3:33:59no-defun-allowedI have a very silly function (defun f (y) (let ((x (/ y))) 2)). Evidently the variable X is never used, but shouldn't it not be removed because it could signal a condition if I call (y 0)?
3:35:56pjbiarebatm`: (find-if-not #'null list) ; ie. write it just as you say it!
3:36:00no-defun-allowedOn SBCL it does optimizing for (safety 3), but not on default declaimations.
3:42:42edgar-rftno-defun-allowed: I think signalling conditions for wrong aguments should be done with CHECK-TYPE or ASSERT, I wouldn't expect that the compiler cares about silly code.
3:43:52no-defun-allowedTrue, but should a compiler be able to optimise out useless code that may signal a condition like that?
3:44:36LdBeth#'no-defun-allowed: without default safety some conditions might not be signaled
3:45:08no-defun-allowedLdBeth: this was on SBCL's default safety AFAIK
3:57:19no-defun-allowedBut fair enough, pretty sure (/ 0) is UB.
3:59:29White_Flameif it's UB, then there's no mandate for it to be checked
4:59:08Godel[m]Hi, does anybody know why alambda macro is not included in the "anaphora" package on quicklisp? Is there in any other package on quicklisp that includes it?
5:43:10p_l(/ 0) is "unspecified", but with proposed error checking