Control over initial SNES func/norm in one place; inner ASPIN fix.

Fixed the behavior of left-pc SNESes preconditioning right-pc SNESes.

Doc fixes in snesob.c, anderson.c, snesngmres.c

PetscObjectQueryFunction: use macro to hide PetscVoidFunction* cast

Rationale is the same as the parent. This makes all of
PetscFunctionListAdd(), PetscFunctionListFind(),
PetscObjectComposeFunction

PetscObjectQueryFunction: use macro to hide PetscVoidFunction* cast

Rationale is the same as the parent. This makes all of
PetscFunctionListAdd(), PetscFunctionListFind(),
PetscObjectComposeFunction(), and PetscObjectQueryFunction() consistent
regarding function pointer return types.

show more ...

removed string version of function name for XXXRegister(), PetscFunctionListAdd() and PetscObjectComposeFunction()

removed path argument to XXXInitializePackage() and XXXRegister()

removed XXXRegisterDynamic() but kept the APIs for everything else underneath the same
phase I of the update to handling registering function pointers

Trivial NPC logging fix for NCG and QN.

removed excessive use of CHKMEMQ;

got value back from webclient to application

Added a damping parameter to SNESNASM.

This involved adding the public functions:
SNESNASM{Set/Get}Damping

And the command option:
-snes_nasm_damping

Added -snes_nasm_finaljacobian_type; selects where jacobian is computed at.

fixed calls to PetscObjectAMSTake/GrantAccess() to cast first argument to PetscObject

work on AMS interface, consistent naming, ams.h not included in petscviewer.h, no ifdefs

Hg-commit: 642571a1dc7109fdb7e5b0abb23bf260d10eb970

added KSPSetPreSolve() and KSPSetPostSolve() and removing SNES_KSPSolve() having the Eisenstat-Walker code use the SetPre and Post

Hg-commit: e42d7b26aaba7877248896e0fb4089fe619da155

made public several private KSP and SNES methods that may be used by plugin KSP and SNES classes written by users

Hg-commit: 0d6205a1c0feb0cbd5a40a08327fe43345c19792

FAS isn't funny anymore.

Hg-commit: d75822f87abb43bbbcdb3f6c06cf53eb04f2bc0c

fixed function names with Default to match new standard (which I just wrote)
fixed a few more functions that should be PETSC_INTERN

Hg-commit: 87889fe4f6d4e8d0963adcc79d438e75330601c6

fixed many more private functions as PETSC_INTERN
renamed a few badly named private functions

Hg-commit: b9e1405ad332ccf13f675f0aa2256f1c838a5595

Change all PETSC_EXTERN_C to PETSC_EXTERN

Hg-commit: 8d2ebbb193fb583bccc64015e35640c4e08c3426

updated from EXTERN_C_BEGIN/END through sys directory

Hg-commit: ee6b7ef2f48ce66402e7e807302de19506ce15df

began removing unneeded EXTERN_C_BEGIN/END and converting needed ones to PETSC_EXTERN_C

Hg-commit: e0c157606a3283d5a4ba8902a0da5cad0e2aa531

changed use of PetscObjectComposeFunctionDynamic() to PetscObjectComposeFunction() to allow use of static for functions for standard use in PETSc
PetscObjectComposeFunctionDynamic() is still availabl

changed use of PetscObjectComposeFunctionDynamic() to PetscObjectComposeFunction() to allow use of static for functions for standard use in PETSc
PetscObjectComposeFunctionDynamic() is still available for use if needed
also fixed calls to PetscObjectComposeFunction() to not wrap lines (per PETSc coding style)

Hg-commit: 822f9ddaac95a8ff6c2a9ad77fbf07d02d2c20d9

show more ...

nonlinear solvers that generate NaN or Inf fnorm should return with diverged reason not generate error

Hg-commit: 5591d03c0bbe46350918e670562ca4785343cd3d

Compute fresh func in ASPIN Jac. eval; works with TSTheta+Coloring now

Hg-commit: 00c55d5515d8cc5fd2e11b07fe8ccc23758a83b7