[R] weird error from MASS::eqcsplot with postscript driver

[Rolf Turner]

The problem is the ``paper="special" '' specification in your eps() 
function.

This makes par("pin"), which eqcsplot() uses in its calculations, undefined.
The ``value'' thus produced is:

[1] NaN NaN

Which makes perfect sense if you think about it.

If you're going to use paper=special, you need to specify "height" and 
"width".

     cheers,

             Rolf Turner

On 02/06/11 07:17, Michael Friendly wrote:
> [Env: R 2.12.2, Win XP]
>
> I'm creating figures using MASS::eqcsplot to provide equal scaling of 
> the axes.  My figures work OK
> when I plot to the screen, but when I try to do the same plot as a 
> postscript file, I get an
> unexplicable error,
>
> > figframe()
> Error in if (yuin > xuin * ratio) yuin <- xuin * ratio else xuin <- 
> yuin/ratio :
>   missing value where TRUE/FALSE needed
> > traceback()
> 2: eqscplot(x = 0, y = 0, xlim = xlim, ylim = ylim, xlab = xlab,
>        ylab = ylab, type = "n")
> 1: figframe()
> >
>
> Here is a minimal source script exhibiting the problem:
>
> library(MASS)
>
> eps <- function(file="Rplot.eps", horizontal=FALSE, paper="special", 
> ...) {
>     postscript(file=file, onefile=FALSE, horizontal=horizontal, 
> paper=paper, ...)
>   }
>
>
> # Figure frame
> figframe <- function(  xlim = c(-3,3), ylim = c(-3,3), xlab = '', ylab 
> = '') {
>     eqscplot( x=0,y=0, xlim = xlim, ylim = ylim, xlab = xlab, ylab = 
> ylab, type = 'n')
>     abline( v=0, col="gray")
>     abline( h=0, col="gray")
> }
>
> # screen: OK
> op <- par(mar=c(3, 3, 1, 1) + 0.4)
> figframe()
> par(op)
>
> # eps: fails
> eps(file="test1.eps")
> op <- par(mar=c(3, 3, 1, 1) + 0.4)
> figframe()
> par(op)
> dev.off()
>
> I know this worked under an earlier version, probably ~ R 2.9.x
>