[ESS-bugs] R arguments showing functions

[Martin Maechler]

Hi Sven,

>>>>> "Sven" == Sven Hartenstein <mail at svenhartenstein.de>
>>>>>     on Mon, 30 Apr 2007 21:23:19 +0200 writes:

    Sven> Dear ESS-core,
    Sven> first, is it a good idea to write to ESS-bugs at stat.math.ethz.ch when I
    Sven> want to inform you about changes to my elisp functions that show some
    Sven> R function's arguments? Please tell me where to write to if not.

Thanks a lot, Sven.  ESS-bugs was a good place to send this.

    Sven> Also, please let me know when I can drop making fixes
    Sven> to my code because you have already inserted it into
    Sven> ESS and fixed everything that I would probably want to
    Sven> fix. As I mentioned, I am everything but an elisp
    Sven> expert and imagine things become more elegant when you
    Sven> put it into ESS.

;-)  Well, that would depend on how much time we'd invest in
     trying to beautify your code. 

    Sven> Well, I've improved my functions today, mainly error handling stuff:

    Sven> * ess-r-args-current-function now checks whether (up-list -1) signals
    Sven> an error (namely because it is called when point is not within
    Sven> parentheses) and gives a nice message if so.

good [I'm still not sure that (up-list .) is  ``good style'' to
      move inside R code though;  but as long as I am too lazy
      to investigate...]

    Sven> * ess-r-args-current-function also checks (by regexp match) whether
    Sven> what has been extracted as function name can be a function name.
    Sven> This e.g. prevents emacs locking up when point is between
    Sven> parentheses after the comment character "#".

thanks

    Sven> * Usage of elisp's let function so that variable names are not altered
    Sven> outside the respective elisp function.

yes, that was one thing needed;  well done!

    Sven> * Nice message when no R process is running.

    Sven> * Minor changes.

    Sven> Please find the new version at:
    Sven> http://www.svenhartenstein.de/emacs-ess.php

Excellent!  Note that you have finally given me the extra small
push to do a bit more about your contribution --- in order to
make sure we have this or something better in the next version
of ESS. 

---> 

 - I have added another "d" to the file name 
 - I have changed one part of the code should it should
    supposedly work with Xemacs.

 - gotten rid of "make" warnings by using (ess-eval-(and|when)-compile)

   {oops; I will also to see what emacs-22.x or even 23.x ``warn about'';
    or then, maybe an other ESS-core member will now have an
    easy way to do this :  " cd <ESS>/lisp ; make "}

 
 - comment now says "this file is part of ESS"
   (instead of "is intended to be ...")

and now committed it -- so we can all look at it easily
(and ESS-core can update it):
https://svn.r-project.org/ESS/trunk/lisp/essd-r-args.el

It's integration into ESS is clearly not yet finished;
and but last time this came up, all those of ESS-core who voiced
an opinion where quite enthused with it.

So let me thank a lot for your great idea and diligent work for
this. Let's see how this goes further and in particular if there
are more configuration things that we should / could do in
ess-cust.el [I mention in the current code that the defvar's
should be moved to ess-cust.el and be changed to defcust's..]

Martin