Neale,
Thanks so much for your help.

The problem I am facing is that some of the development code relies on
return values that should be used as a size_t being declared as an
unsigned int.

In other cases it is a reliance on a pointer value being 32 bits and
casted to an unsigned or signed int to do pointer arithmetic.  While
that works fine in a 32-bit environment it causes problems in a 64-bit
environment.

The immediate solution is to use the correct unit types where called
for.  I am attempting to identify those places where this occurs and
use the correct typedef for the values used.

If this is too burdensome a last resort will be to recompile all the
3rd party libraries as -m31 along with the application

Any other ideas?


On Wed, Nov 5, 2008 at 11:05 AM, Neale Ferguson <[EMAIL PROTECTED]> wrote:
> What type of compiler issues are you having? In System z both for 32 and
> 64-bit size_t is typedef'd as _SIZE_TYPE_ which is unsigned long. On 32 bit
> this is a 32-bit field, on 64-bit a 64-bit field.
>
>
> On 11/5/08 11:11 AM, "Paul Meier" <[EMAIL PROTECTED]> wrote:
>
>> They did use the switch; and it forces all of the 3rd party libraries
>> to be recompiled that way as well, so either they stay with 64 bit for
>> now and try to bull through the issues or redo everything as -m31
>> (including 3rd Party stuff).
>
> ----------------------------------------------------------------------
> For LINUX-390 subscribe / signoff / archive access instructions,
> send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
> http://www.marist.edu/htbin/wlvindex?LINUX-390
>

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to