Then the codebase needs to fully (or as much as feasible) transition to the use of lexical bindings. It's not like we don't know if it will work; Common Lisp su

Then the codebase needs to fully (or as much as feasible) transition to the use of lexical bindings. It's not like we don't know if it will work; Common Lisp supports dynamic variables but you don't pay for them unless you actually use them, and good implementations of CL are fast enough. So what exactly is it that prevents the Emacs Lisp code base from becoming fast enough by abandoning dynamic variables wherever it can? By all rights it should not be any slower than Common Lisp at that point, unless some other obstacle has somehow eluded my attention.