See #2 (comment) item 1 for some background.
Something that should probably be worked on in the near future is some system that prevents processes from segfaulting in weird ways if the entry point passed to loom::ConfigureStack somehow returns. The current solution is to abort() when this happens, and that's not a good thing for fault tolerant code.
The ideal goal would be to log the event (probably to Abseil Logging) and jump back into the worker to clean that fiber up and identify a new fiber to continue execution with.
See #2 (comment) item 1 for some background.
Something that should probably be worked on in the near future is some system that prevents processes from segfaulting in weird ways if the entry point passed to
loom::ConfigureStacksomehow returns. The current solution is toabort()when this happens, and that's not a good thing for fault tolerant code.The ideal goal would be to log the event (probably to Abseil Logging) and jump back into the worker to clean that fiber up and identify a new fiber to continue execution with.