Bug in closed-source, proprietary software that I do not have access to

Steve Kargl sgk@troutmask.apl.washington.edu
Sat May 25 17:01:00 GMT 2019

On Sat, May 25, 2019 at 02:52:53PM +0200, Thomas Koenig wrote:
> consider this: There is a bug, confirmed by several people. This occurs
> in closed-source, proprietary software, and appears to be due to one
> of my commits.
> Despite considerable help from somebody who has access to the source,
> and putting in quite a few (volunteer) hours myself, there is no
> test case.
> So, what to do? Close the PR as INVALID? This would be our standard
> policy, correct?
> FYI, the proprietary, closed-source software is SPEC, the corresponding
> PR is 90539, and the friendly helper is Martin Liska.

Put the PR in WAITING as Richard suggested with a note that
SPEC is proprietary and without a testcase, it is not possible
in investigate further. 

For WRF, I suppose you or Martin could be a good citizen and
contact the project to report a bug.


For netcdf, I suppose you of Martin could report a bug in
their interface at


Martin's comment #23 shows the interface as

And Fortran interface is defined in netcdf/include/netcdf.inc:

      integer         nf_put_vara_double
!                         (integer             ncid,
!                          integer             varid,
!                          integer             start(1),
!                          integer             count(1),
!                          doubleprecision     dvals(1))
      external        nf_put_vara_double

start(1), count(1), and dvals(1) are each an element of an array.
i.e, scalar things.  If one looks at the NetCDF sources, the code
in question (netcdf-fortran-4.4.5/fortran/nf_varaio.F90) is 

 Function nf_put_vara_double(ncid, varid, start, counts, dvals) &

! Write out real(RK8) variable to dataset for given start and count vectors

 USE netcdf_nc_interfaces

 Implicit NONE

 Integer,   Intent(IN) :: ncid, varid
 Integer,   Intent(IN) :: start(*), counts(*)
 Real(RK8), Intent(IN) :: dvals(*)

So, the comment #23 and the implementation have a mismatch. I'll note
that NetCDF has a module fortran/module_netcdf_nf_interfaces.F90, but
munging a direct call from C into the Fortran code could mess up the
interface.  Now, returning to Martin's comment #20 where I wrap the code

  A0 = nc_put_vara_double (*fncid, *fvarid + -1,
       (const size_t *) f2c_coords (*fncid, *fvarid + -1, (const int *) A3,
       (size_t *) &B3),
       (const size_t *) f2c_counts (*fncid, *fvarid + -1, (const int *) A4,
       (size_t *) &B4),

Depending on the declaration of A5, that naked A5 which is dvals(*),
likely  to work.


More information about the Gcc mailing list