| Re: compile a routine wich inlude a commun [message #47139 is a reply to message #47055] |
Mon, 23 January 2006 14:24   |
![Paul Van Delst[1] is currently offline](theme/default/images/xoffline.png.pagespeed.ic.XRkd1fkXye.png "Paul Van Delst[1] is currently offline")
Paul Van Delst[1]
Messages: 1157
Registered: April 2002
|
Senior Member |
|
|
David Fanning wrote:
> To avoid pointers!? Are you a Luddite? Pointers
> are the coolest thing *in* IDL. Global, sticky, variables
> that act *exactly* like any other IDL variables. Fantastic!
> I think almost everyone would agree it is one thing RSI got
> *exactly* right.
Well, they're called pointers but they're not, really. You can't actually "point" to
anything - just make copies. But, not being a pointer expert, let me ask the question:
How *do* you use a pointer in IDL to, uh, well, "point" to an already created variable? Or
just parts of an already created array?
E.g.
IDL> x=indgen(4,4)
IDL> print, x
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14 15
IDL> p=ptr_new(x[1:2,1:2])
IDL> print, *p
5 6
9 10
IDL> *p=*p+100
IDL> print, *p
105 106
109 110
IDL> print, x
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14 15
If pointers in IDL worked right (and by "right", I mean how *I* intuitively expect them to
work <take grain of salt here>) I would expect the "print, x" command to output the following:
IDL> print, x
0 1 2 3
4 105 106 7
8 109 110 11
12 13 14 15
In my pointer-naivete, it seems to me that p should point to the memory that x occupies.
But that's not what happens. I expect it to work like the following Fortran95 program:
program testptr
integer,parameter::n=4
integer,target::x(0:n-1,0:n-1)
integer,pointer::p(:,:)
x=reshape((/(i-1,i=1,n*n)/),(/n,n/))
write(*,'("Print x:")')
write(*,'(4i5)')x
p=>x(1:2,1:2)
write(*,'("Print p:")')
write(*,'(2i5)')p
p=p+100
write(*,'("Print p added to:")')
write(*,'(2i5)')p
write(*,'("Print x:")')
write(*,'(4i5)')x
end program testptr
lnx:scratch : lf95 testptr.f90
Encountered 0 errors, 0 warnings in file testptr.f90.
lnx:scratch : a.out
Print x:
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14 15
Print p:
5 6
9 10
Print p added to:
105 106
109 110
Print x:
0 1 2 3
4 105 106 7
8 109 110 11
12 13 14 15
Maybe PTR_NEW() should be renamed to something else? Like, um, HEAPVAR_NEW() ?
I'm sure all of this has something to do with the pass-by-reference/pass-by-value nature
of certain things in IDL.
paulv
--
Paul van Delst
CIMSS @ NOAA/NCEP/EMC
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
