05:43
<annevk>
Hmm TabAtkins still awake?
05:43
<annevk>
TabAtkins: would still like type=grammar vs for= advice
07:37
annevk
finds https://github.com/tabatkins/bikeshed/blob/master/docs/definitions-autolinks.md
09:56
<nox>
Given when URLUtils' set the input's algorithm is invoked, cannot HTMLAnchorElement's accessors be out of sync when the element's base URL changes?
10:25
<annevk>
nox: that's why we now have "reset the input"
10:26
<annevk>
nox: note that element's don't have a base URL anymore, only documents
10:26
<annevk>
nox: HTML doesn't reflect that yet
10:27
<nox>
annevk: Ok.
10:28
<annevk>
nox: also, if you're implementing this, I noticed the Location object is quite butchered with the current set of specifications :-/
10:29
<nox>
annevk: Yeah.
10:29
<nox>
annevk: URLUtils implementors in general have few in common.
10:29
<annevk>
Which is my fault...
10:30
<annevk>
That is less of my fault, and I'm hoping that we can get browsers to slowly align for most bits.
13:02
<nox>
annevk: So HTMLAnchorElement accessors should always be up-to-date with regard to the Document base URL, right?
14:45
<annevk>
nox: yes, HTMLAreaElement getters too
14:46
<nox>
annevk: Ok.
14:46
<nox>
annevk: I wanted to cache the parsed URL but that sounds like the wrong thing to do then.
14:46
<annevk>
I guess I should make the setters reset the input too...
14:46
<annevk>
Yeah, perhaps the current setup needs to be restructured in light of this...
14:47
<annevk>
nox: file a bug on perhaps removing URLUtils' url concept?
14:47
<annevk>
s/bug/issue/
14:47
<nox>
annevk: To me it looks like URLUtils try to abstract over different things that behave too differently to be abstracted in such a way.
14:48
<nox>
URL never has an input, Location never uses its input and never store anything anyway because it says goodbye the moment you change the URL, etc.
14:48
<annevk>
nox: Location might actually store the url
14:48
<nox>
annevk: Doesn't it load the new one immediately?
14:48
<annevk>
nox: but you might be correct that we need a different approach
14:49
<annevk>
nox: only for the setters, the getters keep returning the old values
14:49
<nox>
annevk: To me, it looks like the whole abstraction is just to support URLUtilsSearchParams.
14:49
<annevk>
nox: the abstraction is there to define the properties in a single place
16:32
<annevk>
nox: you might still want to cache URLs btw if they're frequently accessed and the base URL doesn't change much, and have some global base URL dirty counter that the getters check
17:01
<annevk>
MikeSmith: https://url.spec.whatwg.org/#syntax-url it now has absolute URL with fragment
17:01
<annevk>
MikeSmith: if you wanted to refer to such a thing
17:01
<annevk>
MikeSmith: remaining syntax enhancements are outlined here: https://github.com/whatwg/url/issues/33#issuecomment-131400647
17:02
<annevk>
MikeSmith: if you needed something else would be nice to hear about that soonish
17:02
<annevk>
MikeSmith: trying to resolve most open URL bugs while not much else is going on
19:07
<jojo->
hello
23:12
<TabAtkins>
annevk: Sorry, attending a funeral this weekend