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

[netCDF #GMP-676686]: Potential bug in ncdump v4.3.3.1



> Full Name: Jim Enright
> Package Version: 4.3.3.1
> Operating System: CentOS 6.5

Hi Jim,

Thanks for reporting this bug. I've reproduced it here and will have it fixed 
soon.
It's great that you identified a version in which it worked and a later version 
that
has the bug, which makes it easy to find the exact git commit that introduced
the bug, using "git bisect".

--Russ

> We have come across a potential bug in the ncdump utility from version 
> 4.3.3.1 of NetCDF when inspecting the output of an application we are working 
> with.
> 
> When one views this file using the full annotations for Fortran indices (the 
> "-f f" option) there seems to be an off-by-1 error on the indexing. This 
> causes some issue with the labelling of the last element in each dimension.
> I have reproduced this with the file created by the "simple_xy_wr.f90" 
> example from the NetCDF-Fortran source tarball as shown below.
> 
> $ ncdump -f f simple_xy.nc
> data =
> 0,   // data(1,1)
> 1,   // data(2,1)
> 2,   // data(3,1)
> 3,   // data(4,1)
> 4,   // data(5,1)
> 5,   // data(6,1)
> 6,   // data(7,1)
> 7,   // data(8,1)
> 8,   // data(9,1)
> 9,   // data(10,1)
> 10,   // data(11,1)
> 11,  // data(1,1) <== INCORRECT - should be data(12,1)
> 12,   // data(1,2)
> ...
> ...
> 21,   // data(10,2)
> 22,   // data(11,2)
> 23,  // data(1,2) <== INCORRECT - should be data(12,2)
> 24,   // data(1,3)
> 
> For an older version ncdump, from version NetCDF version 4.1.1, the correct 
> behaviour is seen:
> $ ncdump -f f simple_xy.nc
> data =
> 0,   // data(1,1)
> 1,   // data(2,1)
> 2,   // data(3,1)
> 3,   // data(4,1)
> 4,   // data(5,1)
> 5,   // data(6,1)
> 6,   // data(7,1)
> 7,   // data(8,1)
> 8,   // data(9,1)
> 9,   // data(10,1)
> 10,   // data(11,1)
> 11,  // data(12,1) <== CORRECT
> 12,   // data(1,2)
> ...
> ...
> 21,   // data(10,2)
> 22,   // data(11,2)
> 23,  // data(12,2) <== CORRECT
> 24,   // data(1,3)
> 
> The same behaviour is seen when using C indices for a NetCDF file created by 
> C code.
> 
> Can you provide any advice on this please?
> 
> Thanks for your help,
> Jim
> 
> 
> 

Russ Rew                                         UCAR Unidata Program
address@hidden                      http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: GMP-676686
Department: Support netCDF
Priority: Normal
Status: Closed


NOTE: All email exchanges with Unidata User Support are recorded in the Unidata inquiry tracking system and then made publicly available through the web. If you do not want to have your interactions made available in this way, you must let us know in each email you send to us.