[petsc-users] TSSetExactFinalTime Behavior Without TSAdapt

[Barry Smith]

  Paul,

   I don't think this was an expected change. Can you send code or walk us through the steps that led you to this?

  Barry


> On Jun 3, 2016, at 10:23 AM, Paul Urbanczyk <gomer at stanford.edu> wrote:
> 
> Hello,
> 
> I recently upgraded to the latest version of PETSc (3.7.1) from the git repository.
> 
> Running my code gave me an error message about TSSolve being in the wrong state and needing to call TSSetExactFinalTime.
> 
> I added this function call with TS_EXACTFINALTIME_STEPOVER, but when I re-ran my code, I got a seg fault and error messages pointing to TSAdaptChoose.
> 
> After fiddling around for a while, I found that I now needed to explicitly set the time step adaptation through TSAdaptSetType.
> 
> Can anyone confirm that this is so? If so, it might be helpful to note that in documentation or error messages somewhere.
> 
> Thanks,
> 
> Paul
> 
>