[petsc-dev] PetscOptionsBegin etc all use global variables

[Barry Smith]

On Nov 3, 2013, at 5:53 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> Barry Smith <bsmith at mcs.anl.gov> writes:
> 
>>    Is there anything good we can tell our current user Ed who wants to do something now?
> 
> Can he serialize the configuration step across his threads?  Does ho
> need to dynamically reconfigure at any old time?

   The problem is the damn PetscOptionsBegin … all over the stinking place instead of in XXSetFromOptions alone!  Just run a linear solver with a variety of solvers and you’ll see some of them are calling PetscOptionsBegin…. stuff.  For example (I)LU factorization is calling it to decide if the user wants i-nodes on the damn factored matrix! Or inside PCSetUpOnBlocks() that is called inside KSPSolve …..