Thanks. I'll take a look at your patch later. As for the edit buffer, you can always kill the frame and opening it again. That's always completely safe, as it never performs any destructive operation until you press C-c C-c.
You can also undo as far back as you want. I suppose I could add a revert operation that replaces the content of the edit buffer with the current function definition from the interpreter. The effect would be identical to killing the buffer and opening it again. I am curious though as to what specific workflow you have where you need it. :-) Regards, Elias On 15 May 2014 02:43, "David B. Lamkins" <dlamk...@gmail.com> wrote: > I have two feature requests for gnu-apl-mode: > > 1. Allow an escape to IBM-style function listing and editing. IOW, if > you see something like ∇fname[⎕]∇ (in the general case: ∇fname[<stuff>), > then let GNU APL handle the request directly. (At present, attempting > this in gnu-apl-mode yields `Error when parsing function definition > command' in the minibuffer.) > > 2. (This is unrelated to 1.) When editing a function in a gnu-apl-mode > buffer, provide a way to revert the definition without closing the edit > buffer. > > Finally, a question somewhat related to 2: Is it safe to use kill-buffer > to abandon an edit? > > > >
