Re: C-f in isearch's non-incremental minibuffer edit

[Stefan Monnier <monnier@iro.umontreal.ca>]

> I see your point; however, since C-f at the end of the buffer would
> otherwise be an error, this feature seems to be harmlessly
> upward-compatible.  Does it actually cause trouble or inconvenience,
> or was it just surprise?

I found it both surprising and inconvenient, since I then had to use undo
and be careful trying to figure out which change was "mine" and which change
was from C-f.
I can definitely live with it, but I don't think the extra feature is
important enough.

>     While I'm at it, I also suggest we make isearch-edit-string use the usual
>     history functionality, so it behaves even more normally.

> Could you tell me, at the level of user-visible behavior, how its
> currently differs from the usual history functionality?

No.  It basically works the same, although maybe it doesn't skip
things in quite the same order.  But it obeys user's choice to rebind M-p
and M-n in minibuffer-local-map (I have them bound to things similar to
next-complete-history-element).  And it's less code.


        Stefan