Why a flag? In what case would you turn that analysis off separately from other optimizations?
What we need is a Rust interpreter, or having something like Cranelift properly integrated, to only depend on LLVM when doing release builds on the CI/CD.
The largest single bit tends to be codegen, which is why cargo check is so much faster than cargo build.
Example, development effort wasted on Scala 3 Python like syntax.
There are several performance reports that show that the major culprit is the pile of LLVM IR sent to the backend.