Feat/auto activate emsdk

[bulenty584]

As discussed with @ryanking13, this pr automatically activates emscripten for users (e.g. running source emsdk.sh) and removes the need for them to do it manually. Changes are as follows:

Added activation and error handling in xbuildenv.py
Removed erroneous output after installtion in cli/xbuildenv.py

Previous PR was closed due to a branch mistake

[ryanking13]

Thanks for working on this @bulenty584! I have a few comments.

1. Calling source inside a subprocess does not affect the local shell. If you see the emsdk_env.sh you can find that it uses the eval trick to set the env variable for the local shell.

2. I am okay with running emsdk_env.sh after running install-emscripten command, but since it is volatile after the user closes the shell and reopens it, it would be nice to make a separate command for it, such as pyodide xbuildenv activate-emscripten.

3. Actually, what I really prefer as a long term solution is to automatically adding emcc to the path when a user runs pyodide build or pyodide build-recipes without requiring the user to activate emcc manually. Of course, we can have both options (activating emcc explicitly or injecting path implicitly). But, the problem in (1) can be solved by storing the env variables as an internal variable in Python and passing them every time we call subprocess.

But you don't need to handle (3) in this PR. I am happy to accept this PR with (1) and (2) solved.

[bulenty584]

Hey @ryanking13 so sorry I haven't been very active as I have had some midterms. I've been looking at the codebase to see how I could directly do #3 you outlined above. My question is exactly what is the problem? My understanding is the user must add relevant Emscripten env vars and adjust their path to point to the new emsdk but I am confused as to how adding emcc to their path achieves the same goal or functionality? Thank you!

[ryanking13]

sorry I haven't been very active as I have had some midterms.

No worries :) We are all volunteers, and each person should work according to their own schedule.

My understanding is the user must add relevant Emscripten env vars and adjust their path to point to the new emsdk but I am confused as to how adding emcc to their path achieves the same goal or functionality?

What source emsdk_env.sh does is adding some env vars to the shell. We are also modifying the env vars that we pass to the subprocess call here. So if we update the env vars inside the Python code there, it basically has the same effect.