[R] cannot plot too much (using Mac interface)

[Alessandro Gagliardi]

Thanks.  And sorry for abusing the word crash.  I understand the
difference.  However, now that you mention it, there is a case where I
think R is crashing.  I'm not sure if it's R or if it's the Mac OS
interface, but if I try to plot something complicated (or a series of
complicated things) while rendering them at the same time, R will
freeze and eventually fail.  I've been trying to avoid doing that,
creating an object separately and then ploting that object, for
example, or plotting one thing at a time instead of using a for loop.
But sometimes I forget and end up loosing my work because of it.  Next
time it happens, I will copy down what the command was that does it,
though I'm not sure how useful that will be because it usually depends
upon some of my own functions which are, in turn, built on the
tseriesChaos package.  I figure it's R running out of memory or
something.  Any ideas?

-Alessandro

On 8/9/06, Prof Brian Ripley <ripley at stats.ox.ac.uk> wrote:
> Hmm.  'time' is in package stats, and you seem to have saved an object
> that is pulling in the namespace 'tseries'.
>
> Loading the workspace is done before loading the standard set of packages.
> This appears to be a bug in tseries, for if its namespace needs 'time', it
> should import it from 'stats', and it is not even depending on 'stats'.
>
> I can confirm that:
>
> gannet% env R_DEFAULT_PACKAGES=NULL R
> ...
> > loadNamespace("tseries")
> Error: object 'time' not found whilst loading namespace 'tseries'
>
> That is a matter for the 'tseries' maintainer (Cc:ed here).  If you
> try library(tseries) at that point you find
>
> > library(tseries)
> Loading required package: zoo
> Error: object 'aggregate' not found whilst loading namespace 'zoo'
> Error: package 'zoo' could not be loaded
>
> so package zoo has a similar problem (maintainer Cc:ed).
>
>
> What can you do?  The simplest is to rename the workspace and load()
> afterwards as you have done.  But before saving, remove any objects which
> have a dependence on tseries.
>
>
> On Tue, 8 Aug 2006, Alessandro Gagliardi wrote:
>
> > Lately, when I try to open R I get the following error message:
> >
> > Error: object 'time' not found whilst loading namespace 'tseries'
> > Fatal error: unable to restore saved data in .RData
> >
> > If I rename .RData to RData.RData and then try opening R again it
> > works.  Then I can load("RData.RData") without a problem.  But if I
> > try saving my workspace (as the default, ".RData") and reload R it
> > crashes all over again.
>
> R does NOT crash.  It is objecting (correctly) to your saved workspace.
> Please see the posting guide, which specifically asked you not to abuse
> that word.
>
> > I don't know how to get this 'time' object back.  (I must have removed
> > it by accident at some point.)  Any ideas?
>
>
> --
> Brian D. Ripley,                  ripley at 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
>


-- 
Alessandro Gagliardi
Integrative Neuroscience Program
Rutgers University Mind Brain Analysis
alessandro at gagliardi.name

"The opposite of a correct statement is a false statement.
But the opposite of a profound truth may well be another profound truth."
-Niels Bohr