R devel - Nov 2003 - eigen of matrix of Inf hangs (PR#5406)

Full_Name: Patrick Burns
Version: 1.8.1
OS: Windows 2000
Submission from: (NULL) (81.129.3.151)


The command:

eigen(matrix(Inf, 2, 2))

causes version 1.8.1 to hang in both Windows 2000 and SuSe 8.2.
(Hang defined as apparently needing to kill the process in order
to get a prompt back.)


This is similar to bug #4366 where the same command used NaN
instead of Inf.

Yes, and it applies to svd() too (and the previous fix for NaNs was
applied to eigen but not to La.eigen).

I think this should be an error (with Inf or NaN): it was in the days 
of LIN/EISPACK-based code, and I don't understand why it has been chnaged 
to returning NAs in the case of NaN inputs.

Now all of eigen, La.eigen, svd and La.svd give an error on matrix(Inf,2,2).

On Sat, 29 Nov 2003 pburns@pburns.seanet.com wrote:
> Full_Name: Patrick Burns
> Version: 1.8.1
> OS: Windows 2000
> Submission from: (NULL) (81.129.3.151)
> 
> 
> The command:
> 
> eigen(matrix(Inf, 2, 2))
> 
> causes version 1.8.1 to hang in both Windows 2000 and SuSe 8.2.
> (Hang defined as apparently needing to kill the process in order
> to get a prompt back.)
> 
> 
> This is similar to bug #4366 where the same command used NaN
> instead of Inf.
> 
> ______________________________________________
> R-devel@stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
> 
> 
-- 
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