[RFA] libstdc++/22203, aka PowerPC vs numeric_limits<integer>::traps

Paolo Carlini pcarlini@suse.de
Thu Nov 10 01:21:00 GMT 2005


Benjamin Kosnik wrote:

>    if (!std::numeric_limits<value_type>::traps)
>      value_type r = i/j;
>  
>
Thanks.

For the integer types at least, when traps is true (everywhere, besides
powerpc), we are not testing anything. In principle, we should look for
the signal...

Paolo.



More information about the Libstdc++ mailing list