Re: System Variables: defining and using

[mallors(at)crazyhorse.msfc.nasa.gov (Robert S. Mallozzi)]

In article <6ihdnf$rc8$1@ratatosk.uio.no>,
	steinhh@ulrik.uio.no (Stein Vidar Hagfors Haugan) writes:
> 
>> Apparently IDL
>> requires the system variable to be defined at compile time.
> 
> Yes, that's right. Which means they should preferably be defined
> in the IDL_STARTUP file. There is one way of getting around this,
> though. In your case, write 
> 
>     dummy = execute("endian = !little_endian")
>     return,endian
> 
> instead of simply return,!little_endian. This will allow your
> routine to compile, and define the system variable before it's
> used.
> 

>From the file 51new.txt that was distributed with 5.1b3:

    "Note also that system variables can now be defined 
    after code that references them is compiled (or even 
    in the same routine) and IDL will correctly resolve the
    reference as long as the variable exists at the time the
    referencing code executes."
    
I tried the is_little_endian() function, and it works fine 
with 5.1b3.  I think 5.1 is due to be released any day now.


Regards,

-bob mallozzi   


-- 
Robert S. Mallozzi
http://cspar.uah.edu/~mallozzir/