(j3.2006) Integration of co-arrays with the intrinsic shift functions

[Aleksandar Donev]

On Tuesday 17 July 2007 10:00, Craig Rasmussen wrote:
> assume this means that one can never overload procedures on co-array ?
> and array arguments.
If by never you mean not in this revision, yes. One cannot even overload based 
on an argument being a pointer array versus a normal array (since F90). This 
is part of the design to make pointer arrays and co-arrays and explicit shape 
arrays and all arrays in general as similar as possible, so one can switch 
from one to the other without changing all the calls.

Kurt was proposing a long time ago a "preference-based" generic resolution 
that would allow one to overload based on some specialization attributes. I 
think everyone agreed it sounded better than what we have now, but complex to 
handle and then Kurt left the committee so the momentum died...

> What is the problem with synchronization occurring in CO_ functions?
Dick gave some nice examples.

On Tuesday 17 July 2007 07:21, dick.hendrickson at att.net wrote:
> Better to make them into subroutines with syncs.
You mean keep them as they are :-)

> Since all of the collective routines are "really" reduction functions,
> we'll have to do it for all of them.
Yes, of course.

Aleks