moved non-Vec portions of PetscSection to vec/is subpackageHg-commit: 9e3e9ff2b4d00a78c47133c617a9f7e4efa62be9
moved PetscLayout include info to petscis.h and source code to vec/is/utilspart of goal of moving non-Vec stuff down from VecHg-commit: b3287daffebf7c4cd59893136d387f2ed3b835bf
unneeded includes of petsc.h in examplesHg-commit: 4a5f10caf1d0c33404a6d49f220e90178eae2f45
removed PetscViewerBinaryMatlabOpen() and friends, incorporated functionality into binary PETSc viewerHg-commit: 1676211ea1509edcdd7a044ca5b8b543f443b8ab
change islocaltoglobalmappingapply_() from auto to custom.Perhaps it was custom because ISLocalToGlobalMappingApply() was a macro atsome point - and couldn't be directly used in fortranstubsHg-co
change islocaltoglobalmappingapply_() from auto to custom.Perhaps it was custom because ISLocalToGlobalMappingApply() was a macro atsome point - and couldn't be directly used in fortranstubsHg-commit: 917347d8af115cff7f95de5d9cfda1f51da158eb
show more ...
removed some isimpl.h dependencies from some Mat impls and documented othersmoved CUSP flag from vecimpl.h to petscimpl.hHg-commit: 2c48aa9e76aa67a361ac3f6b034d78e9e3c7b825
removed dependency of VecScatter on isimpl.hHg-commit: 5c06794de8d6cbbc22248eb247593e202a08ed30
put in a few more PetscObjectComm() I missed before due to not general enough regexpwho the hell puts a w in front of comm for no reasonHg-commit: 7bcda6053dee43d0c90c32cc693ca5bf2fdf4c99
added PetscObjectComm() and used it to replace (((PetscObject)obj)->comm)Hg-commit: 3da37c458124ad48ae939f4e9823e4430ee0b8be
PetscSection: When cloning, do not recalculate offsetsHg-commit: b15b30d25f41b46850e362b15ea41ec90dbf3bbc
removed typeid from PetscHeaderCreate() args since it is not know at create time anywayHg-commit: 368e928141e31cf1ff6cf108913006e5886dc6cc
VecGetArrayRead now falls back to ops->getarray() when no ops->getarrayread()This fixes ce5ba4f43dbec9b9832148a287e5c109e88d34d1 compatibility fornon-native Vec types that do not implement ops->ge
VecGetArrayRead now falls back to ops->getarray() when no ops->getarrayread()This fixes ce5ba4f43dbec9b9832148a287e5c109e88d34d1 compatibility fornon-native Vec types that do not implement ops->getarrayread.Hg-commit: cfa7a4364378d72110be93ff1267e8b3c7bbc8b7
Make petscimpl.h privateHg-commit: 910e8b03a54a915953944da6f75eb81c807d7428
Move VecGetArray to rvector.c, removes dependence of petscvec.h on vecimpl.hHg-commit: 2b7fb7663be0517b628b8912084d19baad118c82
Add VecGetLayout() and remove some needless dependencies on vecimpl.hHg-commit: 0d0721e64fa8861a4e7168e2a936ca714ebeebcb
Add vec->ops->getarrayread and restorearrayread so implementations can use cachingHg-commit: db0823f4f9cb0108b7977a0d249d7b245498a249
petscvec.h no longer includes petsc-private/vecimpl.hThis makes dependence on Vec internals more explicit. Almost allimpacted dependencies fall in two categories:* reference to vec->map: could b
petscvec.h no longer includes petsc-private/vecimpl.hThis makes dependence on Vec internals more explicit. Almost allimpacted dependencies fall in two categories:* reference to vec->map: could be eliminated by adding VecGetLayout(), which is probably a good idea anyway.* dependence on VecScatter_MPI_General: this is low-level anyway, so I don't think there is much to gain by abstraction.Hg-commit: 612ef938483b9f0137f3e95bc1d14775547d74f1
More sensible names for IS embedding/covering functions.Hg-commit: 0d337f8ea59ecb3866f87211bd87cbfbcc3a9358
added include stddef.h to petscsys.hremoved many duplicate includes of system includesHg-commit: 26ea335c035572bcbd495f9993b3270632d5d3c6
removed a bunch of duplicate unneeded includes of std include filesHg-commit: b6c9f3e2b0eb8b7277fe131cc83beede5d495522
source code need not include math.h since it is managed by the PETSc include filesa couple other unneeded included files also removedHg-commit: f47a2dda441cb818ab3be604570bccc0c8f15a89
removed PETSC_NULL from C and Fortran (except declaration in C for backward compatibility). Kept PETSC_NULL_xxx for FortranFixed a few bugs where PETSC_NULL had been used incorrectly.Hg-commit: 05
removed PETSC_NULL from C and Fortran (except declaration in C for backward compatibility). Kept PETSC_NULL_xxx for FortranFixed a few bugs where PETSC_NULL had been used incorrectly.Hg-commit: 054705a517d7f4388a8a084415d7478cbe95dff4
Fix broken bfort commentsHg-commit: 3869e6bef39385e053b7eb1d69d693755334d323
wrapped LAPACK calls in PetscStackCall()Hg-commit: 9cee7ca55a1199d641b77bd38e3a2ece2cb4320b
added back in /*I "petscvec.h" I*/ in several files I mistakenly removed it fromwrapped all the BLAS calls in PetscStackCall()Hg-commit: 18be25a7ff539aa5230355449bf40d90e959b06a
1...<<141142143144145146147148149150>>...278