02:46
<MikeSmith>
http://www.w3.org/2010/02/rdfa/meetings/2013-02-28#Mike_Smith__28_tm__29___27_s_Complaint
02:47
<MikeSmith>
"q+ to ask why tool authors who work for us get to wag the dog here"
02:47
<MikeSmith>
gotta love that kind of response
02:52
<Hixie>
hahaha
02:52
<Hixie>
"damn these people sending us valid feedback, why do we have to make a good technology that considers things we hadn't thought of"
02:53
<Hixie>
this is as opposed to when _I_ disagree with you, when that happens i just try to drown you in a wall of text, amirite? ;-)
02:56
<zewt>
Hixie: but they thought of everything, therefore the feedback must be invalid, qed
03:07
<MikeSmith>
not even clear to me from those minutes what point he was trying to make
03:07
<MikeSmith>
"I think that the tool is broken."
03:08
<MikeSmith>
broken in that it actually implements what the spec says, instead of, I guess, what they intended for it to say, or what somebody should be expected to assume from reading it
03:08
<MikeSmith>
I've had some many cases of that with the RDFa specs
03:09
<MikeSmith>
e.g., before the spec never explicitly stated that the values of the various RDFa attributes could be the empty string
03:11
<MikeSmith>
so beack then I commented about that and I think I got the same kind of response; i.e., something like "Well of course they can be the empty string. ANy fool can see that. So the spec doesn't actually have to say it."
03:11
<Hixie>
to be fair, a lot of spec authors are of the mindset that specs don't have to be thorough and precise
03:11
<Hixie>
a _lot_ of w3c and ietf specs are like that
03:12
<Hixie>
they give a general idea, and leave the details to implementors
03:12
<Hixie>
HTML4 being a classic example
03:12
<Hixie>
but by no means either the worst or the first
03:13
<MikeSmith>
yeah
03:13
<MikeSmith>
but I would hope the bar is a lot higher these days
03:14
<Hixie>
a lot of people were never asked if they agreed with the kind of spec quality css2.1 started
03:14
<Hixie>
and so they're still in the old world
03:14
<Hixie>
where spec writing was easy :-)
03:14
<MikeSmith>
heh
03:15
<zewt>
wow, irssi is terrible
03:15
<zewt>
it takes _foo_ and underlines it
03:16
<zewt>
you know you've worked on crappy terminal irc client too long if you're implementing that and actually think it's not horrible gimmicky nonsense
03:16
<othermaciej>
implementors can't be trusted to figure out the details in a consistent way
03:17
<Hixie>
othermaciej: don't forget, this is the same mindset that thinks that non-conforming documents are irrelevant because interop doesn't matter if the author screwed up
03:18
<zewt>
(which itself assumes that a nonconforming document is the author's fault)
03:20
<Hixie>
it often is, to a first approximation, but that's of no consolation to the users when 97% of documents are trivially non-conforming in some way or other :-)
03:21
<zewt>
but 97% of documents are nonconforming != 97% of authors screwed up :P
03:21
<zewt>
(no, we're not disagreeing in any material way)
07:04
<Ms2ger>
annevk: I saw a Gecko patch for autoRevoke, btw
07:25
<zcorpan>
TabAtkins: do you have any URLs to sites that do smooth scrolling?
08:00
<zcorpan>
Hixie: https://www.w3.org/Bugs/Public/show_bug.cgi?id=16920 seems good to me
08:01
<zcorpan>
hmm, except for "This should be done even if the video is paused, except if video playback hasn't ever started yet."
08:34
<annevk>
Why can https://cgi.w3.org/member-bin/MailingListQuery.pl?queryList=www-dom and such no longer be accessed with Member credentials?
08:36
<Ms2ger>
jgraham, so how do I get my pr mirrored?
08:40
<MikeSmith>
annevk: dunno but maybe not intentional
08:40
<MikeSmith>
annevk: best to e-mail sysreq⊙wo to ask
08:40
<jgraham>
Ms2ger: In theory, if you have push access, do nothing
08:40
<jgraham>
Ms2ger: Since the code is totally untested and therefore broken, tell dom to start looking for the stacktrace in the log
09:18
<annevk>
I wonder why filing bugs no longer works on HTML?
09:18
<annevk>
Or did I just file dozens?
09:23
<hsivonen>
zcorpan: there's a harness and then almost no tests
09:24
<hsivonen>
I know one should not read Internet comments, but there are just so many ill-informed people commenting on the Netflix HTML5 news on The Verge.
09:29
<zcorpan>
hsivonen: pointer to the harness?
09:29
<a-ja>
hsivonen: read article, but not comments....pretty silly ones, eh?
09:30
<zcorpan>
anyone have API design opinions on https://www.w3.org/Bugs/Public/show_bug.cgi?id=20328 and https://www.w3.org/Bugs/Public/show_bug.cgi?id=19753 ?
09:31
<annevk`>
Looks like a future
09:35
<zcorpan>
allElementsFromPoint(x,y, 'document') ?
09:36
<Ms2ger>
'document'?
09:37
<annevk>
lol, who is annevk`?
09:38
<annevk>
ah, should have known
09:38
<Ms2ger>
:)
09:39
<hsivonen>
a-ja: even the article was just press release journalism summarising the Netflix post without checking other sources for relevant info
09:39
<zcorpan>
Ms2ger: https://www.w3.org/Bugs/Public/show_bug.cgi?id=20328 wants to use these methods outside the viewport, in the canvas area
09:39
<hsivonen>
zcorpan: https://bitbucket.org/validator/validator/src/0d0e5f1771339b55adda64089b12a32c5dacb78c/test-harness?at=default
09:40
<annevk>
zcorpan: document.hitTest({x:2, y:5}) -> static list (same as the one to be returned from document.find()
09:42
<zcorpan>
annevk: and what about choosing which area you want (viewport vs document)?
09:42
zcorpan
bbiab
09:42
<annevk>
context:"page" / "client" or some such?
09:42
<hsivonen>
the coverage at Ars is a bit better, but is still pretty close to mere press release journalism
09:46
<hsivonen>
at least over at Ars, the commenters have more clue than over at The Verge
09:48
<annevk>
Oh that is interesting. Netflix made a bet on DRM ending up in HTML?
09:53
<hsivonen>
annevk: well, at this point, EME shipping on Chrome OS is not a bet anymore
10:10
<marcosc>
darobin: in respec, how do I escape [[HasOwnProperty]] without respect trying to associate it with a biblio entry?
10:11
<marcosc>
s/respect/respec
10:11
<Ms2ger>
Bwahahaha
10:11
<marcosc>
heh
10:11
<Ms2ger>
Try [<!-- -->[
10:12
<marcosc>
no, Ms2ger, I'm sure Respec will handle this in a clever way
10:12
<marcosc>
:)
10:18
<Ms2ger>
Bwahahaha
10:19
<annevk>
You want to write a spec. You think, I'll use ReSpec. Now you have two problems.
10:25
<zcorpan>
annevk: do you have a fancy name for the new caretPositionFromPoint that returns a list?
10:25
<annevk>
caretHitTest?
10:27
<annevk>
note though that without defining hit testing this will create a ton of issues down the road :(
10:31
<annevk>
it also seems somewhat unclear what kind of list you want to return
10:31
<annevk>
e.g. if an element is transparent to hit testing, should it still be in the list?
10:31
<annevk>
where should it be in the list?
10:31
<marcosc>
annevk, one day we will convince mounir that there is only one true tool for spec writing :)
10:32
<annevk>
mounir is a co-chair now, he doesn't have to entertain the thought of spec writing anymore :p
10:32
heycam
wonders where this idea of co-chairing and not having to do spec writing comes from :)
10:32
<marcosc>
annevk pioneered that, heycam
10:33
<zcorpan>
it seems the stated problem can be solved by setting pointer-events:none on the top-most element and calling elementFromPoint or caretPositionFromPoint again
10:33
<marcosc>
heycam: annevk also told me that if we got on the TAG, we could lounge around all day by the beach and just get paid.
10:33
<marcosc>
:)
10:36
<annevk>
zcorpan: on the element returned?
10:36
<zcorpan>
annevk: yeah
10:37
<annevk>
zcorpan: yeah, I guess at some point you might want to have things like "ignore pointer-events, ignore transparancy, ..."
10:37
<zcorpan>
annevk: similar to the current workarounds that remove or hide the element, but without any visual glitch
10:37
<annevk>
zcorpan: which the dictionary will help with
10:37
<heycam>
marcosc, ha
10:38
<zcorpan>
annevk: do you mean that you would want to get an element even if it is transparent or has pointer-events:none ?
10:40
<annevk>
zcorpan: yeah maybe, would have to study hit testing APIs out there
10:43
zcorpan
finds http://ie.microsoft.com/testdrive/HTML5/HitTest/
11:00
<hsivonen>
wondering if I should file a bug to make Thunderbird and SeaMonkey always send email as UTF-8
11:01
<hsivonen>
surely everyone can receive UTF-8 these days
11:01
<hsivonen>
(except maybe Hixie on pine)
11:01
<annevk>
yeah I'm not sure why that hasn't happened, but I'm thinking there might be some legacy out there in Japan and such on phones
11:02
<annevk>
so that if you want your friend to be able to read the email, it better use the right encoding...
11:04
<hsivonen>
I think we should find out
11:04
hsivonen
is generally annoyed by not doing stuff because of *maybe* weird stuff in Japan
11:05
jgraham
also uses (al)pine (sometimes) fwiw
11:05
<hsivonen>
(actual weird stuff anywhere is another thing)
11:05
<hsivonen>
jgraham: how does it deal with incoming UTF-8?
11:05
<jgraham>
I don't reacall having character encoding issues
11:05
<jgraham>
But I haven't exactly tested
11:05
<hsivonen>
alpine was written in the mid-2000s, so it would be shameful for it not to support UTF-8
11:06
<hsivonen>
I guess I'll file the bug and see what happens to it
11:07
<annevk>
+1
11:08
<marcosc>
annevk: any chance http://*.spec.whatwg.org could be CORS enabled?
11:10
<jgraham>
hsivonen: Incoming UTF-8 seems to work fine
11:10
<annevk>
marcosc: yeah I guess
11:10
<odinho>
marcosc: Or at least fetch? :P
11:10
<hsivonen>
jgraham: great
11:10
<marcosc>
odinho: all would be nice :)
11:11
<marcosc>
annevk: should I file a bug somewhere?
11:11
<odinho>
I was just thinking about doogfooding.
11:11
<odinho>
Minus one o.
11:11
<marcosc>
fetching fetch?
11:11
<marcosc>
:)
11:15
<annevk>
so what's the way to make this work?
11:15
<annevk>
well, is there a simple way to add a header to the output of an entire domain
11:16
<annevk>
in fact, does .htaccess take effect if you put it above the directory the site is served from?
11:18
<jgraham>
annevk: No, but you can put it in the root, or change the apache config
11:18
<jgraham>
(the documentroot I mean)
11:20
<annevk>
stupid console
11:20
<annevk>
shows 304, but does not give a way to show what was actually returned from the cache
11:21
<annevk>
jgraham: so it does actually
11:22
<annevk>
marcosc: everything but dom.spec.whatwg.org now has Access-Control-Allow-Origin:*
11:22
<hsivonen>
bug filed https://bugzilla.mozilla.org/show_bug.cgi?id=862292
11:22
<marcosc>
thanks annevk!
11:22
<marcosc>
how come not dom?
11:22
<annevk>
by adding "Header set Access-Control-Allow-Origin *" to a .htaccess file shared across the domains
11:23
<annevk>
marcosc: I need to fix that one separately unfortunately
11:23
<annevk>
but should be easy
11:23
<marcosc>
np, thanks for doing that :)
11:24
<marcosc>
annevk: works well (confirmed with http://marcoscaceres.github.io/bib_entry_maker/)
11:24
<jgraham>
Holy crap it does
11:24
<jgraham>
That's insane
11:25
<annevk>
DOM has it too now
11:26
<marcosc>
need to add all these references to Respec ... normative references to Living Standards For. The. Win...!
11:42
<hsivonen>
when do blobs get garbage collected if revokeObjectURL is not called?
11:57
<zcorpan>
i wonder how to define offset* for <col>/<colgroup>
12:02
<zcorpan>
is it sufficient to say that elements with computed value of 'display' of 'table-column' and 'table-column-group' must be considered to be CSS layout boxes for the purpose of the offset* attributes?
12:04
<MikeSmith>
hsivonen, annevk : I did used to be that the mail clients on a lot of Japanese keitai did not support UTF-8 but I don't think that's the case any longer
12:05
<MikeSmith>
it used to be that many did not even have fonts with glyphs for latin characters that have diacritical marks and such, non-ascii
12:05
<MikeSmith>
but I think that's not that case any longer either
12:06
<MikeSmith>
lacking those glyphs it didn't matter much how a message was encoded anyway, because you wouldn't be able to read it regardless
12:16
<darobin>
marcosc: you use [[\Foo]]
12:17
<annevk>
hsivonen: I suspect when everything else is collected
12:17
<annevk>
hsivonen: although that might break fastback?
12:48
<annevk>
okay, so merging redirect fetch and CORS fetch was not a good idea :/
12:51
<karlcow>
[07:07] * hsivonen is generally annoyed by not doing stuff because of *maybe* weird stuff in Japan
12:52
<karlcow>
I have the feeling is a combination of technology/culture at a certain point in space/time.
12:52
<karlcow>
I wonder what are the constraints on low-end phones in other territories
12:54
<karlcow>
maybe less an issue now given that the low end phones anyway are a lot more powerful than 10 years ago.
12:54
<karlcow>
bah
12:54
<annevk>
bah?
12:57
<karlcow>
ah. my 「bah」in a 「can't be helped」way.
12:58
<karlcow>
Shikata ga nai
12:58
<karlcow>
https://en.wikipedia.org/wiki/Shikata_ga_nai
14:02
<RobbertAtWork>
hsivonen: you did a lot of research on the HTML doctype in different browsers; is it a known effect that in IE6 the doctype is returned as comment node by getElementsByTagName()?
14:02
<RobbertAtWork>
hsivonen: freaks me out
14:02
<RobbertAtWork>
hsivonen: http://robbertbroersma.nl/demo/html5-ie6/ alerts "#comment" in IE6
14:03
<marcosc>
darobin: thanks!
14:03
<marcosc>
Ms2ger: see! darobin made something clever :)
14:04
<RobbertAtWork>
hsivonen: it's not IE6 only by the way, it's IE8 as well
14:05
<annevk>
RobbertAtWork: I think that was a known quirk of their old parser
14:07
<RobbertAtWork>
annevk: I see. Never knew I had to filter getElementsByTagName() results by nodeType
14:08
<annevk>
RobbertAtWork: http://ln.hixie.ch/?start=1137740632&count=1
14:08
<annevk>
RobbertAtWork: has some related fun stuff
14:11
<RobbertAtWork>
annevk: thanks!
14:20
<Ms2ger>
marcosc, unpossible
14:59
<karlcow>
"Use the glass://map URI to request a map." — developers.google.com/glass/location :(
15:03
<annevk>
It renders as an image?
15:07
<karlcow>
annevk: it seems
15:07
<karlcow>
<img src="glass://map?w=width&h=height&marker=0;latitude,longitude&marker=1;latitude,longitude&polyline=;latitude,longitude,latitude,longitude"
15:08
<karlcow>
width="width"
15:08
<karlcow>
height="height"/>
15:08
<annevk>
I wonder if they're using a full browser or a subset of sorts
15:09
<bholley>
Hixie: yt?
15:09
<annevk>
oh Java and Python
15:11
<bholley>
Hixie: am reading the spec correctly to mean that enumerating cross-origin DOM objects (Window and Location) shouldn't throw, but should instead return the appropriate cross-origin-accessible properties?
15:11
<bholley>
Hixie: that is to say, "get the raw properties and then filter"?
15:13
<annevk>
oh so everything is JSON
15:13
<annevk>
and JSON can include some raw HTML
15:13
<annevk>
that's some weird shit
15:19
<annevk>
smaug____: yo
15:19
<annevk>
smaug____: so upload progress events
15:19
<annevk>
smaug____: combined with a 307
15:19
<smaug____>
meeting
15:19
<annevk>
smaug____: your argument is invalid
15:23
<annevk>
So Gecko does not seem to dispatch a progress event in that sequence...
15:23
<annevk>
Lets try setting up some stuff to make a redirect work...
15:43
<nimbu>
Dear whatwg http://tjvantoll.com/2013/04/15/list-of-pseudo-elements-to-style-form-controls/ is terrible is this specced somewhere?
15:43
<nimbu>
or are there plans to?
15:44
<annevk>
only long term
15:45
<annevk>
nimbu: latest is https://groups.google.com/a/chromium.org/forum/?fromgroups=#!topic/blink-dev/ZAdZJWahyF8
15:45
<nimbu>
annevk: thnx!
15:45
<nimbu>
long term!!
15:45
<nimbu>
long term we will have 100 new nasty pseudo elms
15:46
<nimbu>
ok i will see what i can do.
15:46
<nimbu>
thnx annevk
15:46
<annevk>
well you know, HTML has been around for almost 20 years now and parsing got interoperable about a year ago
15:47
<annevk>
but yeah, I'd love for this to get fixed
15:48
<nimbu>
yeah no thanks to some super human efforts
15:48
<nimbu>
we dont want to be there again
15:51
<marcosc>
darobin: might be a bug, but using [[\Foo]] shows up as [[\Foo]] in the spec :(
15:55
<darobin>
marcosc: it works here — pointer?
15:56
<marcosc>
darobin: http://runtime.sysapps.org/
15:56
<marcosc>
search for [[\
15:56
<Ms2ger>
0x7fdde645398
15:56
marcosc
waits for robin to point out some dumbass mistake that marcos made
15:57
marcosc
will find a way to blame Ms2ger
15:58
<darobin>
marcosc: shows up as [[HasOwnProperty]] here...
15:59
<marcosc>
darobin: what browser?
15:59
<darobin>
I blame whatever benighted browser you must be using
15:59
<darobin>
Firefox
15:59
<darobin>
what else?
15:59
<marcosc>
um,
15:59
<marcosc>
that one with the funny color circle?
15:59
<darobin>
more seriously I doubt that there's a browser difference here
15:59
<marcosc>
checking
15:59
<darobin>
or rather, not in processing
16:00
<marcosc>
yeah, that would be weird on the Web
16:00
<darobin>
I suspect you have a rather old ReSpec aggressively cached maybe?
16:00
<darobin>
it would be weird for something this trivial :)
16:00
<darobin>
yeah, works in Chrome here
16:00
<darobin>
I suggest you reload with extreme prejudice
16:00
<marcosc>
ok, thanks
16:00
<marcosc>
will do that
16:01
<darobin>
looks like your Chrome is pulling off a classic Opera
16:01
<marcosc>
aren't they the same? :)
16:01
marcosc
drumroll please
16:01
<darobin>
ka-ching
16:03
<marcosc>
darobin: one more tiny question... in section 3.3, why is the algorithm counting from 14? (i.e., picking up from counter in section 3)
16:03
<darobin>
hah, that's funny
16:03
<darobin>
I don't believe that's something that ReSpec is doing...
16:04
<darobin>
it's... weird
16:04
<marcosc>
mkay
16:05
<darobin>
fascinating
16:06
<darobin>
actually someone did add support for algorithm styling, for whatever reason
16:10
<blink-bot>
good morning, Whatwg!
16:11
<darobin>
marcosc: I'm very puzzle about why the counter-reset doesn't kick in here
16:12
<marcosc>
darobin: yeah, I thought the markup was ok. About to look at the CSS
16:12
<darobin>
feel free to use a different class name
16:13
<darobin>
but there's something really weird going on in the styling there
16:13
<smaug____>
annevk: back
16:13
<smaug____>
haa haa
16:13
<smaug____>
good morning blink-bot
16:13
<marcosc>
darobin: yeah, gets more funky if you toggle "counter-reset: numsection"
16:13
<marcosc>
then it goes to 21?
16:14
<smaug____>
is that what dglazkov uses his 20% time for, being a blink-bot :p
16:14
<Ms2ger>
He's off half of the time
16:14
<dglazkov>
smaug____: I have no idea how that happened. Must've been drunk last week or something
16:15
<darobin>
marcosc: I have no idea who added that horrible style or why
16:15
<darobin>
I have half a mind to yank it
16:15
<darobin>
whoever did it didn't add a test
16:15
<marcosc>
ok, now it makes sense
16:15
<marcosc>
I'll change the class name
16:15
<darobin>
marcosc: that's what's weird: the counter-reset works once, but not the second time
16:16
<marcosc>
it's like because it's in a sub section it won't reset
16:17
<marcosc>
darobin: it breaks anyway for sub steps
16:17
<marcosc>
darobin: you should probably yank it
16:18
<marcosc>
darobin: see any step in "teps for processing an application manifest "
16:18
<marcosc>
"steps for processing an application manifest"
16:18
<marcosc>
even
16:18
<darobin>
marcosc: yeah, I'm going to pull it
16:18
<darobin>
just looking through history to figure out who added it
16:19
<marcosc>
git blame!!! git blame!!! :D
16:19
<marcosc>
we need a "git name and shame "
16:22
<darobin>
ok, I don't think anyone uses that
16:22
marcosc
updates the spec
16:31
<darobin>
marcosc: that particular weird piece of madness is now gone
16:31
<darobin>
it's a change grandfathered from something that came from v2 in HG
16:31
<marcosc>
darobin: thanks!!!
16:31
<darobin>
so, I blame hg
16:31
<jsbell>
darobin: new respec.js push? I note the list of "undefined definition" errors in the IDB spec is way down today.
16:31
<Ms2ger>
marcosc, did you mess with its references database?
16:32
<darobin>
jsbell: yes, I found that the missing definition detection code had an issue
16:32
<marcosc>
heh, Ms2ger not today :)
16:32
<darobin>
which I plugged
16:32
<darobin>
jsbell: I still need to look at your other problems
16:32
<darobin>
I wish people stopped having issues with the current WebIDL code though, it would give me time to actually make the WebIDL bits modern and nice and all :)
16:33
<marcosc>
darobin: users are the worst!
16:33
<marcosc>
:)
16:33
<darobin>
yeah, they want all sorts of shit :)
16:33
<marcosc>
takes all the fun out of writing code
16:33
<jsbell>
darobin: cool, thanks. The remaining errors list looks mostly legitimate; still interested in how to correctly reference an enum or interface member.
16:33
<jsbell>
WANT WANT WANT
16:33
<Ms2ger>
<dfn>?
16:34
<darobin>
jsbell: I'll look at that tomorrow, right now I've just finished doing my taxes and I need a beer :)
16:34
<Ms2ger>
darobin, also, IDB has issues with TypeError claiming to be a DOMException
16:34
<jsbell>
no rush
16:34
<jsbell>
ms2ger: in an "Exceptio table" ?
16:34
<jsbell>
+n
16:34
<jsbell>
or in the prose?
16:34
<Ms2ger>
Table
16:35
<jsbell>
We need to kill the tables
16:35
<darobin>
KILL THE TABLES
16:35
<Ms2ger>
Yes please
16:35
<darobin>
isn't that what Anssi's legacyThingie does?
16:36
<jsbell>
I'm waiting on https://www.w3.org/Bugs/Public/show_bug.cgi?id=21555 before replacing the exception tales with prose.
16:36
<jsbell>
tables. Can't type today.
16:36
<darobin>
jsbell: look at http://lists.w3.org/Archives/Public/public-sysapps/2013Apr/0132.html
16:36
<jsbell>
Ugh, wrong bug link too...
16:36
<jsbell>
https://www.w3.org/Bugs/Public/show_bug.cgi?id=17681
16:37
<darobin>
jsbell: so yeah, it looks like noLegacyStyle: true in the config will kill the tables
16:38
<darobin>
of course, that'll kill whatever content you were expecting to put there, too :)
16:38
<Ms2ger>
darobin, ... again
16:38
<darobin>
Ms2ger: ?
16:39
<Ms2ger>
ReSpec made the tables disappear a while back
16:39
<jsbell>
right, that's the problem with the IDB spec. Not all of the exceptions are defined in prose.
16:39
<darobin>
Ms2ger: I think you're thinking about exceptions
16:39
<darobin>
completely different problem :)
16:40
<darobin>
IDB suffered when ReSpec dropped support for WebIDL exceptions
16:41
<Ms2ger>
darobin, I am talking about exceptions, yes
16:41
<jsbell>
Yep, we're relying on { idlOldStyleExceptions: true }
16:41
<Ms2ger>
darobin, what are you talking about, if not exceptions?
16:41
<jsbell>
I was talking about exceptions too, so confused. :)
16:41
<darobin>
I think I should add a feature to do spec linting, that would also collect the editors' emails, and would email them insults when they're doing stuff wrong :)
16:41
<darobin>
Ms2ger: there are other tables, parameters and shit like that
16:42
<darobin>
those need to go as well
16:42
<darobin>
in the grand scheme of things
16:42
<darobin>
which is moving everyone to better specs and all
16:42
Ms2ger
suggests anolis for that
16:43
darobin
giggles
16:43
<Hixie>
bholley: i hope so, but wouldn't guarantee it
16:43
<Hixie>
annevk: how does filing bugs not work?
16:43
<darobin>
anyway, not that you fine people don't make for enjoyable conversation
16:43
<darobin>
but as I was saying -> beer
16:43
<darobin>
thanks for the laughs though Ms2ger ;)
16:44
<bholley>
Hixie: can you confirm on https://bugzilla.mozilla.org/show_bug.cgi?id=862380 when you get the chance?
16:45
<Hixie>
send me mail - i'm mostly afk for the next few days
16:46
<Ms2ger>
Tired from the sprint the last few days? :)
16:48
<Hixie>
you could say that
17:03
<annevk>
Hixie: dunno, I filed one manually
17:04
<Hixie>
what wasn't working?
17:09
<annevk>
Hixie: using the form from within Firefox
17:21
<annevk>
Hixie: seems to happen in http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html
17:22
<annevk>
Hixie: tried to report a bug about blink
17:30
<annevk>
smaug____: see public-webapps I suppose
17:31
<smaug____>
k
17:33
<annevk>
I'm still a bit stuck modelling http://fetch.spec.whatwg.org/#fetching
17:34
<annevk>
E.g. synchronous / asynchronous sharing a code path is hard.
17:34
<annevk>
The user being able to cancel the request at any point, regardless of synchronous, is hard.
17:35
<annevk>
Upload progress tasks being able to spain several requests, is hard. Although I guess the same goes for normal progress tasks...
17:36
<annevk>
I'm very much cheating my way out of this engineering challenge by having a set of encompassing requirements that accompany the algorithm.
17:36
<annevk>
Sorry, meant to so I'm close to doing that cheating.
18:12
<Ms2ger>
Nothing new on http://tc39memes.tumblr.com/ lately...
18:17
<jsbell>
Ms2ger: https://twitter.com/FakeAlexRussell isn't enough funny for you?
18:32
<karlcow>
A proposal for unprefixing https://groups.google.com/a/chromium.org/forum/m/?fromgroups#!topic/blink-dev/E0OkYhvF7q8
18:51
<Ms2ger>
TabAtkins, people still haven't figured out that you can have requirements for multiple conformance classes?
22:43
<TabAtkins>
zcorpan: Having to one-by-one strip out all the elements for hit-testing until you hit the bottom is not a good solution. :/
22:43
<TabAtkins>
zcorpan: Plus, it's actually technically wrong, since a child can be below a parent in stacking order.