On Tue, Oct 01, 2024 at 11:03:10AM +0100, Phillip Wood wrote:

I'm afraid I still don't think that changing the default is a good idea as
it is often very difficult to correct a badly edited hunk.

This series isn't about how hard it is to fix a badly edited hunk.

In [1] you say you discarded that idea because the wording was too verbose

Not really.  I still believe that regaining the original intention of
"no" is a better option than adding new options to the interface.

I am not opposed to that change, I just think it's an unnecessary
complication.

A user who experiences problems with a badly edited hunk, edited by
themselves, will probably encounter similar issues as they would when
editing the original hunk.  I think.

I don't think reconstructing a patch is a realistic (sensible)
scenario that we should be concerned about.

The small change in the message, in this iteration, adds a bit of
clarity for them, I think:

quoted

@@ -1286,11 +1287,17 @@ static int edit_hunk_loop(struct add_p_state *s,
  		 * of the word "no" does not start with n.
  		 */
  		res = prompt_yesno(s, _("Your edited hunk does not apply. "
-					"Edit again (saying \"no\" discards!) "
+					"Edit again (saying \"no\" discards your edits!) "
  					"[y/n]? "));

[1]
https://lore.kernel.org/git/6f392446-10b4-4074-a993-97ac444275f8@gmail.com (local)