[Vis5d-users] holes, but no missing data ...

Hello,

I have a very bizarre problem. I generate cdf data with the
fortran subroutine v5dwrite successfully. 
v5dstats confirms the extreme values and reports no missing
data. A test by writing in an ascii file confirms that
everything is normal. However, when I view this data with
vis5d 5.2, I see a big irregular hole around level 35 (out
of 50), where the data is nearly homogeneous. The probe 
confirms that there is some data here. A time step later, 
everything is normal.
My problem is perfectly reproductible.

I'm running vis5d-5.2 on an SGI IRIX64, after having compiled
it myself. Has anybody experienced this kind of difficulty ?
Would you have a clue for testing and finding out what is
wrong ?

My file is nearly 1Mb, so I did not want to attach it to the 
e-mail. But you can find a copy of this e-mail and the file
at the address
http://www.oc.nps.navy.mil/~pascale/data/VIS5D

Thanks a lot,
Pascale.
-- 
Pascale Lherminier		| Tel.: (int) (1) (831) 656 3260
Dept of Oceanography		| Fax.: (int) (1) (831) 656 2712
Naval Postgraduate School	| E-mail : pa...@np...
Monterey, CA93943, USA          | http://www.oc.nps.navy.mil/opbl