PatrickHaecker

I hope that we get an update with this simple fix, soon. But until we get it, you can copy `~/.julia/packages/PackageCompiler/nT5sD/src/embedding_wrapper.c` into your project, modify it accordingly and include the...

Fair enough :-). I need to go through some process because Enterprise, but if it works out I'll make a PR.

Thanks for doing the PR! I am a bit puzzled: To my understanding the code change in embedding_wrapper.c should not have any influence on Julia

Yes, that would be nice. But I suggest that we track that in a different topic. ``` `$(get_julia_cmd()) --sysimage=$sysimage -e 'using Pkg; Pkg.precompile()'` ``` probably has a non-zero return value...

I agree, that a debugger-based workflow will most often be quicker, however we do not have that without #3416 (at least not well integrated). The REPL-based workflow without #3416 inherently...

> You can execute arbitrary code in the local scope via the Debug Console (which admittedly is not quite as nice as the REPL, but it exists). Yes, you are...

Thanks for having a look at the proposal. > Considering the real-time requirements, the design philosophy of the rclc library is to minimize runtime dynamic memory allocation as much as...

> Could the difference in behaviour between setting `JULIA_NUM_THREADS` (lockup) and using `-t ` (works) be caused by the former **also** passing that env to the Julia subprocess actually running...

I can confirm that passing `JULIA_NUM_THREADS` to the process compiling the final system image is the problem. I'll prepare a targeted PR to only deactivate it there.

For all I know (not being the OP), this was fixed by #1042.