Differentr Values for intval(float)

[Bruce]

Consider the following code:

$BB = -2181087916;
$AA = (int)$BB;
$AA = intval($BB);

On some systems, $AA will be int(-2147483648), which is actually
consistent with the documentation.

On most systems, however, $AA will be int(2113879380) , which is the
same value truncated at 32 bits.

I actually need the latter behavior, as it needs to duplicate Delphi
code, which, like C, behaves the same way.

This is encryption code that does a bunch of calculations on signed
32-bit numbers and ignores any overflows.

Any ideas how I can do this?

--Bruce

[Jerry Stuckle]

Re: Differentr Values for intval(float)

Bruce wrote:[color=blue]
> Consider the following code:
>
> $BB = -2181087916;
> $AA = (int)$BB;
> $AA = intval($BB);
>
> On some systems, $AA will be int(-2147483648), which is actually
> consistent with the documentation.
>
> On most systems, however, $AA will be int(2113879380) , which is the
> same value truncated at 32 bits.
>
> I actually need the latter behavior, as it needs to duplicate Delphi
> code, which, like C, behaves the same way.
>
> This is encryption code that does a bunch of calculations on signed
> 32-bit numbers and ignores any overflows.
>
> Any ideas how I can do this?
>
> --Bruce[/color]

It's never a good idea to depend on hardware dependent features like the
handling of overflow. They have a tendency to break when you move to a
different hardware platform.

I guess you could do some testing and bit manipulation on it, but that's
adding more complexity to the code (and more chances for it to break).

--
=============== ===
Remove the "x" from my email address
Jerry Stuckle
JDS Computer Training Corp.
jstucklex@attgl obal.net
=============== ===

[Jasen Betts]

Re: Differentr Values for intval(float)

On 2006-02-18, Bruce <bvanderweb@gma il.com> wrote:[color=blue]
> Consider the following code:
>
> $BB = -2181087916;
> $AA = (int)$BB;
> $AA = intval($BB);
>
> On some systems, $AA will be int(-2147483648), which is actually
> consistent with the documentation.
>
> On most systems, however, $AA will be int(2113879380) , which is the
> same value truncated at 32 bits.
>
> I actually need the latter behavior, as it needs to duplicate Delphi
> code, which, like C, behaves the same way.
>
> This is encryption code that does a bunch of calculations on signed
> 32-bit numbers and ignores any overflows.
>
> Any ideas how I can do this?[/color]

after that do this.

$AA= (( $AA + 2147483648 ) & 4294967295) - 2147483648 ;

it should give the result you want on 64-bit machines
and be harmless on 32-bit machines.

Bye.
Jasen