02:45
<zewt_>
neat, i got mojibake out of amazon jp's account page
02:45
<zewt_>
don't think i've ever had that happen on an amazon page before
02:45
<zewt_>
(tracking info, probably a bad conversion between the shipping service in japan and amazon)
04:15
<MikeSmith>
where is the canonical definition for el.find and .findAll and .matches these days?
04:15
<MikeSmith>
I thought it was supposed to be moved into the DOM spec
04:17
<caitp>
google finds them mentioned on https://blog.whatwg.org/2011/11, seems to indicate that they're just aliases for qsa
04:17
<caitp>
and qs*
04:17
<caitp>
and presumably not implemented anywhere
04:19
<caitp>
only seeing it in mail
04:19
<caitp>
https://dom.spec.whatwg.org/#dom-parentnode-query maybe that?
04:25
<MikeSmith>
caitp: ah yeah
04:25
<MikeSmith>
thanks
04:25
<MikeSmith>
man I should know this
04:26
<MikeSmith>
way too many things to keep up with
04:27
<caitp>
the web is a really, really big framework
04:27
<caitp>
nobody could keep on top of it all all the time :O
04:28
<MikeSmith>
well right now my immediate need is just to review some test cases that Ms2ger refactored, for .findAll, etc.
04:29
<MikeSmith>
but I have no non-obsoleted spec to review them against
04:29
<MikeSmith>
so I guess I'll just pretend
05:20
<MikeSmith>
I notice IE11 still doesn't support .matches
05:21
<MikeSmith>
ah I guess they still have prefixed msMatchesSelector
08:30
<foolip>
MikeSmith: does any engine except Blink implement matches without a prefix?
08:31
<Ms2ger>
foolip, I thought Gecko did
08:31
<Ms2ger>
https://bugzilla.mozilla.org/show_bug.cgi?id=886308 says so
08:34
<MikeSmith>
yeah
08:35
<foolip>
Ms2ger: nice, https://developer.mozilla.org/en-US/docs/Web/API/Element.matches says only IE hasn't unprefixed yet
08:35
<foolip>
that all happened fast, I'm pretty sure when I unprefixed it in Blink it was the first
08:38
<MikeSmith>
http://w3c-test.org/selectors-api/tests/submissions/Opera/Element-matches.html says so now too
08:38
<MikeSmith>
gecko passes more of those than blink :-)
08:38
<MikeSmith>
caniuse seems to have up-to-date info
08:40
<MikeSmith>
http://caniuse.com/#search=matches
09:06
<foolip>
yay Gecko :)
09:07
<Ms2ger>
foolip, since you're interested, want to review that test? :)
10:14
<annevk>
Formalizing internal slots <3
10:14
<annevk>
Is anyone reviewing my single byte decoder tests?
10:16
<Ms2ger>
I probably could
10:28
<MikeSmith>
annevk: zcorpan did already, and jgraham too I think
10:28
<MikeSmith>
annevk: you need to look in critic
10:28
<Domenic_>
Formalizing internal slots <3 indeed
10:29
<Domenic_>
Hmm I seem to not be identified with nickserv on my iPad, annoying
11:35
<annevk>
Ooh, critic
11:35
<annevk>
Ooh, jgraham reviewed some stuff from the Realm thingie
11:36
<annevk>
jgraham: I'm still not intimately familiar with how apply/call/bind work; guess I'd have to read up...
11:37
<Ms2ger>
annevk, actually, zcorpan has reviewed https://critic.hoppipolla.co.uk/r/3155
11:38
<annevk>
Ms2ger: that's not the only set of tests I submitted though
11:38
Ms2ger
looks
11:39
<Ms2ger>
annevk, where?
11:39
<annevk>
Ms2ger: https://critic.hoppipolla.co.uk/r/3147
11:40
<Ms2ger>
annevk, well, you asked about decoder tests :)
11:40
<Ms2ger>
I think I'm staying out of this one
11:40
<annevk>
Ms2ger: yes, but MikeSmith made me look at critic and I found some other stuff...
11:45
<annevk>
jgraham: why is it t.step_func_done and not just t.done taking a callback?
12:00
<jgraham>
annevk: Ask zcorpan or Ms2ger. I didn't implement that one.
12:00
<jgraham>
But I guess "because we didn't think of that"
12:00
<annevk>
zcorpan: ^^
12:01
<jgraham>
Although I guess it could be argued that it's less confusing
12:01
<jgraham>
A callback in done() could be post-test cleanup or something
12:01
<annevk>
The current Pythonic naming and superfluous "func" don't really make it attractive
12:04
<jgraham>
The "func" isn't superfluous
12:04
<zcorpan>
i guess i went from function() { t.step(function() { ... t.done(); }; } to t.step_func(function() { ... t.done(); }); to t.step_func_done(function() { ... });
12:04
<jgraham>
It returns a function that when called runs a step
12:04
<jgraham>
It doesn't actually run a step
12:08
<annevk>
I see
12:08
<annevk>
zcorpan: ideas for testing XML and CSS and such?
12:09
<jgraham>
I guess soon onw will be able to write t.step_func_done(() => {}), which seems like a nice improvement
12:09
<annevk>
zcorpan: note that if you are interested mostly in label testing I think that should probably be covered separately as well
12:10
<annevk>
jgraham: yeah, I was kind of wondering if I could use those already
12:10
<annevk>
jgraham: though they have different this handling, is that going to be signifnicant?
12:10
<jgraham>
annevk: I don't know how widespread the support is
12:12
<jgraham>
The function is called like func.apply(this_obj, Array.prototype.slice.call(arguments, 2))
12:12
<jgraham>
Where the this_obj defaults to the tests object, or can be supplied
12:12
<jgraham>
s/tests/Test/
12:13
<jgraham>
So I guess it will be different in that that won't work, but if you don't use this inside the function it won't matter?
12:13
<annevk>
"t.step_funk_done is not a function" oops
12:16
<jgraham>
Sounds more like a Daft Punk album or something.
12:27
<annevk>
jgraham: I would buy it
12:37
<annevk>
rubys: I created a url team on GitHub
12:37
<annevk>
rubys: and invited you
12:38
<annevk>
rubys: currently url.spec.whatwg.org only pulls in url.html
12:38
<annevk>
rubys: if that setup needs changing let me know
12:39
<rubys>
Thanks! So to be clear, I can make direct edits to the url spec, and simply by committing url.html the whatwg URL Standard will be updated?
12:39
<annevk>
rubys: I think you might need to update your bikeshed since it's still not adding allcaps
12:40
<annevk>
rubys: to the <h1>
12:40
<annevk>
rubys: yeah, though I would appreciate some communication around major changes
12:40
<rubys>
Bikeshed itself needs to be updated as it generates the <h1>
12:40
<annevk>
rubys: yeah, TabAtkins did that
12:41
<rubys>
oh, so you mean I simply need to update and regenerate? Let me do that.
12:41
<annevk>
https://github.com/tabatkins/bikeshed/issues/278
12:41
<annevk>
rubys: it seems the same might be true for your RFC database
12:41
<annevk>
rubys: since tobie_ updated those too to point towards tools.ietf.org
12:42
<rubys>
I'm updating everything.
12:47
<GPHemsley>
Is it true that names of characters introduced into Unicode before 1992 cannot be changed?
12:47
<gsnedders>
AFAIK Unicode character names are guaranteed constant.
12:47
<annevk>
Except they still changed some
12:47
<gsnedders>
http://www.unicode.org/policies/stability_policy.html
12:50
<gsnedders>
annevk: AFAIK they only changed stuff from Unicode 1.0 to 2.0, and never since.
12:50
<gsnedders>
Like there's a documented list of errors, some pretty major, in the canonical names
12:51
<rubys>
annevk: https://rawgit.com/rubys/url/anolis2bikeshed/url.html -- better?
12:52
<annevk>
rubys: unless I'm missing something it looks identical
12:52
<annevk>
rubys: still www.ietf.org rather than tools.ietf.org links and no allcaps
12:53
<rubys>
https://github.com/rubys/url/commit/364e29c04ff2e911e78f5fac5d10ac238dc9e9ef
12:54
<annevk>
rubys: I see, I think you might have pushed that to a different branch
12:54
<zcorpan>
annevk: add_cleanup callback also runs if the test fails or times out
12:55
<zcorpan>
annevk: also if document.body.removeChild(frame) would throw for some reason it would fail to produce a result for the test
12:56
<annevk>
zcorpan: if that throws logging an error somewhere would be good
12:57
<rubys>
trying again: https://github.com/rubys/url/commit/86049dad29e79c1ac14753c67704cdf00f403a6f
12:57
<zcorpan>
annevk: putting it inside step does that
12:58
<rubys>
annevk: https://rawgit.com/rubys/url/anolis2bikeshed/url.html -- better?
13:02
<annevk>
rubys: yeah
13:02
<zcorpan>
annevk: as for xml, i imagine loading an iframe and have the server return <x>...</x> (maybe also with xml decl specifying the encoding), and then checking documentElement.textContent
13:02
<rubys>
annevk: is the bikeshed conversion ready to merge now, or is there more work that needs to be done?
13:02
<annevk>
zcorpan: it's a bit trickier than that, since technically XML has to fail on certain bytes
13:03
<zcorpan>
annevk: for css, #foo:before { content:"..." } and getComputedStyle
13:03
<zcorpan>
annevk: ah yeah
13:03
<annevk>
rubys: go for it
13:03
annevk
still has to install bikeshed
13:04
<zcorpan>
annevk: i guess for xml we'd have to separate the valid bytes and use one document per invalid byte
13:04
<annevk>
zcorpan: also, for CSS it's trickier too, due to certain bytes that have meaning in CSS
13:04
<rubys>
annevk: cool. I plan to complete setters on my clone of the url repository, and then begin work on merging in a branch in the whatwg repository.
13:05
<annevk>
rubys: by setters you mean what the spec currently uses the override state argument for, right?
13:05
<rubys>
annevk: no urgency, but I would like to move my interop results and livedom url parser demo to someplace on whatwg.org.
13:05
<zcorpan>
annevk: only " \ CR LF i think, but yes
13:05
<rubys>
annevk: yes, overriding state. Once I push, I'll discuss what I have done with you before proceeding.
13:06
<annevk>
rubys: I can change the setup of url.spec.whatwg.org now; I'll make it copy all of master instead of copying a single file and set the DirectoryIndex to url.html index.html
13:07
<zcorpan>
annevk: alternatively DOMParser instead of iframe
13:07
<rubys>
just a sec, let me check how I do my push
13:07
<annevk>
rubys: then we can use url.spec.whatwg.org/dir/ for non-spec stuff
13:07
<annevk>
zcorpan: doesn't DOMParser take a string?
13:07
<rubys>
url.spec.whatwg.org works for me. Again, just a sec
13:07
<zcorpan>
annevk: ...yes. nevermind
13:08
<annevk>
zcorpan: so it sounds like both CSS and XML are easyish, but would require a distinct setup from what I have now
13:09
<annevk>
zcorpan: I mostly added the <iframe> thingy because I already had the infrastructure
13:09
<zcorpan>
yeah
13:09
<annevk>
zcorpan: same for TextDecoder()
13:09
<zcorpan>
annevk: are you done with the PR?
13:09
<annevk>
So I would prefer doing those separately. And for the XML case it's not clear that anyone really cares
13:09
<annevk>
zcorpan: I should probably do the cleanup thingy you mentioned
13:10
<zcorpan>
k
13:11
<rubys>
annevk: I'm not currently checking in the outputs of the process, but I can clearly do so. More specifically, three subissues:
13:11
<annevk>
zcorpan: well actually, if you can verify what bytes to avoid in CSS I guess I can add CSS tests
13:11
<zcorpan>
xml could be attack surface if it uses different decoders
13:12
<rubys>
1) I have a CGI, and the shebang comment at the top of the file needs to point to a version of Ruby on the host
13:12
<annevk>
zcorpan: with I meant does anyone care about not-well-formed vs well-formed-with-U+FFFD
13:13
<rubys>
2) punycode.js, unorm.js, etc are downloaded from various places and not checked in.
13:13
<zcorpan>
annevk: ah. yeah possibly we could change that if we wanted to
13:13
<rubys>
3) the urltest-results directory is produced locally and then rsynced.
13:16
<rubys>
annevk: I guess most importantly, url.html in my branch is incomplete (it only contains parsing, not the rest of the spec).
13:16
<rubys>
annevk: I think the net of this is that this is something we should revisit when that branch is ready to merge.
13:16
<annevk>
rubys: issues 1-3 seem like something we could address by giving you access to the server
13:17
<rubys>
annevk: that works to. Dreamhost?
13:17
<annevk>
rubys: however, currently access to this domain is shared with a bunch of others, so I think I'd prefer it if Hixie created an isolated account for url.spec.whatwg.org if we're going to be running more advanced scripts
13:18
<rubys>
I'm in no particular rush, and that works for me.
13:18
<rubys>
I only have one script, that's the one that allows people to upload browser results.
13:18
<rubys>
everything else is static html, javascript, and css.
13:18
<annevk>
so yeah, let's keep the url.html setup for now and then get Hixie to make an account for that domain we can share SSH access to
13:19
<annevk>
rubys: hmm okay
13:20
<annevk>
rubys: just message me your public key then
13:20
<rubys>
https://people.apache.org/keys/committer/rubys.asc
13:23
<annevk>
rubys: you have it in a format somewhere that works for ssh-rsa?
13:25
<annevk>
rubys: https://apache.org/ seems bad btw
13:25
<rubys>
annevk: any of these work http://intertwingly.net/public/ ?
13:26
<annevk>
ta
13:27
<annevk>
hopefully no man-in-the-middle...
13:30
<rubys>
yea, yea, I do need to fix https on my site.
13:31
<zcorpan>
annevk: also U+000C
13:40
<annevk>
zcorpan: so my setup is going to be using a single <iframe> that keeps loading a new HTML page that points to a different CSS file, based on <iframe> load I know the CSS file has loaded
13:40
<aleray>
hi, if I only have one section in my page, should I enclose it in a section tag or is it unnecessary (I'm already using a main element)
13:40
<annevk>
zcorpan: both HTML and CSS file are generated; CSS file doesn't generate problematic bytes, and on the result of getComputedStyle we don't check the problematic bytes
13:40
<annevk>
aleray: not needed
13:43
<zcorpan>
annevk: you can use a backslash before the problematic bytes in css
13:44
<zcorpan>
annevk: and expect empty string for 0x10 0x12 0x13 in getComputedStyle
13:44
<annevk>
zcorpan: why?
13:46
<zcorpan>
annevk: for completeness i guess. but maybe it's not interesting to test those
13:46
<annevk>
zcorpan: no why I should I expect the empty string?
13:47
<zcorpan>
annevk: oh, because that's what \LF means in css
13:50
<aleray>
annevk, thanks
13:50
<annevk>
zcorpan: I see
13:58
<rubys>
Bikeshed produced version of URL Living Standard is now live: https://url.spec.whatwg.org/
13:58
<annevk>
zcorpan: you meant 0x0A, 0x0C, and 0x0D right?
13:59
<zcorpan>
annevk: yes
13:59
<Ms2ger>
Looks like it regressed the FileAPI reference
13:59
<Ms2ger>
And WebIDL
13:59
<rubys>
I'll fix. Any others?
14:00
<Ms2ger>
Looks like that's it
14:10
<rubys>
Ms2ger: done
14:10
<rubys>
annevk: FYI, if you have anolis installed, you have all the prereqs necessary for bikeshed. You can start here: https://github.com/tabatkins/bikeshed/blob/master/docs/install.md#common-steps
14:10
<Ms2ger>
Thanks!
14:11
<annevk>
rubys: ta
14:17
<annevk>
zcorpan: euhm so CSS is quite a bit harder
14:17
<annevk>
zcorpan: it turns out this also depends on the particular serialization rules of computed style
14:18
<annevk>
zcorpan: e.g. I don't get back U+008D, but \8D followed by a space...
14:19
<annevk>
zcorpan: e.g. this is what I get back for x-mac-ukrainian
14:19
<annevk>
assert_equals: expected "x" but got "\"�\\1 \2 \3 \4 \5 \6 \7 \8 \9 \B \E \F \10 \11 \12 \13 \14 \15 \16 \17 \18 \19 \1A \1B \1C \1D \1E \1F !\\"#$%&\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~\7F АБВГДЕЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯ†°Ґ£§•¶І®©™Ђђ≠Ѓѓ∞±≤≥іµґЈЄєЇїЉљЊњјЅ¬√ƒ≈∆«»… ЋћЌќѕ–—“”‘’÷„ЎўЏџ№Ёё
14:19
<annevk>
яабвгдежзийклмнопрстуфхцчшщъыьэю\""
14:19
<zcorpan>
annevk: hmm, yeah. would need to implement http://dev.w3.org/csswg/cssom/#serialize-a-string in js. and assume that the browser conforms to that
14:20
<annevk>
("x" was my expected test value)
14:20
<rubys>
annevk: I'm going to move my name in the acknowledgements to follow yours. I'm also going to fix a pet peeve of mine: the following statement is semantically meaningless: Except where different objects implementing URLUtilsReadOnly are identical to objects implementing URLUtils.
14:20
<rubys>
what I think you mean is that like name attributes in these two interfaces have the same definition for getters.
14:21
<Ms2ger>
annevk, ha
14:21
<annevk>
rubys: yeah, that wording is ugly
14:22
<annevk>
rubys: it should probably become part of the note as well, there's nothing normative there
14:22
<rubys>
annevk: ack
14:24
<annevk>
zcorpan: so yeah, I'll fix the cleanup thing, but leave this CSS stuff for another time
14:25
<zcorpan>
annevk: i'm hacking together a css string serializer if that helps
14:26
<annevk>
zcorpan: yeah, I hadn't actually considered building a string, that might be nicer for the XHR case too
14:26
<annevk>
zcorpan: instead of doing code point comparison
14:26
<annevk>
hmm
14:28
<zcorpan>
annevk: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/3308
14:28
<annevk>
zcorpan: however, we could push https://critic.hoppipolla.co.uk/r/3155 now
14:28
<annevk>
zcorpan: hmm let me look
14:29
<annevk>
zcorpan: throw new Error('InvalidCharacterError') should return U+FFFD, no?
14:30
<annevk>
zcorpan: this would be even better btw if it took an array of code points and returned a string
14:30
<annevk>
zcorpan: the array of code points I have...
14:30
<zcorpan>
annevk: yeah it should. i just implemented the spec literally :-)
14:43
<annevk>
zcorpan: I can put what I have for CSS decoding into a branch so you can play with it if you want
14:43
<zcorpan>
annevk: sure
14:45
<annevk>
zcorpan: https://github.com/w3c/web-platform-tests/tree/encoding-and-css
15:10
<annevk>
Domenic_: https://github.com/tc39/ should offer a template for TC39 proposals
15:10
<annevk>
Domenic_: found it, never mind
15:11
<annevk>
https://github.com/tc39/ecma262 has steps at the end
15:21
<annevk>
rubys: seems to be not as easy for me to get that working that way, I haven't been using MacPorts for a while
15:22
<rubys>
you don't need macports
15:23
<rubys>
annevk: git clone, pip install, git pull --rebase, and bikeshed update should be all you need.
15:23
<rubys>
annevk: https://github.com/tabatkins/bikeshed/blob/master/docs/install.md#common-steps
15:28
<annevk>
https://www.eff.org/deeplinks/2014/11/certificate-authority-encrypt-entire-web \o/
15:30
<annevk>
rubys: -bash: pip: command not found
15:30
<Ms2ger>
easy_install pip
15:30
<annevk>
rubys: I tried installing through running setup.py install but that did not work
15:30
<annevk>
rubys: well, it seemed to work, but then it didn't
15:31
darobin
thinks before "easy_install pip" you have to "setup.py easy_install" :)
15:31
<annevk>
easy_install seems to be there
15:31
<darobin>
that should work, otherwise brew install pip ought to too
15:33
<annevk>
thanks all
15:34
<annevk>
rubys: once I run it on whatwg/url a <main> element is moved...
15:34
annevk
wonders why it would use <main>
15:34
<zcorpan>
hmm, why does gecko have toString, valueOf and toJSON on location? http://software.hixie.ch/utilities/js/live-dom-viewer/saved/3309
15:37
<annevk>
rubys: a newline is also added before <body>
15:37
<Ms2ger>
zcorpan, [Unforgeable]
15:37
<Ms2ger>
zcorpan, everything moves off of the prototype
15:39
<zcorpan>
Ms2ger: but Location doesn't inherit from any other interface, and those members are not specified in the IDL. http://heycam.github.io/webidl/#dfn-consequential-interfaces
15:40
<annevk>
rubys: I suggest we reword the abstract to this: "The URL Standard defines URLs, domains, IP addresses, the <code title>application/x-www-form-urlencoded</code> format, and their API."
15:40
<Ms2ger>
zcorpan, they're on Object.prototype
15:40
<annevk>
rubys: I haven't committed that since it would be interesting to know why our bikeshed does subtly different things for url.html
15:41
<gsnedders>
where does Object.prototype.toJSON come from? is that some new thing in ES6?
15:41
<Ms2ger>
Is it?
15:41
<gsnedders>
It's certainly not in ES5.
15:41
<gsnedders>
Ms2ger: I do like how the solution to "I don't have pip!" is "install pip using another Python package manager" :)
15:41
<zcorpan>
Ms2ger: hmm looks like the requirement is in http://heycam.github.io/webidl/#es-platform-objects
15:42
<Ms2ger>
gsnedders, install easy_install with pip first ;)
15:42
<rubys>
annevk: what are those differences (and, I'm fine with that abstract)
15:42
<zcorpan>
why are these enumerable?
15:42
<zcorpan>
seems annoying
15:42
<Ms2ger>
No idea
15:43
<rubys>
annevk: if you do commit, github will helpfully show those diffs :-)
15:43
<gsnedders>
zcorpan: they predate the concept of enumerability
15:43
<Ms2ger>
Find bholley? :)
15:43
<annevk>
rubys: enjoy the diff
15:45
<annevk>
gsnedders: toJSON was in ES5
15:45
<rubys>
annevk: looks like it moved the </main>
15:45
<gsnedders>
annevk: where? only Date.prototype.toJSON was
15:46
<annevk>
gsnedders: JSON.stringify invokes it
15:46
<gsnedders>
No it doesn't
15:46
<annevk>
I'm not sure if toJSON should be on each object though, only on objects that make sense
15:46
<gsnedders>
Oh, it does
15:46
<gsnedders>
I'm blind
15:47
<gsnedders>
But Object.prototype.toJSON shouldn't exist per ES5.
15:47
<gsnedders>
Well, I mean, it /can/ exist, but it's a non-standard extension.
15:48
<annevk>
Yeah, that part hasn't changed in ES6
15:48
<zcorpan>
filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=27361
16:19
<smaug____>
is whatwg.org down
16:19
<smaug____>
apparently so
16:19
<smaug____>
"Whatwg.org is DOWN for everyone."
16:22
<smaug____>
up
16:29
<Domenic_>
annevk: rubys: we should move https://github.com/whatwg/streams/blob/master/index.bs#L25-L77 into the common WHATWG CSS so that you get the linkable section headers that http://streams.spec.whatwg.org/ has
16:29
<Domenic_>
I would PR it myself but apparently airplane wifi is horribly shitty
16:29
<annevk>
Domenic_: WHATWG CSS is not in GitHub so that wouldn't work
16:29
<annevk>
Domenic_: requires Hixie
16:30
<annevk>
Domenic_: we could put it on resources.whatwg.org perhaps as bikeshed.css since it only applies to bikeshed specs
16:30
<Domenic_>
annevk: or we could put it in the Bikeshed repo
16:30
<Domenic_>
ah yeah or that
16:31
<rubys>
Domenic_, annevk: I'm OK with any of these approaches.
16:31
<Domenic_>
resources sounds easiest
16:32
<rubys>
Domenic_, annevk: I don't mean the following to be defensive, but just as an accurate description of what I have done to date: what I did was push through to see what could be done, now it is time to refactor that work into the best "home" for each piece.
16:33
<rubys>
Domenic_, annevk: nevermind, I just realized that Domenic_ was talking about *his* work, not mine. Oopsie. :-)
16:54
<annevk>
Ah okay, so Let's Encrypt is signed by IdenTrust until it gets into browsers
18:21
<annevk>
foolip: I'm a bit reluctant to change the attribute stuff around again
18:21
<annevk>
foolip: unless we have some coherent plan for why we should be different
18:25
<TabAtkins>
Domenic: Yeah, please either put stuff in a common css file, or inline into the header-whatwg.include file.
19:09
<Hixie>
i'm here now
19:09
<Hixie>
looks like some people need something from me?
19:10
<TabAtkins>
Nah, we get along fine without you.
19:12
<TabAtkins>
But really, there's some styling common to all the Bikeshedded WHATWG specs that should either go into the header boilerplate, or into a stylesheet on resources.whatwg.org.
19:17
<annevk>
Hixie: I think we solved the problems some other way, though you might want to open the developer console and fix some of the WHATWG stylesheet errors
19:22
<Hixie>
which spec?
19:22
<Hixie>
https://html.spec.whatwg.org/multipage/ doesn't give me any messages in chrome
19:26
<Hixie>
annevk: i can't work out which problems you mean
19:26
<Hixie>
annevk: all the warnings firefox gives me are bogus
19:27
<annevk>
Hixie: oh it seems they are
19:27
<annevk>
Hixie: perhaps you fixed the non-bogus one I saw a while back
19:27
<annevk>
sorry
19:28
<Hixie>
did you want something with domains or something?
19:29
<TabAtkins>
Hixie: The aural stuff is bogus, though.
19:29
<TabAtkins>
You can just kill that whole thing.
19:30
<Hixie>
last i checked, emacs still supports that
19:30
<TabAtkins>
emacs is supporting deprecated specs, then, and violating current MQ ("aural" should always be false)
20:03
<hsivonen>
annevk: pretty interesting how the use case for restoring DOS code pages turned out to be better served by the DOS code page that we kept rather than the DOS code pages that we removed
20:04
<hsivonen>
I guess the lesson is to always push for use cases instead of just taking "OMG you broke my workflow" at face value
20:32
<Hixie>
MikeSmith: is cssom not in https://platform.html5.org/ ?
20:37
<Hixie>
resize events then scroll events? or scroll events then resize events?
20:37
<Hixie>
i'm thinking resize first, so you can react to the resize by scrolling
21:40
<foolip>
annevk: ok, if you WONTFIX https://www.w3.org/Bugs/Public/show_bug.cgi?id=27344 it'll at least be a clear signal that you hope that Blink changes instead of the spec
21:43
<foolip>
I suggested what I think would be the shortest and least riskiest path for Blink, since it basically amounts to removing the redundant and almost unused setAttributeNodeNS
22:18
<annevk>
foolip: I'm not really here, but I just wanted to say that I don't really have an opinion, if you could convince a non-Webkit/Chromium browser to chime in that'd help
22:46
<foolip>
annevk: ok :)
22:47
<TabAtkins>
rubys: I'm noodling on a more compact syntax for custom anchor data: https://gist.github.com/tabatkins/b9b07594dbcde338fd86
22:48
<TabAtkins>
(I really like whitespace-based nesting syntaxes. They're so natural and easy to use!)
22:49
<TabAtkins>
(And importantly, easy to edit when you rearrange things. Far better than enclosing-character-based nesting syntaxes.)
23:09
<Hixie>
(and so friendly to copy-paste across whitespace-ignoring environments like HTML)
23:43
<TabAtkins>
Hixie: Luckily, don't care! That's what <pre> is for. ^_^
23:43
<TabAtkins>
Or <plaintext> ^_^