02:00
<Hixie>
hm. opera and webkit render this poorly due to an HTML parsing issue: http://www.thezorklibrary.com/history/00-encyclopedia.html
02:00
<Hixie>
gecko doesn't
02:00
<Hixie>
gecko bug? spec bug?
02:00
<Hixie>
opera and webkit bug?
02:12
<zewt>
fine in ie9
07:06
<zcorpan>
Hixie: did they talk to you before removing the license paragraph?
07:07
<Hixie>
no
07:26
<zcorpan>
booo
07:28
<MarkusDBX>
Should html/css validation be used in the year of 2012? In that case... what is the best validator to date? That follows the WHATWG version of html? Thank you.
07:30
<zcorpan>
why would the year make a difference? a validator is a quality assurance tool (like a spell checker)
07:31
<zcorpan>
http://validator.nu/ for html validation
07:32
<zcorpan>
i can't recommend a css validator but browsers usually flag things they think are invalid in css in their error consoles, so that's what i usually use to find errors in my css
07:35
<MarkusDBX>
zcorpan: sounds reasonable. I just mentioned the year, since the html-spec has changed somewhat i the last years. I used to use the w3c validator, but in the last years we have just used the browser-errors.
07:36
<MarkusDBX>
zcorpan: I now look for even better QA to make our clients even more happy.
07:36
<zcorpan>
yeah the validator isn't perfectly up-to-date with the spec, but then again browsers also aren't
07:37
<zcorpan>
if you find bugs in the validator, file it here: http://bugzilla.validator.nu/ :-)
07:37
<MarkusDBX>
Is chrome or ff the better validator?
07:37
<zcorpan>
for css?
07:38
<MarkusDBX>
for css.
07:38
<zcorpan>
i'd recommend checking in multiple browsers since they might not all support the same set of features
07:39
<MarkusDBX>
ah. I see
07:39
<MarkusDBX>
zcorpan: any other good advice to improve QA?
07:39
<zcorpan>
MarkusDBX: http://blog.whatwg.org/quality-assurance-tools-for-html5
07:40
<MarkusDBX>
great
07:40
<MarkusDBX>
zcorpan: thanks alot
07:42
<zcorpan>
MarkusDBX: on a different level, i'd recommend user testing to get better user experience, and automated regression testing so you get to know when something stops working before your users find out and someone tells you instead of going to your competitor :-)
07:45
<a-ja>
Hixie: anyone bringing up issues with <track default> ?
07:49
<a-ja>
Hixie: asking cuz I was trying out a webvtt polyfill that would only use chapters if default'ed...so one would have to have two defaults specified....one for subtitles/captions and another for chapters.
07:51
<a-ja>
Hixie: which suggests a use case......to select language for chapters as well as subtitles/captions
07:51
<a-ja>
Hixie: thoughts?
07:52
<zcorpan>
a-ja: there's a bug
07:53
<zcorpan>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=17877
07:53
<a-ja>
zcorpan:ah....musta missed it. #?
07:53
<a-ja>
k...tks
07:54
<zcorpan>
a-ja: for now, i'd recommend explicitly setting .mode instead of relying on default=""
07:55
<a-ja>
.mode ??? guess i have some more reading to do.
07:57
<zcorpan>
http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#dom-texttrack-mode
07:58
<zcorpan>
(if you want to expose the chapters yourself and don't want the browser to render them somehow, use "hidden")
08:06
<a-ja>
zcorpan: mode's not an attribute of the track element, but changeable in DOM via js?
08:06
<zcorpan>
right
08:06
<zcorpan>
track_element.track.mode = 'hidden'
08:06
<a-ja>
hrm
08:07
<zcorpan>
setting mode triggers the track to load
08:08
<zcorpan>
see http://simon.html5.org/presentations/foms-2012/ for an example
08:11
<a-ja>
zcorpan: leanback player's the one that required default on chapter track for it to show up in its UI, iirc
08:12
a-ja
is really trying to not have to use js
08:12
<zcorpan>
a-ja: ok, well, until the spec has fixed its rules around default="", i'd recommend leanback player to hook in to a custom data- attribute and set mode explicitly :-)
08:13
<a-ja>
zcorpan: that'd make sense.....esp since it's already using data- for fallback stuff
08:16
<zcorpan>
a-ja: is http://discussion.leanbackplayer.com/categories/development their bug tracker?
08:16
<a-ja>
no biggie for me...strictly for my intranet use...so i don't have to rely on polyfills or even flash fallback. just thought perhaps i'd stumbled on a possibly unknown use case for default-per-type
08:19
<a-ja>
zcorpan: dunno...just tried a few different polyfills briefly (before abandoning them as superfluous for my needs)
08:40
<zcorpan>
a-ja: filed http://discussion.leanbackplayer.com/discussion/153/default-attribute-for-chapters
08:43
<a-ja>
zcorpan: fwiw, there's a new beta out this month....i likely have the prior beta
08:43
<zcorpan>
ok. i haven't tried it at all, i've just stolen their captions for my demo :-P
10:09
<annevk>
karlcow: watching them more or less in chronological order
10:09
<annevk>
karlcow: at least the purported trilogy (still need to see 2046)
11:09
<zcorpan>
Hixie: minimized to http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1757
11:11
<zcorpan>
seems it's related to the "don't reconstruct more than three identical elements" thing
11:14
<zcorpan>
i can't wrap my head around AAA so i can't tell if it's a spec bug or not
11:14
<zcorpan>
but i'll file a opera bug
12:18
<zcorpan>
sigh. why do web shops send me my password in plain text in email?
12:20
<zcorpan>
and also "security tips" like i should change my password every 6 months
12:34
<annevk>
sweet
12:34
<annevk>
the github setup works great
13:16
<annevk>
Hixie: when you're awake again, can you give me encoding.spec.whatwg.org / xhr.spec.whatwg.org / fullscreen.spec.whatwg.org
13:53
<zcorpan>
Hixie: apparently the spec is right (re AAA)
14:04
<zcorpan>
(filed https://bugs.webkit.org/show_bug.cgi?id=96385 )
14:34
<annevk>
Hixie: might as well create url.spec and fetch.spec (for CORS, later fetch) while you're at it
14:35
<annevk>
Hixie: thinking about merging Progress Events into XMLHttpRequest
14:35
<annevk>
dunno about Web Notifications
14:35
<annevk>
anyone with opinions?
14:36
<darobin>
annevk: would it change anything other than document organisation?
14:36
<annevk>
darobin: the merging?
14:37
<darobin>
yeah
14:37
<annevk>
not really, mostly easier for me
14:37
<darobin>
then *shrug*
14:37
<annevk>
well yeah, I actually meant about Web Notifications
14:38
<annevk>
I guess we can fork that too, but maybe wait a bit to see what happens first
14:38
<annevk>
there's no hurry
14:38
<darobin>
ah, gotcha
14:39
<annevk>
Hixie: interested in using https://github.com/whatwg too?
14:39
<annevk>
I suppose it should handle things as large as HTML
14:40
<darobin>
annevk: it handles them just fine :)
14:41
<darobin>
in fact, it's even fucking fast, considering
14:41
<annevk>
ah yeah, you guys are using github too
14:41
<annevk>
sweet
14:42
<jgraham>
Hmm, I tried importing the HTML spec into git once, with the full history
14:42
<darobin>
annevk: to make your life easier you could even just clone https://github.com/w3c/html/tree/feature/whatwg ;-)
14:42
<jgraham>
and it worked fine
14:42
<jgraham>
But it was *big*
14:42
<jgraham>
and github didn;t like it
14:42
<darobin>
the first import takes a while
14:42
<darobin>
jgraham: the W3C repo has all the SVN history from WHATWG
14:43
<darobin>
in fact, in theory we could all use the same repo just with different branches
14:43
darobin
puts flowers in his head
14:43
<annevk>
heh, fork the W3C copy of the WHATWG copy of HTML to WHATWG
14:43
<darobin>
:)
14:47
<jgraham>
darobin: Pretty sure you are only suppose to put the flowers in your hair, not try to penetrate the skull
14:47
<darobin>
jgraham: you're such a boring square
15:43
<scor>
where in the microdta spec does it say that a property cannot be used in its ful URI form if the vocabulary specifying it does not allow it?
16:16
<dglazkov>
good morning, Whatwg!
16:19
<david_carlisle>
TabAtkins: Yes please we (mathematicians everywhere) want to be able to do commutative diagrams by just naturally nesting mathml in svg in html, it almost but not quite "just works"
16:49
<Hixie>
scor: properties don't have "full URI forms" unlsess a vocabulary specifies it
16:49
<Hixie>
annevk: what username should i use for you, assuming we're not going through the hoops we went through last time?
16:50
<Hixie>
annevk: re github, i have no reason to use github
16:50
<scor>
Hixie: ok, so the vocab author would have to specify both the short form and the full URI for both of them to be usable and valid?
16:50
<Hixie>
scor: yeah, though i wouldn't recommend it. You could end up with items that have two values for the property.
16:51
<Hixie>
scor: you'd really have to define them as separate properties and say which one takes preference, etc
16:51
<jgraham>
Hixie: (FWIW if you used github it would be better for me in a few small ways e.g. the revision list would be better than annevk's thing)
16:51
<scor>
for example one could define both 'http://example.com/fn'; and 'fn'
16:51
<scor>
and have the same definition for both
16:51
<jgraham>
(so that is one small reason)
16:51
<Hixie>
jgraham: i think darobin's crew has that covered already
16:52
<Hixie>
scor: say you defined both of those to mean "biological mother"
16:52
<jgraham>
Hixie: Only to the extent that they pull in all the cvhanges you make
16:52
<scor>
Hixie: but I understand that even if both were defined, values using each would not be equivalent
16:52
<Hixie>
scor: and an item had both set, pointing at different other items...
16:52
<jgraham>
Which theya re sure not to
16:52
<Hixie>
scor: you'd just have to say which takes precedence (and say that it's not conforming to have both, so validators of that vocabulary could complain)
16:52
<scor>
ok
16:52
<Hixie>
jgraham: they have a branch on which they are doing so, i believe
16:53
<scor>
Hixie: now, where in the spec does it say that it is not valid to use a full URI for a property if the vocabulary does not define it?
16:53
<jgraham>
So they have one branch that is precisely the WHATWG spec?
16:53
<jgraham>
Interesting
16:53
<scor>
or is that implicit?
16:53
<Hixie>
scor: the spec says it's not valid to use a property that isn't defined, iirc
16:53
<Hixie>
scor: if you have a vocabulary type
16:54
<scor>
Hixie: I'm looking at http://www.whatwg.org/specs/web-apps/current-work/multipage/microdata.html#selecting-names-when-defining-vocabularies
16:54
<scor>
(hope it's the right place)
16:54
<Hixie>
That section is non-normative
16:54
<Hixie>
you want http://www.whatwg.org/specs/web-apps/current-work/multipage/microdata.html#names:-the-itemprop-attribute for the nitty gritty
16:55
<Hixie>
hm, i guess it does allow URLs always
16:55
<Hixie>
interesting
16:55
<scor>
is it a bug?
16:55
<scor>
should it disallow them?
16:55
<scor>
(if not defined in the vocab)
16:55
<Hixie>
yeah, it's a bug. It should only allow URLs that are defined to be properties.
16:56
<Hixie>
same as itemtype=""
16:57
<Hixie>
filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=18837
16:57
Hixie
pokes annevk
17:00
<annevk>
Hixie: if annevk is taken, anne or annevankesteren
17:01
<annevk>
Hixie: or avankest for some irony
17:01
<Hixie>
i think annevk is taken by you
17:02
<Hixie>
anne is taken too, now trying annevankesteren
17:02
<Hixie>
ok, created user. now the domains
17:02
<Hixie>
annevk: which subdomains did you want again? can you list them all, comma separated?
17:03
<annevk>
url.spec.whatwg.org, xhr.spec.whatwg.org, encoding.spec.whatwg.org, fullscreen.spec.whatwg.org, fetch.spec.whatwg.org
17:03
<Hixie>
url, xhr, encoding, fullscreen, and fetch. ok
17:04
<Hixie>
after this i'll be up to 80 subdomains and domains on my plan, sheesh
17:04
<annevk>
heh
17:04
<annevk>
these could've been directories :p
17:04
<Hixie>
do you have a preference for directory structure in /home/annevankesteren/?
17:04
<Hixie>
just .../url.spec.whatwg.org/... ?
17:05
<annevk>
yeah that's fine
17:05
<annevk>
just defaults
17:05
<annevk>
only need to put a small script there
17:05
<Hixie>
roger
17:07
<Hixie>
freaking captchas
17:08
<annevk>
yeah, you'd think if you have a registered account there's no need for captchas
17:09
<Hixie>
nothing stops spammers from registering accounts and then generating 10,000 spammy domains
17:10
<Hixie>
ok any others while i'm at it?
17:11
<zewt>
well, assuming a captcha is required for creating accounts, that means the spammy domains can all be associated with the user and bulk deleted
17:11
<annevk>
Hixie: I used to edit Notifications, but I guess I'll wait to see if the W3C finishes that
17:11
<Hixie>
zewt: once you notice them
17:12
<Hixie>
ok. i'll mail you the e-mails once i get them.
17:12
<Hixie>
anne⊙an?
17:13
<annevk>
Hixie: annevk⊙an
17:13
<Hixie>
lol. http://www.whatwg.org/specs/README -- and people say i was unrealistic with my 2022 date.
17:15
<Hixie>
annevk: no dom.specs.whatwg.org?
17:15
<annevk>
Hixie: http://dom.spec.whatwg.org exists
17:15
<Hixie>
oh right that's the one we jumped through hoops for
17:15
<annevk>
Hixie: yeah, we could move it back, but might as well leave as is for now
17:16
<Hixie>
please yes let's not move that again!
17:16
<annevk>
maybe we should have redirects from specs to spec seeing how you already make that mistake
17:17
<Hixie>
not sure how to do that
17:21
<annevk>
Philip` did something like that once for his philip.html5.org domain
17:21
<Hixie>
on dreamhost?
17:21
<Hixie>
wonder how
17:21
<annevk>
can set up some kind of wildcard DNS match for *.specs.whatwg.org
17:21
<annevk>
but where you catch that I'm less sure about
17:21
<annevk>
Philip`?
17:22
<annevk>
gotta go
17:24
<scor>
Hixie: so just to be sure, is this statement correct? http://lists.w3.org/Archives/Public/public-vocabs/2012Sep/0055.html
17:25
<scor>
in particular, which vocab decides whether it's valid to use full URIs, the vocab used in @itemtype, or the vocab defining the full URI?
17:25
<Hixie>
scor: i don't believe that's right. you can add any random property you want if it's got a URL-based name.
17:26
<Hixie>
scor: the idea is you use your own proprietary URL property names to add random data to items using other people's vocabularies
17:26
<Hixie>
scor: (which is probably why the spec doesn't require a spec to define them, come to think of it)
17:32
<Hixie>
didn't anne also work on CORS?
17:32
<Hixie>
i wonder where that's gonna live
18:21
<annevk>
Hixie: that's "fetch"
18:21
<annevk>
Hixie: bit forward looking maybe...
18:22
<annevk>
Hixie: but since CORS is done (imo) I thought I put it there with some placeholder text to see if anyone is interested in merging CORS and HTML fetch
18:23
<annevk>
Hixie: and I also maintained Progress Events which I plan to fold into xhr
18:35
<heycam>
yes, flexbox (or generally applying CSS layout models) to SVG content!
18:35
<heycam>
I think that is the way to go
18:40
<Hixie>
annevk: k
20:04
<smaug____>
gavin: help. can't connect to moznet. too many connection
20:04
<smaug____>
or sicking
20:05
<sicking>
smaug____: how can i help?
20:10
<smaug____>
gavin: sicking: nm, I filed a bug
22:51
<Hixie>
annevk: yt?
23:14
<padenot>
l/b 51