[petsc-users] [FORGED] Re: NULL section fortran

[Justin Pogacnik]

Any word on this? Thanks again!

-Justin
________________________________________
From: Justin Pogacnik
Sent: Tuesday, July 18, 2017 8:38 AM
To: Barry Smith
Cc: petsc-users at mcs.anl.gov
Subject: Re: [FORGED] Re: [petsc-users] NULL section fortran

Thanks for the response Barry. Matt, is this something that will be pushed out soon(ish)?

Thanks guys!

Justin
________________________________________
From: Barry Smith <bsmith at mcs.anl.gov>
Sent: Thursday, July 13, 2017 7:59 AM
To: Justin Pogacnik
Cc: petsc-users at mcs.anl.gov
Subject: [FORGED] Re: [petsc-users] NULL section fortran

  Matt needs to add something like

     PetscSection, parameter :: PETSC_NULL_SECTION = tPetscSection(-1)
     PetscSectionSym, parameter :: PETSC_NULL_SECTIONSYM = tPetscSectionSym(-1)

to src/vec/f90-mod/petscis.h in the master branch.

   Barry


> On Jul 12, 2017, at 2:52 PM, Justin Pogacnik <j.pogacnik at auckland.ac.nz> wrote:
>
> Hello,
>
> I'm assembling a global finite element stiffness matrix and using the DMPlexMatSetClosure command in fortran. Long ago, in the past, I've used to set the section input parameters to PETSC_NULL_OBJECT. That now returns an error -- no implicit type -- during compilation. Looking at the defined null objects for fortran I don't see a section option. What should I use to invoke the default sections?
>
> Hope this makes sense. Thanks!
>
> Justin