[GiNaC-list] Possible size issue in test in numeric.cpp

[Pierangelo Masarati]

Jens Vollinga writes: 

> thanks for your patch! We had a discussion about it on the other GiNaC 
> mailing list (ginac-devel) and we are going to apply it. It will be part 
> of the forthcoming 1.3.5 release then. 
> 
> @Richy:
> You wondered why the other ctor for unsigned int doesn't raise such a 
> warning. After looking at the code I now wonder whether the code there is 
> optimal. The argument is compared against 2^(cl_value_len-1) and not 
> 2^(cl_value_len)-1 as one might expect. Maybe this should be changed? 
> Could you comment on this soon, because I'd like to roll the release this 
> evening?

I agree it should be 

       i < ((1UL << cl_value_len)-1) 

In this case, a fix similar to the "int" case on 64 bit architectires would 
apply, since a "unsigned" will always be less than 2^32.  A patch is 
available at 

<http://mbdyn.aero.polimi.it/~masarati/Download/mbdyn/ginac-numeric-uint.pat 
ch> 

Cheers, p. 

Dr. Pierangelo Masarati               | voice: +39 02 2399 8309
Dip. Ing. Aerospaziale                | fax:   +39 02 2399 8334
Politecnico di Milano                 | mailto:pierangelo.masarati at polimi.it
via La Masa 34, 20156 Milano, Italy   | http://www.aero.polimi.it/~masarati 

Please avoid sending me Word or PowerPoint attachments.
See http://www.gnu.org/philosophy/no-word-attachments.html