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

19990310: gempak: dchrly dumps core

>From: David Wojtowicz <address@hidden>
>Organization: .
>Keywords: 199903101839.LAA17712

>
>Hi,
>
> I've  been having problems with dchrly dumping core.   I've found
>a particular METAR report that causes dchrly to dump core when
>running under Linux.    I'm using the precompiled 5.4PL8 Linux
>binary distribution from the unidata ftp server under vanilla RedHat Linux
>5.2 on a PII400 with plenty of free memory and disk.
>
>I've narrowed down the bulletins to a particular one and then one by
>one removed the reports in that until this particular one was left.
>
>The exact command I used to invoke dchrly can be found in:
>  http://uiatma.atmos.uiuc.edu/~davidw/y4.csh
>
>The segment of the DDPLUS feed that the above script uses is here:
> http://uiatma.atmos.uiuc.edu/~davidw/y4.dat
>
>The HP-UX version does not dump core on this same case, but I've had
>similar core dumping problems with it at other times.
>
>
>
>
>--------------------------------------------------------
> David Wojtowicz, Systems Manager
> Department of Atmospheric Sciences Computer Services
> University of Illinois at Urbana-Champaign
> email: address@hidden  phone: (217)333-8390
>--------------------------------------------------------
>

David,

The problem with the report is that SNOINCR is supposed to be followed
by a group x/x which is inches/hour and amount on ground. KUIN did
not include the group, so the decoder is jumping off into oblivion.
I put a recompiled dchrly in /gbuddy/nawips-5.4/binary/linux
which should solve your problems on linux.

The dchrly decoder uses NWS routines to parse the meter, and they
don't check for NULL pointers in their code- presumably they always
have perfect reports and never have to deal with what comes across
FOS or NOAAport. As I find these glitches I accumulate them and will
update the distribution.


Steve Chiswell