Re: how do I make ncl abort on fatal errors

From: Mary Haley <haley_at_nyahnyahspammersnyahnyah>
Date: Tue, 21 Apr 2009 10:32:44 -0600 (MDT)


Technically, I think NCL should abort on "fatal" type of errors,
and not "warning" type of errors. I'm afraid there are places
where we may not have been consistent with this.

In this case, it seems that a non-existent file should cause NCL to
abort. In other words, the following script should not get to
the "print" statement:

   a = addfile("baaaadtest.grb","r")
   print("I don't think this line should be printed")

However, I've been wrong about my assumptions before!

Dave B, can you weigh in here?


On Tue, 21 Apr 2009, Keith Lindsay wrote:

> Hi,
> Bob Tomas has informed me that adding begin and end statements to the
> beginning and end of the ncl command sequence, will cause ncl to abort
> on some fatal errors. When I add these statements to my previous script,
> and remove the quit statement, ncl is now aborting when it attempts to
> read the non-existent variable. The modified shell script demonstrating
> this is attached. Thanks Bob.
> A caveat is that even with the begin and end statements, ncl does not
> abort on all fatal errors. ncl is classifying the addfile of the
> non-existent file as fatal, but it is not aborting. I don't know what
> fatal errors ncl will abort on and which ones it will not.
> Thanks, Keith
> --
> ******************************************************************
> Keith Lindsay
> email: phone: 303-497-1722 fax: 303-497-1700
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
Received on Tue Apr 21 2009 - 10:32:44 MDT

This archive was generated by hypermail 2.2.0 : Tue Apr 21 2009 - 11:04:41 MDT