13:48
<asmodai>
holy leaking memory batman *eyes firefox curiously*
17:18
<Hixie>
why was 'mod' removed from calc()?
21:19
<rniwa>
AryehGregor: are you here?
22:04
<rniwa>
annevk: hi anne
22:07
<annevk>
hey rniwa
22:07
<rniwa>
annevk: how are you doing?
22:08
<annevk>
pretty good, bit sleepy :)
22:08
<rniwa>
annevk: I see. I'm a bit confused about the whole discussion on HTMLCollection
22:08
<annevk>
still trying to get used to the cold after my shortish break
22:09
<rniwa>
annevk: oh oops, that sucks.
22:09
<rniwa>
annevk: hopefully it gets better.
22:09
<annevk>
well I think sicking wants to add new members to e.g. NodeList at some point
22:09
<rniwa>
annevk: yeah, that makes sense.
22:09
<rniwa>
annevk: but do you think he's opposed to make getElement*by* return HTMLCollection?
22:10
<annevk>
HTMLCollection has a named getter which would make that difficult if e.g. someone used an ID anywhere on the web and that matches the new member name while also using HTMLCollection in some way
22:10
<rniwa>
annevk: as I noted on the thread, I filed a WebKit bug and was intending to work on it this weekend or so
22:10
<rniwa>
but it sounded as if jonas doesn't want that.
22:11
<annevk>
I think he was somewhat opposed to getElementsByClassName, not the others
22:12
<rniwa>
annevk: I see.
22:12
<rniwa>
annevk: FWIW, we're never got complaints about named getters so maybe we can disable that feature?
22:12
<rniwa>
we've*
22:13
<annevk>
I think even WebKit has a named getter for getElementsByTagName
22:13
<annevk>
bz pointed that out
22:13
<rniwa>
annevk: ah, you're right.
22:13
<rniwa>
annevk: we had that extension on NodeList :(
22:13
<annevk>
yeah so this change would mean you can try to clean that up
22:14
<annevk>
gonna go watch some parks and recreation
22:14
<rniwa>
annevk: I will definitely do that assuming there's no compat. issues
22:15
<rniwa>
annevk: alright. ttyl.
22:16
rniwa
is so confused by Objective-C syntax
22:25
<zewt>
hmm, would be nice to get HTML's use of "default action" updated to use DOM4 terms (canceled flag) ... low editorial priority, i assume
22:26
<zewt>
rniwa: objc is one of the few languages that simply doesn't make even the most basic intuitive sense to me
22:26
<zewt>
putting aside weirdo functional languages, I guess
22:27
<rniwa>
zewt: like prolog?
22:28
<rniwa>
zewt: I mean it's a declarative language. but I find declarative programming languages even less intuitive than functional programming languages.
22:29
<zewt>
annevk: is the preferred way to look at "canceled flag" after dispatch to refer to the flag directly, or to use the return value of the "dispatch" algo
22:29
<zewt>
the latter seems more natural to me (but then there'd be no reason to have the return value at all, I think)
22:30
<zewt>
er
22:30
<zewt>
the former looks more natural (using the flag directly)
22:36
<annevk>
zewt: you'd use fire so there's no return value I think
22:38
<annevk>
could maybe move the return value to dispatchEvent()
22:38
<zewt>
"fire" only lets you create a basic Event, for events with other interfaces, etc. you want dispatch directly, right?
22:38
<zewt>
that's what i was thinking
22:38
<annevk>
no fire is also used for events with other interfaces
22:38
<annevk>
you'd use a modified fire definition
22:38
<zewt>
how? sounds contradictory
22:39
<annevk>
http://dev.w3.org/2006/webapi/progress/#firing-events-using-the-progressevent-interface-for-http
22:39
<annevk>
because otherwise you also need to set the trusted flag, etc.
22:39
<zewt>
yeah that looks contradictory
22:39
<zewt>
it's saying "dispatch an event using the Event interface using the ProgressEvent interface" :)
22:40
<zewt>
not that it's not obvious what's really meant
22:40
<annevk>
not if you read it in the way that it's overriding
22:40
<annevk>
which is more logical imo
22:43
<zewt>
replying to clipboard thread suggesting:
22:43
<zewt>
7. Fire an event using the ClipboardEvent interface, with its type attribute initialized to e and its clipboardData attribute initialized to data transfer, at event target. If the event's canceled flag is set then abort these steps.
22:47
<annevk>
the if statement should prolly be step 8 for clarity
22:47
<zewt>
it was before, i moved it into 7 based on http://www.whatwg.org/specs/web-apps/current-work/multipage/association-of-controls-and-forms.html#form-submission-algorithm step 6
22:48
<annevk>
hmm
22:48
<zewt>
seems like having a concise pattern for this is good ... not sure if it's unclear or not
22:49
<annevk>
maybe I should postpone encoding stuff a bit more and work on DOM next week too
22:51
<zewt>
let me know if you think strongly one way or the other; don't want to bait people like paul by going back and forth
22:52
<annevk>
I don't, it's something that can be mended easily over time
22:53
<annevk>
especially if we make specs more hackable
22:53
<zewt>
("this spec hasn't been set in stone so it can start getting out of date, so everyone should use a useless, confusing, decade-old spec" sigh)
22:56
<annevk>
yeah i don't know what's up with that
22:56
<annevk>
i can sort of see the i want-to-play-by-silly-rules-so-i-cannot-reference-what-implementors-actually-need, but if you need to reference the newer draft anyway, also referencing crappy material is just wrong
22:57
<annevk>
that you even need to argue such a point because of some hearsay about "hot debate"
22:57
<annevk>
ugh