Debian Codegraph @

doc/source/ChangeLog.rst

raw
1 Change log
2 ==========
3
4 2019.2.0.dev0
5 -------------
6
7 - No changes yet.
8
9 2019.1.0 (2019-04-19)
10 ---------------------
11
12 - Support for shape calculus operations, see file:
13 ``python/test/unit/fem/test_coordinatederivative.py``
14 - Support for PETSc and SLEPc 3.10.
15 - Documented Python elastodynamics demo.
16 - Various bug fixes, e.g. memory leaks, build issues.
17 - Enhancements and additions to pybind11-based Python interface.
18
19 2018.1.0 (2018-06-14)
20 ---------------------
21
22 - Remove python2 support.
23 - Remove SWIG (now use pybind11 for wrapping C++)
24 - Rename ``mpi_comm_world()`` to ``MPI.comm_world``.
25 - Removed ``from fenics import *``, use ``from dolfin import *``
26 - Rename ``ERROR``, ``CRITICAL`` etc. to ``LogLevel.ERROR``, ``LogLevel.CRITICAL``.
27
28 2017.2.0 (2017-12-05)
29 ---------------------
30
31 - Remove ``UnitQuadMesh`` and ``UnitHexMesh``. Now use ``UnitSquareMesh`` and
32 ``UnitCubeMesh`` with cell type qualifiers.
33 - Remove ``MeshEditor::open`` without cell type. Now you must explicitly
34 specify CellType when opening a ``Mesh`` with ``MeshEditor``.
35 - Rename ``Mesh::size_global`` to ``Mesh::num_entities_global``.
36 - Remove ``Mesh::size``. Use ``Mesh::num_entities`` instead.
37 - Improved mesh topology computation performance.
38 - Remove excessive calls to MPI init. It may now be necessary in some
39 cases to explicitly intialise MPI.
40 - Improvements to sparsity pattern computation.
41 - Addition of some interfaces using ``Eigen::Map/ref`` in addition to
42 ``dolfin::Array(View)``. ``dolfin::Array(View)``interfaces will be
43 removed in favour of Eigen interfaces in the next release.
44 - Update pkg-config (dolfin.pc) file.
45 - CMake modernisations, with more use of exported targets.
46 - Add experimental pybind11 generated Python interface. Will replace
47 the SWIG generated interface in the 2018.1 release.
48 - Remove redundant SLEPc interfaces. Fixes issue `#908
49 <https://bitbucket.org/fenics-project/dolfin/issues/908>`_.
50 - Bump required Boost version to 1.55.
51 - Remove PETScUserPreconditioner (was unused and untested).
52 - Remove VTK plotting backend. Plotting is no longer available from
53 the C++ interface. Basic plotting is available using ``matplotlib``
54 and ``x3dom`` backends via the ``plot()`` free function in the
55 Python interface. Users are advised to move to e.g. Paraview for
56 more demanding plotting needs.
57 - Updates for ``ufc::finite_element::evaluate_vertex_values``
58 interface change.
59 - Add new methods ``XDMFFile::write_checkpoint``,
60 ``XDMFFile::read_checkpoint`` to write and read (checkpoint)
61 functions.
62 - Implement marking vertex and edge mesh functions by
63 ``SubDomain::mark()`` using ``on_boundary`` flag.
64 - Fix quadratic scaling in PETSc matrix allocation with global dofs;
65 assembly with ``Real`` space now exhibits linear scaling in number
66 of dofs.
67 - Add assembly for quadrilateral and hexahedral meshes with CG and DG
68 elements.
69 - Updates for some demos and tests to show usage of quadrilateral and
70 hexahedral meshes.
71 - Deprecate ``CellSize`` (equivalent to ``2*Circumradius``)
72 in favour of new ``CellDiameter``; add ``MinCellEdgeLength``
73 and ``MaxCellEdgeLength``
74 - Deprecate subclassing of ``Expression`` in Python; new Python class
75 ``UserExpression`` introduced for user overloads
76 - Deprecate ``VertexFunction``, ``EdgeFunction``, ``FaceFunction``,
77 ``FacetFunction``, ``CellFunction``; use ``MeshFunction`` instead
78
79
80 2017.1.0 (2017-05-09)
81 ---------------------
82
83 - Refactor PETScLUSolver to use functionality from
84 PETScKrylovSolver. Simplify interface for solving transposed
85 systems. Fixes #815.
86 - Switch default Python version to Python 3. Use
87 `-DDOLFIN_USE_PYTHON3=off` to build with Python 2.
88 - Remove redundant ``solve_transpose`` functions (use solve with bool
89 argument instead)
90 - Remove OpenMpAsssmebler
91 - Remove MPI communicator as argument in GenericVector::init functions
92 (communicator should be passed via constructor)
93 - Remove ``Function::operator[+-*/]`` to prevent memory corruption problems
94 (does not affect Python interface)
95 - Fix XDMF3 output of time series. The default output method is now to assume
96 that all functions have different meshes, and that the meshes change from
97 time step to time step. Two parameters control the output, one limits each
98 function to only one mesh for the whole time series, turn off the default
99 on parameter ``rewrite_function_mesh`` to enable this. You can also make
100 all functions share the same mesh and time series, which currently is better
101 supported in Paraview than the alternative, turn on ``functions_share_mesh``
102 for this. These two parameters can also be combined in case all functions
103 share the same mesh at all time steps. This creates minimal size files.
104 - Add ``PETScSNESSolver`` and ``PETScTAOSolver`` constructor accepting
105 both communicator and type
106 - Expression("f[0]*f[1]", f=obj) notation now supported for non-scalar
107 GenericFunction obj
108 - Expression("f", f=obj) notation now supports obj of MeshFunction types
109 (only cell based)
110 - Fix MPI deadlock in case of instant compilation failure
111 - Allow using ``Timer`` as context manager and add ``timed`` decorator
112 to measure timings of functions and methods
113 - Add ``NonlinearProblem::J_pc`` and support preconditioning matrix in
114 ``NewtonSolver``, ``PETScSNESSolver`` and ``PETScTAOSolver``
115
116 2016.2.0 [2016-11-30]
117 ---------------------
118
119 - Updates to XDMFFile interface, now fully supporting MeshFunction and
120 MeshValueCollection with multiple named datasets in one file (useful for
121 volume/boundary markers). Time series now only when a time is explicitly
122 specified for each step. Full support for ASCII/XML XDMF.
123 - Improved X3DOM support
124 - Improved detection of UFC
125 - Add CMake option `-DDOLFIN_USE_PYTHON3` to create a Python 3 build
126 - Require CMake version 3.5 or higher
127 - Add pylit to generate demo doc from rst
128 - More careful checks of Dirichlet BC function spaces
129 - Change definition of FunctionSpace::component()
130 - Adaptive solving now works for tensor-valued unknowns
131 - Improve logging of PETSc errors; details logged at level TRACE
132
133
134 2016.1.0 [2016-06-23]
135 ---------------------
136 - Remove support for 'uint'-valued MeshFunction (replaced by 'size_t')
137 - Major performance improvements and simplifications of the XDMF IO.
138 - Remove Zoltan graph partitioning interface
139 - Add new algorithm for computing mesh entiites. Typical speed-up of
140 two with gcc and four with clang. Reduced memory usage for meshes
141 with irregularly numbered cells.
142 - Remove STLMatrix, STLVector, MUMPSLUSolver and PastixLUSolver
143 classes
144 - Remove PETScPreconditioner::set_near_nullspace and add
145 PETScMatrix::set_near_nullspace
146 - Build system updates for VTK 7.0
147 - Remove XDMF from File interface. XDMF is XML based, and has many
148 possibilities for file access, which are not accessible through the
149 limited File interface and "<<" ">>" operators. Instead of File, use
150 XDMFFile, and use XDMFFile.read() and XDMFFile.write() for
151 I/O. Demos and tests have been updated to show usage. XDMF now also
152 supports ASCII I/O in serial, useful for compatibility with users
153 who do not have the HDF5 library available.
154 - Require polynomial degree or finite element for Expressions in the
155 Python interface (fixes Issue #355,
156 https://bitbucket.org/fenics-project/dolfin/issues/355)
157 - Switch to Google Test framwork for C++ unit tests
158 - Fix bug when reading domain data from mesh file for a ghosted mesh
159 - Add interface for manipulating mesh geometry using (higher-order) FE
160 functions: free functions set_coordinates, get_coordinates,
161 create_mesh
162 - Fix bug when reading domain data from mesh file for a ghosted mesh.
163 - Remove reference versions of constructors for many classes that
164 store a pointer/reference to the object passed to the
165 constructor. This is an intrusive interface change for C++ users,
166 but necessary to improve code maintainabilty and to improve memory
167 safety. The Python interface is (virtually) unaffected.
168 - Remove class SubSpace. Using FunctionSpace::sub(...) instead
169 - Remove reference versions constructors of NonlinearVariationalSolver
170 - Remove setting of bounds from NonlinearVariationalSolver (was
171 already available through NonlinearVariationalProblem)
172 - Update Trilinos support to include Amesos2, and better support from
173 Python
174 - Rewrite interface of TensorLayout and SparsityPattern;
175 local-to-global maps now handled using new IndexMap class;
176 GenericSparsityPattern class removed
177 - Remove QT (was an optional dependency)
178 - PETScTAOSolver::solve() now returns a pair of number of iterations
179 (std::size_t) and whether iteration converged (bool)
180 - Better quality refinement in 2D in Plaza algorithm, by choosing
181 refinement pattern based on max/min edge ratio
182 - Removed refine_cell() method in CellTypes
183 - Enable marker refinement to work in parallel for 1D meshes too
184 - Add std::abort to Python exception hook to avoid parallel deadlocks
185 - Extend dof_to_vertex_map with unowned dofs, thus making
186 dof_to_vertex_map an inverse of vertex_to_dof_map
187 - Clean-up in PyDOLFIN function space design, issue #576
188 - Deprecate MixedFunctionSpace and EnrichedFunctionSpace in favour of
189 initialization by suitable UFL element
190 - Add experimental matplotlib-based plotting backend, see mplot demo
191 - Remove method argument of DirichletBC::get_boundary_values()
192 - Change return types of free functions adapt() to shared_ptr
193
194
195 1.6.0 [2015-07-28]
196 ------------------
197 - Remove redundant pressure boundary condition in Stokes demos
198 - Require Point in RectangleMesh and BoxMesh constructors
199 - Remove BinaryFile (TimeSeries now requires HDF5)
200 - Add (highly experimental) support for Tpetra matrices and vectors
201 from Trilinos, interfacing to Belos, Amesos2, IfPack2 and Muelu.
202 - Enable (highly experimental) support for Quadrilateral and
203 Hexahedral meshes, including some I/O, but no assembly yet.
204 - Enable UMFPACK and CHOLMOD solvers with Eigen backend
205 - Add an MPI_Comm to logger, currently defaulted to MPI_COMM_WORLD
206 allowing better control over output in parallel
207 - Experimental output of quadratic geometry in XDMF files, allows more
208 exact visualisation of P2 Functions
209 - Remove GenericMatrix::compressed (see Issue #61)
210 - Deprecate and PETScKryloveSolver::set_nullspace() and add
211 PETScMatrix::set_nullspace()
212 - Remove uBLAS backend
213 - Remove UmfpackLUSolver and CholmodSolver
214 - Add EigenMatrix/Vector::data()
215 - Remove GenericMatrix/Vector::data() and GenericMatrix/Vector::data()
216 (to use backends that support data(), cast first to backend type,
217 e.g. A = A.as_backend_type()
218 - Remove cmake.local, replaced by fenics-install-component.sh
219 - Make interior facet integrals define - and + cells ordered by
220 cell_domains value.
221 - Remove deprecated arguments *_domains from assemble() and Form().
222 - Change measure definition notation from dx[mesh_function] to
223 dx(subdomain_data=mesh_function).
224 - Set locale to "C" before reading from file
225 - Change GenericDofMap::cell_dofs return type from const
226 std::vector<..>& to ArrayView<const ..>
227 - Add ArrayView class for views into arrays
228 - Change fall back linear algebra backend to Eigen
229 - Add Eigen linear algebra backend
230 - Remove deprecated GenericDofMap::geometric_dim function (fixes Issue
231 #443)
232 - Add quadrature rules for multimesh/cut-cell integration up to order
233 6
234 - Implement MPI reductions and XML ouput of Table class
235 - list_timings() is now collective and returns MPI average across
236 processes
237 - Add dump_timings_to_xml()
238 - Add enum TimingType { wall, user, system } for selecting wall-clock,
239 user and system time in timing routines
240 - Bump required SWIG version to 3.0.3
241 - Increase default maximum iterations in NewtonSolver to 50.
242 - Deprecate Python free function homogenize(bc) in favour of member
243 function DirichletBC::homogenize()
244
245
246 1.5.0 [2015-01-12]
247 ------------------
248 - DG demos working in parallel
249 - Simplify re-use of LU factorisations
250 - CMake 3 compatibility
251 - Make underlying SLEPc object accessible
252 - Full support for linear algebra backends with 64-bit integers
253 - Add smoothed aggregation AMG elasticity demo
254 - Add support for slepc4py
255 - Some self-assignment fixes in mesh data structures
256 - Deprecated GenericDofMap::geometric_dimension()
257 - Experimental support for ghosted meshes (overlapping region in
258 parallel)
259 - Significant memory reduction in dofmap storage
260 - Re-write dofmap construction with significant performance and
261 scaling improvements in parallel
262 - Switch to local (process-wise) indexing for dof indices
263 - Support local (process-wise) indexing in linear algerbra backends
264 - Added support for PETSc 3.5, require version >= 3.3
265 - Exposed DofMap::tabulate_local_to_global_dofs,
266 MeshEntity::sharing_processes in Python
267 - Added GenericDofmap::local_dimension("all"|"owned"|"unowned")
268 - Added access to SLEPc or slepc4py EPS object of SLEPcEigenSolver
269 (requires slepc4py version >= 3.5.1)
270 - LinearOperator can now be accessed using petsc4py
271 - Add interface (PETScTAOSolver) for the PETSc nonlinear
272 (bound-constrained) optimisation solver (TAO)
273 - Add GenericMatrix::nnz() function to return number of nonzero
274 entries in matrix (fixes #110)
275 - Add smoothed aggregation algerbraic multigrid demo for elasticity
276 - Add argument 'function' to project, to store the result into a
277 preallocated function
278 - Remove CGAL dependency and mesh generation, now provided by mshr
279 - Python 2.7 required
280 - Add experimental Python 3 support. Need swig version 3.0.3 or later
281 - Move to py.test, speed up unit tests and make tests more robust in
282 parallel
283 - Repeated initialization of PETScMatrix is now an error
284 - MPI interface change: num_processes -> size, process_number -> rank
285 - Add optional argument project(..., function=f), to avoid superfluous
286 allocation
287 - Remove excessive printing of points during extrapolation
288 - Clean up DG demos by dropping restrictions of Constants: c('+') -> c
289 - Fix systemassembler warning when a and L both provide the same
290 subdomain data.
291 - Require mesh instead of cell argument to FacetArea, FacetNormal,
292 CellSize, CellVolume, SpatialCoordinate, Circumradius,
293 MinFacetEdgeLength, MaxFacetEdgeLength
294 - Remove argument reset_sparsity to assemble()
295 - Simplify assemble() and Form() signature: remove arguments mesh,
296 coefficients, function_spaces, common_cell. These are now all found
297 by inspecting the UFL form
298 - Speed up assembly of forms with multiple integrals depending on
299 different functions, e.g. f*dx(1) + g*dx(2).
300 - Handle accessing of GenericVectors using numpy arrays in python
301 layer instead of in hard-to-maintain C++ layer
302 - Add support for mpi groups in jit-compilation
303 - Make access to HDFAttributes more dict like
304 - Add 1st and 2nd order Rush Larsen schemes for the
305 PointIntegralSolver
306 - Add vertex assembler for PointIntegrals
307 - Add support for assembly of custom_integral
308 - Add support for multimesh assembly, function spaces, dofmaps and
309 functions
310 - Fix to Cell-Point collision detection to prevent Points inside the
311 mesh from falling between Cells due to rounding errors
312 - Enable reordering of cells and vertices in parallel via SCOTCH and
313 the Giibs-Poole-Stockmeyer algorithm
314 - Efficiency improvements in dof assignment in parallel, working on
315 HPC up to 24000 cores
316 - Introduction of PlazaRefinement methods based on refinement of the
317 Mesh skeleton, giving better quality refinement in 3D in parallel
318 - Basic support for 'ghost cells' allowing integration over interior
319 facets in parallel
320
321
322 1.4.0 [2014-06-02]
323 ------------------
324 - Feature: Add set_diagonal (with GenericVector) to GenericMatrix
325 - Fix many bugs associated with cell orientations on manifolds
326 - Force all global dofs to be ordered last and to be on the last
327 process in parallel
328 - Speed up dof reordering of mixed space including global dofs by
329 removing the latter from graph reordering
330 - Force all dofs on a shared facet to be owned by the same process
331 - Add FEniCS ('fenics') Python module, identical with DOLFIN Python
332 module
333 - Add function Form::set_some_coefficients()
334 - Remove Boost.MPI dependency
335 - Change GenericMatrix::compresss to return a new matrix (7be3a29)
336 - Add function GenericTensor::empty()
337 - Deprecate resizing of linear algebra via the GenericFoo interfaces
338 (fixes #213)
339 - Deprecate MPI::process_number() in favour of MPI::rank(MPI_Comm)
340 - Use PETSc built-in reference counting to manage lifetime of wrapped
341 PETSc objects
342 - Remove random access function from MeshEntityIterator (fixes #178)
343 - Add support for VTK 6 (fixes #149)
344 - Use MPI communicator in interfaces. Permits the creation of
345 distributed and local objects, e.g. Meshes.
346 - Reduce memory usage and increase speed of mesh topology computation
347
348
349 1.3.0 [2014-01-07]
350 ------------------
351 - Feature: Enable assignment of sparse MeshValueCollections to
352 MeshFunctions
353 - Feature: Add free function assign that is used for sub function
354 assignment
355 - Feature: Add class FunctionAssigner that cache dofs for sub function
356 assignment
357 - Fix runtime dependency on checking swig version
358 - Deprecate DofMap member methods vertex_to_dof_map and
359 dof_to_vertex_map
360 - Add free functions: vertex_to_dof_map and dof_to_vertex_map, and
361 correct the ordering of the map.
362 - Introduce CompiledSubDomain a more robust version of
363 compiled_subdomains, which is now deprecated
364 - CMake now takes care of calling the correct generate-foo script if
365 so needed.
366 - Feature: Add new built-in computational geometry library
367 (BoundingBoxTree)
368 - Feature: Add support for setting name and label to an Expression
369 when constructed
370 - Feature: Add support for passing a scalar GenericFunction as default
371 value to a CompiledExpression
372 - Feature: Add support for distance queries for 3-D meshes
373 - Feature: Add PointIntegralSolver, which uses the MultiStageSchemes
374 to solve local ODEs at Vertices
375 - Feature: Add RKSolver and MultiStageScheme for general time integral
376 solvers
377 - Feature: Add support for assigning a Function with linear
378 combinations of Functions, which lives in the same FunctionSpace
379 - Added Python wrapper for SystemAssembler
380 - Added a demo using compiled_extension_module with separate source
381 files
382 - Fixes for NumPy 1.7
383 - Remove DOLFIN wrapper code (moved to FFC)
384 - Add set_options_prefix to PETScKrylovSolver
385 - Remove base class BoundarCondition
386 - Set block size for PETScMatrix when available from TensorLayout
387 - Add support to get block compressed format from STLMatrix
388 - Add detection of block structures in the dofmap for vector equations
389 - Expose PETSc GAMG parameters
390 - Modify SystemAssembler to support separate assembly of A and b
391
392
393 1.2.0 [2013-03-24]
394 ------------------
395 - Fixes bug where child/parent hierarchy in Python were destroyed
396 - Add utility script dolfin-get-demos
397 - MeshFunctions in python now support iterable protocol
398 - Add timed VTK output for Mesh and MeshFunction in addtion to
399 Functions
400 - Expose ufc::dofmap::tabulate_entity_dofs to GenericDofMap interface
401 - Expose ufc::dofmap::num_entity_dofs to GenericDofMap interface
402 - Allow setting of row dof coordinates in preconditioners (only works
403 with PETSc backed for now)
404 - Expose more PETSc/ML parameters
405 - Improve speed to tabulating coordinates in some DofMap functions
406 - Feature: Add support for passing a Constant as default value to a
407 CompiledExpression
408 - Fix bug in dimension check for 1-D ALE
409 - Remove some redundant graph code
410 - Improvements in speed of parallel dual graph builder
411 - Fix bug in XMDF output for cell-based Functions
412 - Fixes for latest version of clang compiler
413 - LocalSolver class added to efficiently solve cell-wise problems
414 - New implementation of periodic boundary conditions. Now incorporated
415 into the dofmap
416 - Optional arguments to assemblers removed
417 - SymmetricAssembler removed
418 - Domains for assemblers can now only be attached to forms
419 - SubMesh can now be constructed without a CellFunction argument, if
420 the MeshDomain contains marked celldomains.
421 - MeshDomains are propagated to a SubMesh during construction
422 - Simplify generation of a MeshFunction from MeshDomains: No need to
423 call mesh_function with mesh
424 - Rename dolfin-config.cmake to DOLFINConfig.cmake
425 - Use CMake to configure JIT compilation of extension modules
426 - Feature: Add vertex_to_dof_map to DofMap, which map vertex indices
427 to dolfin dofs
428 - Feature: Add support for solving on m dimensional meshes embedded in
429 n >= m dimensions
430
431
432 1.1.0 [2013-01-08]
433 ------------------
434 - Add support for solving singular problems with Krylov solvers (PETSc
435 only)
436 - Add new typedef dolfin::la_index for consistent indexing with linear
437 algebra backends.
438 - Change default unsigned integer type to std::size_t
439 - Add support to attaching operator null space to preconditioner
440 (required for smoothed aggregation AMG)
441 - Add basic interface to the PETSc AMG preconditioner
442 - Make SCOTCH default graph partitioner (GNU-compatible free license,
443 unlike ParMETIS)
444 - Add scalable construction of mesh dual graph for mesh partitioning
445 - Improve performance of mesh building in parallel
446 - Add mesh output to SVG
447 - Add support for Facet and cell markers to mesh converted from
448 Diffpack
449 - Add support for Facet and cell markers/attributes to mesh converted
450 from Triangle
451 - Change interface for auto-adaptive solvers: these now take the goal
452 functional as a constructor argument
453 - Add memory usage monitor: monitor_memory_usage()
454 - Compare mesh hash in interpolate_vertex_values
455 - Add hash() for Mesh and MeshTopology
456 - Expose GenericVector::operator{+=,-=,+,-}(double) to Python
457 - Add function Function::compute_vertex_values not needing a mesh
458 argument
459 - Add support for XDMF and HDF5
460 - Add new interface LinearOperator for matrix-free linear systems
461 - Remove MTL4 linear algebra backend
462 - Rename down_cast --> as_type in C++ / as_backend_type in Python
463 - Remove KrylovMatrix interface
464 - Remove quadrature classes
465 - JIT compiled C++ code can now include a dolfin namespace
466 - Expression string parsing now understand C++ namespace such as
467 std::cosh
468 - Fix bug in Expression so one can pass min, max
469 - Fix bug in SystemAssembler, where mesh.init(D-1, D) was not called
470 before assemble
471 - Fix bug where the reference count of Py_None was not increased
472 - Fix bug in reading TimeSeries of size smaller than 3
473 - Improve code design for Mesh FooIterators to avoid dubious down cast
474 - Bug fix in destruction of PETSc user preconditioners
475 - Add CellVolume(mesh) convenience wrapper to Python interface for UFL
476 function
477 - Fix bug in producing outward pointing normals of BoundaryMesh
478 - Fix bug introduced by SWIG 2.0.5, where typemaps of templated
479 typedefs are not handled correctly
480 - Fix bug introduced by SWIG 2.0.5, which treated uint as Python long
481 - Add check that sample points for TimeSeries are monotone
482 - Fix handling of parameter "report" in Krylov solvers
483 - Add new linear algebra backend "PETScCusp" for GPU-accelerated
484 linear algebra
485 - Add sparray method in the Python interface of GenericMatrix,
486 requires scipy.sparse
487 - Make methods that return a view of contiguous c-arrays, via a NumPy
488 array, keep a reference from the object so it wont get out of scope
489 - Add parameter: "use_petsc_signal_handler", which enables/disable
490 PETSc system signals
491 - Avoid unnecessary resize of result vector for A*b
492 - MPI functionality for distributing values between neighbours
493 - SystemAssembler now works in parallel with topological/geometric
494 boundary search
495 - New symmetric assembler with ability for stand-alone RHS assemble
496 - Major speed-up of DirichletBC computation and mesh marking
497 - Major speed-up of assembly of functions and expressions
498 - Major speed-up of mesh topology computation
499 - Add simple 2D and 3D mesh generation (via CGAL)
500 - Add creation of mesh from triangulations of points (via CGAL)
501 - Split the SWIG interface into six combined modules instead of one
502 - Add has_foo to easy check what solver and preconditioners are
503 available
504 - Add convenience functions for listing available
505 linear_algebra_backends
506 - Change naming convention for cpp unit tests test.cpp -> Foo.cpp
507 - Added cpp unit test for GenericVector::operator{-,+,*,/}= for all la
508 backends
509 - Add functionality for rotating meshes
510 - Add mesh generation based on NETGEN constructive solid geometry
511 - Generalize SparsityPattern and STLMatrix to support column-wise
512 storage
513 - Add interfaces to wrap PaStiX and MUMPS direct solvers
514 - Add CoordinateMatrix class
515 - Make STLMatrix work in parallel
516 - Remove all tr1::tuple and use boost::tuple
517 - Fix wrong link in Python quick reference.
518
519
520 1.0.0 [2011-12-07]
521 ------------------
522 - Change return value of IntervalCell::facet_area() 0.0 --> 1.0.
523 - Recompile all forms with FFC 1.0.0
524 - Fix for CGAL 3.9 on OS X
525 - Improve docstrings for Box and Rectangle
526 - Check number of dofs on local patch in extrapolation
527
528
529 1.0-rc2 [2011-11-28]
530 --------------------
531 - Fix bug in 1D mesh refinement
532 - Fix bug in handling of subdirectories for TimeSeries
533 - Fix logic behind vector assignment, especially in parallel
534
535
536 1.0-rc1 [2011-11-21]
537 --------------------
538 - 33 bugs fixed
539 - Implement traversal of bounding box trees for all codimensions
540 - Edit and improve all error messages
541 - Added [un]equality operator to FunctionSpace
542 - Remove batch compilation of Expression (Expressions) from Python
543 interface
544 - Added get_value to MeshValueCollection
545 - Added assignment operator to MeshValueCollection
546
547
548 1.0-beta2 [2011-10-26]
549 ----------------------
550 - Change search path of parameter file to
551 ~/.fenics/dolfin_parameters.xml
552 - Add functions Parameters::has_parameter,
553 Parameters::has_parameter_set
554 - Added option to store all connectivities in a mesh for TimeSeries
555 (false by default)
556 - Added option for gzip compressed binary files for TimeSeries
557 - Propagate global parameters to Krylov and LU solvers
558 - Fix OpenMp assemble of scalars
559 - Make OpenMP assemble over sub domains work
560 - DirichletBC.get_boundary_values, FunctionSpace.collapse now return a
561 dict in Python
562 - Changed name of has_la_backend to has_linear_algebra_backend
563 - Added has_foo functions which can be used instead of the HAS_FOO
564 defines
565 - Less trict check on kwargs for compiled Expression
566 - Add option to not right-justify tables
567 - Rename summary --> list_timings
568 - Add function list_linear_solver_methods
569 - Add function list_lu_solver_methods
570 - Add function list_krylov_solver_methods
571 - Add function list_krylov_solver_preconditioners
572 - Support subdomains in SystemAssembler (not for interior facet
573 integrals)
574 - Add option functionality apply("flush") to PETScMatrix
575 - Add option finalize_tensor=true to assemble functions
576 - Solver parameters can now be passed to solve
577 - Remove deprecated function Variable::disp()
578 - Remove deprecated function logging()
579 - Add new class MeshValueCollection
580 - Add new class MeshDomains replacing old storage of boundary markers
581 as part of MeshData. The following names are no longer supported:
582 - boundary_facet_cells
583 - boundary_facet_numbers
584 - boundary_indicators
585 - material_indicators
586 - cell_domains
587 - interior_facet_domains
588 - exterior_facet_domains
589 - Rename XML tag <meshfunction> --> <mesh_function>
590 - Rename SubMesh data "global_vertex_indices" -->
591 "parent_vertex_indices"
592 - Get XML input/output of boundary markers working again
593 - Get FacetArea working again
594
595
596 1.0-beta [2011-08-11]
597 ---------------------
598 - Print percentage of non-zero entries when computing sparsity
599 patterns
600 - Use ufl.Real for Constant in Python interface
601 - Add Dirichlet boundary condition argument to Python project function
602 - Add remove functionality for parameter sets
603 - Added out typemap for vector of shared_ptr objects
604 - Fix typemap bug for list of shared_ptr objects
605 - Support parallel XML vector io
606 - Add support for gzipped XML output
607 - Use pugixml for XML output
608 - Move XML SAX parser to libxml2 SAX2 interface
609 - Simplify XML io
610 - Change interface for variational problems, class VariationalProblem
611 removed
612 - Add solve interface: solve(a == L), solve(F == 0)
613 - Add new classes Linear/NonlinearVariationalProblem
614 - Add new classes Linear/NonlinearVariationalSolver
615 - Ad form class aliases ResidualForm and Jacobian form in wrapper code
616 - Default argument to variables in Expression are passed as kwargs in
617 the Python interface
618 - Add has_openmp as utility function in Python interface
619 - Add improved error reporting using dolfin_error
620 - Use Boost to compute Legendre polynolials
621 - Remove ode code
622 - Handle parsing of unrecognized command-line parameters
623 - All const std::vector<foo>& now return a read-only NumPy array
624 - Make a robust macro for generating a NumPy array from data
625 - Exposing low level fem functionality to Python, by adding a Cell ->
626 ufc::cell typemap
627 - Added ufl_cell as a method to Mesh in Python interface
628 - Fix memory leak in Zoltan interface
629 - Remove some 'new' for arrays in favour of std::vector
630 - Added cell as an optional argument to Constant
631 - Prevent the use of non contiguous NumPy arrays for most typemaps
632 - Point can now be used to evaluate a Function or Expression in Python
633 - Fixed dimension check for Function and Expression eval in Python
634 - Fix compressed VTK output for tensors in 2D
635
636
637 0.9.11 [2011-05-16]
638 -------------------
639 - Change license from LGPL v2.1 to LGPL v3 or later
640 - Moved meshconverter to dolfin_utils
641 - Add support for conversion of material markers for Gmsh meshes
642 - Add support for point sources (class PointSource)
643 - Rename logging --> set_log_active
644 - Add parameter "clear_on_write" to TimeSeries
645 - Add support for input/output of nested parameter sets
646 - Check for dimensions in linear solvers
647 - Add support for automated error control for variational problems
648 - Add support for refinement of MeshFunctions after mesh refinement
649 - Change order of test and trial spaces in Form constructors
650 - Make SWIG version >= 2.0 a requirement
651 - Recognize subdomain data in Assembler from both Form and Mesh
652 - Add storage for subdomains (cell_domains etc) in Form class
653 - Rename MeshData "boundary facet cells" --> "boundary_facet_cells"
654 - Rename MeshData "boundary facet numbers" -->
655 "boundary_facet_numbers"
656 - Rename MeshData "boundary indicators" --> "boundary_indicators"
657 - Rename MeshData "exterior facet domains" -->
658 "exterior_facet_domains"
659 - Updates for UFC 2.0.1
660 - Add FiniteElement::evaluate_basis_derivatives_all
661 - Add support for VTK output of facet-based MeshFunctions
662 - Change default log level from PROGRESS to INFO
663 - Add copy functions to FiniteElement and DofMap
664 - Simplify DofMap
665 - Interpolate vector values when reading from time series
666
667
668 0.9.10 [2011-02-23]
669 -------------------
670 - Updates for UFC 2.0.0
671 - Handle TimeSeries stored backward in time (automatic reversal)
672 - Automatic storage of hierarchy during refinement
673 - Remove directory/library 'main', merged into 'common'
674 - dolfin_init --> init, dolfin_set_precision --> set_precision
675 - Remove need for mesh argument to functional assembly when possible
676 - Add function set_output_stream
677 - Add operator () for evaluation at points for Function/Expression in
678 C++
679 - Add abs() to GenericVector interface
680 - Fix bug for local refinement of manifolds
681 - Interface change: VariationalProblem now takes: a, L or F, (dF)
682 - Map linear algebra objects to processes consistently with mesh
683 partition
684 - Lots of improvemenst to parallel assembly, dof maps and linear
685 algebra
686 - Add lists supported_elements and supported_elements_for_plotting in
687 Python
688 - Add script dolfin-plot for plotting meshes and elements from the
689 command-line
690 - Add support for plotting elements from Python
691 - Add experimental OpenMP assembler
692 - Thread-safe fixed in Function class
693 - Make GenericFunction::eval thread-safe (Data class removed)
694 - Optimize and speedup topology computation (mesh.init())
695 - Add function Mesh::clean() for cleaning out auxilliary topology data
696 - Improve speed and accuracy of timers
697 - Fix bug in 3D uniform mesh refinement
698 - Add built-in meshes UnitTriangle and UnitTetrahedron
699 - Only create output directories when they don't exist
700 - Make it impossible to set the linear algebra backend to something
701 illegal
702 - Overload value_shape instead of dim for userdefined Python
703 Expressions
704 - Permit unset parameters
705 - Search only for BLAS library (not cblas.h)
706
707
708 0.9.9 [2010-09-01]
709 ------------------
710 - Change build system to CMake
711 - Add named MeshFunctions: VertexFunction, EdgeFunction, FaceFunction,
712 FacetFunction, CellFunction
713 - Allow setting constant boundary conditions directly without using
714 Constant
715 - Allow setting boundary conditions based on string ("x[0] == 0.0")
716 - Create missing directories if specified as part of file names
717 - Allow re-use of preconditioners for most backends
718 - Fixes for UMFPACK solver on some 32 bit machines
719 - Provide access to more Hypre preconditioners via PETSc
720 - Updates for SLEPc 3.1
721 - Improve and implement re-use of LU factorizations for all backends
722 - Fix bug in refinement of MeshFunctions
723
724
725 0.9.8 [2010-07-01]
726 ------------------
727 - Optimize and improve StabilityAnalysis.
728 - Use own implementation of binary search in ODESolution (takes
729 advantage of previous values as initial guess)
730 - Improve reading ODESolution spanning multiple files
731 - Dramatic speedup of progress bar (and algorithms using it)
732 - Fix bug in writing meshes embedded higher dimensions to M-files
733 - Zero vector in uBLASVector::resize() to fix spurious bug in Krylov
734 solver
735 - Handle named fields (u.rename()) in VTK output
736 - Bug fix in computation of FacetArea for tetrahedrons
737 - Add support for direct plotting of Dirichlet boundary conditions:
738 plot(bc)
739 - Updates for PETSc 3.1
740 - Add relaxation parameter to NewtonSolver
741 - Implement collapse of renumbered dof maps (serial and parallel)
742 - Simplification of DofMapBuilder for parallel dof maps
743 - Improve and simplify DofMap
744 - Add Armadillo dependency for dense linear algebra
745 - Remove LAPACKFoo wrappers
746 - Add abstract base class GenericDofMap
747 - Zero small values in VTK output to avoid VTK crashes
748 - Handle MeshFunction/markers in homogenize bc
749 - Make preconditioner selectable in VariationalProblem (new parameter)
750 - Read/write meshes in binary format
751 - Add parameter "use_ident" in DirichletBC
752 - Issue error by default when solvers don't converge (parameter
753 "error_on_convergence")
754 - Add option to print matrix/vector for a VariationalProblem
755 - Trilinos backend now works in parallel
756 - Remove Mesh refine members functions. Use free refine(...) functions
757 instead
758 - Remove AdapativeObjects
759 - Add Stokes demo using the MINI element
760 - Interface change: operator+ now used to denote enriched function
761 spaces
762 - Interface change: operator+ --> operator* for mixed elements
763 - Add option 'allow_extrapolation' useful when interpolating to
764 refined meshes
765 - Add SpatialCoordinates demo
766 - Add functionality for accessing time series sample times:
767 vector_times(), mesh_times()
768 - Add functionality for snapping mesh to curved boundaries during
769 refinement
770 - Add functionality for smoothing the boundary of a mesh
771 - Speedup assembly over exterior facets by not using BoundaryMesh
772 - Mesh refinement improvements, remove unecessary copying in Python
773 interface
774 - Clean PETSc and Epetra Krylov solvers
775 - Add separate preconditioner classes for PETSc and Epetra solvers
776 - Add function ident_zeros for inserting one on diagonal for zero rows
777 - Add LU support for Trilinos interface
778
779
780 0.9.7 [2010-02-17]
781 ------------------
782 - Add support for specifying facet orientation in assembly over
783 interior facets
784 - Allow user to choose which LU package PETScLUSolver uses
785 - Add computation of intersection between arbitrary mesh entities
786 - Random access to MeshEntitiyIterators
787 - Modify SWIG flags to prevent leak when using SWIG director feature
788 - Fix memory leak in std::vector<Foo*> typemaps
789 - Add interface for SCOTCH for parallel mesh partitioning
790 - Bug fix in SubDomain::mark, fixes bug in DirichletBC based on
791 SubDomain::inside
792 - Improvements in time series class, recognizing old stored values
793 - Add FacetCell class useful in algorithms iterating over boundary
794 facets
795 - Rename reconstruct --> extrapolate
796 - Remove GTS dependency
797
798
799 0.9.6 [2010-02-03]
800 ------------------
801 - Simplify access to form compiler parameters, now integrated with
802 global parameters
803 - Add DofMap member function to return set of dofs
804 - Fix memory leak in the LA interface
805 - Do not import cos, sin, exp from NumPy to avoid clash with UFL
806 functions
807 - Fix bug in MTL4Vector assignment
808 - Remove sandbox (moved to separate repository)
809 - Remove matrix factory (dolfin/mf)
810 - Update .ufl files for changes in UFL
811 - Added swig/import/foo.i for easy type importing from dolfin modules
812 - Allow optional argument cell when creating Expression
813 - Change name of Expression argument cpparg --> cppcode
814 - Add simple constructor (dim0, dim1) for C++ matrix Expressions
815 - Add example demonstrating the use of cpparg (C++ code in Python)
816 - Add least squares solver for dense systems (wrapper for DGELS)
817 - New linear algebra wrappers for LAPACK matrices and vectors
818 - Experimental support for reconstruction of higher order functions
819 - Modified interface for eval() and inside() in C++ using Array
820 - Introduce new Array class for simplified wrapping of arrays in SWIG
821 - Improved functionality for intersection detection
822 - Re-implementation of intersection detection using CGAL
823
824
825 0.9.5 [2009-12-03]
826 ------------------
827 - Set appropriate parameters for symmetric eigenvalue problems with
828 SLEPc
829 - Fix for performance regression in recent uBLAS releases
830 - Simplify Expression interface: f = Expression("sin(x[0])")
831 - Simplify Constant interface: c = Constant(1.0)
832 - Fix bug in periodic boundary conditions
833 - Add simple script dolfin-tetgen for generating DOLFIN XML meshes
834 from STL
835 - Make XML parser append/overwrite parameter set when reading
836 parameters from file
837 - Refinement of function spaces and automatic interpolation of member
838 functions
839 - Allow setting global parameters for Krylov solver
840 - Fix handling of Constants in Python interface to avoid repeated JIT
841 compilation
842 - Allow simple specification of subdomains in Python without needing
843 to subclass SubDomain
844 - Add function homogenize() for simple creation of homogeneous BCs
845 from given BCs
846 - Add copy constructor and possibility to change value for DirichletBC
847 - Add simple wrapper for ufl.cell.n. FacetNormal(mesh) now works again
848 in Python.
849 - Support apply(A), apply(b) and apply(b, x) in PeriodicBC
850 - Enable setting spectral transformation for SLEPc eigenvalue solver
851
852 0.9.4 [2009-10-12]
853 ------------------
854 - Remove set, get and operator() methods from MeshFunction
855 - Added const and none const T &operator[uint/MeshEntity] to
856 MeshFunction
857 - More clean up in SWIG interface files, remove global renames and
858 ignores
859 - Update Python interface to Expression, with extended tests for value
860 ranks
861 - Removed DiscreteFunction class
862 - Require value_shape and geometric_dimension in Expression
863 - Introduce new class Expression replacing user-defined Functions
864 - interpolate_vertex_values --> compute_vertex_values
865 - std::map<std::string, Coefficient> replaces generated CoefficientSet
866 code
867 - Cleanup logic in Function class as a result of new Expression class
868 - Introduce new Coefficient base class for form coefficients
869 - Replace CellSize::min,max by Mesh::hmin,hmax
870 - Use MUMPS instead of UMFPACK as default direct solver in both serial
871 and parallel
872 - Fix bug in SystemAssembler
873 - Remove support for PETSc 2.3 and support PETSc 3.0.0 only
874 - Remove FacetNormal Function. Use UFL facet normal instead.
875 - Add update() function to FunctionSpace and DofMap for use in
876 adaptive mesh refinement
877 - Require mesh in constructor of functionals (C++) or argument to
878 assemble (Python)
879
880
881 0.9.3 [2009-09-25]
882 ------------------
883 - Add global parameter "ffc_representation" for form representation in
884 FFC JIT compiler
885 - Make norm() function handle both vectors and functions in Python
886 - Speedup periodic boundary conditions and make work for mixed
887 (vector-valued) elements
888 - Add possibilities to use any number numpy array when assigning
889 matrices and vectors
890 - Add possibilities to use any integer numpy array for indices in
891 matrices and vectors
892 - Fix for int typemaps in PyDOLFIN
893 - Split mult into mult and transpmult
894 - Filter out PETSc argument when parsing command-line parameters
895 - Extend comments to SWIG interface files
896 - Add copyright statements to SWIG interface files (not finished yet)
897 - Add typemaps for misc std::vector<types> in PyDOLFIN
898 - Remove dependencies on std_vector.i reducing SWIG wrapper code size
899 - Use relative %includes in dolfin.i
900 - Changed names on SWIG interface files dolfin_foo.i -> foo.i
901 - Add function interpolate() in Python interface
902 - Fix typmaps for uint in python 2.6
903 - Use TypeError instead of ValueError in typechecks in typmaps.i
904 - Add in/out shared_ptr<Epetra_FEFoo> typemaps for PyDOLFIN
905 - Fix JIT compiling in parallel
906 - Add a compile_extension_module function in PyDOLFIN
907 - Fix bug in Python vector assignment
908 - Add support for compressed base64 encoded VTK files (using zlib)
909 - Add support for base64 encoded VTK files
910 - Experimental support for parallel assembly and solve
911 - Bug fix in project() function, update to UFL syntax
912 - Remove disp() functions and replace by info(foo, true)
913 - Add fem unit test (Python)
914 - Clean up SystemAssembler
915 - Enable assemble_system through PyDOLFIN
916 - Add 'norm' to GenericMatrix
917 - Efficiency improvements in NewtonSolver
918 - Rename NewtonSolver::get_iteration() to NewtonSolver::iteration()
919 - Improvements to EpetraKrylovSolver::solve
920 - Add constructor Vector::Vector(const GenericVector& x)
921 - Remove SCons deprecation warnings
922 - Memory leak fix in PETScKrylovSolver
923 - Rename dolfin_assert -> assert and use C++ version
924 - Fix debug/optimise flags
925 - Remove AvgMeshSize, InvMeshSize, InvFacetArea from SpecialFunctions
926 - Rename MeshSize -> CellSize
927 - Rewrite parameter system with improved support for command-line
928 parsing, localization of parameters (per class) and usability from
929 Python
930 - Remove OutflowFacet from SpecialFunctions
931 - Rename interpolate(double*) --> interpolate_vertex_values(double*)
932 - Add Python version of Cahn-Hilliard demo
933 - Fix bug in assemble.py
934 - Permit interpolation of functions between non-matching meshes
935 - Remove Function::Function(std::string filename)
936 - Transition to new XML io
937 - Remove GenericSparsityPattern::sort
938 - Require sorted/unsorted parameter in SparsityPattern constructor
939 - Improve performance of SparsityPattern::insert
940 - Replace enums with strings for linear algebra and built-in meshes
941 - Allow direct access to Constant value
942 - Initialize entities in MeshEntity constructor automatically and
943 check range
944 - Add unit tests to the memorycheck
945 - Add call to clean up libxml2 parser at exit
946 - Remove unecessary arguments in DofMap member functions
947 - Remove reference constructors from DofMap, FiniteElement and
948 FunctionSpace
949 - Use a shared_ptr to store the mesh in DofMap objects
950 - Interface change for wrapper code: PoissonBilinearForm -->
951 Poisson::BilinearForm
952 - Add function info_underline() for writing underlined messages
953 - Rename message() --> info() for "compatibility" with Python logging
954 module
955 - Add elementwise multiplication in GeneriVector interface
956 - GenericVector interface in PyDOLFIN now support the sequence
957 protocol
958 - Rename of camelCaps functions names: fooBar --> foo_bar Note:
959 mesh.numVertices() --> mesh.num_vertices(), mesh.numCells() -->
960 mesh.num_cells()
961 - Add slicing capabilities for GenericMatrix interface in PyDOLFIN
962 (only getitem)
963 - Add slicing capabilities for GenericVector interface in PyDOLFIN
964 - Add sum to GenericVector interface
965
966
967 0.9.2 [2009-04-07]
968 ------------------
969 - Enable setting parameters for Newton solver in VariationalProblem
970 - Simplified and improved implementation of C++ plotting, calling
971 Viper on command-line
972 - Remove precompiled elements and projections
973 - Automatically interpolate user-defined functions on assignment
974 - Add new built-in function MeshCoordinates, useful in ALE simulations
975 - Add new constructor to Function class, Function(V, "vector.xml")
976 - Remove class Array (using std::vector instead)
977 - Add vector_mapping data to MeshData
978 - Use std::vector instead of Array in MeshData
979 - Add assignment operator and copy constructor for MeshFunction
980 - Add function mesh.move(other_mesh) for moving mesh according to
981 matching mesh (for FSI)
982 - Add function mesh.move(u) for moving mesh according to displacement
983 function (for FSI)
984 - Add macro dolfin_not_implemented()
985 - Add new interpolate() function for interpolation of user-defined
986 function to discrete
987 - Make _function_space protected in Function
988 - Added access to crs data from python for uBLAS and MTL4 backend
989
990
991 0.9.1 [2009-02-17]
992 ------------------
993 - Check Rectangle and Box for non-zero dimensions
994 - ODE solvers now solve the dual problem
995 - New class SubMesh for simple extraction of matching meshes for sub
996 domains
997 - Improvements of multiprecision ODE solver
998 - Fix Function class copy constructor
999 - Bug fixes for errornorm(), updates for new interface
1000 - Interface update for MeshData: createMeshFunction -->
1001 create_mesh_function etc
1002 - Interface update for Rectangle and Box
1003 - Add elastodynamics demo
1004 - Fix memory leak in IntersectionDetector/GTSInterface
1005 - Add check for swig version, in jit and compile functions
1006 - Bug fix in dolfin-order script for gzipped files
1007 - Make shared_ptr work across C++/Python interface
1008 - Replace std::tr1::shared_ptr with boost::shared_ptr
1009 - Bug fix in transfinite mean-value interpolation
1010 - Less annoying progress bar (silent when progress is fast)
1011 - Fix assignment operator for MeshData
1012 - Improved adaptive mesh refinement (recursive Rivara) producing
1013 better quality meshes
1014
1015 0.9.0 [2009-01-05]
1016 ------------------
1017 - Cross-platform fixes
1018 - PETScMatrix::copy fix
1019 - Some Trilinos fixes
1020 - Improvements in MeshData class
1021 - Do not use initial guess in Newton solver
1022 - Change OutflowFacet to IsOutflowFacet and change syntax
1023 - Used shared_ptr for underling linear algebra objects
1024 - Cache subspaces in FunctionSpace
1025 - Improved plotting, now support plot(grad(u)), plot(div(u)) etc
1026 - Simple handling of JIT-compiled functions
1027 - Sign change (bug fix) in increment for Newton solver
1028 - New class VariationalProblem replacing LinearPDE and NonlinearPDE
1029 - Parallel parsing and partitioning of meshes (experimental)
1030 - Add script dolfin-order for ordering mesh files
1031 - Add new class SubSpace (replacing SubSystem)
1032 - Add new class FunctionSpace
1033 - Complete redesign of Function class hierarchy, now a single Function
1034 class
1035 - Increased use of shared_ptr in Function, FunctionSpace, etc
1036 - New interface for boundary conditions, form not necessary
1037 - Allow simple setting of coefficient functions based on names (not
1038 their index)
1039 - Don't order mesh automatically, meshes must now be ordered
1040 explicitly
1041 - Simpler definition of user-defined functions (constructors not
1042 necessary)
1043 - Make mesh iterators const to allow for const-correct Mesh code
1044
1045
1046 0.8.1 [2008-10-20]
1047 ------------------
1048 - Add option to use ML multigrid preconditioner through PETSc
1049 - Interface change for ODE solvers: uBLASVector --> double*
1050 - Remove homotopy solver
1051 - Remove typedef real, now using plain double instead
1052 - Add various operators -=, += to GenericMatrix
1053 - Don't use -Werror when compiling SWIG generated code
1054 - Remove init(n) and init(m, n) from GenericVector/Matrix. Use resize
1055 and zero instead
1056 - Add new function is_combatible() for checking compatibility of
1057 boundary conditions
1058 - Use x as initial guess in Krylov solvers (PETSc, uBLAS, ITL)
1059 - Add new function errornorm()
1060 - Add harmonic ALE mesh smoothing
1061 - Refinements of Graph class
1062 - Add CholmodCholeskySlover (direct solver for symmetric matrices)
1063 - Implement application of Dirichlet boundary conditions within
1064 assembly loop
1065 - Improve efficiency of SparsityPattern
1066 - Allow a variable number of smoothings
1067 - Add class Table for pretty-printing of tables
1068 - Add experimental MTL4 linear algebra backend
1069 - Add OutflowFacet to SpecialFunctions for DG transport problems
1070 - Remove unmaintained OpenDX file format
1071 - Fix problem with mesh smoothing near nonconvex corners
1072 - Simple projection of functions in Python
1073 - Add file format: XYZ for use with Xd3d
1074 - Add built-in meshes: UnitCircle, Box, Rectangle, UnitSphere
1075
1076
1077 0.8.0 [2008-06-23]
1078 ------------------
1079 - Fix input of matrix data from XML
1080 - Add function normalize()
1081 - Integration with VMTK for reading DOLFIN XML meshes produced by VMTK
1082 - Extend mesh XML format to handle boundary indicators
1083 - Add support for attaching arbitrarily named data to meshes
1084 - Add support for dynamically choosing the linear algebra backend
1085 - Add Epetra/Trilinos linear solvers
1086 - Add setrow() to matrix interface
1087 - Add new solver SingularSolver for solving singular (pressure)
1088 systems
1089 - Add MeshSize::min(), max() for easy computation of smallest/largest
1090 mesh size
1091 - LinearSolver now handles all backends and linear solvers
1092 - Add access to normal in Function, useful for inflow boundary
1093 conditions
1094 - Remove GMRES and LU classes, use solve() instead
1095 - Improve solve() function, now handles both LU and Krylov +
1096 preconditioners
1097 - Add ALE mesh interpolation (moving mesh according to new boundary
1098 coordinates)
1099
1100
1101 0.7.3 [2008-04-30]
1102 ------------------
1103 - Add support for Epetra/Trilinos
1104 - Bug fix for order of values in interpolate_vertex_values, now
1105 according to UFC
1106 - Boundary meshes are now always oriented with respect to outward
1107 facet normals
1108 - Improved linear algebra, both in C++ and Python
1109 - Make periodic boundary conditions work in Python
1110 - Fix saving of user-defined functions
1111 - Improve plotting
1112 - Simple computation of various norms of functions from Python
1113 - Evaluation of Functions at arbitrary points in a mesh
1114 - Fix bug in assembling over exterior facets (subdomains were ignored)
1115 - Make progress bar less annoying
1116 - New scons-based build system replaces autotools
1117 - Fix bug when choosing iterative solver from Python
1118
1119
1120 0.7.2 [2008-02-18]
1121 ------------------
1122 - Improve sparsity pattern generator efficiency
1123 - Dimension-independent sparsity pattern generator
1124 - Add support for setting strong boundary values for DG elements
1125 - Add option setting boundary conditions based on geometrical search
1126 - Check UMFPACK return argument for warnings/errors
1127 - Simplify setting simple Dirichlet boundary conditions
1128 - Much improved integration with FFC in PyDOLFIN
1129 - Caching of forms by JIT compiler now works
1130 - Updates for UFC 1.1
1131 - Catch exceptions in PyDOLFIN
1132 - Work on linear algebra interfaces GenericTensor/Matrix/Vector
1133 - Add linear algebra factory (backend) interface
1134 - Add support for 1D meshes
1135 - Make Assembler independent of linear algebra backend
1136 - Add manager for handling sub systems (PETSc and MPI)
1137 - Add parallel broadcast of Mesh and MeshFunction
1138 - Add experimental support for parallel assembly
1139 - Use PETSc MPI matrices when running in parallel
1140 - Add predefined functions FacetNormal and AvgMeshSize
1141 - Add left/right/crisscross options for UnitSquare
1142 - Add more Python demos
1143 - Add support for Exodus II format in dolfin-convert
1144 - Autogenerate docstrings for PyDOLFIN
1145 - Various small bug fixes and improvements
1146
1147
1148 0.7.1 [2007-08-31]
1149 ------------------
1150 - Integrate FFC form language into PyDOLFIN
1151 - Just-in-time (JIT) compilation of variational forms
1152 - Conversion from from Diffpack grid format to DOLFIN XML
1153 - Name change: BoundaryCondition --> DirichletBC
1154 - Add support for periodic boundary conditions: class PeriodicBC
1155 - Redesign default linear algebra interface (Matrix, Vector,
1156 KrylovSolver, etc)
1157 - Add function to return Vector associated with a DiscreteFunction
1158
1159
1160 0.7.0-1 [2007-06-22]
1161 --------------------
1162 - Recompile all forms with latest FFC release
1163 - Remove typedefs SparseMatrix and SparseVector
1164 - Fix includes in LinearPDE
1165 - Rename DofMaps -> DofMapSet
1166
1167
1168 0.7.0 [2007-06-20]
1169 ------------------
1170 - Move to UFC interface for code generation
1171 - Major rewrite, restructure, cleanup
1172 - Add support for Brezzi-Douglas-Marini (BDM) elements
1173 - Add support for Raviart-Thomas (RT) elements
1174 - Add support for Discontinuous Galerkin (DG) methods
1175 - Add support for mesh partitioning (through SCOTCH)
1176 - Handle both UMFPACK and UFSPARSE
1177 - Local mesh refinement
1178 - Mesh smoothing
1179 - Built-in plotting (through Viper)
1180 - Cleanup log system
1181 - Numerous fixes for mesh, in particular MeshFunction
1182 - Much improved Python bindings for mesh
1183 - Fix Python interface for vertex and cell maps in boundary
1184 computation
1185
1186
1187 0.6.4 [2006-12-01]
1188 ------------------
1189 - Switch from Python Numeric to Python NumPy
1190 - Improved mesh Python bindings
1191 - Add input/output support for MeshFunction
1192 - Change Mesh::vertices() --> Mesh::coordinates()
1193 - Fix bug in output of mesh to MATLAB format
1194 - Add plasticty module (experimental)
1195 - Fix configure test for Python dev (patch from Åsmund Ødegård)
1196 - Add mesh benchmark
1197 - Fix memory leak in mesh (data not deleted correctly in MeshTopology)
1198 - Fix detection of curses libraries
1199 - Remove Tecplot output format
1200
1201
1202 0.6.3 [2006-10-27]
1203 ------------------
1204 - Move to new mesh library
1205 - Remove dolfin-config and move to pkg-config
1206 - Remove unused classes PArray, PList, Table, Tensor
1207 - Visualization of 2D solutions in OpenDX is now supported (3D
1208 supported before)
1209 - Add support for evaluation of functionals
1210 - Fix bug in Vector::sum() for uBLAS vectors
1211
1212
1213 0.6.2-1 [2006-09-06]
1214 --------------------
1215 - Fix compilation error when using --enable-petsc
1216 (dolfin::uBLASVector::PETScVector undefined)
1217
1218
1219 0.6.2 [2006-09-05]
1220 ------------------
1221 - Finish chapter in manual on linear algebra
1222 - Enable PyDOLFIN by default, use --disable-pydolfin to disable
1223 - Disable PETSc by default, use --enable-petsc to enable
1224 - Modify ODE solver interface for u0() and f()
1225 - Add class ConvectionMatrix
1226 - Readd classes LoadVector, MassMatrix, StiffnessMatrix
1227 - Add matrix factory for simple creation of standard finite element
1228 matrices
1229 - Collect static solvers in LU and GMRES
1230 - Bug fixes for Python interface PyDOLFIN
1231 - Enable use of direct solver for ODE solver (experimental)
1232 - Remove demo bistable
1233 - Restructure and cleanup linear algebra
1234 - Use UMFPACK for LU solver with uBLAS matrix types
1235 - Add templated wrapper class for different uBLAS matrix types
1236 - Add ILU preconditioning for uBLAS matrices
1237 - Add Krylov solver for uBLAS sparse matrices (GMRES and BICGSTAB)
1238 - Add first version of new mesh library (NewMesh, experimental)
1239 - Add Parametrized::readParameters() to trigger reading of values on
1240 set()
1241 - Remove output of zeros in Octave matrix file format
1242 - Use uBLAS-based vector for Vector if PETSc disabled
1243 - Add wrappers for uBLAS compressed_matrix class
1244 - Compute eigenvalues using SLEPc (an extension of PETSc)
1245 - Clean up assembly and linear algebra
1246 - Add function to solve Ax = b for dense matrices and dense vectors
1247 - Make it possible to compile without PETSc (--disable-petsc)
1248 - Much improved ODE solvers
1249 - Complete multi-adaptive benchmarks reaction and wave
1250 - Assemble boundary integrals
1251 - FEM class cleaned up.
1252 - Fix multi-adaptive benchmark problem reaction
1253 - Small fixes for Intel C++ compiler version 9.1
1254 - Test for Intel C++ compiler and configure appropriately
1255 - Add new classes DenseMatrix and DenseVector (wrappers for ublas)
1256 - Fix bug in conversion from Gmsh format
1257
1258
1259 0.6.1 [2006-03-28]
1260 ------------------
1261 - Regenerate build system in makedist script
1262 - Update for new FFC syntax: BasisFunction --> TestFunction,
1263 TrialFunction
1264 - Fixes for conversion script dolfin-convert
1265 - Initial cleanups and fixes for ODE solvers
1266 - Numerous small fixes to improve portability
1267 - Remove dolfin:: qualifier on output << in Parameter.h
1268 - Don't use anonymous classes in demos, gives errors with some
1269 compilers
1270 - Remove KrylovSolver::solver()
1271 - Fix bug in convection-diffusion demo (boundary condition for
1272 pressure), use direct solver
1273 - LinearPDE and NewonSolver use umfpack LU solver by default (if
1274 available) when doing direct solve
1275 - Set PETSc matrix type through Matrix constructor
1276 - Allow linear solver and preconditioner type to be passed to
1277 NewtonSolver
1278 - Fix bug in Stokes demos (wrong boundary conditions)
1279 - Cleanup Krylov solver
1280 - Remove KrylovSolver::setPreconditioner() etc. and move to
1281 constructors
1282 - Remove KrylovSolver::setRtol() etc. and replace with parameters
1283 - Fix remaining name changes: noFoo() --> numFoo()
1284 - Add Cahn-Hilliard equation demo
1285 - NewtonSolver option to use residual or incremental convergence
1286 criterion
1287 - Add separate function to nls to test for convergence of Newton
1288 iterations
1289 - Fix bug in dolfin-config (wrong version number)
1290
1291
1292 0.6.0 [2006-03-01]
1293 ------------------
1294 - Fix bug in XML output format (writing multiple objects)
1295 - Fix bug in XML matrix output format (handle zero rows)
1296 - Add new nonlinear PDE demo
1297 - Restructure PDE class to use envelope-letter design
1298 - Add precompiled finite elements for q <= 5
1299 - Add FiniteElementSpec and factor function for FiniteElement
1300 - Add input/output of Function to DOLFIN XML
1301 - Name change: dof --> node
1302 - Name change: noFoo() --> numFoo()
1303 - Add conversion from gmsh format in dolfin-convert script
1304 - Updates for PETSc 2.3.1
1305 - Add new type of Function (constant)
1306 - Simplify use of Function class
1307 - Add new demo Stokes + convection-diffusion
1308 - Add new demo Stokes (equal-order stabilized)
1309 - Add new demo Stokes (Taylor-Hood)
1310 - Add new parameter for KrylovSolvers: "monitor convergence"
1311 - Add conversion script dolfin-convert for various mesh formats
1312 - Add new demo elasticity
1313 - Move poisson demo to src/demo/pde/poisson
1314 - Move to Mercurial (hg) from CVS
1315 - Use libtool to build libraries (including shared)
1316
1317 0.5.12 [2006-01-12]
1318 -------------------
1319 - Make Stokes solver dimension independent (2D/3D)
1320 - Make Poisson solver dimension independent (2D/3D)
1321 - Fix sparse matrix output format for MATLAB
1322 - Modify demo problem for Stokes, add exact solution and compute error
1323 - Change interface for boundary conditions: operator() --> eval()
1324 - Add two benchmark problems for the Navier-Stokes solver
1325 - Add support for 2D/3D selection in Navier-Stokes solver
1326 - Move tic()/toc() to timing.h
1327 - Navier-Stokes solver back online
1328 - Make Solver a subclass of Parametrized
1329 - Add support for localization of parameters
1330 - Redesign of parameter system
1331
1332 0.5.11 [2005-12-15]
1333 -------------------
1334 - Add script monitor for monitoring memory usage
1335 - Remove meminfo.h (not portable)
1336 - Remove dependence on parameter system in log system
1337 - Don't use drand48() (not portable)
1338 - Don't use strcasecmp() (not portable)
1339 - Remove sysinfo.h and class System (not portable)
1340 - Don't include <sys/utsname.h> (not portable)
1341 - Change ::show() --> ::disp() everywhere
1342 - Clean out old quadrature classes on triangles and tetrahedra
1343 - Clean out old sparse matrix code
1344 - Update chapter on Functions in manual
1345 - Use std::map to store parameters
1346 - Implement class KrylovSolver
1347 - Name change: Node --> Vertex
1348 - Add nonlinear solver demos
1349 - Add support for picking sub functions and components of functions
1350 - Update interface for FiniteElement for latest FFC version
1351 - Improve and restructure implementation of the Function class
1352 - Dynamically adjust safety factor during integration
1353 - Improve output Matrix::disp()
1354 - Check residual at end of time step, reject step if too large
1355 - Implement Vector::sum()
1356 - Implement nonlinear solver
1357 - New option for ODE solver: "save final solution" --> solution.data
1358 - New ODE test problem: reaction
1359 - Fixes for automake 1.9 (nobase_include_HEADERS)
1360 - Reorganize build system, remove fake install and require make
1361 install
1362 - Add checks for non-standard PETSc component HYPRE in NSE solver
1363 - Make GMRES solver return the number of iterations
1364 - Add installation script for Python interface
1365 - Add Matrix Market format (Haiko Etzel)
1366 - Automatically reinitialize GMRES solver when system size changes
1367 - Implement cout << for class Vector
1368
1369 0.5.10 [2005-10-11]
1370 -------------------
1371 - Modify ODE solver interface: add T to constructor
1372 - Fix compilation on AMD 64 bit systems (add -fPIC)
1373 - Add new BLAS mode for form evaluation
1374 - Change enum types in File to lowercase
1375 - Change default file type for .m to Octave
1376 - Add experimental Python interface PyDOLFIN
1377 - Fix compilation for gcc 4.0
1378
1379 0.5.9 [2005-09-23]
1380 ------------------
1381 - Add Stokes module
1382 - Support for arbitrary mixed elements through FFC
1383 - VTK output interface now handles time-dependent functions
1384 automatically
1385 - Fix cout for empty matrix
1386 - Change dolfin_start() --> dolfin_end()
1387 - Add chapters to manual: about, log system, parameters, reference
1388 elements, installation, contributing, license
1389 - Use new template fenicsmanual.cls for manual
1390 - Add compiler flag -U__STRICT_ANSI__ when compiling under Cygwin
1391 - Add class EigenvalueSolver
1392
1393 0.5.8 [2005-07-05]
1394 ------------------
1395 - Add new output format Paraview/VTK (Garth N. Wells)
1396 - Update Tecplot interface
1397 - Move to PETSc 2.3.0
1398 - Complete support for general order Lagrange elements in triangles
1399 and tetrahedra
1400 - Add test problem in src/demo/fem/convergence/ for general Lagrange
1401 elements
1402 - Make FEM::assemble() estimate the number of nonzeros in each row
1403 - Implement Matrix::init(M, N, nzmax)
1404 - Add Matrix::nz(), Matrix::nzsum() and Matrix::nzmax()
1405 - Improve Mesh::disp()
1406 - Add FiniteElement::disp() and FEM::disp() (useful for debugging)
1407 - Remove old class SparseMatrix
1408 - Change FEM::setBC() --> FEM::applyBC()
1409 - Change Mesh::tetrahedrons --> Mesh::tetrahedra
1410 - Implement Dirichlet boundary conditions for tetrahedra
1411 - Implement Face::contains(const Point& p)
1412 - Add test for shape dimension of mesh and form in FEM::assemble()
1413 - Move src/demo/fem/ demo to src/demo/fem/simple/
1414 - Add README file in src/demo/poisson/ (simple manual)
1415 - Add simple demo program src/demo/poisson/
1416 - Update computation of alignment of faces to match FFC/FIAT
1417
1418 0.5.7 [2005-06-23]
1419 ------------------
1420 - Clean up ODE test problems
1421 - Implement automatic detection of sparsity pattern from given matrix
1422 - Clean up homotopy solver
1423 - Implement automatic computation of Jacobian
1424 - Add support for assembly of non-square systems (Andy Terrel)
1425 - Make ODE solver report average number of iterations
1426 - Make progress bar write first update at 0%
1427 - Initialize all values of u before solution in multi-adaptive solver,
1428 not only components given by dependencies
1429 - Allow user to modify and verify a converging homotopy path
1430 - Make homotopy solver save a list of the solutions
1431 - Add Matrix::norm()
1432 - Add new test problem for CES economy
1433 - Remove cast from Parameter to const char* (use std::string)
1434 - Make solution data filename optional for homotopy solver
1435 - Append homotopy solution data to file during solution
1436 - Add dolfin::seed(int) for optionally seeding random number generator
1437 - Remove dolfin::max,min (use std::max,min)
1438 - Add polynomial-integer (true polynomial) form of general CES system
1439 - Compute multi-adaptive efficiency index
1440 - Updates for gcc 4.0 (patches by Garth N. Wells)
1441 - Add Matrix::mult(const real x[], uint row) (temporary fix, assumes
1442 uniprocessor case)
1443 - Add Matrix::mult(const Vector& x, uint row) (temporary fix, assumes
1444 uniprocessor case)
1445 - Update shortcuts MassMatrix and StiffnessMatrix to new system
1446 - Add missing friend to Face.h (reported by Garth N. Wells)
1447
1448 0.5.6 [2005-05-17]
1449 ------------------
1450 - Implementation of boundary conditions for general order Lagrange
1451 (experimental)
1452 - Use interpolation function automatically generated by FFC
1453 - Put computation of map into class AffineMap
1454 - Clean up assembly
1455 - Use dof maps automatically generated by FFC (experimental)
1456 - Modify interface FiniteElement for new version of FFC
1457 - Update ODE homotopy test problems
1458 - Add cross product to class Point
1459 - Sort mesh entities locally according to ordering used by FIAT and
1460 FFC
1461 - Add new format for dof maps (preparation for higher-order elements)
1462 - Code cleanups: NewFoo --> Foo complete
1463 - Updates for new version of FFC (0.1.7)
1464 - Bypass log system when finalizing PETSc (may be out of scope)
1465
1466 0.5.5 [2005-04-26]
1467 ------------------
1468 - Fix broken log system, curses works again
1469 - Much improved multi-adaptive time-stepping
1470 - Move elasticity module to new system based on FFC
1471 - Add boundary conditions for systems
1472 - Improve regulation of time steps
1473 - Clean out old assembly classes
1474 - Clean out old form classes
1475 - Remove kernel module map
1476 - Remove kernel module element
1477 - Move convection-diffusion module to new system based on FFC
1478 - Add iterators for cell neighbors of edges and faces
1479 - Implement polynomial for of CES economy
1480 - Rename all new linear algebra classes: NewFoo --> Foo
1481 - Clean out old linear algebra
1482 - Speedup setting of boundary conditions (add MAT_KEEP_ZEROED_ROWS)
1483 - Fix bug for option --disable-curses
1484
1485 0.5.4 [2005-03-29]
1486 ------------------
1487 - Remove option to compile with PETSc 2.2.0 (2.2.1 required)
1488 - Make make install work again (fix missing includes)
1489 - Add support for mixing multiple finite elements (through FFC)
1490 - Improve functionality of homotopy solver
1491 - Simple creation of piecewise linear functions (without having an
1492 element)
1493 - Simple creation of piecewise linear elements
1494 - Add support of automatic creation of simple meshes (unit cube, unit
1495 square)
1496
1497 0.5.3 [2005-02-26]
1498 ------------------
1499 - Change to PETSc version 2.2.1
1500 - Add flag --with-petsc=<path> to configure script
1501 - Move Poisson's equation to system based on FFC
1502 - Add support for automatic creation of homotopies
1503 - Make all ODE solvers automatically handle complex ODEs: (M) z' =
1504 f(z,t)
1505 - Implement version of mono-adaptive solver for implicit ODEs: M u' =
1506 f(u,t)
1507 - Implement Newton's method for multi- and mono-adaptive ODE solvers
1508 - Update PETSc wrappers NewVector, NewMatrix, and NewGMRES
1509 - Fix initialization of PETSc
1510 - Add mono-adaptive cG(q) and dG(q) solvers (experimental)
1511 - Implementation of new assebly: NewFEM, using output from FFC
1512 - Add access to mesh for nodes, cells, faces and edges
1513 - Add Tecplot I/O interface; contributed by Garth N. Wells
1514
1515 0.5.2 [2005-01-26]
1516 ------------------
1517 - Benchmarks for DOLFIN vs PETSc (src/demo/form and src/demo/test)
1518 - Complete rewrite of the multi-adaptive ODE solver (experimental)
1519 - Add wrapper for PETSc GMRES solver
1520 - Update class Point with new operators
1521 - Complete rewrite of the multi-adaptive solver to improve performance
1522 - Add PETSc wrappers NewMatrix and NewVector
1523 - Add DOLFIN/PETSc benchmarks
1524
1525 0.5.1 [2004-11-10]
1526 ------------------
1527 - Experimental support for automatic generation of forms using FFC
1528 - Allow user to supply Jacobian to ODE solver
1529 - Add optional test to check if a dependency already exists (Sparsity)
1530 - Modify sparse matrix output (Matrix::show())
1531 - Add FGMRES solver in new format (patch from eriksv)
1532 - Add non-const version of quick-access of sparse matrices
1533 - Add linear mappings for simple computation of derivatives
1534 - Add check of matrix dimensions for ODE sparsity pattern
1535 - Include missing cmath in Function.cpp
1536
1537 0.5.0 [2004-08-18]
1538 ------------------
1539 - First prototype of new form evaluation system
1540 - New classes Jacobi, SOR, Richardson (preconditioners and linear
1541 solvers)
1542 - Add integrals on the boundary (ds), partly working
1543 - Add maps from boundary of reference cell
1544 - Add evaluation of map from reference cell
1545 - New Matrix functions: max, min, norm, and sum of rows and columns
1546 (erik)
1547 - Derivatives/gradients of ElementFunction (coefficients f.ex.)
1548 implemented
1549 - Enable assignment to all elements of a NewArray
1550 - Add functions Boundary::noNodes(), noFaces(), noEdges()
1551 - New class GaussSeidel (preconditioner and linear solver)
1552 - New classes Preconditioner and LinearSolver
1553 - Bug fix for tetrahedral mesh refinement (ingelstrom)
1554 - Add iterators for Edge and Face on Boundary
1555 - Add functionality to Map: bdet() and cell()
1556 - Add connectivity face-cell and edge-cell
1557 - New interface for assembly: Galerkin --> FEM
1558 - Bug fix for PDE systems of size > 3
1559
1560 0.4.11 [2004-04-23]
1561 -------------------
1562 - Add multigrid solver (experimental)
1563 - Update manual
1564
1565 0.4.10
1566 ------
1567 - Automatic model reduction (experimental)
1568 - Fix bug in ParticleSystem (divide by mass)
1569 - Improve control of integration (add function ODE::update())
1570 - Load/save parameters in XML-format
1571 - Add assembly test
1572 - Add simple StiffnessMatrix, MassMatrix, and LoadVector
1573 - Change dK --> dx
1574 - Change dx() --> ddx()
1575 - Add support for GiD file format
1576 - Add performance tests for multi-adaptivity (both stiff and
1577 non-stiff)
1578 - First version of Newton for the multi-adaptive solver
1579 - Test for Newton for the multi-adaptive solver
1580
1581 0.4.9
1582 -----
1583 - Add multi-adaptive solver for the bistable equation
1584 - Add BiCGSTAB solver (thsv)
1585 - Fix bug in SOR (thsv)
1586 - Improved visual program for OpenDX
1587 - Fix OpenDX file format for scalar functions
1588 - Allow access to samples of multi-adaptive solution
1589 - New patch from thsv for gcc 3.4.0 and 3.5.0
1590 - Make progress step a parameter
1591 - New function ODE::sparse(const Matrix& A)
1592 - Access nodes, cells, edges, faces by id
1593 - New function Matrix::lump()
1594
1595 0.4.8
1596 -----
1597 - Add support for systems (jansson and bengzon)
1598 - Add new module wave
1599 - Add new module wave-vector
1600 - Add new module elasticity
1601 - Add new module elasticity-stationary
1602 - Multi-adaptive updates
1603 - Fix compilation error in LogStream
1604 - Fix local Newton iteration for higher order elements
1605 - Init matrix to given type
1606 - Add output of cG(q) and dG(q) weights in matrix format
1607 - Fix numbering of frames from plotslab script
1608 - Add png output for plotslab script
1609 - Add script for running stiff test problems, plot solutions
1610 - Fix bug in MeshInit (node neighbors of node)
1611 - Modify output of sysinfo()
1612 - Compile with -Wall -Werror -pedantic -ansi -std=c++98 (thsv)
1613
1614 0.4.7
1615 -----
1616 - Make all stiff test problems work
1617 - Display status report also when using step()
1618 - Improve adaptive damping for stiff problems (remove spikes)
1619 - Modify Octave/Matlab format for solution data (speed improvement)
1620 - Adaptive sampling of solution (optional)
1621 - Restructure stiff test problems
1622 - Check if value of right-hand side is valid
1623 - Modify divergence test in AdaptiveIterationLevel1
1624
1625 0.4.6
1626 -----
1627 - Save vectors and matrices from Matlab/Octave (foufas)
1628 - Rename writexml.m to xmlmesh.m
1629 - Inlining of important functions
1630 - Optimize evaluation of elements
1631 - Optimize Lagrange polynomials
1632 - Optimize sparsity: use stl containers
1633 - Optimize choice of discrete residual for multi-adaptive solver
1634 - Don't save solution in benchmark proble
1635 - Improve computation of divergence factor for underdamped systems
1636 - Don't check residual on first slab for fixed time step
1637 - Decrease largest (default) time step to 0.1
1638 - Add missing <cmath> in TimeStepper
1639 - Move real into dolfin namespace
1640
1641 0.4.5
1642 -----
1643 - Rename function.h to enable compilation under Cygwin
1644 - Add new benchmark problem for multi-adaptive solver
1645 - Bug fix for ParticleSystem
1646 - Initialization of first time step
1647 - Improve time step regulation (threshold)
1648 - Improve stabilization
1649 - Improve TimeStepper interface (Ko Project)
1650 - Use iterators instead of recursively calling TimeSlab::update()
1651 - Clean up ODESolver
1652 - Add iterators for elements in time slabs and element groups
1653 - Add -f to creation of symbolic links
1654
1655 0.4.4
1656 -----
1657 - Add support for 3D graphics in Octave using Open Inventor (jj)
1658
1659 0.4.3
1660 -----
1661 - Stabilization of multi-adaptive solver (experimental)
1662 - Improved non-support for curses (--disable-curses)
1663 - New class MechanicalSystem for simulating mechanical systems
1664 - Save debug info from primal and dual (plotslab.m)
1665 - Fix bug in progress bar
1666 - Add missing include file in Components.h (kakr)
1667 - New function dolfin_end(const char* msg, ...)
1668 - Move numerical differentiation to RHS
1669 - New class Event for limited display of messages
1670 - Fix bug in LogStream (large numbers in floating point format)
1671 - Specify individual time steps for different components
1672 - Compile without warnings
1673 - Add -Werror to option enable-debug
1674 - Specify individual methods for different components
1675 - Fix bug in dGqMethods
1676 - Fix bug (delete old block) in ElementData
1677 - Add parameters for method and order
1678 - New test problem reaction
1679 - New class FixedPointIteration
1680 - Fix bug in grid refinement
1681
1682 0.4.2
1683 -----
1684 - Fix bug in computation of residual (divide by k)
1685 - Add automatic generation and solution of the dual problem
1686 - Automatic selection of file names for primal and dual
1687 - Fix bug in progress bar (TerminalLogger)
1688 - Many updates of multi-adaptive solver
1689 - Add class ODEFunction
1690 - Update function class hierarchies
1691 - Move functions to a separate directory
1692 - Store multi-adaptive solution binary on disk with cache
1693
1694 0.4.1
1695 -----
1696 - First version of multi-adaptive solver working
1697 - Clean up file formats
1698 - Start changing from int to unsigned int where necessary
1699 - Fix bool->int when using stdard in Parameter
1700 - Add NewArray and NewList (will replace Array and List)
1701
1702 0.4.0
1703 -----
1704 - Initiation of the FEniCS project
1705 - Change syntax of mesh files: grid -> mesh
1706 - Create symbolic links instead of copying files
1707 - Tanganyika -> ODE
1708 - Add Heat module
1709 - Grid -> Mesh
1710 - Move forms and mappings to separate libraries
1711 - Fix missing include of DirectSolver.h
1712
1713 0.3.12
1714 ------
1715 - Adaptive grid refinement (!)
1716 - Add User Manual
1717 - Add function dolfin_log() to turn logging on/off
1718 - Change from pointers to references for Node, Cell, Edge, Face
1719 - Update writexml.m
1720 - Add new grid files and rename old grid files
1721
1722 0.3.11
1723 ------
1724 - Add configure option --disable-curses
1725 - Grid refinement updates
1726 - Make OpenDX file format work for grids (output)
1727 - Add volume() and diameter() in cell
1728 - New classes TriGridRefinement and TetGridRefinement
1729 - Add iterators for faces and edges on a boundary
1730 - New class GridHierarchy
1731
1732 0.3.10
1733 ------
1734 - Use new boundary structure in Galerkin
1735 - Make dolfin_start() and dolfin_end() work
1736 - Make dolfin_assert() raise segmentation fault for plain text mode
1737 - Add configure option --enable-debug
1738 - Use autoreconf instead of scripts/preconfigure
1739 - Rename configure.in -> configure.ac
1740 - New class FaceIterator
1741 - New class Face
1742 - Move computation of boundary from GridInit to BoundaryInit
1743 - New class BoundaryData
1744 - New class BoundaryInit
1745 - New class Boundary
1746 - Make InitGrid compute edges
1747 - Add test program for generic matrix in src/demo/la
1748 - Clean up Grid classes
1749 - Add new class GridRefinementData
1750 - Move data from Cell to GenericCell
1751 - Make GMRES work with user defined matrix, only mult() needed
1752 - GMRES now uses only one function to compute residual()
1753 - Change Matrix structure (a modified envelope/letter)
1754 - Update script checkerror.m for Poisson
1755 - Add function dolfin_info_aptr()
1756 - Add cast to element pointer for iterators
1757 - Clean up and improve the Tensor class
1758 - New class: List
1759 - Name change: List -> Table
1760 - Name change: ShortList -> Array
1761 - Make functions in GridRefinement static
1762 - Make functions in GridInit static
1763 - Fix bug in GridInit (eriksv)
1764 - Add output to OpenDX format for 3D grids
1765 - Clean up ShortList class
1766 - Clean up List class
1767 - New class ODE, Equation replaced by PDE
1768 - Add Lorenz test problem
1769 - Add new problem type for ODEs
1770 - Add new module ode
1771 - Work on multi-adaptive ODE solver (lots of new stuff)
1772 - Work on grid refinement
1773 - Write all macros in LoggerMacros in one line
1774 - Add transpose functions to Matrix (Erik)
1775
1776 0.3.9
1777 -----
1778 - Update Krylov solver (Erik, Johan)
1779 - Add new LU factorization and LU solve (Niklas)
1780 - Add benchmark test in src/demo/bench
1781 - Add silent logger
1782
1783 0.3.8
1784 -----
1785 - Make sure dolfin-config is regenerated every time
1786 - Add demo program for cG(q) and dG(q)
1787 - Add dG(q) precalc of nodal points and weights
1788 - Add cG(q) precalc of nodal points and weights
1789 - Fix a bug in configure.in (AC_INIT with README)
1790 - Add Lagrange polynomials
1791 - Add multiplication with transpose
1792 - Add scalar products with rows and columns
1793 - Add A[i][j] index operator for quick access to dense matrix
1794
1795 0.3.7
1796 -----
1797 - Add new Matlab-like syntax like A(i,all) = x or A(3,all) = A(4,all)
1798 - Add dolfin_assert() macro enabled if debug is defined
1799 - Redesign of Matrix/DenseMatrix/SparseMatrix to use Matrix as common
1800 interface
1801 - Include missing cmath in Legendre.cpp and GaussianQuadrature.cpp
1802
1803 0.3.6
1804 -----
1805 - Add output functionality in DenseMatrix
1806 - Add high precision solver to DirectSolver
1807 - Clean up error messages in Matrix
1808 - Make solvers directly accessible through Matrix and DenseMatrix
1809 - Add quadrature (Gauss, Radau, and Lobatto) from Tanganyika
1810 - Start merge with Tanganyika
1811 - Add support for automatic documentation using doxygen
1812 - Update configure scripts
1813 - Add greeting at end of compilation
1814
1815 0.3.5
1816 -----
1817 - Define version number only in the file configure.in
1818 - Fix compilation problem (missing depcomp)
1819
1820 0.3.4
1821 -----
1822 - Fix bugs in some of the ElementFunction operators
1823 - Make convection-diffusion solver work again
1824 - Fix bug in integration, move multiplication with the determinant
1825 - Fix memory leaks in ElementFunction
1826 - Add parameter to choose output format
1827 - Make OctaveFile and MatlabFile subclasses of MFile
1828 - Add classes ScalarExpressionFunction and VectorExpressionFunction
1829 - Make progress bars work cleaner
1830 - Get ctrl-c in curses logger
1831 - Remove <Problem>Settings-classes and use dolfin_parameter()
1832 - Redesign settings to match the structure of the log system
1833 - Add vector functions: Function::Vector
1834 - Add vector element functions: ElementFunction::Vector
1835
1836 0.3.3
1837 -----
1838 - Increased functionality of curses-based interface
1839 - Add progress bars to log system
1840
1841 0.3.2
1842 -----
1843 - More work on grid refinement
1844 - Add new curses based log system
1845
1846 0.3.1
1847 -----
1848 - Makefile updates: make install should now work properly
1849 - KrylovSolver updates
1850 - Preparation for grid refinement
1851 - Matrix and Vector updates
1852
1853 0.3.0
1854 -----
1855 - Make poisson work again, other modules still not working
1856 - Add output format for octave
1857 - Fix code to compile with g++-3.2 -Wall -Werror
1858 - New operators for Matrix
1859 - New and faster GMRES solver (speedup factor 4)
1860 - Changed name from SparseMatrix to Matrix
1861 - Remove old unused code
1862 - Add subdirectory math containing mathematical functions
1863 - Better access for A(i,j) += to improve speed in assembling
1864 - Add benchmark for linear algebra
1865 - New definition of finite element
1866 - Add algebra for function spaces
1867 - Convert grids in data/grids to xml.gz
1868 - Add iterators for Nodes and Cells
1869 - Change from .hh to .h
1870 - Add operators to Vector class (foufas)
1871 - Add dependence on libxml2
1872 - Change from .C to .cpp to make Jim happy.
1873 - Change input/output functionality to streams
1874 - Change to new data structure for Grid
1875 - Change to object-oriented API at top level
1876 - Add use of C++ namespaces
1877 - Complete and major restructuring of the code
1878 - Fix compilation error in src/config
1879 - Fix name of keyword for convection-diffusion
1880
1881 0.2.11-1
1882 --------
1883 - Fix compilation error (`source`) on Solaris
1884
1885 0.2.11
1886 ------
1887 - Automate build process to simplify addition of new modules
1888 - Fix bug in matlab_write_field() (walter)
1889 - Fix bug in SparseMatrix::GetCopy() (foufas)
1890
1891 0.2.10-1
1892 --------
1893 - Fix compilation errors on RedHat (thsv)
1894
1895 0.2.10
1896 ------
1897 - Fix compilation of problems to use correct compiler
1898 - Change default test problems to the ones in the report
1899 - Improve memory management using mpatrol for tracking allocations
1900 - Change bool to int for va_arg, seems to be a problem with gcc > 3.0
1901 - Improve input / output support: GiD, Matlab, OpenDX
1902
1903 0.2.8
1904 -----
1905 - Navier-Stokes starting to work again
1906 - Add Navier-Stokes 2d
1907 - Bug fixes
1908
1909 0.2.7
1910 -----
1911 - Add support for 2D problems
1912 - Add module convection-diffusion
1913 - Add local/global fields in equation/problem
1914 - Bug fixes
1915 - Navier-Stokes updates (still broken)
1916
1917 0.2.6 [2002-02-19]
1918 ------------------
1919 - Navier-Stokes updates (still broken)
1920 - Output to matlab format
1921
1922 0.2.5
1923 -----
1924 - Add variational formulation with overloaded operators for systems
1925 - ShapeFunction/LocalField/FiniteElement according to Scott & Brenner
1926
1927 0.2.4
1928 -----
1929 - Add boundary conditions
1930 - Poisson seems to work ok
1931
1932 0.2.3
1933 -----
1934 - Add GMRES solver
1935 - Add CG solver
1936 - Add direct solver
1937 - Add Poisson solver
1938 - Big changes to the organisation of the source tree
1939 - Add kwdist.sh script
1940 - Bug fixes
1941
1942 0.2.2:
1943 ------
1944 - Remove curses temporarily
1945
1946 0.2.1:
1947 ------
1948 - Remove all PETSc stuff. Finally!
1949 - Gauss-Seidel cannot handle the pressure equation
1950
1951 0.2.0:
1952 ------
1953 - First GPL release
1954 - Remove all of Klas Samuelssons proprietary grid code
1955 - Adaptivity and refinement broken, include in next release