| 07:50 | <annevk> | GPHemsley: you have write access |
| 08:15 | <annevk> | At least back in the CSS 2.1 I heard claims that layout wasn't that hard (modulo some tricky bits), but that dynamic layout is. That is, there were independent layout engines, but they didn't allow node tree manipulation and adjust layout accordingly. But that was like fifteen years ago so... |
| 12:51 | <Luca Casonato> | What exactly is the spec basis for the "Revoke blob URL after creating Request, will fetch" WPT from FileAPI/url/url-with-fetch.any.js succeeding? |
| 12:53 | <Luca Casonato> | The test says that "Request should take a reference to the blob as soon as it receives it in open()". There is no open() in spec or in the test though, and there is no relevant steps in the Request constructor as far as I can tell. |
| 12:53 | <Andreu Botella (he/they)> | The URL parser sets the url's "blob URL entry" at the time of parsing (https://url.spec.whatwg.org/#url-parsing step 4) |
| 12:54 | <Luca Casonato> | Ok, TIL. Thats magical |
| 12:54 | <Andreu Botella (he/they)> | yep |
| 12:55 | <Luca Casonato> | I feel like this is a test directly taken from XMLHttpRequest (which does have .open()) - that would explain the comment in the test at least |
| 13:04 | <smaug> | hmm, https://html.spec.whatwg.org/#focus-processing-model:shadow-including-inclusive-ancestor why top level bc has such special case? |
| 14:00 | <annevk> | Luca Casonato: seems likely, comment can be updated |