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

Re: 20031021: netcdf 3.5.1-beta13 run time error

>To: <address@hidden>
>From: "WangSG" <address@hidden>
>Subject: netcdf 3.5.1-beta13 run time error
>Organization: ?
>Keywords: 200310211410.h9LEArOb010638

Hi sg,

> i installed the above version of netcdf on a 32bit redhat linux. i used =
> ifc to compile the netcdf source program.=20
> 
> i encounter "SEGV" when calling routine "nf_put_var_real_". the =
> following is the error message i get from intel debugger.
> 
> >
> >0  0x80967b8 in swapn4b(...) in test.linux.x
> #1  0x809927d in ncx_putn_float_float(...) in test.linux.x
> #2  0x809b678 in putNCvx_float_float(...) in test.linux.x
> #3  0x809c0dd in putNCv_float(...) in test.linux.x
> #4  0x80a2263 in nc_put_var_float(...) in test.linux.x
> #5  0x8091042 in nf_put_var_real_(...) in test.linux.x
> #6  0x806bd41 in write_netcdf(netcdf_file=3D(...), option=3D1) =
> "src/putcdf.n":584
> #7  0x805d7fa in main() "src/main.f":554
> #8  0x42017589 in __libc_start_main(...) in /lib/i686/libc.so.6
> #9  0x80498c1 in _start(...) in test.linux.x
> #10 0x805a168 in main() in test.linux.x
> -- end
> 
> basically the error occurs when my program tries to write out a real =
> value with nf_put_var_real.=20
> 
> i ran the program on both sun and sgi machines, it worked fine. is it =
> because of floating point exception?=20
> 
> pls do tell me if more information is needed.

Yes, we will need more information in order to see if we can duplicate
the problem here, which is a first step to determining the cause of the
problem.

After building and installing the netCDF library from source, did you
run "make test" and did all the tests succeed?  

If not, please send us the output listed at the end of the
installation instructions:

  http://www.unidata.ucar.edu/packages/netcdf/INSTALL.html#ReportingProblems

If the "make test" works, it includes tests of many calls to
nf_put_var_real that didn't result in a SEGV, so we would need a test
case that demonstrates the error.  If you can isolate the failing call
to the simplest code that still gets the error and send it to us, you
would be doing a great service and we could fix the bug.  Often in the
process of trying to simplify what looks like a bug, a different cause
is discovered, which is another good reason for trying to isolate the
error to a simple test case.

Thanks in advance for any efforts you put into discovering the source
of the problem.

--Russ

_____________________________________________________________________

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