| 11:12 | <hsivonen> | hmm. looks like the html5lib tree construction test format may have changed againg while I wasn't paying attention |
| 11:16 | <hsivonen> | it's the #new-errors directive |
| 11:49 | <zcorpan> | hsivonen: yeah that's part of the effort to standardize the error codes, which was done for tokenizer and in progress for tree-builder |
| 12:13 | <nox> | https://twitter.com/nokusu/status/955412996775399424 |
| 15:08 | <wanderview> | annevk: am I missing something about `requestStorageAccess()`? won't every 3rd party ad/tracking iframe just call this whenever it can? |
| 15:08 | <wanderview> | without a prompt there is no dis-incentive to spamming it |
| 15:09 | <annevk> | wanderview: the reply I got suggested some kind of complicated heuristic might work |
| 15:09 | <annevk> | wanderview: I'm fairly skeptical of this entire setup though |
| 15:10 | <wanderview> | annevk: so uh, why not just grant permissions automatically when that heuristic matches? |
| 15:10 | <wanderview> | why have a method that gets spammed at all? |
| 15:10 | <annevk> | wanderview: I'm not really the right person to try to defend this proposal I think |
| 15:11 | <wanderview> | annevk: I wasn't try to ask you to... more just... I thought maybe I was missing something |
| 15:13 | <wanderview> | annevk: I asked on the issue |
| 20:52 | <smaug____> | igrigorik: there aren't resource timing on workers wpt tests, right? |
| 20:53 | smaug____ | is looking at https://wpt.fyi/resource-timing but perhaps there are some tests elsewhere too? |