| 00:52 | <MikeSmith> | gotta like the term *special agent cluster* |
| 01:35 | <tantek> | MikeSmith: sounds like a bad spy parody with Mr. Bean |
| 03:36 | <MikeSmith> | tantek: :) |
| 03:37 | <MikeSmith> | mkwst: not tests anywhere yet for Credential Management API? |
| 03:37 | <MikeSmith> | *no tests |
| 04:26 | <annevk> | Domenic: that or there is some kind of security issue |
| 04:27 | <annevk> | Domenic: cannot think of anything though, but shared memory is complex |
| 04:38 | <annevk> | Domenic: the rationale for not changing the definition of agent/agent clusters btw is that if you change those they're no longer useful as process boundaries, since setting document.domain would end up moving you into a different process |
| 04:38 | <Domenic> | yeah, but i think their role as shared memory boundaries is more useful than their role as process boundaries, as the latter is just an implementation detail |
| 04:39 | <Domenic> | "An agent cluster is a maximal set of agents that can communicate by operating on shared memory." |
| 04:40 | <annevk> | It's not an implementation detail though, it actually matters for document.domain |
| 04:40 | <annevk> | But we'll just need a different term for that then... |
| 04:40 | <Domenic> | I don't see how agent clusters matter for document.domain. |
| 04:41 | <annevk> | Agents, not agent clusters |
| 04:41 | <Domenic> | However the definition I found maybe argues for including similar-origin since such agents can operate on shared memory by just sharing reference to an ArrayBuffer. |
| 04:42 | <annevk> | Only if document.domain is set though |
| 04:42 | <Domenic> | Sure, they can operate on shared memory by e.g. posting a message to each other telling each other to set document.domain |
| 04:52 | <annevk> | Domenic: I'm going to squash the commits if that's okay with you |
| 04:52 | <Domenic> | sure |
| 05:09 | <annevk> | Domenic: in https://github.com/whatwg/html/pull/2521#issuecomment-294333732 you mean "the two globals" |
| 05:09 | <Domenic> | annevk: thanks, updated |
| 10:48 | <annevk> | Domenic: I wish we had discussed how to define StructuredSerializeForStorage since you're prolly asleep now |