Closed matthiaskrgr closed 1 year ago
Hmm is this some kind of "an error occurred that should not have happened" panic? I find many of those that seem to have different causes.
I may have fixed the issue in 9165542, but please verify. If I got the root cause correctly, this caused when a worker thread reports an error and in doing so races with the main compilation thread finishing with the creation of new jobs for the worker threads. I moved an abort_if_errors()
call to after all worker threads are finished to potentially fix this issue.
Looks good to me!
rustc cranelift with debug assertions from 984eab57f708e62c09b3d708033fe620130b5f39 (rustc repo)
rustc -Zcodegen-backend=cranelift <file>