00:38
<smaug____>
could someone remind me what is the use case for <iframe srcdoc="..."> which can't be handled with data urls?
00:39
<TabAtkins>
srcdoc has much easier escaping requirements.
00:40
<TabAtkins>
To securely escape @srcdoc, you only need to escape your quoting character. (You also want to escape &, but that's not necessary for security.)
00:41
<smaug____>
but no different use cases
00:42
<smaug____>
just thinking whether to implement it
01:29
<Hixie_>
smaug____: srcdoc="" is basically equivalent to data:, but has the huge advantage that it is actually readable in the source (e.g. you don't have to escape ' ').
01:33
<TabAtkins>
Also, if you're using DOM to set it, it's literally nothing more than "iframe.srcdoc = text;". No escaping at all is required.
01:34
<Hixie_>
indeed
01:34
<Hixie_>
so in conclusion, it's just convenience, but it's _really_ much more convenient. :-)
01:36
<smaug____>
k
05:10
<annevk>
krijn: server does not appear to be responding
05:15
<annevk>
TabAtkins: not sure that being nr. 1 on the IRC stats is actually an accomplishment or just an indication I don't do much at all :-)
06:05
<annevk>
hmm
06:05
<annevk>
should TextEncoder support encoding to utf-16/ut16-be?
06:05
<annevk>
utf-16be doh
06:32
<annevk>
optional DOMString label = "utf-8" // this is fine right?
06:36
<hsivonen>
sigh. code is so broken
06:37
<hsivonen>
The HTML WG generates too much bug spam these days. I wonder if it’s possible to create a filter that shows only the resolutions that need objecting to.
06:38
<hsivonen>
maybe I should send it all to /dev/null
06:38
<hsivonen>
and public-html, too
07:02
<Stevef__>
hsivonen: morning, on list you wrote "if the <main>element is used in a conforming way, it won't have a <p> element preceding it anyway." what effect does having the p element preceding have on parsing?
07:04
<annevk>
Stevef__: <p> is normally closed by block-level elements
07:04
<annevk>
euh I mean, whatever we call block-level these days :)
07:05
<Stevef__>
annevk: ah OK understand now, thanks!
07:05
<hsivonen>
Stevef__: what annevk said
07:07
<annevk>
I don't really understand the fear of changing the parser btw
07:07
<annevk>
I kinda thought the whole point of having it defined was so we could make changes
07:11
<hsivonen>
annevk: it doesn’t really help authors to get subtly different results in parsers of different vintage
07:11
<hsivonen>
annevk: with with the UA stylesheet, you can always deal with differences by putting main { display: none; } in your author stylesheet
07:12
<hsivonen>
annevk: I thought the point of having it defined was making it the same everywhere
07:12
<hsivonen>
so I search for list:(public-html.w3.org) in Gmail and it doesn’t find anything older than September 28
07:13
<hsivonen>
bug or feature?
07:13
<Stevef__>
hsivonen: :-)
07:13
<annevk>
I kinda see it as making the baseline much stronger, but I don't like clinging to the baseline that much. If it makes sense to introduce a parser change, I'm all for another five years of instability around that particular feature.
07:14
<annevk>
Introducing <intent> for instance does not seem that harmful to me and long term it's probably better. (Assuming we want intents at all, I don't really have an opinion on that.)
07:16
<hsivonen>
annevk: making <main> close <p> would be relatively harmless and would have a transition story
07:17
<hsivonen>
annevk: <intent> has no sane transition story
07:17
<hsivonen>
annevk: it worries me that you think <intent> would not be harmful
07:17
<annevk>
https://twitter.com/miketaylr/status/258613311057309696 this is great
07:18
<annevk>
hsivonen: now all browsers are updating having a transition story seems less necessary, breaking compat is still to be avoided
07:20
<hsivonen>
annevk: Android stock and IE are updating but not cleaning up after themselves yet
07:21
hsivonen
also expects time warp-stuck copies of Presto to live on all sorts of devices
07:21
<annevk>
obviously it's better to not introduce breakage, but the reality is that a lot of things will be broken on those older browsers already
07:22
<annevk>
and the sites where these new features would see adoption first are likely to already be rendering badly
07:24
<annevk>
The main reason I think breakage is okay is that I think usability and aesthetics of the APIs we introduce matter
07:28
<annevk>
Anyone know why http://wiki.whatwg.org/wiki/StringEncoding#TextDecoder decode()'s first argument would be optional?
07:29
<annevk>
gonna change that for now
07:31
<hsivonen>
I wonder why cwilso rejoined the HTML WG
07:34
<annevk>
distributed extensibility!
07:38
<MikeSmith_>
hsivonen: if you have suggestions for changing the bugmail config for public-html, please tell the chairs
07:39
<MikeSmith_>
right now as far as I can recall it's set up so that only messages for new bugs go to the list
07:39
<MikeSmith_>
I'd rather no bugmail at all went to the list
07:41
<MikeSmith_>
btw I thought there was user option in modern bugzilla for subscribing to bugmail for specific components
07:42
<MikeSmith_>
I can't find any option for it in the version we're running now
07:42
<MikeSmith_>
which I think is 4.2.3
07:55
<hsivonen>
MikeSmith_: I have already told the chairs.
07:55
<hsivonen>
MikeSmith_: My communication with the chairs does not seem to be useful.
07:55
<MikeSmith_>
ah ok
07:55
<hsivonen>
MikeSmith_: It seems it would be rational to stop communicating with the Chairs and go write some software.
07:56
<MikeSmith_>
the last time I tried to have a discussion about it the response I got was "nobody else has complained"
08:14
<zcorpan>
krijn: down?
08:14
<krijn>
Meh, yeah
08:17
hsivonen
realizes public-webapps has bugspam, too
08:18
<jgraham>
I quite like that bugspam
08:18
<krijn>
I'm more into hugspam
08:18
hsivonen
creates a filter to make IndexedDB disappear
08:20
<hsivonen>
no wonder why my Bugzilla label was always so full of stuff I didn’t want to see
08:20
<hsivonen>
(I already had a generic filter to move discussion list bugspam among all other bugmail)
08:22
<zcorpan>
Hixie_: i read logs, so you can just pretend i'm here and i'll respond 12h later :-P
08:23
<zcorpan>
(well i *usually* reads logs; email is a safer bet)
08:24
<hsivonen>
zcorpan: logs are not responding
08:24
<zcorpan>
hsivonen: logbot is
08:29
<zcorpan>
anyone object to me removing this dir? http://w3c-test.org/webapps/tests/
08:29
<jgraham>
Not me
08:30
<zcorpan>
a bit annoying that html and webapps have different structure
08:30
<zcorpan>
both broken
08:30
<jgraham>
I was going to say :)
08:30
<Ms2ger>
Yes
08:30
<zcorpan>
can we fix it? what should it be like instead?
08:31
<Ms2ger>
I added that dir back because links were broken
08:31
<zcorpan>
i can put in a htaccess that redirects
08:31
<Ms2ger>
If that works, sure
08:31
<jgraham>
Ms2ger: Can't you fix the links?
08:34
<annevk>
zcorpan: wait, is that the dir with XHR in it?
08:34
<annevk>
zcorpan: because then you're breaking links on w3.org
08:34
<annevk>
zcorpan: but maybe I should stop caring about that
08:36
<zcorpan>
annevk: look closer
08:37
<zcorpan>
which links are we talking about?
08:40
<annevk>
the link towards http://w3c-test.org/webapps/tests/XMLHttpRequest/info.htm
08:40
<annevk>
from the XHR CR
08:40
<annevk>
which is kinda impossible to find right now, but it did exist
08:41
<annevk>
ah, http://www.w3.org/TR/XMLHttpRequest1/
08:41
<annevk>
http://www.w3.org/TR/2010/CR-XMLHttpRequest-20100803/ "test suite"
08:51
<SimonSapin>
"There is a good chance a test suite for XMLHttpRequest will be placed around here." :)
08:54
<zcorpan>
ok, i'll put in a 301 redirect from http://w3c-test.org/webapps/tests/XMLHttpRequest/info.htm to http://w3c-test.org/webapps/XMLHttpRequest/tests/submissions/Opera/info.htm
09:01
<annevk>
the original location was much better :-)
09:04
<zcorpan>
sigh. why does the redirect not work?
09:16
<zcorpan>
annevk: maybe say "Implementations that do not implement the DOM" instead of "Non-browser implementations"?
09:17
<annevk>
hmm, maybe non-JavaScript implementations but I would not want to require Node for instance...
09:21
<annevk>
Ms2ger: liked my hack to Anolis?
09:21
<annevk>
Ms2ger: I think something went wrong again with branches btw
09:21
<annevk>
Ms2ger: sucky branches
09:22
<zcorpan>
annevk: is the BOM not consumed in streaming mode?
09:23
<annevk>
zcorpan: it is
09:23
<annevk>
zcorpan: there's a subtle difference between the streaming flag and the stream variable and I abuse it
09:24
<zcorpan>
annevk: i'm confused
09:25
<annevk>
the first time you invoke decode() with stream set to true the streaming flag will be unset
09:25
<annevk>
the second time you invoke decode() you don't want a BOM to be removed anymore
09:26
<annevk>
unless I guess you feed it one byte at a time
09:26
<Ms2ger>
annevk, haven't looked yet
09:26
<zcorpan>
oh, i see. it would be less confusing if that was called something like "decode has been invoked before" or some such
09:26
<annevk>
if feeding it one byte at a time is a reasonable scenario we'll need a "seen BOM flag" too
09:27
<zcorpan>
ah. yeah.
09:27
<zcorpan>
that should just work, really
09:27
<zcorpan>
i mean, stream should work with arbitrary chunks
09:28
<annevk>
I guess yeah, but then I need even more changes I just realised :/
09:30
<jgraham>
Yeah, you have to work feeding it one byte at a time
09:31
<zcorpan>
annevk: "If encoding is failure, or is none of utf-8, utf-16, and utf-16be, throw a "EncodingError"." doesn't NotSupportedError make more sense here?
09:32
<annevk>
the annoying thing is that then I need to define the actual buffer semantics
09:33
<annevk>
zcorpan: that comment seems appropriate for the list
09:34
<zcorpan>
annevk: is there a bugzilla component?
09:34
<annevk>
zcorpan: doesn't test.w3.org exist for cross-origin tests?
09:34
<annevk>
zcorpan: yes
09:35
<annevk>
zcorpan: not sure jsbell will see that though
09:35
<jgraham>
Cross origin tests?
09:35
<zcorpan>
annevk: www1.w3c-test.org exists for cross-origin tests. w3c-test.org was introduced because tests can't be trusted to not do damage to w3.org.
09:35
<jgraham>
Ah, what zcorpan said
09:35
<jgraham>
Gazumped
09:36
<zcorpan>
annevk: can you add the file bug script?
09:36
<annevk>
it's there
09:36
<jgraham>
Oh, that word doesn't actually seem to have the sense I thought it did
09:36
<annevk>
oops, doesn't link to the right place
09:38
<zcorpan>
jgraham: what did you think it means? (i haven't heard that word before)
09:39
<jgraham>
I was thinking it could be used in a generic sense to mean that you were working on something, and at the last minute someone else came in first with something better than your offering
09:40
<annevk>
zcorpan: fixed link
09:40
<annevk>
will fix the decode situation later
09:40
<jgraham>
But it seems that it is more narrowly defined to be about commercial transactions
09:40
<annevk>
need some kind of decode algorithm that keeps track of the buffer
09:40
<annevk>
:( :( :(
09:44
<zcorpan>
annevk: filed a bug
09:47
<krijn>
Hmpf, power outage, again
09:49
<zcorpan>
krijn: you should install a backup battery
09:58
<hsivonen>
long ISSUE is long: http://www.w3.org/2001/tag/group/track/issues/57
10:06
<zcorpan>
i read the description and i don't understand what the issue is
10:21
<hsivonen>
The Microdata API has revealed multiple incompatibilities with Web content after Firefox 16 went to release
10:22
<hsivonen>
I wonder why these weren’t already discovered and reported by Opera
10:22
<jgraham>
hsivonen: Examples?
10:22
<hsivonen>
https://bugzilla.mozilla.org/show_bug.cgi?id=763229
10:23
<hsivonen>
https://bugzilla.mozilla.org/show_bug.cgi?id=801988
10:23
<hsivonen>
https://bugzilla.mozilla.org/show_bug.cgi?id=802548
10:23
<hsivonen>
https://bugzilla.mozilla.org/show_bug.cgi?id=800850
10:23
<hsivonen>
https://bugzilla.mozilla.org/show_bug.cgi?id=802874
10:31
<jgraham>
hsivonen: Not sure if/why we didn't find those (maybe they got site-patched?), but they mostly seem to have been solved by evangelism
10:31
<jgraham>
Would be really nice if people could be convinced that making up their own attributes/elements is badness though
10:34
<zcorpan>
as well as setting properties on elements, right?
10:57
<jgraham>
Yeah, generally messing with platfoem objects except in the narrow list of safe ways is badness
11:09
<krijn>
annevk: heb jij mijn html5.org pass nog ergens?
11:09
<divya>
jgraham: likely they just square away opera with ie
11:28
<smaug____>
control.labels isn't live, right?
11:54
<zcorpan>
smaug____: why would it not be?
11:58
<smaug____>
zcorpan: I don't see anything in the spec saying it would be
12:00
<smaug____>
ah, by default it is live
12:00
<zcorpan>
smaug____: "Unless otherwise stated, a collection must be live." http://dom.spec.whatwg.org/#concept-collection
12:00
<smaug____>
readability is less than good
12:20
<MikeSmith_>
zcorpan: http://w3c-test.org/quirks-mode/
12:20
<MikeSmith_>
sorry it took me so long to get around to it
12:21
<MikeSmith_>
looks like everything is there now as expected
12:21
<MikeSmith_>
http://w3c-test.org/quirks-mode/tests/
12:21
<zcorpan>
MikeSmith_: thanks!
12:21
<MikeSmith_>
zcorpan: and now I'm trying to figure out how I can get the redirect set up on test.w3.org
12:22
<zcorpan>
MikeSmith_: cool
12:25
<annevk>
krijn: kan 'm resetten
12:25
<annevk>
krijn: doe ik nu even
12:25
<hsivonen>
oh great. Gecko detects BOMless UTF-16 for XML even though the spec says it’s prohibited
12:25
<hsivonen>
I wonder what kind of YSoD doom would ensue if I fixed that
12:26
<annevk>
krijn: wil je ftp houden of liever ssh?
12:28
<annevk>
noticing encoding issues on that w3c-test.org/quirks-mode URL
12:29
<hsivonen>
annevk: btw, the .txt files for the Encoding Standard need charset=UTF-8 (or BOM)
12:30
<zcorpan>
hsivonen: does the html spec support bomless utf-16 for html? last i checked i couldn't find anything
12:31
<hsivonen>
zcorpan: it doesn’t (except potentially via the heuristic detector black box)
12:31
<zcorpan>
ok
12:31
<zcorpan>
do you know what the browser landscape looks like offhand?
12:32
<hsivonen>
I don’t
12:32
<annevk>
hsivonen: but the files seem to work fine?
12:32
<annevk>
ooh, only in Opera
12:32
<hsivonen>
Gecko tries to detect the case where the file is all-Basic Latin BOMless UTF-16
12:32
<hsivonen>
probably an overreaction
12:32
<zcorpan>
hsivonen: before fixing gecko to match the xml spec, might be worth it to see what happens in other browsers and in html. maybe the xml spec is just wrong?
12:33
<hsivonen>
if we reacted like that to Microdata, we’d have nuked the Microdata API by now
12:33
<hsivonen>
zcorpan: yeah. testing other browsers is just more annoying than looking at the Gecko code
12:35
<zcorpan>
annevk: opera detects utf-8. maybe we should drop that
12:35
<hsivonen>
in other news, Gecko looks broken in the case where the BOM or the XML declaration isn’t fully within the first network buffer
12:35
<annevk>
got some very bad latency over ssh :/
12:36
<hsivonen>
Same thing with the CSS parser, it looks like, unless I’m missing some buffering
12:36
<hsivonen>
looks like I’m the only one in the habit of making my BOM sniffing resilient against buffer boundaries
12:37
<annevk>
hsivonen: I think I fixed it by adding AddDefaultCharset utf-8
12:38
<annevk>
whoa, AddDefaultCharset defaults to iso-8859-1
12:38
<hsivonen>
annevk: thanks
12:39
<hsivonen>
annevk: IIRC, Gecko works around that Apache stubbornness
12:39
<annevk>
zcorpan: it's causing breakage in other browsers!
12:42
<odinho>
annevk: mosh -- just installed it on client and server because I've been doing lots'a 3G lately. It makes remote ssh-ing smooth. :-)
12:44
<MikeSmith_>
zcorpan: OK set up http://test.w3.org/* to 301 to http://w3c-test.org/*
12:44
<zcorpan>
MikeSmith_: thanks. seems the redirect prepends a slash to the new url
12:44
<zcorpan>
in the path
12:45
<MikeSmith_>
yeah saw that I dunno why
12:45
<zcorpan>
how did you implement it?
12:45
<MikeSmith_>
RedirectMatch permanent ^(.*)$ http://w3c-test.org/$1
12:45
<zcorpan>
I'd do: Redirect 301 / http://w3c-test.org/
12:46
<zcorpan>
but otherwise dropping the slash before $1 would work i guess
12:46
<MikeSmith_>
ok lemme try that
12:46
<MikeSmith_>
ah yeah
12:47
<MikeSmith_>
OK slash shoudl be gone now
12:48
<zcorpan>
thanks!
12:48
<zcorpan>
MikeSmith_: i had sent an email about this (and the other redirect issue) http://www.w3.org/mid/op.wmdajyzjidj3kv⊙sl
12:49
<MikeSmith_>
yeah saw that
12:49
<MikeSmith_>
plh pinged me about it
12:50
<MikeSmith_>
hmm the link in that spec is still busted
12:51
<MikeSmith_>
since the tests subdirectory is gone
12:52
<MikeSmith_>
seems it should just get redirected to http://w3c-test.org/webapps/XMLHttpRequest/tests/ now, right
12:52
<MikeSmith_>
?
12:52
<MikeSmith_>
I think if you add a redirect in an .htaccess in the repo, that should work
13:01
<zcorpan>
i did that, it didn't work
13:02
<zcorpan>
http://dvcs.w3.org/hg/webapps/rev/f5791dfb4e7b
13:02
<MikeSmith_>
ok
13:02
<MikeSmith_>
lemme check
13:03
<zcorpan>
maybe redirects are not allowed except for the root .htaccess? dunno if apache has such a setting
13:03
<MikeSmith_>
ah
13:03
<MikeSmith_>
yeah, probably
13:04
<MikeSmith_>
AllowOptions setting
13:07
<hsivonen>
about:blank is defined to be UTF-8, right?
13:09
<jgraham>
http://tools.ietf.org/html/draft-holsten-about-uri-scheme-06 says so, at least
13:12
<hsivonen>
ok. thanks
13:14
<MikeSmith_>
zcorpan: if you change it to Redirect 301 /webapps/tests/XMLHttpRequest/info.htm http://w3c-test.org/webapps/XMLHttpRequest/tests/submissions/Opera/info.htm
13:14
<MikeSmith_>
that should work
13:15
<MikeSmith_>
absolute path
13:28
<annevk>
zcorpan: how do you feel about TypeError?
13:29
<zcorpan>
annevk: you mean using an enum?
13:29
<annevk>
no just throwing TypeError
13:30
<annevk>
as if it was an enum, basically
13:30
<zcorpan>
i guess that works
13:30
<annevk>
cool
13:54
<annevk>
i fixed the stream thingie too I think
13:55
<annevk>
it should prolly be done in a nicer way though, if someone sees a way to rewrite the conceptual interfaces that'd be great
14:12
<jgraham>
http://programmers.stackexchange.com/questions/168751/is-the-use-of-utf8-preferable-to-utf8-true
14:23
<annevk>
interesting answer
14:31
<zcorpan>
MikeSmith_: seems the redirect works now. thanks!
14:32
<MikeSmith_>
cool
14:37
<zcorpan>
how was it again, setting event handler idl attributes to non-functions means it gets set to null?
14:51
<annevk>
think so
15:25
<annevk>
zewt: optional might then make sense, not sure about nullable
15:27
<annevk>
but I guess making input optional makes sense
16:25
<krijn>
annevk: maakt niet uit, ftp is goed genoeg denk ik
17:36
<zcorpan>
annevk: did specification-data move?
18:30
<annevk>
zcorpan: yes
18:30
<annevk>
zcorpan: https://github.com/whatwg/xref
18:35
<zcorpan>
ok, thanks. cloned.
18:37
<annevk>
zcorpan: you want to be a team member?
18:37
<zcorpan>
does that give me write access?
18:37
<annevk>
yeah
18:37
<zcorpan>
sure
18:37
<annevk>
gives you access to all things WHATWG on GitHub
18:37
<annevk>
simonp is you?
18:37
<zcorpan>
no
18:38
<zcorpan>
zcorpan
18:38
<annevk>
zcorpan: zcorpan gives a 500?
18:38
<zcorpan>
yeah i noticed
18:39
<zcorpan>
no idea why
18:39
<zcorpan>
i'm logged in as "zcorpan"
18:39
<annevk>
well you should have access now
18:40
<annevk>
zcorpan: I also pushed an update to Anolis so you can now reuse the same database for several specifications
18:40
<annevk>
zcorpan: all of the WHATWG standards have already been updated to that effect
18:41
<smaug____>
anyone know if there is js version of Graphviz' dot
19:26
<pablof>
is there a repo somewhere for the eventsource spec?
19:46
<Hixie_>
pablof: svn.whatwg.org, it's just part of HTML
19:47
<pablof>
thanks
19:58
<JonathanNeal>
In the Media Query language, if we're matching anything 30em we use (width: 30em), and if we're matching anything 30em or smaller we use (max-width: 30em), but what if we're matching anything less than 30em?
20:00
<Philip`>
(max-width: 29.999em) ?
20:02
<JonathanNeal>
That's the work around, definitely.
20:02
<JonathanNeal>
Has anyone suggested language that targets "less than"?
20:03
<Hixie_>
what's the use case?
20:03
<Hixie_>
(www-style would be the appropriate forum for this discussion, btw)
20:09
<JonathanNeal>
is that an email group, Hixie_?
20:09
<JonathanNeal>
or rather, where is www-style?
20:10
<JonathanNeal>
Hixie_: the usecase is defining multiple breakpoints based on the same value. The problem is that defining anything 30em or smaller is easy, but defining anything less than 30em is hard, you end up appending all these 9's.
20:11
<annevk>
JonathanNeal: http://lists.w3.org/Archives/Public/www-style/
20:11
<JonathanNeal>
@media (max-width: 30em) {} @media (min-width: 30.0001em) && (max-width: 60em) {}
20:11
<Hixie_>
JonathanNeal: what anne said (sorry, was afk)
20:12
<Hixie_>
JonathanNeal: i don't understand why you can't just do a series of max breakpoints
20:12
<Hixie_>
JonathanNeal: like, max 300px, max 600px, max 900px, anything above
20:14
<JonathanNeal>
because then i have to manage more offsets, between the unmq'd stylesheet and now each mq matching stylesheet after that.
20:14
<Hixie_>
come again?
20:14
<Hixie_>
i don't follow
20:24
<JonathanNeal>
Sorry, Hixie_, I think there should be something like < and > (albeit those symbols may conflict with X/HTML). I thought something may have been pitched.
20:32
<JonathanNeal>
Like @media (max-width: 30em) {} @media (gt-width: 30em) && (max-width: 50em) {}
20:38
<JonathanNeal>
alternatively, above-width, below-width, over-width, under-width.
20:41
<JonathanNeal>
I like above and below because they have the name number of letters and fit nicely beside max and min.
21:53
SamB_MacG5
wishes file:// had a CORS-equivalent ...
23:34
Hixie_
deals with some feedback on the spec that makes him think he should require that everyone who comments on the spec should first read the spec and take a test
23:34
<Hixie_>
(though actually, the number of times people suggest something that hte spec already says is remarkably low, all things considered)
23:49
<SamB_MacG5>
Hixie_: that wouldn't be a good plan for clarification suggestions ...
23:49
<SamB_MacG5>
/requests