| 11:14 | <nicolo-ribaudo> | If by "module maps" you mean a referrer+specifier -> evaluatedModule map probably yes, but it's populated lazily when modules are evaluated for the first time |
| 11:15 | <nicolo-ribaudo> | Is it correct that a realm can contain multiple compartments, and a compartment is "owned" by a sigle realm? |
| 11:42 | <yulia> | I would need to see in detail what gets moved, but i think moving things to host hooks may be the appropriate approach here |
| 11:44 | <yulia> | For the july 6 meeting -- can folks think about what their goals are with the module loader api? What capabilities do you want to enable? I will be touching on that, so that we can have a picture as to what problems we are looking to solve. |
| 15:08 | <Kris Kowal> | yuliaMotivating use cases that I know of so far are: bundlers and bundle runtimes, import map generators, import map emulators, hot module replacement for fast reload during development, test and benchmark watchers, and of course sub-realm sandboxing to mitigate supply chain attacks. These are what I’ve captured so far on the proposal readme and I’m very interested in hearing about other cases. |
| 15:49 | <yulia> | Thats also very useful! maybe folks can also think about the motivating cases. What I was thinking of was -- what are peoples goals. For example: Is ergonomics a goal? Is high level configuration a goal? Is support for bundlers a goal etc |
| 22:58 | <Kris Kowal> | At least one design tension stretches over the ergonomics vs web manifesto atomicity dimension! I’m interested to hear which way people lean. |
| 22:59 | <Kris Kowal> | I’ll socialize your request for the folks I know don’t hang out here. |