[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 19990330: c++ interface bug? NcFile::get_var()

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: Re: 19990330: c++ interface bug? NcFile::get_var()
  • Date: Tue, 30 Mar 1999 08:44:28 -0700
Ionutz,

>Date: Tue, 30 Mar 1999 14:35:29 +0900
>From: Ionutz Borcoman <address@hidden>
>Organization: .
>To: Steve Emmerson <address@hidden>
>Subject: Re: 19990329: c++ interface bug? NcFile::get_var()
>Keywords: 199903251256.FAA21940

In the above message, you wrote:

> gcc is obsolete on Debian 2.2. Even my kernel 2.2.3 and my glibc2.1 are
> compiled with egcs. The gcc was completly replaced and I have no
> intention to put it back. :-)
> 
> About the problem with NULL pointers: yes, I think here is the problem.
> We have a similar problem with our VDK library.

The problem might not lie with a null pointer.  I thought that your
netCDF file didn't contain the variable "P"; thus, I misinterpreted
where your program crashed.  If, instead, your netCDF file contains
a "P" variable but not an "xxx" variable, then your program might be
crashing in the "printf" statement of the following:

    var = nc->get_var("xxx");
    if(!var)
        printf("not existent variable\n");

I suggest that you determine where, exactly, your program is crashing.
A debugger that understands both C and C++ might help.

> It seems more not a compiler problem, but that netCDF C++ interface is
> not ANSI C++ compliant.

NcFile::get_var() is compliant.  It's equivalent to the following when
the named variable doesn't exist:

    NcVar* NcFile::get_var( NcToken name ) const
    {
        return 0;
    }

So the issue isn't conversion between pointer types, but rather 
conversion of zero to a null pointer of type NcVar* -- which is always 
allowed.

> I already have this copyright, but I'm not very good at interpreting
> these documents :-)

Who is?  :-)

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