Evaluating code in .Rnw (R/tex) buffers

[Anthony Rossini]

Yes, I've used it for a while.  It's nice, though I wish they'd not used .Rnw for the extension.

The process problem is not going to be easily solved, short of doing the code right (which means probably mmm-mode, not the noweb-mode hacks).  What you suggest may not work, since noweb-mode re-initializes the sub-mode and does something similar to narrow-to-region.  Part of the problem can be realized by actually doing noweb programming in multiple languages.

As far as the Sweave stuff, we probably can add a hot-key to do it.

best,
-tony


On Tue, 26 Mar 2002, Stephen Eglen wrote:

> Hi,
> 
> I have recently started using Sweave, the R/S package by Friedrich
> Leisch that allows R and LaTeX code to be intermixed in the same
> document.  It is a very nice package, if you've not seen it
> (http://www.ci.tuwien.ac.at/~leisch/Sweave).
> 
> Has anyone else started using ESS for writing .Rnw files?  If so, I
> had a couple of questions:
> 
> 1.  Everytime the point jumps out of a code chunk, the major mode
> correctly switches from R to LaTeX.  In the same buffer, if I return
> into a code chunk, and evaluate a line of code (using
> e.g. ess-eval-line-and-step), I am again asked which "Process to load
> into:".
> 
> For example, do "M-x R" then load the following into a file,
> e.g. test.Rnw:
> ----------------------------------------------------------------------
> % -*- mode: noweb; noweb-default-code-mode: R-mode; -*-
> \documentclass[12pt]{article}
> 
> \begin{document}
> 
> Some \LaTeX code
> <<>>=
> rnorm(3)
> @
> 
> Another chunk
> <<>>= 
> 2+5
> @ 
> \end{document}
> ----------------------------------------------------------------------
> 
> If I put point on the "rnorm" line, then M-x ess-eval-line-and-step, I
> am prompted which process to send the line to, before the line is
> evaluated.  If I move out of that chunk and then back in, it will then
> ask me again which process I want to send the line to.
> 
> It seems that everytime I re-enter the code chunk, the variable
> ess-local-process-name is reset to nil.  This happens since the
> variables in S-editing-alist get reset to the cdr.  In the case of
> ess-local-process-name name, there is no "default value", so if it
> already has got a value, perhaps it should keep that value, rather
> than being reset to nil?
> 
> So, my suggestion is to test within ess-setq-vars-local that cdr of
> each alist entry is non-nil before setting the value:
> 
> (defun ess-setq-vars-local (alist &optional buf)
>   "Set language variables from ALIST, in buffer BUF, if desired."
>   (if buf (set-buffer buf))
>   (mapcar (lambda (pair)
> 	    (make-local-variable (car pair))
> 	    (if (cdr pair)	;; add this test?
> 		(set (car pair) (eval (cdr pair)))))
>           alist)
>   (ess-write-to-dribble-buffer
>    (format "(ess-setq-vars-LOCAL): language=%s, dialect=%s, buf=%s, comint..echoes=%s, comint..sender=%s\n"
>            ess-language ess-dialect buf comint-process-echoes comint-input-sender)))
> 
> 
> 2.  Has anyone else got any shortcuts for running Sweave on the
> current .Rnw buffer?  At the moment, once I've finished editing the
> .Rnw file, I switch to the *R* buffer and then run the command:
> 
> > Sweave("chap1.Rnw"); system("latex chap1")
> 
> So, if anyone had any suggestions on keybindings/approaches, I'd be
> glad to hear of them.
> 
> 
> Best wishes,
> Stephen Eglen
> 
> -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> ess-help mailing list -- To (un)subscribe, send
> subscribe	or	unsubscribe
> (in the "body", not the subject !)  To: ess-help-request at stat.math.ethz.ch
> _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> 


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
ess-help mailing list -- To (un)subscribe, send
subscribe	or	unsubscribe
(in the "body", not the subject !)  To: ess-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._