSmith>
"The Town"
11:17
<MikeSmith>
which, it better be good, or I'm going to raise hell
11:17
<MikeSmith>
because I had to travel all the way across town to find a theater showing it at a decent time
11:18
<MikeSmith>
you all don't have too much fun without me
11:19
<deane>
MikeSmith: enjoy your movie
11:44
<annevk>
heycam, so should I file a bug on Web IDL to get some kind of object support?
12:01
<annevk>
http://www.w3.org/2005/06/tracker/webapi/issues/20 Member-only?
12:26
<annevk>
hmm, CustomEvent is not yet in Gecko or WebKit
12:27
<annevk>
so initCustomEvent can probably be dropped then
12:27
<annevk>
if people react positively to the initEvent(object) proposal
12:29
jgraham
doesn't really know why create and init are seperate steps
12:29
<jgraham>
But I havne't read the draft carefully
12:32
<hsivonen>
having create and init as separate steps feels like an artifact of C++ memory management
12:33
<annevk>
maybe at some point we'll make the init implicit
12:33
<annevk>
like new MouseEvent(initObject)
12:33
<annevk>
but for now this is good enough and much simpler
12:33
<zcorpan>
can we jump to that directly?
12:34
<annevk>
I'd rather not
12:34
<zcorpan>
having three ways to create+init seems bad
12:34
<annevk>
I mean everyone here gets excited about simplicity and how complex this is, but it's really not that complex and the use cases are nihil
12:35
<hsivonen>
does IE SmartScreen treat all unsigned .exe files dangerous by default even if they don't match against malware fingerprints?
12:35
<annevk>
zcorpan, one would just be a shorthand for the other in the end...
12:35
<zcorpan>
annevk: if the use cases are nihil, why bother with initEvent(object)?
12:36
<annevk>
zcorpan, as I explained in the email, I want event interfaces to be extensible and initXXXEvent does not accomplish that
12:37
<annevk>
zcorpan, so I want to introduce something as simple as possible that prevents introduction of any new initXXXEvent methods
12:37
<annevk>
(and maybe allows us to remove some of them)
13:17
<annevk>
hmm, newsycombinator discovered canvaspaint.org
13:20
<loucapo>
greetings everyone
13:21
<annevk>
good day
13:21
<loucapo>
i dont believe there is, but does anyone here know for sure about limitations of CORS posting from http to https?
13:25
<annevk>
in theory there are no limitations
13:26
<annevk>
in practice implementations might enforce limitations
13:26
<annevk>
at some point we might formalize those one way or another
13:26
<loucapo>
i see, we are seeing that status 0 issue again for non 200s
13:26
<loucapo>
and 0 for 200s on chrome
13:28
<bga_>
http://d.hatena.ne.jp/masa141421356/20110228/1298904732
13:41
<annevk>
if DOM Core did not have XML production checking the only normative dependencies would be Web IDL and RFC 2119...
13:41
<annevk>
hardcore
13:58
annevk
wonders where AryehGregor is
13:59
<annevk>
guess I should update XHR with the new hooks HTML has now
14:30
<annevk>
http://who-t.blogspot.com/2011/03/how-to-dos-developer.html glad I work on open standards instead...
14:33
<annevk>
updated http://dev.w3.org/2006/webapi/XMLHttpRequest-2/ with the new terms from HTML
14:33
<annevk>
now I guess DOM Level 3 Events review comments
14:36
smaug____
should review events part of web dom core
14:37
<jgraham>
annevk: Sarcasm? (about the "how to DOS…" article)
14:42
<annevk>
only slightly, because we do not nearly have as much people involved as open source projects I think
14:47
<jgraham>
Pretty sure I have seen most of the IRC and email things
14:58
<Rik`>
TabAtkins: http://bonjourmozilla.fr/?post/2011/02/25/Boris-Zbarsky for some BZ facts :)
15:09
<loucapo>
annevk, have you seen something similar to what i was saying earlier?
15:09
<annevk>
strange as it may seem, I have not played much with CORS yet
15:10
<loucapo>
seems like not too many have
15:10
<loucapo>
i hear people still say its not possible to post across domains
15:10
<loucapo>
makes me wanna slap them
15:10
<annevk>
it's getting traction, but it's not exactly simple :/
15:10
<loucapo>
i put my code on github
15:11
<loucapo>
but never shared it
15:11
<annevk>
hopefully Apache will do this by default one day
15:12
<jgraham>
Rik`: I think that is the first photo of bz I ever saw
15:30
<annevk>
I think I raised all relevant points on DOM Level 3 Events: http://lists.w3.org/Archives/Public/www-dom/2011JanMar/
15:33
<zewt>
way too late to invert the return value of dispatchEvent...
15:36
<annevk>
zewt, I did not suggest that
15:38
<annevk>
As a heads up, Shelley Powers will be writing the WHATWG Weekly starting with her first post around March 21
15:38
<annevk>
jgraham volunteered to help out with publishing, though she can publish directly if she wishes in line with what we use for other contributors (everyone is allowed to blog)
15:39
<annevk>
She asked that any suggestions for the WHATWG Weekly are submitted to "Shelley Powers" <shelleyp⊙bn>
15:39
<annevk>
as she will be primarily be focusing on HTML and might not always have all the details on everything else that is going on
15:40
<annevk>
I will include this summary in the next two WHATWG Weekly posts that I write as well and then hopefully the world knows how it works :)
15:42
<karlcow>
there will be redundancy ;) that will cool to see
15:42
<karlcow>
s/will/will be/
15:51
<annevk>
zewt, you misread what I wrote right? otherwise I need to fix something
16:03
<zewt>
ah yeah, k
16:03
<jonathanMH>
This may be a little wordpress specific, but how should I differ between nav and aside in my themes? When I distribute a theme I can't tell which widgets will be put in the widget areas by the user. Would <section> be the best solution for this kind of uncertainty?
16:08
<annevk>
anyone know if html5test.com guy is available on IRC?
16:08
<annevk>
the type=color validation tests are wrong
16:09
<annevk>
it's impossible to make that invalid
16:09
<annevk>
also, the form event tests are wrong
16:09
<annevk>
besides that the events are now removed 'dispatchFormInput()' in form will return false because of the parenthesis
16:11
<Rik`>
annevk: I believe volkmar is in contact with him, by mail
16:12
<volkmar>
Rik`: not really by email, but i have his email
16:13
<volkmar>
i was going to send him a patch to remove forminput, formchange, and dispatch methods
16:13
<volkmar>
but i have to understand why vim doesn't like utf-8 on this computer :(
16:15
<annevk>
the keygen tests are also wrong
16:16
<Rik`>
(and probably talking about forms :) )
16:17
<volkmar>
annevk: who cares about keygen?
16:18
<annevk>
dunno, but it's worth points!
16:18
<annevk>
actually, I used to care about <keygen>, mostly because it existed and I wanted it documented, but now that is done
16:18
<volkmar>
eh... i can't really ask who cares about these kind of tests, sadly :(
16:19
<annevk>
we decided to look through it because it's Friday
16:19
<annevk>
drinking a beer would be more enjoyable however so I'm going to do that now
16:20
<volkmar>
annevk: a beer is a thousand times better :)
16:20
<volkmar>
and cheers ;)
16:27
<TabAtkins>
Rik`: That bz link is funny. ^_^
16:30
<mven>
is it still going ?
16:30
<mven>
beh wrong channel ;p
16:33
<AryehGregor>
annevk, I'm reluctant to subscribe to any more spec mailing lists, they take enough time as-is.
16:33
<jgraham>
http://ie6countdown.com/
16:34
<TabAtkins>
Hah, Opera's nearly stamped out IE6 usage in Norway. ^_^
16:35
<AryehGregor>
Also, yeah, http://clubajax.org/presentations/the-fight-over-html5/ is awesome.
16:36
<AryehGregor>
"12.0% of the world was using Internet Explorer 6, which was 9% lower than the previous year"
16:36
<AryehGregor>
I assume they mean "9 percentage points" and not "9%" there.
16:38
<TabAtkins>
Yes, I assume so.
16:39
<TabAtkins>
This is why percentages are confusing - they don't contain information about what they're relative to.
17:28
<AryehGregor>
You know, if <xmp> isn't made valid, I could always do <script type=text/plain style=display:inline><b>some HTML I don't need to escape&&&!</b></script>
17:28
<AryehGregor>
That would just be semantically incorrect.
17:28
<AryehGregor>
I have to say, it's sad that <xmp> is flow content instead of phrasing content. Makes it much less useful.
17:34
<AryehGregor>
"My main regret in saying this is that WebIDL, having evolved from Corba IDL,
17:34
<AryehGregor>
is notationally almost pessimal for encouraging good JS interface design. We
17:34
<AryehGregor>
have all been suffering the consequences of that. I fear that WebIDL, and
17:34
<AryehGregor>
What's that supposed to mean?
17:34
<AryehGregor>
JS programmers are the only ones we cater to, in real life . . .
17:40
<AryehGregor>
Argh, my toString() definition is wrong too.
17:40
<AryehGregor>
Ranges are bloody confusing.
17:44
<bfrohs>
Yes, yes they are.
17:45
<AryehGregor>
It's so tempting to assume that they assume that they include everything between their start and end nodes.
17:45
<AryehGregor>
That would make things so simple.
17:46
<AryehGregor>
But it's a primrose path. Don't be fooled!
17:46
<bfrohs>
They don't?
17:46
<AryehGregor>
No, not even remotely.
17:46
<AryehGregor>
Consider a range whose start is (head, 0) and whose end is (body, 3).
17:46
<AryehGregor>
It contains a whole bunch of nodes that come after body in the DOM.
17:46
<bfrohs>
Also, I noticed when copying an ordered list, if you don't include the first item, it doesn't realize it is an ordered list and only sees the li and makes it unordered. (at least in ff)
17:47
<AryehGregor>
. . . what?
17:47
<AryehGregor>
Oh, I see.
17:47
<AryehGregor>
That sounds like a Selection.toString() issue!
17:47
<AryehGregor>
Which I'm supposed to be standardizing, but I got demoralized and moved on to things that are saner.
17:47
<bfrohs>
IIRC, there are several other issues when copying and pasting from mid-string (similar to the li issue)
17:48
<AryehGregor>
Yes, this is why I ran away from it.
17:48
<AryehGregor>
I didn't even consider this sort of mid-string issue.
17:48
<AryehGregor>
But I still ran away.
17:48
<bfrohs>
I don't blame you one bit
17:49
<AryehGregor>
Also because it's not clear anyone wants to standardize the API.
17:49
<zot-->
bravely ran away?
17:49
<AryehGregor>
Possibly.
17:49
<bfrohs>
oh, you wouldn't happen to be standardizing double/triple clicking selects, would you?
17:49
<AryehGregor>
Certainly not, those are platform-specific.
17:49
<bfrohs>
Cause it becomes *very* interesting when generated content is involved
17:49
<AryehGregor>
Oh, selecting generated content.
17:49
<AryehGregor>
Yeah, that's a fun one.
17:49
<bfrohs>
:first-letter especially
17:50
<AryehGregor>
Hmm? What generated content would you put in :first-letter?
17:50
<bfrohs>
well, wait
17:50
<bfrohs>
Let me go back to what I was doing and figure out exactly what the issue was--one sec
17:51
<bfrohs>
Take that back, SPAN around first letter in an H1
17:55
<dglazkov>
is there any indication anywhere that meter/progress elements should change their orientation depending on the containing width?
17:55
<dglazkov>
I just scanned http://www.whatwg.org/specs/web-apps/current-work/multipage/the-button-element.html#the-progress-element and http://www.whatwg.org/specs/web-apps/current-work/multipage/the-button-element.html#the-meter-element
17:55
<dglazkov>
and there doesn't seem to be anything there
17:57
<AryehGregor>
dglazkov, http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#the-meter-element-0
17:57
<TabAtkins_>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=12232
17:59
<uf0>
question
17:59
<TabAtkins_>
answer
17:59
<uf0>
what kind of bugs occur if I leave out '<meta charset="utf-8">' in a basic html page
17:59
<TabAtkins_>
Depends. Are you sending a charset header?
17:59
<uf0>
sending, as in in the apache config?
18:00
<TabAtkins_>
If you are, the page uses that. If not, it defaults to, um, Windows-1252 or something? Something that'll make your page look messed up if you use characters too far outside of ascii.
18:00
<TabAtkins_>
Yes.
18:00
<zot-->
depends on the user's environment--typically defaults to their default windows codepage, i think
18:00
<zot-->
(... in windows)
18:00
<uf0>
makes sense, you've answered my question :)
18:00
<uf0>
got another one
18:01
<zot-->
so many japanese pages with no charset defined anywhere, just expecting that everyone will default to sjis, heh
18:01
<uf0>
<meta http-equiv="X-UA-Compatible">
18:01
<zot-->
i almost wish every browser would just default to something totally nonsensical so the problem wouldn't be so obscured
18:01
<uf0>
is that something WhatWG created or microsoft
18:01
<uf0>
?
18:01
<TabAtkins_>
Microsoft.
18:01
<uf0>
so why this..
18:01
<uf0>
<meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1">
18:01
<TabAtkins_>
We wouldn't do something so dumb.
18:01
<uf0>
the chrome=1 part
18:01
<uf0>
which is google
18:02
<TabAtkins_>
That's something that Chrome Frame listens for.
18:02
<uf0>
what's the point of adding that then, I see the point for IE
18:02
<uf0>
because I actually ran into a bug where it rendered in IE7 mode
18:02
<uf0>
and edge fixed
18:02
<TabAtkins_>
It tells Chrome Frame "hey, if you're installed, use yourself to render me".
18:03
<uf0>
you're a fucking guru, thanks!
18:03
<uf0>
case closed.
18:19
<AryehGregor>
Does Opera have a built-in mail client?
18:20
<karlcow>
AryehGregor: yes
18:20
<karlcow>
irc too
18:20
<karlcow>
p2p too
18:20
<zot-->
also tetris and a copy of emacs
18:20
<AryehGregor>
Mmm.
18:21
<AryehGregor>
rniwa is saying he thinks maybe <font> is needed for e-mail clients, because some clients don't understand CSS.
18:21
<AryehGregor>
I'll have to do a survey of e-mail clients to see.
18:21
<annevk>
AryehGregor, if DOM related matters are to be discussed on www-dom I suspect that would make sense for DOM Range too
18:22
<AryehGregor>
Then maybe I can subscribe to www-dom and unsubscribe from public-webapps.
18:22
<AryehGregor>
Is www-dom also a Webapps WG list?
18:22
<karlcow>
AryehGregor: the support of HTML in email clients has been historically bad.
18:22
<karlcow>
there is a very good source for it
18:22
<annevk>
AryehGregor, sort of
18:23
<karlcow>
let me fetch it
18:25
<karlcow>
AryehGregor: http://www.campaignmonitor.com/css/
18:27
<AryehGregor>
Ouch, Blackberry doesn't support font-family or font-size.
18:33
<AryehGregor>
Actually, I remember now, people were complaining about this and Hixie WONTFIXed them.
18:34
<AryehGregor>
They wanted legacy markup to be made valid because Blackberry needed it and he said they should complain to Blackberry.
18:34
<AryehGregor>
I suspect browser implementers whose rendering engines are used for e-mail clients won't be receptive to that suggestion.
18:34
<karlcow>
Email marketers/communications teams/PR people have to rely on tables for their html emails
19:04
<zot-->
html spec vs. chrome 9: html spec wins
19:06
<AryehGregor>
Hmm?
19:08
<zewt>
made the mistake of triple-clicking in the spec, heh
19:24
<AryehGregor>
Argh, that's the second time in two days I wrote "for (var i = 0; i < foo; i++)" instead of "for (var i = 0; i < foo.length; i++)".
19:24
<AryehGregor>
This is why we want foreach-style loops. C-style for loops are an abomination.
19:25
<othermaciej>
what language are you writing in?
19:26
<AryehGregor>
JavaScript.
19:26
<zewt>
gah, i can't copy text from the spec with chrome no matter what I do; it always decides to select the whole document
19:26
<AryehGregor>
As I understand it, for (var entry in arr) is a bad idea because it might pick up added attributes or wacky stuff like that, right?
19:27
<gsnedder1>
AryehGregor: Right
19:29
<jamesr_>
it's a crappy language that lets you do that
19:29
<jamesr_>
without at least some sort of indication of a problem
19:29
<jamesr_>
comparing an int to an array is silly
19:29
<othermaciej>
for (var entry in arr) won't really do what you want
19:29
<AryehGregor>
Yes, the typing here is really too weak.
19:30
<AryehGregor>
It should raise an exception or at least log a warning if you compare a number to an array.
19:30
<AryehGregor>
I guess it's too late for that, though.
19:30
<AryehGregor>
(Maybe in strict mode? :) )
19:30
<zewt>
can we just make strict mode python
19:30
<gsnedder1>
Logging a warning can always be done. And changes in Harmony are always possible.
19:30
<AryehGregor>
(Is there any provision for extending strict mode? "use stricter mode";?)
19:30
<AryehGregor>
zewt++
19:31
<gsnedder1>
AryehGregor: Harmony will have separate MIME-type, as I understand the intention
19:31
<othermaciej>
it will enumerate the property names, not the properties
19:31
<othermaciej>
however, I don't see how forgetting .length undermines for loops
19:31
<othermaciej>
obviously there is much better syntax for enumerating arrays out there, but that's not a very good example of why for loops are grungy
19:31
<AryehGregor>
Given that, will anyone actually use it?
19:31
<AryehGregor>
Well, I'd reply, but you just left.
19:32
<AryehGregor>
gsnedder1, FWIW, I found no Opera bugs with my deleteContents() tests, other than unrelated bugs in Range that I already knew about. I found at least one bug in every other browser.
19:32
<gsnedder1>
AryehGregor: Well you generally have a properly if you ever want to make any syntax change because they're inherently not BC compatible, and will just throw SyntaxError in existing impls
19:32
<AryehGregor>
. . . although the unrelated bugs still make Opera fail more bugs than any other non-IE browser.
19:33
<AryehGregor>
gsnedders, so instead, let's pop up a "would you like to download this file?" dialog in legacy UAs! . . . or actually, I guess they just ignore scripts with different <script type="">, right?
19:33
<gsnedders>
AryehGregor: Just ignore them. They have to, because of legacy VBScript.
19:33
<AryehGregor>
As long as they don't care about the actual MIME type served by the server, that sounds potentially sane.
19:34
<gsnedders>
AryehGregor: There was some talk about both in-band and out-of-band markers
19:35
<gsnedders>
I haven't been massively keeping up-to-date on all of this, and I'm not on the TC39 list
19:49
<uf0>
why are these chosen to display block: article, aside, details, figcaption, figure, footer, header, hgroup, menu, nav, section { display:block; }
19:50
<uf0>
why aren't any others added? is it because those are the only block elements
19:50
<uf0>
and everything else is inline by default?
20:10
<AryehGregor>
IE sometimes randomly doesn't run some of my tests, even with timeout: 60000. :(
20:11
<AryehGregor>
Where by "sometimes" I mean "usually".
20:11
<AryehGregor>
Oh, now it worked.
20:52
<aho>
http://html5doctor.com/html5-simplequiz-6-zeldmans-fat-footer/#comment-14585
20:52
<aho>
is there anything else one could use for image+caption?
20:53
<aho>
dl/dt/dd is kinda icky :>
20:59
<AryehGregor>
<div> and <span>?
20:59
<aho>
<div><img/><span>caption</span></div>... meh :>
21:00
<aho>
with <dl><dt><img/></dt><dd>caption</dd></dl> you got at least some relationship between the image and the caption
21:01
<aho>
i.e. they belong together
21:01
<aho>
whereas the former is just some image and oh there is also some span... might be the caption or some other text
21:02
<AryehGregor>
<div> implies there's some type of grouping relevant to its contents.
21:03
<aho>
that's only as strong as the outer dl in the second example
21:04
<aho>
it's like a pile of things vs a pile of things where two things are glued together :>
21:30
<annevk>
Hixie, no idea why impls do it that way, found it via http://www.w3.org/2006/webapi/track/issues/35
21:31
<annevk>
Hixie, I suspect there might be a few other things that work this way, but I have not tried figuring it out
21:31
<annevk>
Hixie, we could introduce some kind of "default handler" concept in DOM Core for this I suppose
21:47
<AryehGregor>
annevk, okay, I joined www-dom. If DOM discussion stays there, maybe I can drop public-webapps.
21:47
<AryehGregor>
Mostly stuff about IndexedDB and whatnot, it seems.
22:17
<uf0>
what's the using the keyword 'only' in: @media only screen and (max-device-width: 480px)
22:18
<uf0>
what's the use**
22:21
<Hixie>
annevk: can we change implementations?
22:23
<uf0>
I don't see the point of using 'only' media queries
22:23
<uf0>
because browsers that don't support MQ will ignore anyway
22:23
<uf0>
am I wrong here?
22:24
<uf0>
@media only screen and (max-device-width: 480px) vs @media screen and (max-device-width: 480px)
22:24
<Hixie>
uf0: according to html4, browsers that don't support MQ will treat "screen and (foo)" as "screen"
22:24
<stevela>
q
22:24
<stevela>
sorry - ignore me.
22:25
<uf0>
so are you advocating for me to use @media only screen and (max-device-width: 480px)
22:25
<uf0>
?
22:25
<uf0>
that's safer?
22:26
<Hixie>
only way to know what's safer is to test browsers
22:26
<tw2113>
use media queries to alter what would be the default view, and if they're not supported, have the regular version show up
22:27
<uf0>
tw2113 I'm specifically talking about the keyword 'only'
22:27
<uf0>
should it be left in or not
22:29
<uf0>
thanks Hixie.
22:39
<heycam>
annevk, yes file a bug please (for generic structured-cloneable object support)
22:39
<annevk>
Hixie, no idea really, maybe cc jonas and such
22:39
<annevk>
Hixie, and some other "old timers"
22:39
<annevk>
Hixie, they might know :)
22:40
<annevk>
heycam, will do tomorrow
22:40
<heycam>
annevk, how does html5 deal with? are there specific apis that take an object to do a structured clone of?
22:41
<annevk>
heycam, postMessage() and localStorage I guess
22:42
<annevk>
heycam, Jonas says Indexed DB might use it
22:42
<annevk>
heycam, for events I guess we only need the keys and their immediate values to be "sane"
22:42
<annevk>
heycam, as e.g. CustomEvent.details can take an arbitrary object as far as I can tell
22:43
<heycam>
annevk, for details though will it do the structured clone?
22:43
<heycam>
or just hold on a reference to that object?
22:43
<heycam>
if the latter, we don't need anything special
22:43
<heycam>
and the implementation won't be inspecting that object, it's only for script to use, right?
22:44
<annevk>
heycam, right, but that object can be nested in another one that does need some kind of "saneness"
22:44
<annevk>
heycam, see http://lists.w3.org/Archives/Public/www-dom/2011JanMar/0061.html
22:45
<heycam>
annevk, ah
22:45
<annevk>
heycam, so it's just a reference, but it can be inside an object where you need some kind of guarantees
22:46
<heycam>
annevk, so I think the way some of the DAP specs handled this is to define an interface (with [Callback] maybe?) with writable attributes for each of the keys you might expect on the object
22:46
<heycam>
annevk, that might not work if you don't know the set of keys you're expecting
22:46
<heycam>
s/might not/won't/
22:55
<annevk>
heycam, what I want is an object with an unknown set of keys that I can query in any order without side effects of which the values can be basically anything; i.e. numbers, objects, strings, etc.
22:56
<heycam>
annevk, ok
22:56
<heycam>
annevk, please mention that in the bug :)
22:57
<annevk>
yeah, tomorrow I will try to write it down
22:57
<zewt>
with deterministic side-effects, rather than with no side-effects
22:57
<annevk>
and make sure I add the Indexed DB people
22:57
<heycam>
zewt, yeah if the object has getters on it...
22:57
<zewt>
eg. evaluate getters into static properties
22:58
<annevk>
http://my.opera.com/haavard/blog/2011/03/04/doj
22:58
<zewt>
heh
23:00
<zewt>
<mpeg-la> we can't possibly tell you which of our patents you're infringing. we have way too many and they're all hundreds of pages each, you can't expect us to know them all!
23:02
<Hixie>
annevk: make sure to mention that we need to define what happens with custom getters on these things
23:02
<annevk>
yeah, that's the side effect thingie above
23:03
<annevk>
guess I'll just end up enumerating the various cases it needs to address and the constraints and let heycam solve it :)
23:04
<Hixie>
k
23:05
<annevk>
heycam, btw, http://html5.org/specs/dom-parsing.html
23:05
<heycam>
annevk, ah thanks
23:05
<heycam>
annevk, so that parses/serialises as either HTML/XML depending on the document mode
23:05
<annevk>
yeah
23:06
<annevk>
not in DOM Core to prevent lots of dependencies
23:06
<heycam>
I wonder if it would be useful to be able to get a definite parsing mode regardless of the document mode
23:06
<annevk>
DOMParser does that
23:06
<heycam>
annevk, ok. (though less convenient.)
23:06
<annevk>
except not yet it seems
23:08
<heycam>
annevk, when are you away?
23:09
<annevk>
anyway, bedtime
23:09
<annevk>
10 days from now
23:09
<heycam>
k
23:09
<heycam>
nn
23:09
<annevk>
nn