Compatibility of array indexing for gcc 3.4.6 and gcc 4.4.7?

[Richard Biener]

On Fri, Nov 1, 2013 at 5:58 PM, Stefan Schmitt <sschmitt@mail.desy.de> wrote:
>
> Thank you all very much for your valuable comments.
>
> As I said before, the problem is with getting backwards-compatible results
> from a very large piece of code (millions of lines of code), when going from
> the 32-bit to the 64-bit system.

Try -m32?

;)

> I mentioned one of the problems in this forum and was asking for help.
> Apparently there seems to be no easy solution. If course I can fix the
> specific problem by adding the line
>      index=iand(index,z'3fffffff')
> but there may be hundreds of similar problems hidden elsewhere. This is why
> I have been asking for a compatibility flag. Please apologize for having
> wasted your time, I am not going to come back again.

No, there isn't any option that would do what you desire.

Richard.

> Stefan
>
>
>> On 01/11/13 14:07, Stefan Schmitt wrote:
>>>
>>> Still I have the question whether there is a compiler flag to ensure
>>> compatibility with 32 bit addressing.
>>
>>
>> At this point I'm reminded of the XCKD comic "Workflow"; in the comic a
>> "space bar causes processor overheating" bug being fixed caused someone
>> to complain: someone who was using "rapid temperature increase" to be
>> interpreted as a keyboard input.
>>
>> I'm inclined to agree with other repliers here: the "solution" is to
>> fix/rewrite the offending code.
>>
>> ~ Adam