R devel - May 2003 - on.exit(par(old.par)) warnings

I often use something like

  old.par <- par(set someting)
  on.exit(par(old.par))

but in R 1.7.0. I now get warnings:               
          
> old.par <- par()
> par(old.par)
Warning messages: 
1: parameter "cin" can't be set in: par(args) 
2: parameter "cra" can't be set in: par(args) 
3: parameter "csi" can't be set in: par(args) 
4: parameter "cxy" can't be set in: par(args) 
5: parameter "din" can't be set in: par(args) 
6: gamma cannot be modified on this device in: par(args) 

Is this a mistake or a new feature? If it is a new feature, is there a simple
way to avoid the warnings?

Thanks,
Paul Gilbert

>-----Original Message-----
>From: r-devel-bounces@stat.math.ethz.ch 
>[mailto:r-devel-bounces@stat.math.ethz.ch] On Behalf Of Paul Gilbert
>Sent: Monday, May 12, 2003 3:33 PM
>To: R-devel@stat.math.ethz.ch
>Subject: [Rd] on.exit(par(old.par)) warnings
>
>
>I often use something like
>
>  old.par <- par(set someting)
>  on.exit(par(old.par))
>
>but in R 1.7.0. I now get warnings:               
>          
>> old.par <- par()
>> par(old.par)
>Warning messages: 
>1: parameter "cin" can't be set in: par(args) 
>2: parameter "cra" can't be set in: par(args) 
>3: parameter "csi" can't be set in: par(args) 
>4: parameter "cxy" can't be set in: par(args) 
>5: parameter "din" can't be set in: par(args) 
>6: gamma cannot be modified on this device in: par(args) 
>
>Is this a mistake or a new feature? If it is a new feature, is 
>there a simple
>way to avoid the warnings?
>
>Thanks,
>Paul Gilbert

Paul,

You should be using:

old.par <- par(no.readonly=TRUE)
on.exit(par(old.par))

Without the  'no.readonly=TRUE' in the first call, you will store and
then try to set non-user adjustable (read only) parameters, hence the
warning.

HTH,

Marc Schwartz

>From the help page's examples section
     ## Alternatively,
     op <- par(no.readonly = TRUE) # the whole list of settable par's.
     ## do lots of plotting and par(.) calls, then reset:
     par(op)

and no, it's not new and it is not a mistake (in R).

On Mon, 12 May 2003, Paul Gilbert wrote:
> I often use something like
> 
>   old.par <- par(set someting)
>   on.exit(par(old.par))
> 
> but in R 1.7.0. I now get warnings:               
Note: that's not the same thing as the previous usage.
> > old.par <- par()
> > par(old.par)
> Warning messages: 
> 1: parameter "cin" can't be set in: par(args) 
> 2: parameter "cra" can't be set in: par(args) 
> 3: parameter "csi" can't be set in: par(args) 
> 4: parameter "cxy" can't be set in: par(args) 
> 5: parameter "din" can't be set in: par(args) 
> 6: gamma cannot be modified on this device in: par(args) 
> 
> Is this a mistake or a new feature? If it is a new feature, is there a
simple
> way to avoid the warnings?
-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595