(PR#1604) Re: Platform-specific: simple assignments in Rprofile cause core dump (PR#1604)

Prof Brian D Ripley ripley@stats.ox.ac.uk
Tue, 28 May 2002 08:33:24 +0100 (BST)


On Tue, 28 May 2002 NEFTH@pacbell.net wrote:

> This only happens on platforms with a faulty implementation
> of realloc(). Some C libraries do not allow realloc() to be
> called with a NULL pointer to resize.

What other platforms give the problem?

The behaviour R assumes is required by the ISO C89 standard (at least
according to Lewine's book). Would it not be better to fix realloc rather
than R: there are many alternative implementations of malloc out there (and
R uses one for Windows).

One issue is that realloc is used in several R packages on CRAN, and
quite possibly in others.  These are likely to assume standard-conformant
behaviour.  (It is used in other places within the R sources too, but I
assume you checked all of those.)

> The assignment
>
> ISOLatin1 <- 0:255
>
> in Rprofile then causes a core dump. The attached set
> of 4 patches should be very safe to use and fix the problem.
>
> --please do not edit the information below--
>
> Version:
>  platform = i386-pc-sco3.2v4.2
>  arch = i386
>  os = sco3.2v4.2
>  system = i386, sco3.2v4.2
>  status =
>  major = 1
>  minor = 4.1
>  year = 2002
>  month = 01
>  day = 30
>  language = R

That is not current.

-- 
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 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._