Andreas Rossberg

IIUC, the only ambiguity is the case of a try with no handlers. But in that case, both interpretations also happen to be semantically equivalent. Would it not work to...

> It would be difficult to do though in the parser I see. > I suppose another option could be to disallow an empty `try` for the new instructions. It'd...

That mostly seems to match what I defined [here](https://github.com/WebAssembly/exception-handling/blob/exnref-1b/proposals/exception-handling/Exceptions.md) and implemented in the reference interpreter, except for the choice of [opcodes](https://github.com/WebAssembly/exception-handling/blob/exnref-1b/proposals/exception-handling/Exceptions.md#changes-to-the-binary-model). I don't feel strongly about those, but if you...

Extra flag byte sounds fine to me and is extensible. (If we wanted to save the extra byte we could also use a hack and represent the labelidx (or the...

I'm fine with or without dropping versions. I don't think the extra byte is an issue, since catch clauses shouldn't exactly be frequent, and their size is certainly dominated by...

The exnref-1b branch now has the latest instruction format (Option B').

I also updated the spec document for Option B', see #283. That should check all the spec and interpreter boxes for moving Option B' to phase 4. Now it's up...

@keithw, happy to merge this to main once the code review is done. For the test suite, see #284. Agreed regarding more binary tests, but somebody else should perhaps volunteer...

I think we need to distinguish exnref values from host values thrown as exceptions. When you catch a host value, then the exnref you get conceptually _carries_ that value, but...

I think that would still be problematic, because ToWasmValue with target type exnref would then have to convert non-WA.Exception values to exnrefs with the JS tag. But doing so it...