The bytecode VM is currently documented as an alternate execution path but does not yet have a formally demonstrated 1:1 semantic equivalence with the interpreter.
This issue tracks defining the exact conformance scope, building a shared test corpus, and implementing a runner that executes identical programs on both backends and compares results and error shapes.
Goal: clearly document and enforce which language constructs are guaranteed to behave identically on the interpreter and the VM.
The bytecode VM is currently documented as an alternate execution path but does not yet have a formally demonstrated 1:1 semantic equivalence with the interpreter.
This issue tracks defining the exact conformance scope, building a shared test corpus, and implementing a runner that executes identical programs on both backends and compares results and error shapes.
Goal: clearly document and enforce which language constructs are guaranteed to behave identically on the interpreter and the VM.