[R] problems when compiling package 'norm'

ripley@stats.ox.ac.uk ripley at stats.ox.ac.uk
Tue Jan 21 10:23:02 CET 2003

2.95.3 was and is fine, and prior to `2.96' the advice was to use it,
to the frequent posters on R-help who had exactly the symptoms you 
described with `2.96'.  Note that Linux distros usually shipped with
obselete versions (e.g. egcs) at that point.

Did you actually use 2.95.3 (or 2.95.2)?

One problem with `2.96' is that it was a snapshot of the development 
version, and there are many different compilers out there under that 

On 21 Jan 2003, Jari Oksanen wrote:

> On Tue, 2003-01-21 at 10:26, ripley at stats.ox.ac.uk wrote:
> > Also, you are using gcc `2.96' and that is known to be broken in various 
> > ways, notably with g77.
> > 
> I don't want to defend Red Hat's unfortunate choice of their own gcc
> ("2.96") -- and it is even more surprising that other Linux distributors
> have followed their design. However, my experience is that g77-2.96 was
> the first more or less working Fortran compiler in Linux. The previous
> ones simply gave segfaults in R libraries or even crashed R when loading
> a library. Well, I don't use gcc 2.96, so this is an academic defence
> now...

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

More information about the R-help mailing list