11:03
<AutomatedTester>
annevk: hey, if I wanted to define something to not use relative URLs, is it sufficient to just saw "Absolute Urls with Fragment" or should it be 2 links for Absolute URLs and Absolute Urls with Fragments?
11:07
<annevk>
AutomatedTester: hmm I think those terms probably need to say that the fragment is optional
11:07
<annevk>
AutomatedTester: because currently it doesn't make much sense given the definition of URL
11:08
<annevk>
AutomatedTester: maybe file an issue and maybe copy MikeSmith since I think he was involved when we defined these
11:08
<AutomatedTester>
annevk: well, a fragment is define as 0 or more url units
11:08
<AutomatedTester>
which is why I was asking
11:09
<AutomatedTester>
I will raise a bug.
11:09
<AutomatedTester>
https://url.spec.whatwg.org/#syntax-url-fragment
11:11
<annevk>
AutomatedTester: the problem is that # is required
11:11
<AutomatedTester>
ok cool
11:14
<AutomatedTester>
bug raised and MikeSmith cc'ed
11:53
<annevk>
ta
11:55
<annevk>
yhirano_: Domenic: what is happening with upload streams? I feel like we stalled a bit
11:55
<annevk>
yhirano_: Domenic: is there some private discussion going on?
14:24
<Ms2ger>
No love for https://github.com/whatwg/html/issues/465 ?
14:53
<Domenic>
AutomatedTester: if you're specifying an algorithm you should parse the URL with no base and then if it fails, it's not an absolute URL.
14:53
<nox>
Ms2ger: Aren't these already specified in DOM-Parsing?
14:53
<Ms2ger>
nox, that's innerHTML
14:53
<nox>
Ms2ger: Damn I can't read.
14:53
<Domenic>
annevk: I'm hoping to restart the discussion next week when I go to Tokyo... I don't believe there are any real blockers.
14:54
<AutomatedTester>
Domenic: we arent specifying an algorithm, just want to say it needs to be an absolute url
14:55
<Ms2ger>
AutomatedTester, and what if it isn't? :)
14:55
<AutomatedTester>
Ms2ger: then we say we must return an error
14:55
<Ms2ger>
Sounds like an algorithm
14:56
<Domenic>
AutomatedTester: yeah that is definitely an algorithm.
14:56
<AutomatedTester>
ok
15:07
<Domenic>
Ms2ger: should we also do outerText? I didn't hear about Firefox being forced to implement that one. https://github.com/whatwg/html/issues/668
15:07
<Ms2ger>
Domenic, I'd default to no
19:52
<jsbell>
smaug____: entries-api & callback interfaces vs. callback functions; why again did you prefer the former, other than "that's what blink does" ? (foolip@ has opinions)
20:23
<smaug____>
jsbell: just because of "that's what blink does"
20:24
<smaug____>
jsbell: in general Gecko isn't following the entries API if blink does something else.
20:25
<jsbell>
smaug____: Given that the point of entries-api is to document what's necessary for web compat, please let me know if you notice anything else!
20:25
<smaug____>
sure
20:26
<smaug____>
jsbell: it is hard to know what is needed for web compat
20:26
<smaug____>
which is why following what blink does is safer for now
21:01
<Domenic>
smaug____: the issue is that blink's callback interfaces are just callback functions
21:01
<Domenic>
smaug____: except handleEvent and nodeFilter
21:53
<smaug____>
Domenic: oh
21:53
<smaug____>
surprising
21:54
smaug____
is trying to get webkitGetAsEntry() working in chrome, but it seems to return entries which can't be read, just give errors
22:26
<smaug____>
aha, https://bugs.chromium.org/p/chromium/issues/detail?id=463089 is still happening on linux, or some similar bug
23:12
<jsbell>
smaug_____: Did you get it working/
23:12
<jsbell>
?
23:14
<jsbell>
smaug_____: Also, re: what Domenic said above, any objection to me changing them back to callback functions?
23:14
<jsbell>
Or jump in on https://github.com/WICG/entries-api/issues/3