[Bug fortran/45586] [4.6 Regression] ICE non-trivial conversion at assignment

[Joost.VandeVondele at pci dot uzh.ch]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45586

--- Comment #14 from Joost VandeVondele <Joost.VandeVondele at pci dot uzh.ch> 2010-11-26 10:08:11 UTC ---
(In reply to comment #13)
> Ugh.  That might be terrible.  Can you point to some language in the standard
> supporting that (I haven't looked myself and am not too fluent in the fortran
> standard).

I'm afraid that finding this in the standard is not quite easy. I guess the
idea is that you point to something which is a subobject of variable with
pointer attribute. Maybe Tobias can find this? 

The closest I come is actually in IBM's compiler manual:

http://publib.boulder.ibm.com/infocenter/cellcomp/v9v111/index.jsp?topic=/com.ibm.xlf111.cell.doc/xlflr/f90pass.htm

target
    is a variable or expression. If it is a variable, it must have the TARGET
attribute (or be a subobject of such an object) or the POINTER attribute.

It isn't that terrible, I think. you still can't point to allocatable arrays,
if e.g. they are local variables, or just part of derived types that do not
have the pointer attribute.