Change to ISContiguousLocal, ISContiguous was a dumb idea. Specialize VecGetSubVector to be no-copy when possible.

Hg-commit: 7e322e7f6ca20e1f3fcabac1bb98d7648fa30947

Add ISContiguous and ISSetContiguous, to test for the ability to make a fast subvector

Hg-commit: 85f6004cd26354b2acd3658a6c83cb4e8308dc7a

ISSetBlockSize_Stride needs to set is->bs

Hg-commit: 9cb3a14dfcb08af23e98be1334a9cf6c4dfa67a5

Add ISLocalToGlobalMappingGetIndices()

Hg-commit: eea7cc64a88fc9a2e3c83f2e385f95d1e1305e39

Change ISBlockGetBlockSize to ISGetBlockSize, add ISSetBlockSize

Hg-commit: 0a222ea1e9a192436a7e4681121fe012f4928844

Add ISOnComm() to associate an index set with a different communicator, with optional copy

Hg-commit: ed078ad75ff96815ad327a4c8cf7c249ae6ef9d1

silly index translation mistake

Hg-commit: 10bae133f38d41de56e6c09860e02fe3b76aaddb

Remove stale "Developer Note"

Hg-commit: 6bd24e55bf12b075700994320779022edd6a715f

Missed a __FUNCT__

Hg-commit: 0fa1e14243e184d2fbcce16bee518663850e7929

Add ISLocalToGlobalMappingUnBlock()

Hg-commit: 3312f3901584240c926dc971d8bb48dbf9d54cfc

Remove ISAllGatherIndices, use ISCreateGeneral() then ISAllGather() instead

Hg-commit: 792f1190630e5dbe5875a022bdf47ab8b9b7235b

Wrap ISGetNonlocalIndices in an IS.

Hg-commit: c9a9a3ea7e43738b096ec37cb82da6b2031b394f

ISTotalIndices, ISNonlocalIndices added. This causes weird Fortran errors in alltest.

Hg-commit: 24c33df10190a7e2bd2ba036afac383ff8c93632

Merge.

Hg-commit: 408da9d97fba22d7285ddfd7e8a6e622299e2cce

Apparent bugs in asm that allocate not arrays of objects, but arrays of pointers to objects.
Minor doc/comment fixes.

Hg-commit: bf80f5b2d259b82720b566b07eb191b67098cfb0

added MatZeroRowsColumns_MPIAIJ() and modified ex140.c to test parallel case

Hg-commit: 58e4f0a93bc8e8e34c80527848169824711c6a3a

forgot to list ISToGeneral_General() in is ops table for ISGENERAL

Hg-commit: 42bbacd894a8f75b4206e8fb45d564ac1157147f

commit after massive merge, what the hey did people do to PETSc?

Hg-commit: 031d5f8bde1a14a3fe45758d80598ded09e1c94a

added MatZeroRowsColumns() for SeqAIJ and SeqBAIJ, added test code for this
added binary viewer for ISGENERAL and added code to load IS into Matlab
added changes and faq for MatZeroRowsColumns()

Hg-

added MatZeroRowsColumns() for SeqAIJ and SeqBAIJ, added test code for this
added binary viewer for ISGENERAL and added code to load IS into Matlab
added changes and faq for MatZeroRowsColumns()

Hg-commit: 1121f78e90708a5c1c51c1e9ef599ba80ffa6f27

show more ...

added vec x and b to the MatZeroRowsXXX() routines, they are not yet used.

Hg-commit: aeb830e2090107e6c8d4fcf07605e004a9e621cf

Remove ISBlock(), ISStride() as there is namespace clash with ISBLOCK & ISSTRIDE [fortran allcaps]
Suggest using PetscTypeComapre() directly instead.

Hg-commit: aca83e6ea6c9ac1f4c54c4ffc4c6b8ad7be57

Remove ISBlock(), ISStride() as there is namespace clash with ISBLOCK & ISSTRIDE [fortran allcaps]
Suggest using PetscTypeComapre() directly instead.

Hg-commit: aca83e6ea6c9ac1f4c54c4ffc4c6b8ad7be57763

show more ...

Clean up after messy sed in original conversion: (PetscBool ) -> (PetscBool)

Hg-commit: d2297d04e5f4a401f60e8ca64322828d2eec0c0c

PetscOptionsTruth -> PetscOptionsBool and related

Hg-commit: dfea0d194d94180043ddc1c2d9d074a0b8717c77

ISCreateBlock() now takes the PetscCopyMode argument

Hg-commit: af104d8e976a6631e09ed8c035ce3c2897191d87

ISCreateBlock() and ISBlockSetIndices() now takes values relative to blocks not indices. So the third block, for example, is indicated with 3 not 3*bs

Hg-commit: 71ba4ee06e95b5fcfc0ac2ec0bc9dfc756f7

ISCreateBlock() and ISBlockSetIndices() now takes values relative to blocks not indices. So the third block, for example, is indicated with 3 not 3*bs

Hg-commit: 71ba4ee06e95b5fcfc0ac2ec0bc9dfc756f72691

show more ...