This flag allows the developer to see the result of linking even if it fails the verifier, as a step in debugging cases where the linked module fails the verifier. Same as opt -disable-verify.
No test is introduced. I don't want to write a test that relies on any manner of producing an invalid output that is a bug that should be fixed, as then the test would break if a bug is fixed.
There's an alternative that I considered but did not implement. Unlike opt, llvm-link has many intermediate outputs as it links more files and runs the verifier on each one. We could instead have a flag that outputs a module and stop after running the verifier and seeing a failure. This would allow us to find issues closer to the source, but it also requires that the verifier not crash on the invalid input in question.