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

Re: netCDF error returns

> Subject: Re: netCDF error returns
> Hi Glenn,
>
> I'm relieving John of his middleman duties on this.  Hope you don't mind.

In the future, just start with 'address@hidden'.
That way your question, the answer, and follow-ons get tracked properly.
You will still get the same friendly personalized service.

> The point is that I am using the netCDF-3 interface and getting the
> netCDF-2 error behavior.  I've triple checked to make sure that I'm
> using nc_open and not the old ncopen and this is true.  Is there anything
> else that could be causing this behavior?  As far as I know I'm not using
> any part of the netCDF-2 interface.
>
> Thanks,
> Brian

Well, since the message says 'Abort', you may be hitting an
assert() failure. This shouldn't happen, since the default
compilation is to define NDEBUG and turn off the assertions.

At any rate you can use dbx to find out what is going on.
% dbx yourprogram yourcorefile
(dbx) where

-glenn
--- Begin Message ---
  • Subject: Re: netCDF error returns
  • Date: Wed, 19 Aug 1998 09:56:12 -0600 (MDT)
Hi Glenn,

I'm relieving John of his middleman duties on this.  Hope you don't mind.

The point is that I am using the netCDF-3 interface and getting the
netCDF-2 error behavior.  I've triple checked to make sure that I'm
using nc_open and not the old ncopen and this is true.  Is there anything
else that could be causing this behavior?  As far as I know I'm not using
any part of the netCDF-2 interface.

Thanks,
Brian


>Hi John,
>
>Your suggestions are relevant to netCDF-2.x.  In version 3.x that stuff is
>out the window (according to appendix D in the user guide).
>
>I have done some further checking and I get exactly the same behavior on
>both winterpark (SGI-R10000) and antero (CRAY-C90) as I get on our suns, so
>I don't think the problem is a screwy CGD installation.
>
>Could you pass this on the best person for an answer if you don't have one?
>Thanks,
>Brian
>
>
>>
>> are you setting NC_FATAL= false ?
>>
>> ( use CALL NCPOPT(0)   )
>>
>> -----Original Message-----
>> From: Brian Eaton <address@hidden>
>> To: John Caron <address@hidden>
>> Date: Saturday, August 15, 1998 5:44 PM
>> Subject: netCDF error returns
>>
>>
>> >Hi John,
>> >
>> >I am trying to check whether or not a file is a netCDF file by doing an
>> >nc_open expecting it to return with an error status.  Instead I'm getting
>> >the following (under Solaris-2.6):
>> >
>> >Not a netCDF id
>> >Abort (core dumped)
>> >
>> >Do you know why I'm not getting an error return?  Determining that the file
>> >is not a valid netCDF file is surely part of the netCDF library rather than
>> >something happening in the i/o system under it.



> >John:
> >
> >I believe your suggestion to this individual was correct.
> >The netcdf-2 interface is still available in netcdf-3 and still
> >has the old default behaviour on error, which is to print
> >the error message to stderr and abort. This is what the
> >individual is seeing. So, he can do as you say with the
> >netcdf-2 interface or use the netcdf-3 interface.
> >
> >-glenn
> >
> 
>

--- End Message ---