Re: Memory exaustion or size too large for index

[Jaroslav Hajek]

On Thu, Feb 25, 2010 at 2:07 PM, Carlo de Falco <address@hidden> wrote:
>
> On 25 Feb 2010, at 14:06, Jaroslav Hajek wrote:
>
>> On Thu, Feb 25, 2010 at 2:02 PM, Carlo de Falco <address@hidden>
>> wrote:
>>>
>>> On 25 Feb 2010, at 13:58, Jaroslav Hajek wrote:
>>>
>>>> #include <iostream>
>>>> #include <cstddef>
>>>> #define DUMPSIZE(TYPE) std::cout << "sizeof ("#TYPE") = " << sizeof
>>>> (TYPE) << '\n'
>>>> int main()
>>>> {
>>>>  DUMPSIZE(int);
>>>>  DUMPSIZE(unsigned int);
>>>>  DUMPSIZE(long);
>>>>  DUMPSIZE(unsigned long);
>>>>  DUMPSIZE(long long);
>>>>  DUMPSIZE(unsigned long long);
>>>>  DUMPSIZE(size_t);
>>>>  DUMPSIZE(ptrdiff_t);
>>>>  DUMPSIZE(void *);
>>>> }
>>>>
>>>> It seems to me that in your case size_t is of size 4 and void * is of
>>>> size 8, which seems impossible.
>>>
>>> $ /opt/octave/3.3/bin/mkoctfile --link-stand-alone prova.cc
>>> $ ./a.out
>>> sizeof (int) = 4
>>> sizeof (unsigned int) = 4
>>> sizeof (long) = 4
>>> sizeof (unsigned long) = 4
>>> sizeof (long long) = 8
>>> sizeof (unsigned long long) = 8
>>> sizeof (size_t) = 4
>>> sizeof (ptrdiff_t) = 4
>>> sizeof (void *) = 4
>>>
>>
>> Eh? So it's a 32-bit system?
>
> no, see the mail I just sent.
> It seems apple gcc by default compiles for 32 bit systems even when run on
> 64 bit machine...
> c.
>

I see the pitfalls of portability are really unlimited. In any case,
if you build Octave with 64-bit pointers, it should work for you. If
your pointers are 32-bit, you have no chance whatsoever to get big
arrays working. I don't understand why you don't get a std::bad_alloc
in the allocation, though.


-- 
RNDr. Jaroslav Hajek, PhD
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz