01:59
<MikeSmith>
annevk: pushed another commit to fix that. thanks for catching it
07:49
<MikeSmith>
hsivonen: http://bugzilla.validator.nu/show_bug.cgi?id=1010 and https://bugzilla.mozilla.org/show_bug.cgi?id=1096172 are htmlparser bugs I raised today that I think don't affect the parser in gecko. Please let me know if there's some other place I should submit patches for bugs like that -- I mean other than through bugzilla.mozilla.org
08:51
<annevk>
MikeSmith: you don't sleep long
09:02
<MikeSmith>
annevk: I get some extra sleep on the train :) going to the office 3 days a week these days
09:32
annevk
tries to reply to all the emails
09:52
<jgraham>
annevk: Mind doing mine too?
09:53
<annevk>
hah
09:54
<jgraham>
Well about 98% of my email is telling me that a test I wrote is unstable and no one either disabled it or fixed it in the last two weeks
09:56
<annevk>
jgraham: speaking of tests, really amazed with the infrastructure you guys put together for web-platform-tests
10:05
<annevk>
https://twitter.com/larsberg_/status/531641273959333889 :-)
10:22
<Ms2ger>
:)
11:03
<jgraham>
Pretty sure that Ms2ger is actually the one running the whole show
11:04
<annevk>
About time Björn Höhrmann got a replacement
11:04
<Ms2ger>
Ha
11:04
Ms2ger
starts sending rambling emails about rust packages to www-archive
11:04
<jgraham>
Why do you think they are different entities?
11:40
<Ducki>
hi
11:41
<annevk>
jgraham: so I can't edit the hosts file on Windows
11:41
<annevk>
jgraham: the workaround I found was that 10.0.2.2 points to the Mac's localhost
11:41
<annevk>
jgraham: which is a Virtualbox feature
11:41
<annevk>
jgraham: since I'm not doing cross-origin tests, that appears to be working
11:43
<jgraham>
annevk: Hmm, why can't you edit it?
11:43
<annevk>
jgraham: it claims that file is readonly
11:43
<annevk>
jgraham: changing the readonly-ness of the directory talks about admin stuff and then fails...
11:45
<jgraham>
annevk: The internet claims it should be possible to unset the readonly bit on that file :|
11:58
<Ms2ger>
annevk, while you're in a testing mood, https://critic.hoppipolla.co.uk/r/3138 might be of interest to you :)
12:06
<annevk>
jgraham: that "sign in" problem though...
12:06
<annevk>
Ms2ger: done
12:07
<Ms2ger>
I guess I never have sign-in issues because I have critic in a pinned tab
12:07
<Ms2ger>
annevk, thanks
12:50
<annevk>
Does IDL now do nested dictionaries?
12:54
<Ms2ger>
What are those?
12:57
<annevk>
Ms2ger: e.g. new Notification(..., { behavior: { vibrate: [2, 3, 2, 3] }})
12:57
<Ms2ger>
Oh, I think those always existed
12:59
<annevk>
Pretty sure that is not true
13:20
<hsivonen>
Is the item at the end of http://www.w3.org/Consortium/mission.html really everything the W3C proclaims about privacy and mission?
13:31
<annevk>
hsivonen: well, you know how much scrutiny TR/ stuff gets, this is worse
13:35
<matijs>
annevk: edit the host file on windows by running notepad as administrator (or start cmd as administrator if you prefer command line edit)
13:35
<annevk>
matijs: how does that even work?
13:41
<annevk>
hsivonen: if you have some time, would appreciate a look at https://www.w3.org/Bugs/Public/show_bug.cgi?id=19961#c3
13:42
<jgraham>
annevk: You should be able to right click items on the start menu and do "Run as Administrator"
13:43
<matijs>
annevk: Right click in the start menu is the new sudo
14:15
<zcorpan>
annevk: looks like https://www.w3.org/Bugs/Public/show_bug.cgi?id=19961#c3 was cut off. i guess you used a non-BMP character?
14:25
<annevk>
zcorpan: oh god
14:27
<annevk>
zcorpan: added a comment that doesn't have that problem
15:20
<matijs>
1f4a9 :D
15:22
<caitp>
carlooooos!
17:29
<caitp>
answer in one sentence or less, why you think the web is the greatest country on earth
17:30
<danbri>
caitp, because it's the easiest to make a backup of?
17:31
<jgraham>
"who the hell are you and how did you get into my house". Or at least that's what TimBL said when I asked him. (with apologies to xkcd)
17:32
<caitp>
first step in solving any problem is recognizing there is one
17:36
<caitp>
but jeff daniels is right, the web is not the greatest country in the world, and he doesn't know what the eff you're talking about, with all of the crazy nonsense baked into the platform, if only there was some way to make it less terrible
17:37
<gsnedders>
Greatest or least bad? :)
17:38
<caitp>
like five eyes, the web is pretty much in the pockets of the other terrible countries in the world, and as such is not any less bad
17:39
<caitp>
not that that's the webs fault, per se
17:40
<caitp>
anyways, that's your motivational talk for today, now focus on making the web a better place =)
17:40
<jgraham>
(I have literally no idea what you are talking about btw)
17:41
<caitp>
it's vague by design, you could apply it to any particular problem and motivate yourself to solve it, or to cry about how unfixable it is
17:42
<caitp>
but maybe things don't have to be unfixable
18:43
<annevk>
https://twitter.com/jbouie/status/531835851287056385 o_O
20:38
<rubys>
annevk: should javascript:alert("Hello, world!"); be a parse error? See http://intertwingly.net/projects/pegurl/liveview.html#javascript:alert%28%22Hello,%20world!%22%29;
21:21
<annevk>
rubys: yeah, at least " " is not a URL code point
21:23
<annevk>
rubys: note that I have been thinking about relaxing some of these rules, but so far I tried to follow the RFCs on conformance, except for a couple of differences with RFC 3987 that seemed wrong in that RFC
21:24
<annevk>
rubys: dfn.js breaking is really painful
21:24
<annevk>
rubys: but I can fix that next week most likely
21:24
<annevk>
rubys: you were back next week as well, right? Can we just do the bikeshed transition then?
21:27
<jarek>
Hi
21:28
<jarek>
Is there any standardisation effort to introduce cursor manager API like CursorManager in Flash?
21:28
<jarek>
this API is important for rich desktop-like application
21:32
<rubys>
annevk: I should be back next week, yes.
21:34
<tantek>
jarek - what are your specific use-cases for cursors? i.e. what scenario in what kind of "desktop-like application"? can you be more specific?
21:35
<rubys>
annevk: in case you didn't notice, I've tried to implement IDNA processing: http://intertwingly.net/projects/pegurl/idna.js I also found a unicode normalizer https://github.com/walling/unorm
21:36
<tantek>
jarek, for starters, perhaps check out the CSS 'cursor' property and what you need beyond that.
21:37
<jarek>
tantek: simillar API is present in all desktop toolkits
21:38
<jarek>
tantek: for example, I have an app with a canvas widget and a slider widget. When cursor is over canvas I want to show "zoom" cursor, but when user drags the slider I want to change the cursor to "resize" (even if user drags the cursor over canvas)
21:39
<jarek>
tantek: for starters CSS is enough, but for complex desktop-like apps you need abstraction on top of it
21:40
<jarek>
that will allow you to register/unregister cursors and assign weights to them
21:56
<Hixie>
so have you people seen this? https://www.w3.org/wiki/Webizen
21:56
<Hixie>
the w3c wants to come up with a way to even shake non-companies for money
21:57
<annevk>
rubys: oh cool
21:57
<tantek>
Hixie, I think they offer a t-shirt
21:57
<annevk>
rubys: perhaps that makes it easier to work out a syntax production for valid domain names
21:58
<Hixie>
tantek: bloody expensive t-shirt
21:59
<tantek>
Hixie, people pay lots of money for branded clothing & accessories of various sorts
22:02
<jarek>
"Add a line on the CV that shows commitment to the advances of Web standards" :)
22:03
<Hixie>
i can't believe how brazen that is, even for the w3c
22:03
<Hixie>
i love how the goals section has all the w3c's goals first
22:03
<Hixie>
that's so symbolic
22:03
<tantek>
Hixie, IIRC it was in response to *requests* from some number of folks to be able to do that
22:03
<rubys>
annevk: disallowed code points are clearly a parse error. Would ignored and/or mapped code points be a recoverable parse error?
22:04
<annevk>
Hixie: I saw it, but it didn't make much sense...https://twitter.com/annevk/status/445628553556930560 Jeff himself said it best: "Valuable membership forces the price to be too high for this program."
22:04
<annevk>
rubys: what's the difference?
22:05
<annevk>
rubys: oh wait, you're talking conformance
22:06
<rubys>
annevk: yes
22:06
<Hixie>
tantek: man, i wish people would ask _us_ if they could send us money
22:06
<Hixie>
well, except for the tax implications
22:06
<annevk>
rubys: I still had the javascript URL in my mind, not IDNA
22:06
<tantek>
Hixie, you mean people haven't?
22:06
<tantek>
we've definitely gotten requests to accept donations at microformats.org and indiewebcamp.com
22:06
<Hixie>
oh man i have the best idea for a whatwg membership program
22:06
<annevk>
rubys: yeah, that sounds about right for IDNA
22:06
<tantek>
Hixie, you could print ID cards and hand them out ;)
22:06
<Hixie>
we ask people to write their address on an envelope, and to stamp the envelope
22:06
<Hixie>
then put some money on the envelope
22:06
<Hixie>
then post it
22:06
<Hixie>
er, in, not on
22:06
<tantek>
Hixie, what's an address?
22:06
<tantek>
Is that like a URL?
22:07
<Hixie>
it's a url for the physical web
22:07
<tantek>
what scheme does it use?
22:07
<Hixie>
http://en.wikipedia.org/wiki/Address_(geography)
22:07
<Hixie>
("depends on the country")
22:07
tantek
goes to create a wikipedia redirect thanks to bad auto-linker in Colloquy
22:09
<rubys>
annevk: possible exception: uppercase is mapped to lowercase. Would uppercase be conforming or not?
22:09
<annevk>
rubys: I guess we need to allow either ASCII or Unicode as input, perhaps if input is only ASCII, require ToASCII(ToUnicode(ToASCII(input))) === input and when it's Unicode require ToUnicode(ToASCII(input)) === input
22:09
<annevk>
ooh
22:09
<annevk>
rubys: should probably not be
22:10
<annevk>
rubys: because IDNA2008 does not have such a mapping
22:10
<tantek>
Hixie I don't see any address URL schemes defined in that page.
22:10
<annevk>
rubys: so the address would not be portable for "pure" IDNA2008 systems (if those would exist...)
22:10
<tantek>
Hixie, perhaps you meant https://en.wikipedia.org/wiki/Geo_URI ?
22:10
<annevk>
rubys: but if we start taking that into account things get tricky fast
22:11
tantek
checks URL spec to see if it includes geo:
22:11
<tantek>
apparently not.
22:12
<annevk>
tantek: URL specification defines infrastructure, not all URL types
22:12
<tantek>
annevk - seems to define parts of some URL schemes here: https://url.spec.whatwg.org/#urls
22:13
<annevk>
tantek: infrastructure unfortunately relies on knowledge of some scheme names
22:13
<annevk>
tantek: it's somewhere between XML and HTML at this point
22:13
<tantek>
is there a criteria about what needs to go into that table and not?
22:14
<annevk>
tantek: yeah, it's for legacy parsing rules basically
22:15
<annevk>
anyway, you all have fun, nn
22:17
<tantek>
Hixie perhaps you need to define an "adr:" address scheme for your "address" requests. Note that vCard4 has an "ADR:" property that sorta looks like it might fit your needs, perhaps you can extract and repurpose. http://tools.ietf.org/html/rfc6350#section-6.3.1
22:18
<rubys>
tantek: FYI http://intertwingly.net/projects/pegurl/url.html#url
22:18
<jarek>
Is UndoManager spec dead?
22:21
<Hixie>
tantek: i was being deadpan to your facetious comments, sorry to confuse you :-P
22:25
<tantek>
Hixie - it's ok, I'm easily confused, especially where web standards are concerned. :P
22:26
<tantek>
so I just keep asking questions :)
22:26
<tantek>
rubys did you mean http://intertwingly.net/projects/pegurl/url.html#relative-scheme ?
22:27
<tantek>
Hixie, also, addresses are a pain (almost as much as "person" formats), and people keep unnecessarily reinventing them in much worse versions than vCard.
22:32
<Hixie>
tantek: vCard doesn't support chinese addresses :-(
22:32
<tantek>
Hixie, oh? Thought it did.
22:32
<Hixie>
tantek: we had to extend it in html's autocomplete stuff
22:32
<tantek>
any particular Chinese addresses you had trouble encoding?
22:33
<Hixie>
any with more than two locality/region fields
22:33
<tantek>
Hixie, hence the cardinality of those fields changed in vCard4
22:33
<tantek>
you can have more than one
22:33
<Hixie>
ah, cool, same thing we did in autocomplete basically
22:33
<Hixie>
(though limited to 4)
22:33
<tantek>
yes. fixed years ago.
22:34
<tantek>
for an email-list-only based standard, vCard4 is not bad, and definitely a big improvement over vCard3 in many ways.
22:37
<Hixie>
well except that it's versioned and specced in a snapshot format :-P
23:17
<rniwa>
wycats: yt?
23:17
<jarek>
rniwa: what is the status of UndoManager spec?
23:18
<rniwa>
jarek: I don’t know. it’s probably abondoned?
23:18
<rniwa>
jarek: I stopped working on it since I left Google
23:19
<jarek>
rniwa: is there any alternative API proposed?
23:19
<rniwa>
jarek: don’t think
23:19
<rniwa>
jarek: so*
23:19
<rniwa>
jarek: although Microsoft implemented some enhancements to undo via execCommand
23:20
<rniwa>
jarek: see http://blogs.msdn.com/b/ie/archive/2013/10/24/enhanced-rich-editing-experiences-in-ie11.aspx
23:20
<jarek>
Mozilla did implement the spec: https://bugzilla.mozilla.org/show_bug.cgi?id=617532
23:20
<rniwa>
jarek: wow!
23:20
<jarek>
not sure if it's just a stub though...
23:21
<rniwa>
I apploud the effort if they really did
23:21
<rniwa>
the spec was a mess at the end...
23:21
<rniwa>
it got way too complicated