slightly OT: nullifying an array of pointers

[Daniel Franke]

> What on earth are you talking about?

TYPE :: int_holder
  INTEGER :: i
END TYPE
TYPE(int_holder), DIMENSION(n) :: array

array%i = intnull()       ! valid, intnull() a function that returns 0
array   = int_holder(0)   ! valid

!----------

TYPE :: intptr_holder
  INTEGER, POINTER :: i
END TYPE
TYPE(intptr_holder), DIMENSION(n) :: array

array%i = null()                   ! invalid <--
array   = intptr_holder(null())    ! valid


The pointers do not do as the ints do. For someone fluent in fortran this is 
obvious. For someone who coded C(++) for years this deviating behaviour  
(although standard conforming) is ... strange. I spent two, maybe three hours 
this morning trying to figure out what the hell the compiler(s) kept 
complaining about. As quoted by you, this behaviour is described in the 
standard. That's fine for those who have the standards readily available - I 
do not. 

Let's close this discussion, it is off topic and I will avoid this particular 
trap in the future. Thanks nevertheless.

Regards
	Daniel