Re: Allocatable arrays in derived types

relaxmike <michael.baudin@[EMAIL PROTECTED]
> wrote:

> I don't remember exactly where I read it on this forum,
> but I remember that the difference between allocatable and pointer
> arrays is that, when an allocatable array goes out of scope,
> it is automatically deallocated.

That is *A* difference. It is certainly not *THE* difference.

> This is the main reason for the overhead of a allocatable
> array

No. That has pretty much nothing to do with it. Indeed, the size
overhead of an alocatable array is typically the same as that for a
pointer array.

> there is also an additional bookkeeping
> for the derived type itself ?

Not typically. There can be padding in some cases.

> And what if one computes directly the size of the allocatable
> array ? I expect it to be 1, but is that true ?

No. One would actiually have a bit of trouble computing such a thing (I
assume you mean the size of the descriptor or some such thing). If you
did, it would more likely be the same as that of the component. A single
integer is "clearly" insufficient to store the necessary data, which at
a minimum for a rank 1 array is the lower bound, size, and memory
location.

-- 
Richard Maine                    | Good judgement comes from experience;
email: last name at domain . net | experience comes from bad judgement.
domain: summertriangle           |  -- Mark Twain