Re: [Pan-users] Re: GigaNews warns "Usenet Growth Reveals Need for 64-bit Based Article Numbering"

[Beso]

2008/8/25 CSV4ME2 <address@hidden>

On Monday 25 August 2008, Duncan wrote:
> Beso <address@hidden> posted
> address@hidden, excerpted
>
> below, on  Mon, 25 Aug 2008 16:06:13 +0200:
> > this might mean quite a big trouble with klibido too...
>
> Indeed... and probably thunderbird and claws and knode and... and...
>
> BTW, "long" is 32-bit on x86_64 too, right?  Or is it 64-bit?  I know the
> addresses are 64-bit, but if I'm not mistaken, one of the porting issues
> was that a lot of software expected memory addresses to be unsigned long
> and on 64-bit, they're not, but rather unsigned long long, or /something/
> like that.  Did I get it right?  And plain int, is that 16-bit, or 32?

run this litle c file thru your compiler, gcc -o size size.c
and you will know!

C

============


#include <stdio.h>

#define P(x) printf( "%20s = %2d bits\n", #x, sizeof(x)*8 );

int main( int argc, char **argv )
{
       P( char );
       P( unsigned char );
       P( int  );
       P( unsigned int );
       P( short        );
       P( unsigned short );
       P( long );
       P( unsigned long );
       P( long long    );
       P( unsigned long long );
       P( float        );
       P( double       );
       P( long double  );
       return 0;
}

maybe it's time to start porting into knode the klibido features and to fix out the memory address. if someone starts to think on how to fix this issue in pan2 i'd be glad to have a look at the code so that to see if there's any way to port it to klibido also.

--
dott. ing. beso