MATIS: deprecate MatISGetMPIXAIJ in favour of MatConvert()

Updated code and tests

Merge remote-tracking branch 'origin/stefano_zampini/feature-ptatin3d-support'

reverting back to using shared MatSolve interface

Merge branch 'master' into denera/lmvm-mat-and-pc

Merge branch 'stefano_zampini/feature-fieldsplit-innerkspwithpcksp'

PCBDDC: use corners only if present (ptatin support)

PCBDDC: added command line option to filter small entries when detecting disconnected subdomains (damn dolfin)

PCBDDC: allow command line debug of SubSchurs

PCBDDC: use new API ISLocalToGlobalMappingGetNodeInfo

PCBDDC: added command line option to visualize the coarse matrix

Custom solve function for LMVM matrices, disconnected from the common MatSolve interface

Merge remote-tracking branch 'origin/maint'

PCPatch: C89 fix

PCKSPSetKSP: should use TryMethod

PCFieldSplit: adde PCFieldSplitSchurGetSubKSP() method

this method differs from PCFieldSplitGetSubKSP() since it returns all the KSPs used in the Schur factorization

PCFieldSplit: use PCKSP as PC for KSP inv(A00) when an inner KSP for the Schur complement is asked for

Fixed whitespaces

PCKSP: added PCKSPSetKSP() method

cannot call the reset method during PCDestroy_KSP since it will reset the operators of the inner KSP

PCKSP: adhere with PETSc standards about getters method

PCPatch: C89 fix

PCBDDC: minor fixes for sequential runs

PCMG: remove misplaced PetscFunctionBegin

PCMG: Need to copy the finest rhs if the smoother changes it during presolve

PC+Doc: Patch documentation fixes

PCPatch: Updated to new hash table interface

Merge branch 'master' into knepley/feature-pc-patch

* master: (107 commits)
cast in GAMG_HASH so that it works for PetscInt32 near PETSC_INT_MAX without overflow to negative numbers
Add link for

Merge branch 'master' into knepley/feature-pc-patch

* master: (107 commits)
cast in GAMG_HASH so that it works for PetscInt32 near PETSC_INT_MAX without overflow to negative numbers
Add link for DrMemory which may serve as an alternative to valgrind on Windows
Finish Gail's copyedits for developers manual.
openmpi: fix namespace conflict [with static build on ppc64le]
two pages of Gail's edits to developers guide
Another chapter of Gail's copyedits for the developers guide
Gail's copyedits on two more chapters of the developers guide
Fix more %d to %D in GAMG code. Fix code style in GAMG code
DMPlexVTK: vertices is int! need to use %d instead of %D
TS ex18: update outputs (reported by nightly tests)
Use %D in error messages for GAMG hash; fix style issues with GAMG hash
config: Update petsc4py git commit
Fix bad plain extern declaration.
Add Gail's copyedits for the initial portion of the developers guide
Add missing common block to subroutine
fix mpiuni tests with xl compilers
Copy & paste the change on PetscSparseDensePlusDot
Fix a wrong macro
PetscInfo: Add missing errorchecking
Log: various fixes and cleanups
...

Conflicts:
src/dm/impls/plex/plex.c

show more ...