Re: [PATCH, RFC] Macros expansion changes, robust symbol macros, and constant propagation

[Stefan Monnier]

>>>> +(defun byte-optimize-do-constant-propagation (let-form)
> By the way: why do we have byte-optimize forms _and_ compiler macros?

Partly/largely for historical reasons.  But note that there's more to it
than that: I tried to unify them by moving the byte-opt processing to
macroexpand-all time, but it has the downside that you then get
"spurious" warnings that are much harder to silence (e.g. "a" is unused
after optimizations but not before optimizations, and those
optimizations depend on the Emacs that compiles the code, so fixing the
warning without breaking the code on other Emacsen becomes tricky).


        Stefan