changes for solving the strict aliasing issue in PetscObject and derived

Hg-commit: 278faecc91564f5495fa9f7d48380d547995fa45

changed bmake/common everywhere to conf
fixed a little more for new install system

Hg-commit: 1aa781ff09e652ebafd6f016d84e88c15ea650f1

docs

Hg-commit: 7a639698b8f1f79c3c69433d58f6632b62669d33

fix compile errors

Hg-commit: 3894b7441633f67a6bd31af8e686ab6640ef4658

removed Vec x arguement from MatCreateMFFD() and MatCreateSNESMF()
fixed couple of nightly build outputs

Hg-commit: 87488d3d99d32edefe80d688e1de04953f0049ef

removed matmffd debug statements, testexamples runs

Hg-commit: 10603d7aa910f9e77ce72f07c685ec2461fb9bcc

fixed MatCreateSNESMF() so that if MatMFFDSetBase() is called then the base
is no longer ever set by MatAssemblyEnd_SNESMF(). This is the way the old
verson worked and made DMMG for SNES work with ma

fixed MatCreateSNESMF() so that if MatMFFDSetBase() is called then the base
is no longer ever set by MatAssemblyEnd_SNESMF(). This is the way the old
verson worked and made DMMG for SNES work with matrix-free

Hg-commit: 93dfb1436d944e1b8612848f65782f14b0eccf31

show more ...

fixed problem with funvec and function_f usage in reworked MFFD

Hg-commit: 9457db19a553fcd733092243d0947e717eb48f98

removed no longer needed snesmf* files and custom fortran interface

Hg-commit: d3ac7ee04f1b43b3d407675ebd7875c1f77a6eae

stripped MatSNESMF() code to what is needed after moving over MatMFFD stuff

Hg-commit: fae73a7f9491ceaf0f0f812761cf28e5c8138b1a

preliminary move of MatSNESMF stuff to mat/impls directory not tested

Hg-commit: cdc2d6c91cd5f6a4972e0f160887a3eae1b957f5

changed PetscFListDestroy() to again take the pointer and zero
it instead of relying on caller to zero

Hg-commit: 172f34d6b8078eedc0431c8ccfeae639116e17f9

XXXSetType() need not call XXXRegisterAll() since XXXCreate() calls XXXInitializePackage()

Hg-commit: 1d153ffcd810a51e9367f1b0827ad0c0027606cf

MatSNESMFCtx --> MatSNESMF

Hg-commit: 6457ce4e5cd2be3ffc9382e844c5bba133d89328

more minor changes to increase consistency of method calling sequences and naming

Hg-commit: 4d17aa8465a0d1835166ba4c0ee47a08a267b110

fix incref/decref pattern in many places

Hg-commit: 22e103f7957be81eb52288a9ed4ffb27537b6d8f

typo fix. Rm extra 'const'.

Hg-commit: a97526cfd74afb8c087e679c7448461f31034d40

moved many implementation headers to include/private

Hg-commit: aa99826eaac01bf46b6b3b0207d0ecd69d3393cb

changed ksp,snes,ts monitor function names and command line options
to a consistent model like the -ksp_view_.... model.

Hg-commit: 59f4eafae0b5023efcea351555f5a4c24277ba99

Cleanup of MatConvert....
Removal of special AIJ matrix type
Moved setting the matrix type_name from MatSetType() to each MatCreate_

Pretty messy stuff, not tested for external packages
probably l

Cleanup of MatConvert....
Removal of special AIJ matrix type
Moved setting the matrix type_name from MatSetType() to each MatCreate_

Pretty messy stuff, not tested for external packages
probably lots of broken stuff

Hg-commit: 4d6c1df13f9fa0d93bc9d5bb964240852547fe7b

show more ...

added KSP/PCGetOperatorsSet() to determine if the operator is ready
inside the solver object (i.e. KSPSetOperators() was called)
changed all uses of KSP/PCGetOperators() that were used to see
i

added KSP/PCGetOperatorsSet() to determine if the operator is ready
inside the solver object (i.e. KSPSetOperators() was called)
changed all uses of KSP/PCGetOperators() that were used to see
if the operator had been set to use KSP/PCGetOperatorsSet() instead
Minor manual page fixups

Hg-commit: aef2f37a6c88c0a2f393ce75cd2b759d94512c0f

show more ...

vecgreateghost now auto sets VecSetLocalToGlobalMapping()
hg sucks, I don't remember what I changed


Hg-commit: 80900d4fa4bf867df1dc3867199aa585a0afb76c