Martin Liška
Martin Liška
The PR received a couple of fixed that very identified and reduced from running the RISC-V Singlepass compiler on tests for the following Rust crates: - aho-corasick - bstr -...
Good point, that's something we should easily notice in the CI I guess.
@syrusakbary: General question about WASMI and WAMR - why are we using C API bindings instead of their native Rust crates? Similarly, there's even an existing V8 Rust binding: https://github.com/denoland/rusty_v8.
> LGTM, but can we turn the overly verbose `cfg`s into a macro please? Done that. @theduke please check my change hasn't changed the expected build behavior?
`cargo tree --workspace` shows we've got the crate as transitive dependencies for the following top-level ones: ``` │ ├── sha2 v0.10.9 │ │ ├── cfg-if v1.0.3 │ │ ├── cpufeatures...
To be honest, I haven't managed to reproduce it locally: ``` cargo +nightly rustdoc --features cranelift,llvm,singlepass,wasmer-artifact-create,static-artifact-create,wasmer-artifact-load,static-artifact-load Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.15s ``` both for the current...
I agree with @syrusakbary - having a reasonable default backend makes sense. What's the specific reason for disabling default features?
@syrusakbary am also interested in what's missing for the middleware system?
Noticed the progress bar is not displayed if I compile `php`: ``` cargo r --features=singlepass,llvm -p wasmer-cli run --disable-cache --llvm -- php --version ``` If I compare the back-trace leading...
CC: @Arshia001