On Apr 30, 2012, at 9:26 PM, Barry Smith wrote:

On Apr 30, 2012, at 10:37 AM, Blaise Bourdin wrote:

...

Hi,

I know I suggested it before and the idea did not get much traction, but it would be _really_ nice to add Real and Integer arrays to PetscBag. The idea is _not_ to add full fields in a bag. Instead, I would like to be able to pass material properties as command line arguments in a concise way. (Think of a Hooke's law in 3d and its 21 coefficients, passing it as -a 1,1,1,1,1,... is much nicer than having to pass -a_1111 1 -a 1122 1 etc).

I have added PetscBagIntArray() and PetscBagRealArray(). Please report any problems.

Awesome!

...

Also, fortran wrappers over PetscOptionsEnum and PetscBagRegisterEnum would be nice. I understand that it is a bit tricky since it involves passing fortran arrays of strings.

Yup. We'd be willing to support these from FORTRAN but someone else has to do the dirty work of figuring out how to handle those array of strings (maybe F2003?).

I'll go through my files tomorrow. Using the fortran 2003 C interoperability module, I was able to pass an array of strings between fortran and C. The issue is that all work is done from the fortran side through an interface. I am not completely sure how this would work with PETSc's fortran bindings. I'll see if I can at least write a proof of concept example, at least. Thanks again, for the arrays in bag. This will make my life MUCH easier. Blaise -- Department of Mathematics and Center for Computation & Technology Louisiana State University, Baton Rouge, LA 70803, USA Tel. +1 (225) 578 1612, Fax +1 (225) 578 4276 http://www.math.lsu.edu/~bourdin