Problem: Added class PetscProblem
- This is intended to encapsulate a set of discretizations and functions defining the physics
- It calculates total sizes and offsets

PetscDualSpace: Fixed index error in Lagrange creation

PetscFE: Fixed PetscFERefine()
- The refined dual space was not properly duplicated
- Added PetscDualSpaceDuplicate()

PetscFE: Fixed allocation bug

PetscFV: I hate fucking complex

Merge branch 'knepley/feature-fv-class'

* knepley/feature-fv-class: (44 commits)
PetscFV: Another complex fix
DMPlex+TS: Fixing types
PetscFV: Straighten out complex types
TS ex11: Fix casts

Merge branch 'knepley/feature-fv-class'

* knepley/feature-fv-class: (44 commits)
PetscFV: Another complex fix
DMPlex+TS: Fixing types
PetscFV: Straighten out complex types
TS ex11: Fix casts and remove unused variable
PetscFV: Fix Fortran define
PetscFV: Complex fixes
SNES+DM: Added SNESMonitorFields() - Activated by -snes_monitor_fields
PetscFV: Add Fortran types
SNES ex62: Use new DMPlexAddBoundary() interface - Added declaration of DMPlexComputeL2FieldDiff() - Fixed cast
SNES ex12: Fixed Neumann boundary conditions, and moved to new DMPlexAddBoundary interface() - Now do not automatically mark boundary faces
PetscFV: Fixed Fortran bindings
PetscFV+Plex: Refactored to put reconstruction routines in PetscFV_LeastSquares - Added PetscFVComputeGradient() - Added PetscFVLeastSquaresSetMaxFaces() - Moved last squares fitting routines to PetscFV_LeastSquares, and all the work space
PetscLimiter: Add other implementations
TS ex11: Fixed parallel mesh creation - Distribute after physics - Add output for tests
DMPlex: Fixed bug when using refinement from options - Must copy over the point SF
DMPlex: Distribution should copy over BC info
TS ex11: Added refinement tests
TS ex11: Fix boundary conditions after refinement, set types from command line
DMPlex: Added DMPlexCopyBoundary() - Copy boundary conditions after DMRefine() and DMPlexConstructGhostCells()
TS ex11: Specify boundary label
...

Conflicts:
config/builder.py
include/petscdmplex.h
src/dm/impls/plex/plex.c
src/snes/examples/tutorials/ex62.c

show more ...

PetscFV: Another complex fix

PetscFV: Straighten out complex types

PetscFV: Complex fixes

PetscFV: Fixed Fortran bindings

PetscFV+Plex: Refactored to put reconstruction routines in PetscFV_LeastSquares
- Added PetscFVComputeGradient()
- Added PetscFVLeastSquaresSetMaxFaces()
- Moved last squares fitting routines to Pets

PetscFV+Plex: Refactored to put reconstruction routines in PetscFV_LeastSquares
- Added PetscFVComputeGradient()
- Added PetscFVLeastSquaresSetMaxFaces()
- Moved last squares fitting routines to PetscFV_LeastSquares, and all the work space

show more ...

PetscLimiter: Add other implementations

PetscFE: Fix complex casting

PetscFV: Added PETSCFVLEASTSQUARES
- Still need to move in implementation pieces

PetscFV: Added PetscFVGet/SetComputeGradients()

PetscFV+Doc: Fixed some function docs

PetscFE+Doc: Fixed some function docs

PetscLimiter: Initial checkin
- Its not clear that I have the correct interface, we may want to do the transformation to f inside

DMPlex: Fixup type mismatch in PetscFEIntegrate()

PetscFV: Initial checkin
- Added an Upwind implementation

PetscFE: Fixes for discontinuous elements

PetscSpace: Fixed polynomial space view

PetscFE: Changed PetscFECreateDefault()
- Now continuity and reference cell shape are set from the command line

PetscFE: Added domain integration
- Added PetscFEIntegrate(), only implemented for Basic
- Added DMPlexComputeIntegralFEM() and DMPlexComputeL2FieldDiff()
- This should replace the L2Diff functions

PetscFE: I think I finally fixed the cell dof