Malcolm Still

Enough WASI to run doom was implemented in https://github.com/malcolmstill/zware/pull/206. We want a complete implementation. To that end we need to: a) implement the whole spec b) test our implementation...there seems...

I was going to open a new issue for this but since writing / encoding performance is mentioned here latterly, I'll add to this discussion. We have an issue where...

@TUSF thanks for the issue and apologies for the current lack of documentation. There are some examples in the `testrunner`: Define a function taking the `zware.VirtualMachine` as a parameter: ```zig...

> Mind if I ask if there was a reason for that? At least at the moment, there is fixed function pointer type for these host functions. That doesn't leave...

Thinking we might also expand `test/interface` to, given a `.wasm` file, dump out skeleton implementations + function loading

@TUSF I've add your suggestion of a `addHostFunction` function in https://github.com/malcolmstill/zware/pull/196 (currently explicitly taking slices of `ValType` but would be nice to use a tuple). Will look at your other...

@TUSF I've added in https://github.com/malcolmstill/zware/pull/197 a new program `generate_interface` that can take `.wasm` file and generate some stub host functions and a bunch of `addHostFunction` calls.

https://github.com/malcolmstill/zware/pull/199 adds wrappers for _exported_ functions

I've renamed `generate_interface` to `zware-gen` and moved to `tools/zware-gen`

In https://github.com/malcolmstill/zware/pull/206/ `zware-gen` now doesn't generate stubs for wasi functions, rather we now implement (some of) wasi and by default `zware-gen` will point the imports at these implementations