Re: building active perl from scratch

[Anthony Nemmer]

Gregg Morrison wrote:
On Mon, 21 Jun 2004 02:37:42 -0700
"Anthony Nemmer" <[EMAIL PROTECTED]> wrote:
 

$dummy = 2048*1024*1024;
printf("%llu $dummy\n",$dummy);
$dummy *= 2;
printf("%llu $dummy\n",$dummy);
$dummy *= 2;
printf("%llu $dummy\n",$dummy);
with active perl build 804 for win32 I obtain:
2147483648 2147483648
4294967295 4294967296
4294967295 8589934592
and that version is not compiled with use64bitint set.
the only thing that doesn't work seems to be %lld in printf.
   


 Could this be casting Text Value to Floating Point ?
 

Yes, it is hard to tell when perl switches from int or long to float, 
and you won't start losing integer precision until you start getting 
into really large float values... you'd have to experiment to find out 
exactly when on which platform.  But I think if you want true 64 bit 
integer math you need perl compiled on a 64 bit machine (like a Digital 
Alpha), then you need to do a

   use integer;
and even then, it's still going to be SIGNED integer (so really you are 
getting 63 bit integers.)  Someone pointed out that perl is smart 
enought to know when to use unsigned integers as opposed to signed 
integers, but I didn't understand the post.

Tony
 Gregg Morrison
 [EMAIL PROTECTED]
_______________________________________________
ActivePerl mailing list
[EMAIL PROTECTED]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
 


--
SKYKING, SKYKING, DO NOT ANSWER.
_______________________________________________
ActivePerl mailing list
[EMAIL PROTECTED]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs