Sam Clegg

The nice thing about the `var origOnMsg = self.onmessage; self.onmessage = ...` technique is that the user can then decide which message should ultimiately be delivered the emscripten-generated message handler....

If the idea is to allow user code communicate with pthreads vis postMessage, and we should at least add a test for that. Should we also add some kind of...

I think it was actually this change which removes the pinned version of closure compiler: https://chromium-review.googlesource.com/c/emscripten-releases/+/6388656

Looks like this is due to (related to) and upstream issue in closure compiler: https://github.com/google/closure-compiler/issues/4126

@dschuff should we start including this? This tool doesn't seem to exist on my debian-based system, is it new? ``` $ which clang-scan-deps $ apt-cache search clang-scan-deps ```

Ah, I found it via `sudo apt-get install clang-tools-17`. Does that mean that CMake 3.28 with C++20 depends not just on the clang package but also on the clang-tools package?

I think maybe we don't build "clang/tools" at all yet on our builder.

emscripten doesn't really have support for C++20 modules yet. Its something we should probably add tests for, but for now its not something that is easy to do. Just FYI

@Boris-Rasin I'm not seeing any iseeu with simply using C++20. For example, with the above CMakeLists.txt file I can run cmake 3.2.8 just fine: ``` ./emcmake cmake -S . -B...

Thanks! I was able to reproduce now. Uploading a fix now.