Posts tagged with "nukneval.el"

Back when I first really got into writing Emacs Lisp code, one of the first things I got very used to and really fell in love with was being able to eval-last-sexp (C-x C-e) the code I was writing, either to test it right there in the buffer, or to cause it to be bound so I could use it elsewhere. It was so different from any other mode of working I'd used before and it was really addictive as a way of hacking on code.

Also quite quickly I got used to the fact that eval-last-sexp wasn't so helpful with things like a defvar or a defconst, if I was changing them up to try out new ways of doing things with the code I was working on; there I had to remember to get used to using eval-defun (C-M-x). Hardly a great problem, but something to keep in mind¹.

Pretty quickly, as I worked on longer packages, I found myself wanting to, in effect, unload a whole buffer of code and evaluate it again. From this desire came nukneval.el.

The original version of this has been sat around since 2002 or so, perhaps a little earlier, and has served me well every time I've been messing with a new package. While I suspect there is (now, perhaps was then too?) a better way of doing things, the approach used in nukneval helped me learn some things and served me well (and still does). Now it's muscle-memory to run it.

The way it works is quite simple: go to the start of a buffer, read each form, check if the car is of a given list of symbols, decide if it's something I want to unbind, and then pick either makunbound or fmakunbound and use that on the symbol. Finally, once the end of the buffer has been hit: eval-buffer.

I've just released v1.3 as part of my slow wander through my old Emacs Lisp packages, with this release cleaning up a deprecated use of setf to move point, and also rewriting the code so it's a bit cleaner and also gives better feedback.