R

As far as I understand it, `wasi-emulated-pthread` should only be needed if the program actually ends up referencing `pthread_{create|detach|join}`. However, this wouldn't work if you're building a .so (TIL that...

I just compared the size difference between a C++ `std::cout` hello world using 183K) this is not at all a fair comparison, but the real different is _probably_ minor?

Hmm, you are indeed correct. I forgot that you can do that. I'm still not sure how to handle documentation and user expectations for these issues though (especially since linkage...

@alexcrichton stated above that requiring users to explicitly pass `-lwasi-emulated-pthread` probably isn't what users expect. However, I was just reminded of how glibc a long time ago would require passing...

Thanks @whitequark!

The reason I've opened this issue is because every single one of the bugs here can cause GTKWave to crash (even though I agree that the security implications are not...

Nope, no "real" tool that I am aware of generates corrupt files like this. I only stumbled upon these bugs while trying to write my own tooling for them (and...

> However, it requires some more effort to bring libcxx's `std::mutex` and `std::atomic`, so we don't have a fixed timeline in this direction. fwiw, if https://github.com/WebAssembly/wasi-libc/pull/518 manages to achieve consensus...

I believe bjorn3 is indeed correct. However, I made a PR which does the former, making the -threads target behave as if `-pthread` was passed. https://github.com/llvm/llvm-project/pull/129164

I've been playing with [whitequark's port of LLVM to WASI](https://discourse.llvm.org/t/rfc-building-llvm-for-webassembly/79073) which would need not only/specifically pthread.h but instead the C++ runtime library types such as `std::mutex`. In general, LLVM is...