06:41
<annevk>
teehee, twitter profile pages are /{username} again
07:10
Ms2ger
loves how Hixie consistently used dom-cva-checkValidatity
07:17
<Hixie>
i use copy and paste a _lot_ :-)
07:17
<Hixie>
and emacs' dabbrev-expand
07:18
<Hixie>
(i even have that bound to my foot pedal)
07:22
Ms2ger
notes: "Essential to spec editing: a foot pedal"
09:54
<annevk>
can someone please check http://annevankesteren.nl/about for weird mistakes?
09:54
<annevk>
(also not so weird mistakes)
10:20
<deane>
annevk: Hi, try validating it
10:50
<annevk>
hmm, what's 'Unsupported character encoding name: charset=utf-8. Will sniff.'
10:51
<annevk>
sounds like a bug in the validator
10:51
<annevk>
MikeSmith: is that because you recently changed the regular expression the validator uses for encodings?
10:51
<annevk>
hsivonen: ^^
10:54
<annevk>
MikeSmith: hsivonen: I suspect it might be because I have no space between ";" and "charset", which is perfectly legal, but might be a case you have not considered
10:54
<deane>
annevk: yeah, try it at validator.w3.org and it works
10:55
<deane>
which would be an older revision of validator.nu
10:57
<annevk>
nice that I also manage to break validators
11:07
<annevk>
Ms2ger: has Microsoft forked DOM Parsing & Serlialization yet?
11:07
<Ms2ger>
Haven't seen anything yet
11:12
<annevk>
With standards it seems you really need to take a break that lasts longer than a couple of months before anything substantive has happened
11:14
<Ms2ger>
I'd rather you didn't :)
12:25
<MikeSmith>
MikeSmith: I actually a replaced that regular expression with a state machine
12:25
<MikeSmith>
annevk: ↑
12:26
<MikeSmith>
so yeah I probably forgot to deal with the no-space case there
12:26
<MikeSmith>
annevk: this is for the Content-Type header, right?
12:28
<MikeSmith>
annevk: ah actually it's working at http://validator.w3.org/nu/?doc=http%3A%2F%2Fannevankesteren.nl%2F
12:30
<MikeSmith>
which has the latest code, with that regexp replaced by the new parser
12:30
<MikeSmith>
so I think the error at validator.nu is because Henri has not pulled that new code yet
12:35
<annevk>
interesting
12:35
<annevk>
because I think I used to validate on validator.nu
12:35
<annevk>
thanks MikeSmith
12:36
<MikeSmith>
I'll take a look again the old regexp and see
12:37
<MikeSmith>
weird
12:37
<MikeSmith>
"^\\s*charset\\s*=\\s*(?:\"([^\"\\s]+)\")|([^\"\\s]+)\\s*$"
12:37
<MikeSmith>
is what is was
12:38
<MikeSmith>
so it should be working even if Henri's still running the old code
12:40
<MikeSmith>
and the case in the new code for it looks like this: https://gist.github.com/3478622
12:40
<MikeSmith>
so it should work either way as far as I tell
12:49
<annevk>
weird that it doesn't, then
13:08
<smaug____>
annevk: ping
13:09
<smaug____>
(welcome back)
13:09
<annevk>
hey smaug____
13:09
<smaug____>
annevk: I wonder the status of element.prepend() etc new methods
13:10
<smaug____>
there was some discussion to change before and after
13:10
<annevk>
I think they're all good, though I just read a thread between episodes about before/after
13:10
<annevk>
indeed
13:11
<annevk>
not that I agree with everything Alex writes, but the library argument http://infrequently.org/2012/08/inadmissible-arguments/ did not seem particularly compelling
13:11
<annevk>
the whole reason we added these was to have shorter names
13:11
<annevk>
(and ES6-style API)
13:11
<smaug____>
hmm, /me doesn't actually understand what all can be passed to those methods
13:12
<smaug____>
and in which way
13:12
<smaug____>
I guess element.prepend("asdfo", someotherelement, "bar");
13:12
<annevk>
that's like inserting a DocumentFragment consisting of a text, element, and a text node
13:13
<smaug____>
it is not, because the argument is something else
13:13
<smaug____>
I mean, is it an array
13:13
<smaug____>
or is it a list of separate parameters
13:13
smaug____
reads webidl..
13:13
<annevk>
it's not an array
13:14
<smaug____>
yup
13:16
<annevk>
and by stable I mean that everyone agreed to the design when we added it, but I don't think anyone gave the actual text a careful reading
13:16
<annevk>
smaug____: http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#mutation-method-macro takes care of converting stuff to a DocumentFragment
13:17
<smaug____>
I wonder who strings are handled
13:17
<annevk>
smaug____: you don't actually have to do that of course in your implementation
13:17
<smaug____>
how
13:17
<smaug____>
I mean, is toString() called
13:17
<annevk>
smaug____: WebIDL handles that
13:18
<smaug____>
I hope so
13:18
<annevk>
me too :)
13:18
<annevk>
but I'm pretty sure it does
13:18
smaug____
downloads D4 and WebIDL for offline reading... boarding in any minute
13:19
<annevk>
hehe
14:48
<zewt>
annevk: i think i strongly dislike the before/after names too; it sounds like a query, like "search for a node with this CSS selector starting with this element and moving backward", not like a mutator
14:49
<zewt>
fwiw, while i definitely agree that a lot of the old DOM names are way too long (getElementById needs to die), i think the insertBefore name is just fine
14:57
<zewt>
currently writing code to manipulate XML using minidom and it's making me a sad person
15:48
<jgraham>
zewt: Why would you do that?
15:49
<jgraham>
(also the problem I have with insertBefore is that it's on the wrong node, so I can never remember the order of the arguments parent.insertBefore(target, reference) is crazy compared to reference.addBefore(target)
15:49
<jgraham>
)
15:51
<jgraham>
(although I guess it could be target.addBefore(reference). So I kind of underminded my own point there. But at least you only need two nood references around, not 3)
15:51
<jgraham>
*node
15:52
<jgraham>
*undermined
15:52
<zewt>
(jgraham: because all XML implementations suck and from the ones I've used recently it has the right set of bugs)
15:54
<jgraham>
zewt: Wow you must have a weird set of requirements
15:54
<zewt>
well recently i've used it, ET and lxml
15:54
<zewt>
iirc lxml (or at least the ones installed in OSX by default, I think newer ones improve it) wouldn't pass through doctypes, don't recall the issue I hit with ET off-hand
15:56
<jgraham>
I pretty much exclusively use lxml, but happily almost never have to care about doctypes
15:56
<zewt>
i don't personally care, but I'm editing external data in-place so I don't want things messed with that don't have to be
15:56
<jgraham>
I see
15:56
Ms2ger
just uses JS
16:13
<zewt>
am I the only person who can never remember if it's "startswith" or "beginswith"?
16:14
<Ms2ger>
startswith
16:19
<jgraham>
zewt: Well that's one I have never had difficulty with
16:19
Philip`
finds it much easier to remember "$s =~ /^foo/" than "s.startswith('foo')"
16:19
Ms2ger
murders Philip`
16:20
<jgraham>
Philip` was strange
16:21
<zewt>
"starts" and "begins" are in the same hash bucket in my mind; something like "isprefixed" would have been more distinct, if a little uglier