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

20030604: NIDS Storm Relative Velocity (N0S) caused gpmap. GARP and NMAP2(5 .6j) to core dump



>From: Robert Mullenax <address@hidden>
>Organization: UCAR/Unidata
>Keywords: 200306042138.h54LcYLd029141

>I noticed today that both a locally built Linux version of GEMPAK 5.6j and
>the Unidata
>Solaris SPARC binary (GEMPAK5.6j) crash when trying to display N0S with GARP
>or NMAP2 or from gpmap.  It crashes with a coredump.  5.6h works fine.
>
>Thought you'd want to know about this one.
>
>Thanks,
>Robert Mullenax
>
>Robert Mullenax
>Weather Systems Administrator
>Universal Weather and Aviation
>


Robert,

I don't have a general problem displaying the N0S products.
It is possible that you have more than 1 image in your file
(run "od -c file | less", and look for the start of another product,
for example, if you are looking at KFWS, then you might see the pattern
"K   F   W   S"  midway through the file in a second WMO header.

The old version of the code would stop when the zlib portion ended-
but the NWS started sending some products where the end wasn't compressed.
So, the change was made to allow this. I suspect that the program is
running on past the end of the first product and overrunning its buffer.
I'll look into stopping this.

In the mean time, if you are getting more than 1 image in a file
make sure you aren't filing both FNEXRAD floaters and NNEXRAD
products to the same file. And, we'll see if certain station/products
seem to be rebroacast with different data (MD5 checksum).

Steve