[R] using tcltk in R under ESS/XEmacs on Windows

Liaw, Andy andy_liaw at merck.com
Fri Sep 24 23:48:18 CEST 2004


> From: Marc Schwartz
> 
> On Fri, 2004-09-24 at 15:02, Liaw, Andy wrote:
> > Sorry for the cross-post.  Not sure where the problem is...
> > 
> > A while back I posted an R function to R-help:
> > 
> > cd <- function (dir = tclvalue(tkchooseDirectory()), 
> saveOld = FALSE, 
> >     loadNew = TRUE) {
> >     stopifnot(require(tcltk))
> >     if (saveOld) 
> >         save.image(compress = TRUE)
> >     setwd(dir)
> >     rm(list = ls(all = TRUE, envir = .GlobalEnv), envir = 
> .GlobalEnv)
> >     if (loadNew && file.exists(".RData")) {
> >         loaded <- load(".RData", envir = .GlobalEnv)
> >         return(invisible(loaded))
> >     }
> > 
> > where the default value for the `dir' argument is to run 
> the tcltk directory
> > chooser and get the directory name chosen.  (Thanks to 
> Prof. John Fox for
> > the tcltk part!!)  While this function works fine under 
> Rgui on Windows, it
> > doesn't work when running R within ESS (5.2.3) and XEmacs 
> (21.4.13).  The
> > directory chooser never shows up, and dir just gets the 
> empty string.  Does
> > anyone have any idea what could be the problem?  I'd very 
> much appreciate
> > any pointers.
> > 
> > Best,
> > Andy
> 
> Andy,
> 
> This works under FC2 using ESS 5.2.3 with XEmacs version 21.4.15, so
> presumably there is something specific to the Windows implementation?

Given Prof. Fox's follow-up and your obvservation, I guess the problem _is_
Windows-specific. 8-(
 
> Also, two things:
> 
> 1. You are missing a closing brace above, which I presume may be a
> simple copy and paste issue.

Yes.  My apologies.

> 2. If you successfully change the directory, the cd() 
> function itself is
> deleted from the global environment via your rm(...), as you currently
> have it implemented. I am not sure if this is intentional or not.

Well, sort of.  I've placed it in a small package along with other handy
stuff, so that won't be a problem.

Best,
Andy
 
> HTH,
> 
> Marc
> 
> 
> 
>




More information about the R-help mailing list