Re: repeat-complex-command too complex

[despen]

Barry Margolin <barmar@alum.mit.edu> writes:

> In article <ick4l5okgl.fsf@verizon.net>, despen@verizon.net wrote:
>
>> I  often need  to redo  query-replace, usually with slightly different
>> arguments.
>> 
>> I first do M-% aaa bbb.
>> 
>> Then I want to repeat the command to maybe replace xxx by bbb.
>> 
>> I do C-x ESC ESC (repeat-complex-command)
>> 
>> I then see at least 2 lines in the mini-buffer reading:
>> 
>> redo: (query-replace "aaa" "bbb" nil (if
>>  (and transient-mark-mode mark-active)
>>  (region-beginning))
>>  (if (and transient-mark-mode mark-active) (region-end)))
>> 
>> Why is all that ?JUNK? there.  It's not so easy to get back to
>> the "bbb" part and change it.
>
> M-< will get you to the beginning.

I've been doing that but I was looking for something better.

>> I'd really like to see something like:
>> 
>> redo: (query-replace "aaa" "bbb")
>> 
>> I tried this with emacs -q and got the same results.
>
> When you use M-%, it automatically confines its operation to the region 
> if you use transient-mark-mode and have a region enabled.  This is 
> implemented in the function's (interactive) specification.
>
> When you repeat the command, it has preserve that functionality.  Since 
> you're no longer invoking the command interactively, but through the 
> low-level Lisp interface, the expressions to do this defaulting has to 
> be included explicitly.

Yes, some tests verify what you describe.
(I just changed the "redo" command to the shortened form.)

It's pretty unfortunate.  I actually have repeat-complex-command
bound to a key and use it a lot.  Plus I never use transient mark
mode.  It's too visually disturbing.

Thanks.