Re: [Libmesh-users] Ran into the error "invalid time_solver DIE" when trying fem_system_ex3.C

[John P. <jwp...@gm...>]

On Sat, Oct 27, 2018 at 6:14 PM Yuxiang Wang <yw...@vi...> wrote:

> Dear libmesh community,
>
> Sorry for the spam again.
>
> I have compiled my libmesh stable release 1.3.0 with PETSc and it was able
> to run most examples perfectly (for example, the Systems of Equations
> Example 6 which is the implicit version of this example). The environment
> is 64-bit Windows 10, in Windows Subsystem for Linux, Ubuntu 18.04.
>
> However, when I tried to run fem_system_ex3.C, it built fine but gave the
> following error:
>
> yw5aj@Shawn-K430:/mnt/x/libmesh/with_vtk/examples/fem_system/ex3-test$
> make;./example-opt -ksp_type cg -pc_type jacobi
> Can't open perl script "/mnt/x/libmesh/with_vtk/contrib/bin/
> make_dependencies.pl": No such file or directory
>

FYI, I recently released a 1.3.1 update (
https://github.com/libMesh/libmesh/releases/tag/v1.3.1) which should make
the Makefile error above go away.


Could anyone please share any thoughts that you might have, maybe just
> ideas on where to start the troubleshooting? I'd really appreciate it.
>


That error starts with the default value set by this line:

  // Solve this as a time-dependent or steady system
  std::string time_solver = infile("time_solver","DIE!");

and the error actually comes from the end of the if-else statement on line
213. The reason it happens is that this example doesn't run without command
line arguments. See the script used to run this example at:
examples/fem_system/fem_system_ex3/run.sh

-- 
John