bug#46621: Copy line

[Gregory Heytings]

> Bottom line, if we want to consider this command, we should somehow come 
> up with the relevant use cases, and then weigh them against the added 
> complexity and maintenance costs.  I therefore urge people who think 
> they know these details to please speak up and contribute to this 
> discussion.

My 2 cents: a typical case where this could be useful is:

switch (foobar) {
case long_enum_value_name_a: printf ("a"); break;
case long_enum_value_name_b: printf ("b"); break;
case long_enum_value_name_c: printf ("c"); break;
case long_enum_value_name_d: printf ("d"); break;
case long_enum_value_name_e: printf ("e"); break;
case long_enum_value_name_f: printf ("f"); break;
case long_enum_value_name_g: printf ("g"); break;
case long_enum_value_name_h: printf ("h"); break;
case long_enum_value_name_i: printf ("i"); break;
default: printf ("?"); break;
}

Another typical case:

import java.util.Set;
import java.util.TreeSet;
import java.util.HashSet;
import java.util.LinkedHashSet;

I agree with Richard that this can be done with C-a C-k C-k C-y..., which 
is what I've been using in such cases, but that doesn't mean that the 
proposed command wouldn't be welcome.  I agree with Juri that Richard's 
solution "pollutes" the kill-ring with something that the user actually 
did not want to kill.  I would suggest the following (based on Juri's 
code):

(defun duplicate-line (&optional arg)
  (interactive "p")
  (save-mark-and-excursion
    (dotimes (_ arg)
      (forward-line 1)
      (insert "\n")
      (forward-line -1)
      (copy-from-above-command))))
(global-set-key (kbd "C-x r z") 'duplicate-line)