[ESS] differences in performance of ESS versions (using either XEmacs or Emacs)

[Deepayan Sarkar]

On 4/11/08, Robert M Dorazio <bob_dorazio at usgs.gov> wrote:
>
>  I've used ESS in conjunction with Xemacs on various Windows operating
>  systems for several years.  I mainly use Xemacs+ESS for R programming and
>  Xemacs for C++ programming.  Recently, I tried GNU Emacs to compare its
>  operating characteristics with Xemacs.  As part of the comparison, I
>  installed the most recent versions of the 3 software packages:
>
>  ESS 5.3.7
>  Xemacs 21.4.21
>  Emacs 22.2
>
>  Emacs has a slightly more contemporary "look and feel", but is otherwise
>  quite similar to Xemacs.  I like them both.
>
>  An expected discovery was that the current version of ESS has several
>  undesirable operating characteristics compared to the version of ESS that I
>  had been using (ESS 5.2.12).  For example,
>
>  1. The default behavior of ESS 5.3.7 is to echo a buffer's contents to R
>  when sourcing a function that is opened in the  buffer.  This only occurs
>  during the first time a function is sourced, but it is annoying.  Is there
>  a way to avoid this??
>
>  2. While sourcing an R function the first time (as described in #1), ESS
>  5.3.7 is MUCH SLOWER than ESS 5.2.12, even for only moderate sized
>  functions.  Why is that?  Is there a way to speed things up?
>
>  3. When using R2WinBUGS with Emacs+ESS+R, there is a significant delay (a
>  minute or so!!) in regaining access to the R buffer -- or, for that matter,
>  to any Emacs buffer -- once WinBUGS is finished and has returned control to
>  R.   I experience no such delay when using Xemacs+ESS+R.  This is the only
>  major difference I experienced between Emacs and Xemacs that would seem to
>  favor the latter when using ESS[R].
>
>
>  I should add that I made no special changes to Xemacs or Emacs other than
>  the following:

You have not been very clear on this, but I think you have updated to
Emacs22 (from Emacs21) the same time you updated ESS, and many of the
problems you describe are attributable to Emacs22 and not ESS.

If that's the case, then this has been reported before, with no good
solution. I tend to use Emacs21 where I can, and with Emacs22, I use

(setq ess-eval-visibly-p nil)

in my .emacs.

-Deepayan