Daniel Ehrenberg

I don't understand how JSON would be related to this; I thought the idea was that you delay early errors for functions, right? JSON doesn't have functions.

Should we do something parallel to modules here? One part is `` (usable inline?) and another part would be a `nobinjs` modifier to provide the legacy fallback, similar to `nomodule`....

If it were only down to this "political"/layering issue, I would be motivated to work out some kind of general solution. However, when discussing it earlier with @marjakh , I...

Thinking about this a little more: I suspect we could define AbortController and AbortSignal as [serializable objects](https://html.spec.whatwg.org/#serializable-objects) (assuming they are restricted to only in-memory), but they aren't [defined](https://dom.spec.whatwg.org/#interface-abortcontroller) that way...

Rather than forcing developers to take the low-level route of manually compiling and passing around a WebAssembly module, I imagine that the pattern within a Wasm ESM world would be...

I guess there are two questions: - How the integration should work in HTML - How the integration should work in other environments In HTML, implementations don't really have the...

@lars-t-hansen For Option 1, is there more I could read about the goals of Wasm being independent from JS? Is this about JS object semantics accidentally creeping in, or because...

I guess I meant to ask more specifically, is anyone considering that for the web embedding, threads might not give an option to share the memory with JavaScript? (There's a...

To be a little more concrete about a refactoring that could make sense to support Wasm more cleanly: Currently, functions like SetValueInBuffer operate on JS values like Number (and BigInt...

@rossberg-chromium The first two problems need to be solved in formalizing the JS.md specification, which exactly is supposed to hook up Wasm's formalization with JS for single-threaded code. There's some...