Dear PETSc users, The patch release petsc-3.5.4 is now available for download. http://www.mcs.anl.gov/petsc/download/index.html
Some of the changes include: * Mat: Fix issue where the nonzerostate of Mat_MPIAIJ did not monotonically increase when matrix entries are zeroed in B. * Mat: bugfix for MatPtAPSymbolic_MPIAIJ_MPIAIJ() * DMDA: DMDAVTKWriteAll_VTS - Fix handling of higher-dimension coordinates * MatNestGetISs: use custom fortranstub to handle PETSC_NULL_OBJECT * Sys: fix PetscSubcommSetTypeGeneral wrt zero-ing the subsizes array * PetscEnumFind: Fix wrong counting of entries in list * configure: fix failure with --with-batch=1 * MatLoad_SeqDense() always allocated new space for the data even if the user already provided it * Added PetscBagRegister64bitInt() * Removed direct use of HDF5_NATIVE_DOUBLE to support single precision as well * DMShell: destroy local to local scatter * Fix for ISGlobalToLocalMappingApply() for when it has a block size > 1 * bugfix for MatGetRowIJ_SeqSBAIJ() * PetscDrawLGDraw: draw and flush only in process zero * PetscDrawSetType: Check the draw type is registered before destroying * TS: Use %g to draw time in TSMonitorDrawSolution() * PCMG must set initial guess nonzero for down (pre) smoother when using full or Kaskade multigrid because otherwise it zeros out the interpolated solution from the coarser level * Ensure the DMDA PetscViewerBinary writer using MPI-IO functionality respects the skip header flag when wither writing or reading files. * Added checks to enable loading of a vector which was written without the binary header. Satish
