Re: [petsc-dev] Fortran binding for VecGetValuesSection

[Blaise Bourdin]

Hi Jed, I pushed my attempts in https://gitlab.com/petsc/petsc/-/merge_requests/5385  I thought that they would be similar to VecGetArrayF90, but evidently they are not... Strangely, valgrind reports a problem in VecSetValues, but not in VecGetValues. There are a few occurrences of a XXXGetYYY without a matching XXXRestoreYYY (but I can’t think of any at the moment), typically when the Get does not allocate any memory but just returns a pointer. I would indeed prefer if all Get were matched with a Restore, even if the Restore does nothing.  Regards, Blaise On Jun 26, 2022, at 1:20 AM, Jed Brown <j...@jedbrown.org> wrote: Sorry about our late reply. Isn't VecGetValues similar? Note that VecGetValuesSection currently leaks a mutable pointer and really shouldn't exist like this without a matching VecRestoreValuesSection.  PetscCall(VecGetArray(v, &baseArray));  *values = &baseArray[s->atlasOff[p]];  PetscCall(VecRestoreArray(v, &baseArray)); Can you make a draft MR with what you've tried thus far? Blaise Bourdin <bour...@mcmaster.ca> writes: Hi, I’ve been trying to add a fortran90 binding for VecGetValuesSection, without success, and could not find any function with a similar prototype to learn from. Can anybody help? Regards, Blaise -- Professor, Department of Mathematics & Statistics Hamilton Hall room 409A, McMaster University 1280 Main Street West, Hamilton, Ontario L8S 4K1, Canada https://www.math.mcmaster.ca/bourdin | +1 (905) 525 9140 ext. 27243 --  Professor, Department of Mathematics & Statistics Hamilton Hall room 409A, McMaster University 1280 Main Street West, Hamilton, Ontario L8S 4K1, Canada  https://www.math.mcmaster.ca/bourdin | +1 (905) 525 9140 ext. 27243