freenode :: #whatwg

11 Sep 2017
05:46sangwhanI'm not entirely sure which parts of Servo made it into Gecko, so asking here.. Is this https://dxr.mozilla.org/mozilla-central/source/servo/components/script/dom/htmllinkelement.rs used by Gecko?
07:04annevksangwhan: seems highly unlikely
07:04annevksangwhan: for now it's text encodings and part of the style system
07:08sangwhanannevk: Thanks, was just checking if apple-touch-icon is whitelisted in link rel within Gecko, seems like it's dealt with on the Android platform side?
07:25TimothyGuannevk: we could cite https://tc39.github.io/ecma262/#sec-ecmascript-language-types-string-type for a definition of UTF-16
07:26TimothyGuesp. when one of the refs to UTF-16 is DOMstring conversion to ES String
07:26TimothyGu*USVString
07:35annevkTimothyGu: as I said, once the real fix is in all those UTF-16 references go away
07:35annevkTimothyGu: so I don't think we should worry too much about it
07:35TimothyGuok
07:36annevksangwhan: I don't really know anything about that
07:37annevksangwhan: I've at various times suggested we create some kind of icon finder algorithm in HTML that deals with all this crap once and for all, but nobody picked up on it and I haven't made the time
07:40sangwhanI could sit down with wireshark and see what implementations do and document that in the form of a PR if anyone has time to review it..
07:41sangwhanTIL: ids in implementations are weird, and also zcorpan is living proof that I can't read English
07:44zcorpansangwhan: hehe. ids used to be weirder I believe
07:45sangwhanNone of the implementations I checked seem to acknowledge or enforce the uniqueness though, and also let whitespaces through. This probably is worth fixing, but not sure at which layer
07:46annevkWhoa, this latest thread on whatwg@ is rather surprising to me
07:46annevkThird email is from someone I think I haven't seen before, but basically spot on with advice
07:46annevkThat doesn't happen a lot
08:07annevkhayato: https://github.com/w3c/web-platform-tests/pull/4541#issuecomment-328285388
08:13hayatoannevk: thanks. That looks good to me.
09:56smaug____Which spec might define that modifying .style modifies style attribute too...perhaps HTML
10:00smaug____hmm, no
10:00smaug____foolip: perhaps you know
10:01smaug____I guess https://drafts.csswg.org/cssom/#the-elementcssinlinestyle-interface
10:02foolipsmaug____: looks like it, yeah
10:04smaug____foolip: I wonder if DOM attribute should be set, when one sets a declaration to its current value
10:05smaug____That seems to cause https://bugzilla.mozilla.org/show_bug.cgi?id=1395767
10:14nox"Mutating the declarations must set the style content attribute on the context object to the serialization of the declarations."
10:15smaug____nox: exactly
10:15smaug____that is vague
10:15noxI don't see how.
10:16smaug____does setting declaration to its current value mutate it?
10:16noxIf I don't see wording that special cases that, I take it as always doing the thing.
10:18smaug____it is not really a special case, I think. It is more about what the word "mutate" means
10:19noxSetting = mutating to me.
10:20noxAnd AFAIK there would be no precedent for such a behaviour.
10:21noxReplacing an element by itself queues a mutation record, etc.
10:54annevkThat sentence could be clarified, but I agree with nox's sense of what should happen
10:55annevkIn particular it would be better if declaration mutation was a defined algorithm that ended up invoking something that updates the style content attribute
12:18gsneddersbah, no hsivonen
12:26jgrahamHe sent dev.platform an email not long ago, so you could try email?
12:26jgrahamIn unrelated questions
12:27jgrahamDoes a websocket connection really not make a page unsalvagable?
12:27jgrahamHow are you supposed to keep that alive in the bfcache?
12:27jgrahamWhat am I missing?
12:27jgrahamgsnedders: Oh I see.
14:27gsneddersjgraham: yeah, I went for #content instead
14:38gsneddersjgraham: and now replied to his email. was just trying to be fast before I ran out.
19:37czmWith respect to Fetch, is the following statement from RFC 7231 Section 7.1.1.1 germane?
19:37czm"When a sender generates a header field that contains one or more timestamps defined as HTTP-date, the sender MUST generate those timestamps in the IMF-fixdate format."
12 Sep 2017
No messages
   
Last message: 9 days and 9 hours ago