Update implementation and tests for the trap redesign

[kpreisser]

Hi, this updates the implementation and tests for the trap redesign (bytecodealliance/wasmtime#5149, bytecodealliance/wasmtime#5215).

Fixes #173

• The ExitCode property is moved from TrapException/TrapAccessor to WasmtimeException, as the WASI exit code now needs to be retrieved from an wasmtime_error_t*.
• The Frames property is moved from TrapException to WasmtimeException, since both errors and traps can have traces now.
• The functionality of passing the original .NET exception as InnerException (see Pass original .NET callback exception as TrapException's inner exception  #172) is moved from the TrapException to the WasmtimeException, since host-created traps (with wasmtime_trap_new()) are now returned as error (wasmtime_error_t*).
  (While theoretically it could happen that the returned wasmtime_error_t* has actually another cause than the error returned at the host-to-wasm transition, I assume this won't be the case, similar as to a .NET exception that would just bubble up the stack once thrown.)

Currently, tests that are still failing are the following:

• FuelConsumptionTests: Waiting for Fix OutOfFuel trap code not represented in the C API. wasmtime#5230

[peterhuene]

It appears that OutOfFuel trap code is missing from the C API, which is why the fuel tests I think are crashing.

I'll try to fix this upstream.

[peterhuene]

Fix upstream at bytecodealliance/wasmtime#5230.

[kpreisser]

I added the OutOfFuel trap code and fixed the definition of wasmtime_trap_code (see GHSA-h84q-m8rr-3v9q) , so once bytecodealliance/wasmtime#5230 is merged, I think this should be complete 🙂

Note: Previously, wasm_trap_code in wasmtime-dotnet was defined to take a pointer to a int (32-bit) which actually matched the Rust code definition rather than the C header definition (trap.h), so prior versions of wasmtime-dotnet are not affected by the out-of-bounds write.

[peterhuene]

Thanks! Now that it has merged, once the dev artifact is published with that commit, I'll restart CI for this PR and merge.

[kpreisser]

Sorry, I forgot to update the exception messages in the FuelConsumptionTests (checked with a locally built Wasmtime).

[peterhuene]

It looks like the CI is now failing for a legitimate reason in the examples, as we're currently expecting a TrapException and not WasmtimeException.

[kpreisser]

Sorry, I hadn't thought of checking the examples. Wasmtime now seems to return the not enough fuel remaining in store message as wasmtime_error_t* instead of wasm_trap_t*, so the example should probably be updated to catch a WasmtimeException, similar to the change in the FuelConsumptionTests:

wasmtime-dotnet/tests/FuelConsumptionTests.cs

Lines 85 to 89 in b7dbedf

	Action action = () => Store.ConsumeFuel(2000UL);
	action
	.Should()
	.Throw<WasmtimeException>()
	.WithMessage("not enough fuel remaining in store*");

(Note however, that the all fuel consumed by WebAssembly message is still returned as wasm_trap_t*, with the new OutOfFuel code.)

I can take a look at this tomorrow, unless you beat me to it 🙂

[peterhuene]

I'll push up the fix.

[peterhuene]

Thanks again for fixing this!