bug#10906: 24.0.94; `c-mark-function' does set the mark well

[Dani Moncayo <dmoncayo@gmail.com>]

On Mon, Mar 5, 2012 at 19:41, Alan Mackenzie <acm@muc.de> wrote:

>> > Any chance of a quick summary of how c-mark-function differs from
>> > mark-defun?
>
>> AFAIK, they differ at least in two things:
>
>> 1. `mark-defun' saves the original point location in the mark ring,
>> whereas `c-mark-function' does not.  IMO, the point should be saved,
>> because in large defuns it may jump to a remote location and you may
>> want to return back to the original position.  This bug report is
>> about this inconsistency, as you can see in the original post.
>
>> 2. Successive interactive invocations of `mark-defun' extend the
>> region to the next defuns (which I find useful), whereas
>> `c-mark-function' does not have this feature.  Bug #5525 is a request
>> to remove this inconsistency, as you can see in the corresponding
>> thread.
>
>> >> Agreed, and BTW, if `c-mark-function' is going to be revised, please,
>> >> take also this problem into account:
>
>> >> http://debbugs.gnu.org/cgi/bugreport.cgi?bug=5525
>
> Here is a patch which should fix these problems.  The logic around
> transient-mark-mode is somewhat complicated, so I would be grateful if
> you would check the patch does the Right Thing.  I haven't amended the
> doc string yet.  Thanks!

Thank you.

I've tried your patch, and it seems to work[*], at least with my usage
pattern, i.e., with `transient-mark-mode' enabled.

[*] Except for one thing:
 - If you do `C-M-h' several times in a row until the mark reaches the
end of the file, the point ends up at the end of the region, and the
mark at the beginning (which doesn't seem right).
 - If you try the same experiment with `mark-defun', the mark stays at
the end and the point at the start of the region (as expected).

-- 
Dani Moncayo