07:53
<MikeSmith>
hsivonen: about the validator XHTML1 schema, I'm wondering if you remember why the xml:space attribute for the style element in the schema
07:53
<MikeSmith>
https://bitbucket.org/validator/validator/src/7625938e679a27265ae625e57b5ed33c063dfacb/schema/xhtml10/style.rnc?at=default#cl-44
07:54
<MikeSmith>
oh wait maybe I see why
07:56
<MikeSmith>
you added it to some other patterns
07:57
<MikeSmith>
anyway, I ask above if you remember why you zapped it from style
07:58
<MikeSmith>
I assume it's probably because you intended that it be included in some other pattern that the style schema inclluded
08:00
<MikeSmith>
but anyway at this point it doesn't, and the effect is that the validator flags any style element with an xml:space attribute as an error
08:03
<MikeSmith>
which is especially a problem if you turn the "load external entities" option on, because the XHTML1 DTD makes xml:space #FIXED 'preserve' for style
08:04
<MikeSmith>
so the XML parser adds it if it's not there already, and then the validator flags it as an error even though it's not even present in the input document (pre-parsing)
08:04
<MikeSmith>
which from a user perspective is pretty confusing
08:09
<MikeSmith>
OK yeah I see now that you created an "xml.space.attrib" pattern and added it to the "Core.attrib" pattern
08:11
<MikeSmith>
the problem is that because the style.attlist pattern doesn't include Core.attrib, that has the effect of xml:space no longer being allowed for the style element
08:14
<MikeSmith>
Ms2ger: do you know, is Servo still using hubbub for HTML parsing?
08:14
<Ms2ger>
Yup
08:14
<Ms2ger>
Want to finish the translation? :)
08:15
<MikeSmith>
heh
08:15
<MikeSmith>
thanks for the kind offer, but no thanks
08:16
<Ms2ger>
Aww
08:16
<Ms2ger>
Oh hey, tests for marquee
08:17
<MikeSmith>
nice
08:17
<MikeSmith>
who from?
08:17
<MikeSmith>
Ms2ger: btw, <!doctype html><p>Test
08:17
<MikeSmith>
in servo
08:17
<MikeSmith>
with no </p> end tag
08:17
<Ms2ger>
Intel
08:17
<MikeSmith>
hmm
08:18
<MikeSmith>
I think Intel may have some people with too much time on their hands then
08:18
<MikeSmith>
which person at Intel? Not Zhang?
08:19
<Ms2ger>
xiaojunwu, but I don't know if they wrote or just submitted them
08:19
<MikeSmith>
so can I assume that <!doctype html><p>Test fails to render as expected because of broken parsing? or some other reason?
08:20
<Ms2ger>
Looking
08:22
<Ms2ger>
Looks like we don't do a reflow
08:23
<MikeSmith>
well
08:23
<MikeSmith>
that's odd
08:24
<MikeSmith>
I wouldn't think that should be depending on the presence of the end tag in the input document
08:25
<Ms2ger>
Or not
08:26
<MikeSmith>
seems like it must be constructing a different DOM
08:27
<MikeSmith>
so please add an Inspect Element in Servo so I can inspect the DOM
08:27
<Ms2ger>
Heh
08:27
<MikeSmith>
in exchange I'll contribute some marquee tests for Servo
08:27
<Ms2ger>
DOM looks sane
08:27
<MikeSmith>
weird
08:28
<Ms2ger>
Oh no it doesn't
08:28
<Ms2ger>
There's no Text
08:29
<MikeSmith>
that would probably do it I guess
08:30
<Ms2ger>
File an issue?
08:30
<MikeSmith>
hai
08:30
<Ms2ger>
Not that I expect it to be fixed before we get the new parser
08:30
<MikeSmith>
this will be the first non-build-related bug I reported for Servo
08:30
<Ms2ger>
Which brings us to... :)
08:30
<MikeSmith>
heh
08:31
<MikeSmith>
first I'd need to learn Rust
08:31
<MikeSmith>
which I think would require me dosing up more than normal
08:32
<Ms2ger>
It's actually quite nice :)
08:32
<MikeSmith>
btw by file an issue you mean a bugzilla.mozilla bug? or a gh issue?
08:32
<Ms2ger>
Github
08:32
<MikeSmith>
oK
08:33
<Ms2ger>
Unfortunately
08:35
<MikeSmith>
yeah well I guess they're going to fully break new ground they might as well use a broken bug-reporting system too
08:35
<Ms2ger>
Ouch.
08:36
<MikeSmith>
well I was going to say if they're going to make a toy browser engine they might as well use a toy bug-reporting system
08:36
<MikeSmith>
but I decided not to say that
08:36
<MikeSmith>
because that would be really obnoxious
08:39
<Ms2ger>
Hey, Samsung might actually ship it
08:41
<MikeSmith>
good maybe they can have somebody finish the parser work
08:41
<MikeSmith>
it is going to be very cool when it does finally ship
08:42
<MikeSmith>
it's pretty crazy how far along it is after just a year
08:42
<Ms2ger>
Could use more people who know about the DOM, fwiw ;)
08:42
<MikeSmith>
plh!
08:43
<MikeSmith>
I won't tell him but I'll suggest to him that he learn Rust just for fun
08:43
<Ms2ger>
Heh
09:02
<MikeSmith>
yipes Mike West getting a little carried away in the blink "Intent to remove"-fest https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/14q_I06gwg8
09:02
<MikeSmith>
"what we currently do with the tag is strange, and doesn't seem to actually support it's usage in the wild.
09:03
<MikeSmith>
where what they are currently doing with the tag = conforming to what the HTML spec requires them to do with
09:04
<MikeSmith>
which is admittedly pretty strange
09:04
<MikeSmith>
but you'd hope engineers would be reading the specs before proposing to yank stuff
09:06
<MikeSmith>
especially when he's citing the spec in his message
10:11
Ms2ger
wants a version of http://simon.html5.org/tools/js/unicodeize/ that also gives the names of the characters
10:15
<zcorpan>
Ms2ger: fork it :-)
10:15
<Ms2ger>
Sorry
10:15
<Ms2ger>
I want someone else to make it :)
10:19
<zcorpan>
there's http://software.hixie.ch/utilities/cgi/unicode-decoder/character-identifier but i don't know if that helps
10:21
<zcorpan>
i wonder why https://critic.hoppipolla.co.uk/r/437 doesn't show the meta refresh commit
10:21
<Ms2ger>
Oh, that's neat too
10:22
<Ms2ger>
Even though it says (this script is currently broken)
10:22
<Ms2ger>
I'd fix it to match what I need, but it's perl :/
10:22
Ms2ger
specs PI serialization instead
10:26
Ms2ger
notices that Travis didn't serialize the target
10:38
<Ms2ger>
So, a PI with an empty data
10:38
<Ms2ger>
"<?a ?>" or "<?a?>"?
10:39
<jgraham>
zcorpan: Better?
10:39
<zcorpan>
jgraham: yep, thanks
10:39
<zcorpan>
Ms2ger: in xml?
10:40
<Ms2ger>
Yeah
10:41
<jgraham>
08:32 < Ms2ger> It's actually quite nice :)
10:41
<jgraham>
Ms2ger: One word: AbstractNode
10:42
<Ms2ger>
jgraham, we're busy removing that ;)
10:42
<zcorpan>
Ms2ger: what does ie do?
10:43
<jgraham>
Ms2ger: Sure, but the fact that it ever happened :)
10:44
<jgraham>
(I like Rust btw, but the inability to model the DOM without craziness is rather unfortunate)
10:44
<Ms2ger>
IE and Chrome have the space, Gecko doesn't have one
10:46
<zcorpan>
Ms2ger: then include the space :-)
10:46
<zcorpan>
http://www.browserstack.com/screenshots/b7d6431d229dc231b4dbd4948514028a882dfed6 :-|
10:47
<Ms2ger>
How about in HTML?
10:48
<Ms2ger>
Chrome and IE throw when you create a PI, Gecko serializes without a space
10:48
Ms2ger
tries something else
10:49
<zcorpan>
http://www.whatwg.org/specs/web-apps/current-work/multipage/the-end.html#html-fragment-serialization-algorithm defines that already, with space
10:55
<zcorpan>
if i understand the spec correctly then <embed type=image/svg+xml> should create a browsing context that is able to load basically anything into it
10:59
<annevk>
zcorpan: no, because that's not a plugin that the browser supports
11:01
<zcorpan>
annevk: hmm. in that case the resource has to have image/svg+xml content-type to create the browsing context. and then you can navigate to anything
11:01
<annevk>
zcorpan: yeah
11:02
<Ms2ger>
Hmm, HTML is interesting
11:02
<Ms2ger>
The HTML spec and Gecko do <?a >, IE and Chrome do <?a ?>
11:03
<Ms2ger>
zcorpan, so what do you think about that? :)
11:04
<zcorpan>
Ms2ger: i think the spec is right
11:05
<zcorpan>
Ms2ger: html ends the bogus comment on the first > and in sgml PIs end with >
11:06
<Ms2ger>
It's been a while since I heard SGML used as an argument here :)
11:06
<annevk>
Is this for serializing a processing instruction?
11:07
<Ms2ger>
Yeah
11:07
<annevk>
Just do the same as in XML?
11:07
<Ms2ger>
zcorpan disagrees :)
11:08
<annevk>
Either that or serialize as a plain comment
11:08
<annevk>
<!--target contents-->
11:08
Ms2ger
steps back to let the people who care fight it out
11:08
<zcorpan>
annevk: why?
11:08
<annevk>
zcorpan: same path as XML is simpler and more predictable, HTML comments generates less errors on a reparse
11:17
Ms2ger
will post to the list later
11:25
<hsivonen>
MikeSmith: Maybe I added xml:space to all elements and had to remove it from <style> to avoid the double attribute error from Jing
11:27
<MikeSmith>
hsivonen: as far as I can see in the current schema it's not been added to all elements. Because you added it to Common.attrib but <style> doesn't use Common.attrib
11:30
<hsivonen>
MikeSmith: OK. I have no idea what I was thinking, then
11:35
<MikeSmith>
hsivonen: Maybe you were thinking that the schema you started from was probably consistent and logical :)
11:36
<Ms2ger>
Ehehehe
11:37
<MikeSmith>
can't make a silk purse from a sow's ear
11:37
<MikeSmith>
since the specs the schema is modeling aren't consistent and logical
11:38
<Ms2ger>
That's an interesting expression
11:38
<hsivonen>
yeah, I don't think the schema writers are to blame in this case
11:38
<annevk>
Why do we want to allow xml:space though?
11:38
<annevk>
Seems like a bug
11:38
<hsivonen>
(and I don't mean myself but the earlier authors)
11:39
<annevk>
And why would <style> not have Common.attrib?
11:39
<hsivonen>
annevk: I may have been thinking that xml:space should be allowed throughout as part of the XML layer
11:40
<hsivonen>
annevk: I made the changeset MikeSmith is referring to back when I still tried to support xml:id, too
11:40
<annevk>
xml:id :-)
11:41
<hsivonen>
annevk: well, you believed in xml:id once, too
11:41
<annevk>
Oh yeah, was just laughing because I forgot about it
11:43
Ms2ger
goes off for lunch
11:44
<annevk>
Ms2ger: https://github.com/whatwg/domparsing/pull/1
11:44
<Ms2ger>
Huh
11:44
<Ms2ger>
I missed that
11:45
<Ms2ger>
Will merge it after lunch
11:48
<MikeSmith>
annevk: the reason for the schema not allowing Common.attrib for style is that that the XHTML1 spec doesn't allow style to have certain global attributes. Which makes sense because HTML4 didn't allow them for style either, and XHTML1 was supposedly just a reformulation of HTML4 in XML
11:49
<MikeSmith>
but then I notice that the XHTML1 spec allows style to have the id attribute, despite the fact that HTML4 didn't
11:51
<annevk>
Why do we still validate HTML4?
11:51
<MikeSmith>
annevk: and this is all just about XHTML1/HTML4 checking in the validator, so the reason to allow xml:space is because the XHTML1 spec allows it
11:51
<MikeSmith>
annevk: so I can retire the legacy W3C validator for once and all
11:51
<annevk>
Can't we retire both?
11:52
<annevk>
And wow, didn't know that was still around
11:52
<MikeSmith>
yeah, despite me doing my best to kill it
11:53
<MikeSmith>
anyway, the requirement I've been given is that I need to continue to provide support for checking HTML4 and XHTML1 documents
11:54
<borior>
annevk: hey there, I've been pointed in your direction after moaning a lot about the absence of a vendor-agnostic browser extension API and/or packaging format
11:54
<annevk>
MikeSmith: dude, you sound like Steven Pemberton now :p
11:54
<annevk>
"Our charter told us to build XHTML 2.0, so we did."
11:55
<borior>
I've been reading about the webapps wg. Where's the best place to get an overview of the current state of affairs and intended directions?
11:55
<annevk>
borior: I'm not sure anyone is working on standardized packaging at the moment, maybe the SysApps WG?
11:56
<annevk>
borior: I'm not a big fan of packaged apps: http://the-pastry-box-project.net/anne-van-kesteren/2013-december-2/
11:56
<borior>
right, neither am I
11:57
<annevk>
borior: as for extensions, that might make sense, but given that we have enough trouble as it is to agree on public APIs, standardizing internal APIs seems tough
11:57
<borior>
the use case I'm interested in at the moment is annotation
11:57
<borior>
which in technical terms, for me, means I'd like to authorise some code to run in an isolated execution context on every page i visit, with access to the DOM
11:58
<borior>
this doesn't require packaging, per se, but it would require some mechanism to install that intent (and code) into my UA
11:58
<borior>
which seems partially to be the intent of the mozApps stuff
11:59
<borior>
but hey, that's vendor-specific again, and it makes me sad :(
11:59
<MikeSmith>
annevk: I forgot to mention that I'll also be adding XForms support. When I'm done with that I plan to also include modularization and then, decentralized extensibility
12:00
<annevk>
MikeSmith: sounds like you'll be hitting your bonus targets
12:00
<annevk>
borior: yeah, I don't think we're even close on getting to some standardized solution for problems like that
12:01
<borior>
annevk: is is something worth raising in the webapps wg lists? or are you focused on features for specific (i.e. "has-a-url") pages?
12:01
<borior>
*it
12:02
<annevk>
borior: doesn't hurt to ask I suppose, but I don't know anyone trying to tackle that
12:02
<borior>
okay, that's really helpful, anyway
13:01
<annevk>
Domenic_: do we have a way yet to see how many tests a browser passes of the promises test suite?
13:01
<annevk>
Domenic_: even if not run through the Ecma test suite?
13:01
<annevk>
Domenic_: would be nice to see how much Gecko passes before we enable it by default in https://bugzilla.mozilla.org/show_bug.cgi?id=918806
13:33
<zcorpan>
anyone know why <embed name> is obsolete?
13:33
<zcorpan>
just don't want to embrace the browsing context aspect of it?
14:00
<annevk>
Okay, what's the context for the ice cream memes?
14:00
<annevk>
I'm not finding it on public-restrictedmedia :/
14:01
<zcorpan>
annevk: follow the link in the post
14:04
<annevk>
zcorpan: ta
14:05
<annevk>
zcorpan: I'm having a hard time following that
14:25
<zcorpan>
Hixie: the spec status annotation thing could have some more context in the password email so that it's easier to find
14:25
<darobin>
I really likes that ice cream analogy
14:25
<SimonSapin>
annevk: All URLs (including with non-relative scheme) have a fragment, but not a query?
14:25
<SimonSapin>
http://url.spec.whatwg.org/ says "An absolute URL must be a scheme, followed by ":", followed by scheme data, optionally followed by "?" and a query."
14:25
<annevk>
SimonSapin: URLs have a query
14:26
<SimonSapin>
data:text/html,Test?query#fragment shows ?query as part of the content
14:26
<zcorpan>
Hixie: like including terms like "whatwg", "html standard", "Specification annotation system"
14:27
<annevk>
SimonSapin: sure
14:27
<jgraham>
darobin: For the content, or for the fact that it mentions ice cream
14:27
<jgraham>
?
14:27
<darobin>
jgraham: yes
14:27
<annevk>
SimonSapin: content being scheme data, ?, query
14:27
<jgraham>
darobin: s/or/xor/
14:27
<SimonSapin>
hum, ok
14:27
<darobin>
s/yes/quantum superposition/
14:28
<Ms2ger>
s/darobin/cat/
14:28
<darobin>
Ms2ger: that's already a given :)
14:28
<annevk>
I thought it was dahut
14:28
<jgraham>
I couldn't really understand the content
14:28
Ms2ger
puts darobin in an unobservable box
14:28
<annevk>
But ice cream!
14:28
<annevk>
Ms2ger: I thought the box was observable
14:28
<jgraham>
I don't know who are supposed to be the equivalent of ice-cream vendors in this situation
14:28
<darobin>
annevk: no I *hunt* dahuts
14:29
<darobin>
Ms2ger: please include ice cream
14:29
<Ms2ger>
annevk, I was wondering who was going to be pedantic about it :)
14:29
<annevk>
darobin: must be some big ass cat
14:29
<Ms2ger>
darobin, there will be a superposition of ice cream and no ice cream, that's the best I can do
14:29
<darobin>
annevk: hey! I may have put on some weight, but I think "big ass" is stretching it
14:29
<darobin>
Ms2ger: I reckon that's superposably delicious
14:30
<jgraham>
(I also think that argument-by-analogy is very silly)
14:30
<jgraham>
(and generally leads to bad results)
14:31
<annevk>
Yeah, much better to argue through authority
14:31
<darobin>
jgraham: this is TimBL, you have to think hypertextually
14:31
<Ms2ger>
Well, if annevk says so...
14:31
<annevk>
I'm on the TAG, now shut up
14:32
<gsnedders>
jgraham: Do you have any opinion on what should be an attribute and what should be a method in Python?
14:32
<daleharvey>
So, anyone know who I should ask about crazy behaviour with websql on safari?
14:32
<darobin>
his point is linked to the part where he talks about only having Apple content on iOS, only having Sony content on Sony devices, etc.
14:32
<daleharvey>
I mean crazier than usual behaviour, obviously
14:32
<Ms2ger>
daleharvey, why are you using websql?
14:32
<annevk>
daleharvey: maybe ask in #webkit? WebSQL isn't really part of the web
14:32
<darobin>
same with ice cream with limited distribution if the fridges are all controlled by a given company
14:32
<daleharvey>
backwards compat :(
14:33
<jgraham>
darobin: That doesn't seem like a problem at the moment, and doesn't seem like it will become a problem in any scenario where EME could make a difference
14:33
<darobin>
I'm not sure I agree (the argument seems to rely on indie film makers really wanting DRM, I'm not sure that's the case), but within those parameters it has merit
14:33
<jgraham>
gsnedders: Are you actually asking about usedEncoding?
14:33
<gsnedders>
jgraham: Yes.
14:33
<daleharvey>
annevk: cool cheers
14:33
<jgraham>
gsnedders: I think parser.encoding could be fine
14:33
<darobin>
jgraham: yeah, I'm not saying that it reaches the heart of the problem, just that it makes sense as a problem *if* it is a problem
14:35
<jgraham>
gsnedders: (that is, perhaps, a bit misleading; one could argue that it should be documentEncoding or something. But shortness seems desirable)
14:36
<SimonSapin>
jgraham: there is naming, and there is property vs. method
14:37
<SimonSapin>
Properties look like attributes, which are cheap to access. So properties that are expensive (in run time) to access are bad, but this is not the case here.
14:37
<jgraham>
SimonSapin: I don't see ehy it would be a method in this case
14:37
<SimonSapin>
ok
14:37
<jgraham>
It just reflect a lower-level attribute
14:39
<SimonSapin>
for naming, I wanted it to show that it may be different from the given encoding parameter (eg. if there is a BOM)
14:39
<hsivonen>
zcorpan: if you don't like the WebKit/Blink hack for x-user-defined in <meta>, do you prefer my other option at https://bugzilla.mozilla.org/show_bug.cgi?id=213517#c13 ?
14:40
<gsnedders>
SimonSapin: See my recent bug on needing to split the encoding argument up
14:40
<SimonSapin>
yes
14:40
<SimonSapin>
it’s related
14:41
<gsnedders>
hsivonen: +1 for WebKit/Blink behaviour
14:43
<gsnedders>
jgraham: I share SimonSapin's concern about ambiguity given encoding parameter
14:43
<jgraham>
shrug
14:43
<hsivonen>
gsnedders: can you elaborate on why you prefer the WebKit/Blink behavior?
14:44
<jgraham>
If you think it's a big deal, use documentEncoding or something
14:45
<zcorpan>
hsivonen: i think i prefer special-casing XHR in terms of being less insane, but i don't care strongly either way, and i don't have an opinion about the font/UI issue
14:47
<hsivonen>
zcorpan: I think the font/UI issue is the deciding factor
14:47
<hsivonen>
zcorpan: though I dislike the craziness of HTTP-level x-user-defined in the WebKit/Blink case
14:48
<zcorpan>
hsivonen: so for the record let's consider #3 remove mystery UI, make x-user-defined alias of windows-1252 everywhere except XHR and in XHR it maps to the x-user-defined encoding
14:48
<SimonSapin>
gsnedders: I have http://pythonhosted.org/webencodings/ that implements http://encoding.spec.whatwg.org/ , if that helps
14:48
<gsnedders>
SimonSapin: Change it to be an attribute (possibly just a getter?) and probably go for documentEncoding? Either that or follow CSS WG habit and call it "bikeshed"? :)
14:49
<gsnedders>
SimonSapin: Can you file a bug somewhere along those lines?
14:55
<SimonSapin>
gsnedders: done and done
14:57
<gsnedders>
SimonSapin: I've filed a bug on webencodings, btw
15:00
<SimonSapin>
gsnedders: when is getstate/setstate it useful?
15:11
hsivonen
sees TimBL's DRM reply on www-tag
15:11
<hsivonen>
whoa
15:17
<gsnedders>
SimonSapin: I've used it in implementation of a decoder than can change the currently used decoder while decoding.
15:18
<gsnedders>
SimonSapin: For the sake of changing when you have a meta beyond the pre-scan and need to change the encoding of the input stream.
15:20
<Domenic_>
annevk: I think I have time to work on that tonight, should be relatively straightforward.
15:44
<darobin>
Domenic_: if I were a friend I'd tell you to run away fast; but since I don't really know you that well: best of luck in the elections :)
15:46
<jgraham>
darobin: misanthrope
15:47
<darobin>
jgraham: hey, I've been there, I'm allowed to relish the pain of others who stumble down the same path
15:47
<darobin>
though to be honest, the TAG just ain't what it used to be :)
16:49
<JonathanNeal>
Hello!
16:50
<Domenic_>
darobin: thanks/haha :). where was the announcement? Haven't seen it yet.
16:51
<darobin>
Domenic_: it's coming out soon
16:53
<Domenic_>
should probably fire up ye old blog engine then...
17:13
<darobin>
Domenic_: yeah, get the blog warmed up, and ping the band manager when you have something!
17:13
<Domenic_>
installing ruby on work computer.... @_____@ why is everything so hard
17:26
jgraham
wonders what you call a group of managers
17:26
<Ms2ger>
A herd?
17:26
<Ms2ger>
A meeting?
17:27
<jgraham>
I was thinking "Committee"
17:27
<jgraham>
I had an object called a ManagerPool
17:27
<jgraham>
But it isn't really a pool
17:27
<Ms2ger>
ManagerMeeting? Sounds good to me :)
17:57
<annevk>
Domenic_: sweet
19:14
<Hixie>
smaug____: pong
19:14
<Ms2ger>
Hixie, wdyt about PI serialization in HTML, btw?
19:17
<smaug____>
Hixie: the question was about http://logbot.glob.com.au/?c=freenode%23whatwg&s=1%20Dec%202013&e=1%20Dec%202013#c842854
19:22
<Hixie>
Ms2ger: do PIs still exist?
19:22
<Hixie>
smaug____: looking...
19:22
<Hixie>
smaug____: it's defined in various places
19:22
<Hixie>
smaug____: usually when a Document is created
19:23
<Hixie>
smaug____: though i think there's at least one case (document.open, iirc) where the assignment changes on the fly
19:23
<smaug____>
especially "the Window object with which document is currently associated."
19:23
<smaug____>
I couldn't find where it is defined
19:23
<smaug____>
but perhaps I was searching for wrong key words
19:23
<Ms2ger>
Hixie, sure
19:24
<Hixie>
smaug____: "Create a new Window object, and associate it with the Document" at #create-a-document-object is the main place
19:25
<Hixie>
smaug____: looks like the about:blank document never gets explicitly associated, let me file a bug to fix that
19:25
<smaug____>
And data documents aren't associated either, right?
19:25
<smaug____>
which is why I started to wonder http://logbot.glob.com.au/?c=freenode%23whatwg&s=1%20Dec%202013&e=1%20Dec%202013#c842849
19:25
<Hixie>
smaug____: and document.open() says "Replace the Document's singleton objects with new instances of those objects. (This includes in particular the Window [...]"
19:26
<Hixie>
smaug____: some documents don't have an associated window, right
19:26
<Hixie>
smaug____: an interesting question
19:27
<Hixie>
smaug____: looks like "get the current value of the event handler" doesn't check for that case, oops
19:28
<smaug____>
http://www.whatwg.org/specs/web-apps/current-work/#getting-the-current-value-of-the-event-handler would have a null pointer error, effectively
19:28
<Hixie>
yeah
19:28
<Hixie>
ok, i filed a bug on that too
19:28
<Hixie>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=23957 https://www.w3.org/Bugs/Public/show_bug.cgi?id=23956
19:29
<smaug____>
I wonder what behavior we want for the latter one
19:30
smaug____
should test non-Gecko browsers
20:09
<aleray>
Hi, in html5lib python, how to choose between etree, lxml and dom for treewalker?
20:21
<TabAtkins>
aleray: Just look at the libraries and figure out which one has the things you want?
20:22
<TabAtkins>
I use the lxml treewalker. It's not a good API at all (I've had to write several DOM compat shims for missing functions), but it's compatible with CSSSelect, another library that converts CSS Selectors into XPath for me.
20:30
<aleray>
TabAtkins, thanks
20:56
<Domenic_>
Man that @w3ctag twitter account is having a hard time.
20:58
<TabAtkins>
Domenic_: What's wrong?
20:58
<Domenic_>
First it got the wrong date, then the wrong link
20:58
<TabAtkins>
Ah.
22:00
<Hixie>
so does anyone have an opinion on cite=""?
22:00
<Hixie>
the obvious answer is "let's drop it, nobody uses it"
22:01
<Hixie>
but that seems a bit sad
22:01
<Hixie>
even the spec uses them
22:01
<Hixie>
(by "nobody uses it" i meant no UAs make use of it)
22:02
<Hixie>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=22907
22:06
<tantek>
Hixie - I've been leaning more and more negative on cite="" (presuming you mean on q and blockquote)
22:06
<tantek>
cite="" feels very longdescy
22:07
<tantek>
attributes to URLs which don't cause any noticeable effect in the browser tend to either rot or become polluted with noise.
22:07
<tantek>
it's an invisible metadata problem
22:07
<Hixie>
its main difference with longdesc="" is that since there's been no advocacy for it, the data isn't unusually bad
22:07
<Hixie>
but yeah
22:08
<tantek>
Hixie, and I've *tried* to use it too (and often still do) but often wonder what's the point.
22:08
<Hixie>
ditto
22:08
<tantek>
It's like, oh maybe I'll make use of this in some script or JS thing eventually...
22:08
<Hixie>
there's been a note in the spec for many years suggesting adding a <credit> element for <figure>, which could take a similar role if used with a <a> in <q>/<blockquote>
22:08
<tantek>
but that never seems to happen :(
22:09
<Hixie>
that would make it less hidden-metadatay
22:09
<tantek>
nah, since it's a URL it should be something that works with <a href>
22:09
<tantek>
an already visible URL building block
22:09
<Hixie>
right that's what i mean
22:10
<tantek>
I mean not a new element
22:10
<tantek>
maybe someway to tie the q/blockquote with the <a href> to the source.
22:10
<Hixie>
something like <blockquote> ... <credit> Quoted from <a href="...">foo bar</a>, copyright 2004. </credit> </blockquote>
22:10
<tantek>
similar to <label for> maybe
22:10
<tantek>
oh just by nesting? perhaps
22:11
<Hixie>
a lot of people really want some sort of nesting for blockquote
22:11
<Hixie>
(not sure why <figure> isn't enough, but that's another issue -- and credit/a-href could work there too)
22:11
<tantek>
I really don't know what that means "want some sort of nesting for blockquote"
22:11
<tantek>
I've tried using <figure> and it's been awkward enough that I gave up.
22:11
<tantek>
I can try again I suppose.
22:12
<Hixie>
what's awkward about it?
22:12
<Hixie>
by "some sort of nesting" i mean people want the <blockquote> to contain the citation as well somehow
22:12
<tantek>
the markup just didn't seem to fit into how I was authoring the images / text
22:12
<Hixie>
ah
22:12
<Hixie>
would be interesting to study that
22:12
<tantek>
yeah
22:12
<tantek>
I should gather some more detailed data/feedback
22:13
<Hixie>
<figure> was designed to just take the place of the <div> people were using anyway
22:13
<tantek>
but wanted to at least share that rough impression
22:13
<tantek>
as (un)helpful as that is
22:13
<tantek>
typically these are all example I run across when writing blog posts
22:13
tantek
is still sad Hixie has switched to blogging on G+ rather than ln.hixie.ch
22:14
<Hixie>
g+ gave me per-post acls, which i've wanted for years
22:15
<Hixie>
g+ sucks when you want inline links or diagrams though
22:16
<tantek>
Hixie - maybe only post on G+ when you need an ACL then, and otherwise post your "public" posts on ln.hixie.ch?
22:16
<mrbkap>
dglazkov: ping?
22:17
<tantek>
Hixie - you could always add G+ sign-in to ln.hixie.ch to show ACL'd posts there ;)
22:17
<Hixie>
tantek: g+ also gives me other things, like comments, spam filtering on comments, etc
22:17
<tantek>
Hixie, I'll get back to you once we've solved the blog comment spam problem :)
22:19
<Hixie>
heh
22:20
<Hixie>
good luck with that
22:21
<Hixie>
(i'm also pretty sure far more people read my stuff now than did on my blog)
22:23
<tantek>
yeah that may be a harder problem than the spam problem (reach/distribution)
22:35
<Hixie>
there's dozens and dozens of features that people have asked for in HTML that, before I can progress on them, I need to figure out whether anyone cares about
22:58
<tantek>
Hixie - you mean you're not able to pre-occupy them by asking them to research and document existing publishing practices that demonstrate the need? ;)
22:59
<tantek>
(existing publishing practices on *others* websites - helps demonstrate that some *other* people *might* care about it)
22:59
<tantek>
as opposed to every armchair tag proposer
23:06
<Hixie>
tantek: i'm not able to determine if there's people who want the feature or not, because they aren't aware of the bug and therefore haven't commented on it in the first place
23:06
<Hixie>
tantek: i'm oddly in the opposite problem than the one we usually find ourselves in -- not enough people saying they want something :-)
23:11
<tantek>
Hixie - that's a fine situation to just let time accrue interest if there is any. Such things can be left in limbo until there's demand.
23:11
<Hixie>
tantek: my problem is that i don't know if there's demand, because i don't know where the people who want these features are saying so
23:14
<tantek>
Hixie - sure - hopefully if such people exist they are searching for such a feature, and if you've documented the ideas/thoughts on the feature so far somewhere search discoverable, hopefully they'll find it.
23:14
<tantek>
praise be upon web search to see all, know all, find all, solve all. ;)
23:15
<Hixie>
tantek: i don't think bugzilla is searchable from google
23:15
<tantek>
well then that would be a problem
23:15
<Hixie>
tantek: also, it's also possible that the people who want this feature are finding it in some other inferior solution
23:15
<Hixie>
tantek: e.g. suppose someone wanted something from microformats, but found rdfa could do it in a pinch, and so stopped looking for how to do it in microfarmats
23:18
<tantek>
Hixie - that would be fine - if there really was sufficient such interest in such a "something" then eventually we'd see publishing patterns emerge of that kind of RDFa which we could use as data to develop a cleaner/simpler way to do it with microformats, document a transition, and go.
23:18
<Hixie>
would be nice to hear about it earlier on though
23:19
<tantek>
it's ok, we can accept the web as highly organically evolving and somewhat inefficient
23:19
<tantek>
the really important stuff will bubble up to the top
23:19
<Hixie>
i would like it to be more efficient :-)
23:19
<Hixie>
or at least, i'd like to know which of these many features are stuff nobody cares about, and which are things that lots of people are bitching about but i haven't heard enough complaints to know
23:19
<tantek>
me too. hence I ask people to post things places where Google and other search engines will pick them up, index them, and make them easily findable.
23:20
<tantek>
in my experience, amount of bitching does not actually related to utility / usefulness
23:21
<tantek>
and the good thing is, we can encourage people to "bitch" (I prefer brainstorm ;) ) on their on blogs
23:21
<tantek>
*own blogs