20020611: There is an input or output error

Aime,

>Date: Tue, 11 Jun 2002 09:35:15 -0600 (MDT)
>From: "Aime' Fournier" <address@hidden>
>Organization: NCAR/CGD
>To: Steve Emmerson <address@hidden>
>Subject: Re: 20020607: There is an input or output error
>Keywords: 200206072228.g57MSaJ26805

The above message contained the following:

> As I indicated, I am invoking NF_STRERROR, which
> returns the string value 'There is an input or
> output error'.  I don't know the version because
> the blackforest links are opaque, but it's an OLD
> version.  Anyway the problem mysteriously went
> away!

I like those kinds of problems!

> I have upgraded my error handling to
> 
>       SUBROUTINE HANDLE_ERR(STATUS)
> #include "netcdf.inc"
>       INTEGER STATUS
>       IF (STATUS .NE. NF_NOERR) THEN
>          WRITE(*,'(A,/,''STATUS='',I,/,A)')NF_STRERROR(STATUS),STATUS,
>      &      NF_INQ_LIBVERS()
>          STOP 'Stopped by HANDLE_ERR.F'
>       ENDIF
>       RETURN
>       END
> 
> so I'll have the version etc. if problems recur.

Excellent idea!

You might add a test for STATUS being -31, which indicates a system
error (i.e. not a netCDF error) and then print out the value of "errno"
-- or whatever is the Fortran equivalent.

Regards,
Steve Emmerson   <http://www.unidata.ucar.edu>