freenode :: #whatwg

17 Feb 2017
05:08kochiMikeSmith: ping?
05:09kochiyoichio@ is having trouble getting his first PR merged.
05:09kochiDoes it require "w3c" membership?
06:30annevkkochi: seemed rniwa gave some feedback?
06:31rniwakochi, annevk: in fact, we don't want to merge that change until Blink makes the behavioral change
06:31rniwaI think I made that clear in the issue tracker?
06:33annevkrniwa: yeah, the referenced issue seems pretty clear, but sometimes worth restating the reqs in the PR
06:33annevkrniwa: tests would be good too btw
06:34annevkhsivonen: seems the Chrome encoding folks would be game to standardize sniffing
06:35annevkhsivonen: put it in Encoding?
06:37kochiannevk, rniwa: I see, I was just helping to resolve the warning that CI tool was complaining. Yoichi seems aware of it.
07:55hsivonenannevk: I commented on the Chrome bug. I'd like to explore the necessity of sniffing before deciding whether to standardize and in which spec.
08:00annevkhsivonen: sg
08:01zcorpan ... is mozilla intentionally using single slash in URLs now?
08:02zcorpanuh... nevermind
08:38zcorpanAnyone have Edge more recent than 13 to test
09:08MikeSmithkochi: looking now
09:08kochiMikeSmith: thx!
09:12MikeSmithkochi: so from what I have seen so far the tool that handles the IPR check is just broken
09:12MikeSmithI will ask the team contacts for the WG if they know
09:14kochiit's Friday, and that PR is not in a hurry. thanks in any case.
09:17MikeSmithwill try to get it fixed, anyway
09:20kochiMikeSmith: if there is anything yoichio@ has to do, please let him know.
09:22MikeSmithbut I think there here is nothing yoichio can to do resolve it; instead W3C systems team needs to fix the tool I guess
09:25annevkzcorpan: I can look in Browserstack
09:25zcorpanannevk: i thought BrowserStack had Edge 13 but not newer?
09:26annevkzcorpan: mine has 14
09:26annevkzcorpan: object CSSStyleSheet, 1, loaded
09:26zcorpanannevk: thanks
10:15annevkfoolip: the chromium-wpt-export-bot creates a lot of noise on the master commit log
10:16annevkfoolip: e.g., the "Do not guess UTF8 encoding" commit is repeated eight or so times
10:16Ms2gerannevk, yeah, it was turned off
10:16annevkfoolip: master commit log of web-platform-tests
10:16annevk(also, that commit seems wrong, seems better to just remove that test as it's invalid)
11:03annevkmkwst_: I thought I suggested hiding nonces at some point
11:36zcorpananother anecdote that tests+bugs is working well (also cdumez++)
12:38annevkAnyone able to make alert(someWindow.location) return a javascript URL?
12:56zcorpanannevk: var alert = () => new URL('x:');
12:56zcorpanDid not manage to break someWindow.location though
13:04zcorpanmaybe an attack could be to somehow confuse the script about what someWindow is
13:05annevkzcorpan: I wasn't really looking for an attack though
13:06annevkzcorpan: just a scenario where the Location object represents a javascript URL
13:06zcorpanoh a javascript:"something" URL. not an instance of window.URL
13:08annevkzcorpan: apologies if I just wasted a bunch of time due to a poor question
13:08Ms2gerI understood the question :)
13:19zcorpanannevk: this is about the check in .hash getter?
13:20annevkzcorpan: yeah
13:27zcorpanannevk: works in gecko. but seems like a bug per html standard
23:31KiChjangwhen reading the specs, i often come across notes that say something like "these (sub)steps are only done when ..."
23:32KiChjangdoes the note apply to the text above or below the note?
23:52jsbellKiChjang: do you have an example? That sounds like a pattern used in older specs, before we started using a more algorithmic style.
23:53KiChjangyes, one sec
23:54KiChjangstep 10
23:54KiChjangthe part where it says "To process request body..."
23:57jsbellAh... well, after staring at that one, that appears to refer to the preceding steps. That's a rather complex algorithm.
23:58jsbellWe tend to have notes like that annotating what's above it.
23:59jsbell(Usually to decipher some complex normative text into something actually comprehensible.)
18 Feb 2017
No messages
Last message: 4 days ago