[ESS-bugs] ess-eval-deactivate-mark doesn't work (patch included)

[Martin Maechler]

Hi David (and ESS core),

sorry for my slowness to come back to this topic.

>>>>> "DR" == David Reitter <david.reitter at gmail.com>
>>>>>     on Wed, 6 May 2009 09:34:52 -0400 writes:

    DR> Martin,

    >> I cannot confirm the bug that your patch is supposedly fixing.
    >> Could it be that the problem is  Mac or even aquamacs specific?

    DR> Turns out it's not Aquamacs specific, and that sort of thing wouldn't  
    DR> commonly be Mac-specific, either:

    DR> Emacs -Q

    DR> (load "~/sv.aquamacs-emacs.git/aquamacs/src/site-lisp/edit-modes/ess- 
    DR> mode/lisp/ess-site.el")

    DR> M-x R
    DR> C-x b *scratch*
    DR> M-x R-mode

    DR> print("hello")

    DR> (mark line above as region with mouse or keyboard, then C-c C-r)

    DR> ess-eval-deactivate-mark ==> nil
    DR> transient-mark-mode ==> t

    DR> This even reproduces with Emacs -Q -nw, so I really can't see how it  
    DR> would be Mac-specific.

Ok, I believe you.

    >> What's the exact versions of Emacs and ESS you are using?

    DR> emacs-version => 23.0.92.19  (this is a very recent Emacs build from  
    DR> the GNU Emacs trunk)
    DR> ess-version => 5.3.10  (as included in Aquamacs - looking at my Git  
    DR> log, it's vanilla)

    DR> The newer release 5.3.11 does not seem to have changed anything in  
    DR> this department.

that's correct.


    DR> I just tested with the latest Emacs 22 branch build of Aquamacs.  It  
    DR> shows the same behavior: the region is deactivated after C-c C-r.

Well, that's what I could not believe at first, since I could
not reproduce the problem.
But now with your minimal example above I can.

So, infact, something else that I load in my Emacs setup has
always "fixed the bug" in the past ...

    >> For me, switching  ess-eval-deactivate-mark  from default t to
    >> nil, does  indeed change the "transient-mark" behavior; i.e.,
    >> the mark remains transiently visible after C-c C-r == ess-eval-region,
    >> 

    DR> It doesn't do that for me, and the default of ess-eval-deactivate-mark  
    DR> is nil.  This is set so in ess-cust.el:
    DR> https://svn.r-project.org/ESS/trunk/lisp/ess-cust.el

Yes, I know, and yes, you are true, the default is nil; 
I simply have it configured to  t   for so many years, that I
drew a wrong conclusion.

    DR> Hope that's enough info!

Indeed, thank you.
I've now tested ESS without and with your patch, indeed see that
it fixes the problem, also for Emacs 22.3 / 22.4

ESS core:  I am (finally !!)  ready for releasing ESS 5.4.

    are you too?

Regards,
Martin