Irix 64 nc_test error

[Jianwei Li]

Jon,

The "nc_test" is set to run as follows:

# nc_test -c
# nc_test [-v]
(or use "mpirun -np 1" to run the tests)

Except the "ERROR of ncmpi_open of nonexistent file", you should
get no other FAILURE of tests.

Please let us know if there's any problem more than that.

Thanks!
Jianwei

On Tue, 20 Apr 2004, Rhoades.Jon wrote:

> I made the change and the make works building the executables but when I do a make test in the nc_test dir I get the following.
>
> andrew 5# make test
> ./nc_test
> *** Testing ncmpi_strerror ... ok
> *** Testing ncmpi_open ...  0: MPI_File_open error = No such file or directory
>
>         FAILURE at line 94 of test_read.c: ncmpi_open of nonexistent file should have returned system error
>         ### 1 FAILURES TESTING ncmpi_open! ###
> *** Testing ncmpi_close ... ok
> *** Testing ncmpi_inq ... ok
> *** Testing ncmpi_inq_dimid ... ok
> *** Testing ncmpi_inq_dim ... ok
> *** Testing ncmpi_inq_dimlen ... ok
> *** Testing ncmpi_inq_dimname ... ok
> *** Testing ncmpi_inq_varid ... ok
> *** Testing ncmpi_inq_var ... ok
> *** Testing ncmpi_inq_natts ... ok
> *** Testing ncmpi_inq_ndims ... ok
> *** Testing ncmpi_inq_nvars ... ok
> *** Testing ncmpi_inq_unlimdim ... ok
> *** Testing ncmpi_inq_vardimid ... ok
> *** Testing ncmpi_inq_varname ... ok
> *** Testing ncmpi_inq_varnatts ... ok
> *** Testing ncmpi_inq_varndims ... ok
> *** Testing ncmpi_inq_vartype ... ok
> *** Testing ncmpi_get_var_text ...  179 good comparisons. ok
> *** Testing ncmpi_get_var_short ...  682 good comparisons. ok
> *** Testing ncmpi_get_var_int ...  1189 good comparisons. ok
> *** Testing ncmpi_get_var_long ...
>         FAILURE at line 829 of test_get.c: value read not that expected
>         FAILURE at line 829 of test_get.c: value read not that expected
>         FAILURE at line 829 of test_get.c: value read not that expected
>         FAILURE at line 829 of test_get.c: value read not that expected
>         FAILURE at line 829 of test_get.c: value read not that expected
>         FAILURE at line 829 of test_get.c: value read not that expected
>         FAILURE at line 829 of test_get.c: value read not that expected
>         FAILURE at line 829 of test_get.c: value read not that expected
>  903 good comparisons.
>
>
> It continues with the value read not that expected errors.
>
> Thanks,
> -Jon Rhoades
>
> -----Original Message-----
> From: Jianwei Li [mailto:jianwei at cheetah.cpdc.ece.northwestern.edu]
> Sent: Monday, April 19, 2004 7:06 PM
> To: parallel-netcdf at mcs.anl.gov; Rhoades.Jon
> Subject: Re: Irix 64 nc_test error
>
>
> Hi Jon,
>
> Thank you for the interest in using PnetCDF and for pointing out this problem!
>
> The problem occurs because we put different arguement datatype "size_t" for the
> strided access interface than the original netCDF's "ptrdiff_t" that is also
> used in the nc_test. Could you try to replace the "ptrdiff_t" to "size_t" in
> the variable declaration to solve your problem temperarily for now? We'll look
> into the code to further address this problem.
>
> Thanks!
>
> Jianwei
>
> >
> >I am trying to get Parallel Netcdf 0.9.3 installed and when I try to compile
> the test scripts I get the following errors.
> >
> >
> >cc-1164 cc: WARNING File = test_get.c, Line = 2071
> >  Argument of type "ptrdiff_t *" is incompatible with parameter of type
> >          "const size_t *".
> >
> >          err = ncmpi_get_vars_text_all(BAD_ID, i, start, edge, stride, value);
> >
> >I also have compiled MPI V2 and the test scripts for it ran fine. I am running
> on a an SGI O2000 Irix 6.5.  Any help would be appreciated.
> >
> >Thanks,
> >Jon Rhoades, CCNA MCP
> >Deputy Division Manager
> >Information Systems & Technology
> >ENSCO, Inc.
> >
>