| 02:58 | <annevk> | mkwst: so I think I figured out what this document.domain feature policy needs to do for actual isolation. Create a new kind of agent that can only be reused by globals whose document also has the bit set. document.domain isn’t the interesting feature, agents and agent clusters are. That would prevent messaging SAB cross-origin but same-site, for |
| 02:58 | <annevk> | instance. |
| 02:59 | <annevk> | mkwst: that is a lot more involved though |
| 15:10 | <yoav> | Hey! I'm revamping the buffering logic and have a question regarding event firing. Current definition says that that the event should fire with its "bubbles" attribute set to true https://w3c.github.io/resource-timing/#dfn-add-a-performanceresourcetiming-entry |
| 15:11 | <yoav> | the event is attached to the Performance interface, so I'm not sure what bubbling or capturing would actually mean in this context |
| 15:11 | <yoav> | wanted to verify that I'm right and I can simply remove the "bubbles" attribute altogether |
| 15:13 | <yoav> | s/buffering logic/resource timing buffering logic/ |
| 15:25 | <annevk> | yoav: you can, but add a test |
| 15:26 | <annevk> | yoav: it's indeed best to use the defaults in the DOM Standard unless there's a specific reason not to |
| 15:26 | <yoav> | annevk: what should I be testing? where would an event on `window.performance` bubble? to `window`? |
| 15:28 | <annevk> | yoav: on the event object that's dispatched, simply add assert_false(e.bubbles) to an existing test |
| 15:28 | <annevk> | sorry, in a listener |
| 15:28 | <yoav> | oh, ok |