Re: address@hidden: Partial completion]

[Chong Yidong]

martin rudalics <address@hidden> writes:

>> I think it's time to check this patch in.
>
> The last version of this with a correction supplied by Glenn Morris
> and a thick comment layer requested by Stefan is on the thread
>
> address@hidden: EVAL and mouse selection in *Completions*]
>
> please have a look.

The patch looks OK to me.

> It does _not_ handle the buffer erasure problem since I thought you'd
> take care of this in a more general way.  What do you intend to do?
> What shall I do?

The way standard completion handles this now, with my recent checkin,
is to check minibuffer-completing-symbol and not bind completion-base
size if it is non-nil (eval-expression binds
minibuffer-completing-symbol to indicate that it is currently reading
a symbol):

  (setq completion-base-size
    (cond
      ..             
     (minibuffer-completing-symbol nil)