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

Pierangelo Masarati masarati at aero.polimi.it
Thu Aug 17 11:26:46 CEST 2006


[cross-posting for clarification] 

Pierangelo Masarati writes: 

> Richard B. Kreckel writes:  
> 
> [snip]  
> 
> Also, CLN's documentation states that |i| < 2^29, which means that the 
> test in GiNaC for integers could be broken as well, as it does  
> 
>       i >= -(1L << (cl_value_len-1))  
> 
> which allows |i| == 2^29 when i is -2^29 (assuming cl_value_len == 30). 

Actually, CLN documentation in Section 3.1 states that 

<snip>
Small integers (typically in the range -2^29...2^29-1,
</snip> 

while in Section 3.4 it states that 

<snip>
The conversion from `int' works only if the argument is < 2^29 and > -2^29.
</snip> 

This is not necessarily contradictory, but it sounds a bit odd.  Why would 
 -2^29 be legal if no conversion occurs for that value?  I'll post this 
remark to the cln-list as well. 

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 



More information about the GiNaC-list mailing list