01:32
<annevk>
So that is weird... For @WHATWG to stop and html5.org to continue...
01:44
<annevk>
So it seems a Python module has disappeared
01:47
<GPHemsley>
Hixie: How quickly did DreamHost say they were going to upgrade?
02:02
<annevk>
It seems "{'errors': [{'message': 'The Twitter REST API v1 will soon stop functioning. Please migrate to API v1.1. https://dev.twitter.com/docs/api/1.1/overview.';, 'code': 68}]}" might be the problem :/
02:02
<zewt>
nothing encourages using apis like forcing everyone to rewrite things
02:08
<annevk>
Yeah, not really sure what the right approach is here.
02:10
<Hixie>
is there a python twitter module we can rely on?
02:10
<Hixie>
also, woot, an anne online!
02:10
<Hixie>
there was some bug i was supposed to ask you about
02:11
<annevk>
Hixie: the GBK thing is not completely sorted out, basically
02:11
<Hixie>
ah yeah, that's the one
02:11
<Hixie>
ok
02:11
<annevk>
Hixie: there's some outstanding questions as to whether we can merge it into gb18030 or not
02:11
<Hixie>
is there some bug i can follow on that?
02:13
<annevk>
Hixie: https://www.w3.org/Bugs/Public/show_bug.cgi?id=16862
02:14
<Hixie>
awesome
02:14
<annevk>
Hixie: there's a Python module which we're using. However, I'm having a hard time understanding how Python module updates work, especially across Python versions, etc.
02:15
<annevk>
Hixie: versioned software development is not something I'm very capable at. And understanding which Python version executes and which modules it'll pick up... I don't even...
02:15
annevk
prefers browsers
02:15
<Hixie>
yeah i dunno how it works for python
02:16
<Hixie>
ok. based on that bug i'm guessing that GB18030 is fine as a default encoding.
02:16
<Hixie>
gonna call it for now until i get a specific compat complaint.
02:28
<annevk>
Twitter account will remain broken for a bit until I find someone that can fix the Python setup. Not really interested in going through that while sleep deprived.
02:28
<annevk>
Twitter account will remain broken for a bit until I find someone that can fix the Python setup. Not really interested in going through that while sleep deprived.
06:38
<matjas>
Twitter — breaking the web one API endpoint at a time
08:09
<jgraham>
(In principle I can probably help fix the Twitter thing. Or at least I generally understand how Python works)
11:01
<Ms2ger>
reyre, [SetterThrows] is Gecko-only, it isn't supposed to end up in specs
11:14
<mitsuhiko>
hey everybody
11:14
<mitsuhiko>
does anyone have some authoritive answer in regards to http headers and rfc5987?
11:15
<mitsuhiko>
it looks like browsers are currently handling this encoding for their internal usage, but xmlhttprequest returns the raw string?
11:16
<mitsuhiko>
does that mean users of the XMLHttpRequest object are supposed to do that on their own?
13:08
<SimonSapin>
hsivonen: Do you know about this? (See mitsuhiko’s question above.)
13:30
<zewt>
did something happen to the styling of the webvtt spec, heh
13:30
<zewt>
seems hard to read now (but I'm not sure why), and I don't remember getting that impression the last time I opened it
13:32
<zewt>
oh, it's because the underlined text is black, instead of blue (thanks @ web.archive.org)
13:32
<zewt>
or maybe because the spacing changed
13:37
<annevk>
jgraham: still interested?
13:38
<jgraham>
annevk: In?
13:38
<jgraham>
The twitter thing?
13:39
<jgraham>
Sure
13:39
<jgraham>
Also, probably for most other things that I previously expressed interest in
13:39
<SimonSapin>
annevk: Do you know about this? http://krijnhoetmer.nl/irc-logs/whatwg/20130613#l-308
13:39
<MikeSmith>
annevk: http://krijnhoetmer.nl/irc-logs/whatwg/20130613#l-308
13:40
<MikeSmith>
SimonSapin: jinx
13:41
<annevk>
SimonSapin: the specification doesn't care
13:42
<annevk>
SimonSapin: well, it treats values as byte sequences, not strings
13:42
<annevk>
(although it somewhat confusingly exposes them as strings...)
13:42
<SimonSapin>
mitsuhiko: ^
13:43
<mitsuhiko>
annevk: eg: client libraries have to decode the headers by hand?
13:44
<mitsuhiko>
does that also apply to outgoing headers?
13:44
<annevk>
mitsuhiko: yeah
13:44
<annevk>
yeah
13:45
<annevk>
basically the core HTTP header value type is byte sequences on top of which they created a number more specific value types
13:45
<annevk>
however, the API only knows about the core HTTP header value type
13:46
<mitsuhiko>
it's a bit of an unfortunate situation because from what i can see, no javascript high level library actually implements proper header parsing/emitting wrt to unicode
13:49
<annevk>
it'd be quite complicated to do so even
13:50
<annevk>
mitsuhiko: you might want to file a bug on http://xhr.spec.whatwg.org/ with the use case for outgoing and incoming headers you have in mind
13:50
<annevk>
mitsuhiko: it might be worth considering for an upgrade or new API
13:51
<mitsuhiko>
annevk: i will have to think about this first before i propose anything. I am just going through various unicode related issues for a server side http library right now and wanted to compare with browsers.
13:51
<annevk>
fair enough
13:52
<annevk>
in general I think they haven't really designed HTTP headers particularly well
13:53
<mitsuhiko>
it definitely got worse over time. by now 5 rfcs are involved in encoding and headers: http://greenbytes.de/tech/tc2231/ (not counting the cookie headers)
13:53
<SimonSapin>
there might have been more evolution than design involved…
13:53
<mitsuhiko>
SimonSapin: i'm not complaining :)
13:53
<SimonSapin>
I am :)
14:12
<GPHemsley>
mitsuhiko: Hey, that's a useful link. Thanks for that!
14:14
<GPHemsley>
annevk: I'm not too familiar with CSP; I think I got the directive wrong for the regular browsing context
14:14
<GPHemsley>
annevk: (I used 'connect-src')
14:14
<GPHemsley>
annevk: Which makes me wonder if there even is a directive for the regular browsing context (maybe 'default-src'?)
14:15
<GPHemsley>
annevk: Also, it makes me wonder if we need another context for 'connect-src'
14:15
<GPHemsley>
annevk: Also, it makes me wonder if we need another context for 'connect-src'
14:15
<annevk>
GPHemsley: I don't think CSP blocks following links
14:16
<GPHemsley>
perhaps not
14:52
<GPHemsley>
hmm...
14:53
<GPHemsley>
parse natural DNA into natural pre-RNA; remove introns from natural pre-RNA to make natural mRNA; serialize natural mRNA into synthetic cDNA; patent synthetic cDNA
14:55
<GPHemsley>
(that's like saying a potentially-malformed format from the wild is natural and unpatentable, but serializing it into a format that follows the ABNF results in a patentable form)
14:55
GPHemsley
throws his hands up
16:03
<zewt>
today's insanity: in mobile safari, body { 50%; } works, but html body { 50%; } gives 25%
16:04
<Ms2ger>
html, body?
16:05
<zewt>
no, the real rule is html[something] body
16:05
<zewt>
oh i see what's happening
16:05
<zewt>
oh i see what's happening
16:05
<zewt>
body { zoom: 50%; } body { body: 50%; } becomes 25%, instead of just overriding the rule
16:05
<zewt>
that's ... batshit insane
16:05
<zewt>
er, zoom:
16:08
<SimonSapin>
zewt: is zoom supposed to be the same as scale() in a transform?
16:08
<zewt>
i'm not sure of the particulars
16:11
GPHemsley
recalls 'zoom' not being an interoperable property
16:11
<zewt>
i don't think so: a transform lays the contents out the same way, then scales the result; zoom changes the size of the contents, so it affects the actual flow
16:11
<Ms2ger>
I thought it was something you used to... Force hasLayout in IE?
16:15
<GPHemsley>
annevk: I wonder if the 'sandbox' directive might be more appropriate?
16:17
<GPHemsley>
Hixie: Does HTML treat XHR/WebSocket/EventSource requests as different from <script> requests?
16:43
GPHemsley
wonders why CSP uses 400 instead of 403
17:11
<GPHemsley>
hmm... an XHR/WebSocket/EventSource request is sort of a hybrid between a script context and a (nested) browsing context...
17:12
<GPHemsley>
leaning more towards the latter
17:44
<smaug____>
slightlyoff: ping
17:44
<smaug____>
since annevk isn't here
17:46
<smaug____>
do you recall why http://dom.spec.whatwg.org/#promises-model requires then-property to be callabke.
17:47
<smaug____>
it would be simpler to just throw and catch and call reject in case then wasn't callable
17:51
<slightlyoff>
We're going to go for ghetto branding, not what's there now, methinks
17:52
Ms2ger
wonders why we're even implementing it now if we're going to have to rewrite everything anyway
17:56
<smaug____>
Ms2ger: yeah
18:51
<Hixie>
GPHemsley: yes, but maybe not in the way you care about? not sure what you're asking
18:51
<Hixie>
hm, floats are hard to use right.
18:51
<Hixie>
i want something that floats to the right of my paragraph, but comes after the paragraph in the dom
18:51
<Hixie>
not sure that's even possible
18:52
<Ms2ger>
Float your paragraph left?
18:52
<Hixie>
the paragraph is part of a long document
18:52
<Hixie>
the paragraph is part of a long document
18:53
<Hixie>
think <p> vs <p class=note> in the spec
18:53
<Hixie>
i want the .note to float to the right of the p
18:53
<Ms2ger>
!p + p.note {}?
18:53
<Hixie>
that doesn't work if i don't know their relative sizes ahead of time
18:53
<Ms2ger>
Or do you want something that works practically rather than theoretically?
18:53
<Hixie>
practically, yes
18:54
<Ms2ger>
No idea, then
18:54
<Hixie>
(also, really, we're going with the "!" thing? ugh)
18:54
<Hixie>
(:matches is such a better solution.)
18:55
<Ms2ger>
I dunno
18:55
<Ms2ger>
It's not like anybody is implementing
18:58
<GPHemsley>
what's 1?
18:58
<GPHemsley>
err, !
18:58
<GPHemsley>
what's !?
18:58
<GPHemsley>
what's !?
19:01
<Ms2ger>
If you have !a + b { color: green }, the a is green
19:03
<jgraham>
Hasn't Hixie been promoting :matches since like 2000?
19:03
<Hixie>
yeah
19:03
<Hixie>
but i'm not active in the wg
19:03
<Hixie>
so... :-)
19:04
<jgraham>
Well yeah, but it's like "<Hixie> We should solve it using :matches" is a permanent feature of the web
21:10
<GPHemsley>
IDK, ! makes sense to me, now that I remember what it means.
21:10
<GPHemsley>
since it patterns with !important
21:13
<GPHemsley>
plus, you'd have to explain to me how :matches would work
21:40
<Hixie>
GPHemsley: with !, how do you pick an element that has either a <p> next sibling, a <div> next sibling, or an <ol> next sibling?
21:41
<GPHemsley>
isn't there any() for that?
21:41
<Hixie>
any is a variant of matches
21:41
<GPHemsley>
!* :any(p, div, ol)
21:42
<GPHemsley>
or something
21:42
<Hixie>
an even better example would be, how do you match an element that has both an <ul> following subling, and an <ol> following sibling
21:42
<GPHemsley>
when would that happen?
21:42
<GPHemsley>
or you mean generically
21:42
<GPHemsley>
not immediately
21:42
<GPHemsley>
himm
21:42
<GPHemsley>
-i
21:42
<Hixie>
you could do it as a pair, as in: !* ~ ul ~ ol, !* ~ ol ~ ul
21:42
<Hixie>
but that quickly gets out of hand
21:43
<GPHemsley>
!* ul/ol
21:43
<Hixie>
with :matches it's just :matches(# ~ ol):matches(# ~ ul)
21:43
GPHemsley
needs to brush up on his selectors
21:43
<Hixie>
what's "/"?
21:43
<Hixie>
what's "/"?
21:43
<GPHemsley>
dunno
21:43
<GPHemsley>
dunno
21:43
<GPHemsley>
just made it up
21:43
<Hixie>
hah
21:43
<Hixie>
k
21:43
<Hixie>
i guess that's fair
21:43
<Hixie>
i'm making up :matches :-D
21:43
<GPHemsley>
:)
21:44
<Hixie>
anyway :matches() is just a pseudo that matches if its selector matches the node in question, with "#" being a placeholder in the argument for "the current element being checked"
21:47
<GPHemsley>
does :nth-match() help here?
21:47
<GPHemsley>
(the spec is kinda sparse on examples)
21:47
<GPHemsley>
(the spec is kinda sparse on examples)
21:48
<GPHemsley>
oh, it's E! not !E
21:50
<GPHemsley>
*! ( (~ ul) && (~ ol) )
21:51
<GPHemsley>
oh, apparently the spec is not consistent on the placement of the !
21:56
<GPHemsley>
oh, apparently :matches() is already in the spec, too
21:56
<GPHemsley>
but it only takes two arguments?
21:56
<GPHemsley>
I'm not clear on the syntax here
21:57
<GPHemsley>
*|*:matches(:hover, :focus)
21:57
<GPHemsley>
!*:matches(~ ul, ~ol) would probably be the right thing, then
21:57
<GPHemsley>
~ ol
21:57
<GPHemsley>
(are the spaces signficant?)
21:58
<GPHemsley>
+i
21:58
<GPHemsley>
(are the spaces signficant?)
21:58
<GPHemsley>
+i
21:58
<GPHemsley>
oh, hmm
21:58
<GPHemsley>
namespaces are confusing
22:06
<Hixie>
i don't think the :matches() in the spec supports the "#" part i suggested
22:17
<GPHemsley>
nope
22:17
<GPHemsley>
but I'm not sure that it needs to