Re: [Emacs-diffs] master 2c8a7e5: Improve diff-mode navigation/manipulation

[Tino Calancha <tino.calancha@gmail.com>]

Tino Calancha <tino.calancha@gmail.com> writes:

>> Hi. I pushed to master the fix to this and the logic to invoke the
>> auto-refinement only if the preceding motion was successful.
>
> Thank you.
> I think is not working as before.
> Consider following snippet code:
>
> M-! git show SHA1 RET
> C-x b *Shell Command Output* RET
> C-x C-q
> M-x diff-mode
> n
>
> When i use as SHA1
> e5ef59b87da5c2ddfa22f7342efe29b3eea6ed97
> previous commands successfully move point to next hunk.
>
> But if i use commit
> 6b6abe0dba6a9a2e5f78aac3814421886e7a184f
> then they don't work; i got the following error:
> user-error: No next hunk
>
> I am wondering why the second commit behaves differently than
> the first one.

In Emacs-25 the commit message is not considered part of the
first hunk:
So in the previous examples, the point is set in front of
@@ -551,23 +551,7 @@ diff--auto-refine-data
(for e5ef59b87da5c2ddfa22f7342efe29b3eea6ed97)
@@ -768,7 +768,7 @@ diff-beginning-of-file-and-junk
(for 6b6abe0dba6a9a2e5f78aac3814421886e7a184f).

After your changes the commit message is somehow part
of the first hunk; then, from the beginning of buffer
`n' will jump to the second hunk.  This is unconvenient
specially when the commit messages are long.
I prefer the way this is handled in Emacs-25.

Regards,
Tino