2009-11-04 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-11-02

2009-11-03 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated spurious currents test cases

2009-11-02 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated advection test case

2009-11-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Metrics descend from GfsVariableMetric

2009-11-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Macros only work with gawk (i.e. not mawk)

2009-10-29 Stephane Popinet <popinet at users.sf.net>
  * New 'cosine bell' test case

2009-10-29 Stephane Popinet <popinet at users.sf.net>
  * OutputScalarSum, OutputScalarNorm etc... take metric into account

2009-10-29 Stephane Popinet <popinet at users.sf.net>
  * Godunov advection scheme takes metric into account

2009-08-07 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Update of the strouhal testcase

2009-10-28 Stephane Popinet <popinet at users.sf.net>
  * coarse->fine and fine->coarse interpolations take metric into account
  
  in order to conserve quantities.

2009-10-28 Stephane Popinet <popinet at users.sf.net>
  * MetricCubed takes an optional "level" parameter

2009-10-27 Stephane Popinet <popinet at users.sf.net>
  * Extended "Cubed" metric to entire "expanded spherical cube"

2009-10-25 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-10-24

2009-10-24 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases

2009-10-23 Stephane Popinet <popinet at users.sf.net>
  * gfs_domain_cfl() takes metric into account

2009-10-16 Stephane Popinet <popinet at users.sf.net>
  * Optimisation of GfsMetricLonLat and GfsMetricCubed

2009-10-14 Stephane Popinet <popinet at users.sf.net>
  * "Cubed sphere" metric

2009-10-13 Stephane Popinet <popinet at users.sf.net>
  * Adaptivity takes "minlevel" into account properly

2009-10-08 Stephane Popinet <popinet at users.sf.net>
  * New object GfsMapFunction

2009-10-07 Stephane Popinet <popinet at users.sf.net>
  * CFL calculation for GfsRiver takes metric into account

2009-10-07 Stephane Popinet <popinet at users.sf.net>
  * Hydrostatic balance for 2nd-order curvilinear scheme in GfsRiver

2009-09-29 Stephane Popinet <popinet at users.sf.net>
  * General Orthogonal Coordinates

2009-09-25 Stephane Popinet <popinet at users.sf.net>
  * Longitude-latitude metric

2009-10-22 Stephane Popinet <popinet at users.sf.net>
  * New object GfsVariableStreamFunction

2009-10-21 Stephane Popinet <popinet at users.sf.net>
  * Support for "rotated" periodic boundaries

2009-10-20 Stephane Popinet <popinet at users.sf.net>
  * New "relative" option for GfsOutputErrorNorm

2009-10-21 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-10-19

2009-10-20 Stephane Popinet <popinet at users.sf.net>
  * Fixed missing header in OutputSimulation

2009-09-19 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-09-18

2009-09-18 Stephane Popinet <popinet at users.sf.net>
  * Support for "preloaded" modules

2009-09-17 Stephane Popinet <popinet at users.sf.net>
  * Units of VariableTerrain are properly reset

2009-09-17 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-09-16

2009-09-16 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for CGD/GTS files and multiple GfsFunctions

2009-09-16 Matthieu Castellazzi <m.castellazzi at niwa.co.nz>
  * Some fixes to the rpm spec file

2009-09-10 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-09-08

2009-09-09 Stephane Popinet <popinet at users.sf.net>
  * RPM includes man pages

2009-09-04 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-09-02

2009-09-03 Stephane Popinet <popinet at users.sf.net>
  * Cosmetics

2009-09-02 Stephane Popinet <popinet at users.sf.net>
  * New object GfsSourceControlField

2009-08-27 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-08-19

2009-08-20 Stephane Popinet <popinet at users.sf.net>
  * New object GfsSourceFlux

2009-08-26 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for gfs_streamline_new()

2009-08-26 Stephane Popinet <popinet at users.sf.net>
  * Cosmetics for wave model

2009-08-26 Stephane Popinet <popinet at users.sf.net>
  * BcSubcritical now tolerates negative depths

2009-08-20 Stephane Popinet <popinet at users.sf.net>
  * H and P terrain reconstruction for GfsVariableTerrain is optional

2009-08-20 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for gfs_function_expression()
  
  When specifying two GfsFunctions in a row.

2009-08-20 Stephane Popinet <popinet at users.sf.net>
  * GfsRefineTerrain defines two new variables "dmin" and "dmax"

2009-08-06 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-08-05

2009-08-05 Stephane Popinet <popinet at users.sf.net>
  * Added stability correction for wind wave model

2009-08-05 Stephane Popinet <popinet at users.sf.net>
  * GSE alleviation for GfsWave wind wave model

2009-07-30 Stephane Popinet <popinet at users.sf.net>
  * Removed obsolete GfsAdaptCurvature

2009-07-30 Stephane Popinet <popinet at users.sf.net>
  * Important improvements to GfsAdaptError

2009-07-30 Stephane Popinet <popinet at users.sf.net>
  * Minor optimisation of gfs_domain_solid_force()

2009-07-28 Stephane Popinet <popinet at users.sf.net>
  * Fixed negative wave actions in wavewatch module

2009-07-27 Stephane Popinet <popinet at users.sf.net>
  * GfsOutputSolidForce can take an optional weight argument

2009-07-27 Stephane Popinet <popinet at users.sf.net>
  * Fix for incorrect maxtag MPI value

2009-07-24 Stephane Popinet <popinet at users.sf.net>
  * Added option to use "bubble partitioning"

2009-07-24 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed debian changelog generation

2009-07-24 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-07-21

2009-07-22 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Fix for moving second order bug with merged cells

2009-07-23 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_traverse_and_bc()
  
  For overlapping computation and communications.

2009-07-23 Stephane Popinet <popinet at users.sf.net>
  * New interface for gfs_domain_locate() and gfs_domain_boundary_locate()

2009-07-22 Stephane Popinet <popinet at users.sf.net>
  * Overlapping communications and computations
  
  Only using gfs_traverse_and_homogeneous_bc() for now.

2009-07-22 Stephane Popinet <popinet at users.sf.net>
  * Daily script uses the parallel version

2009-07-22 Stephane Popinet <popinet at users.sf.net>
  * Improved parallel example graph

2009-07-14 Stephane Popinet <popinet at users.sf.net>
  * New parallel vortex street example

2009-07-21 Stephane Popinet <popinet at users.sf.net>
  * Simplified divergence redistribution of moving boundaries

2009-07-21 Stephane Popinet <popinet at users.sf.net>
  * Moving cylinder test case

2009-07-20 Stephane Popinet <popinet at users.sf.net>
  * Important bug fix for graph partitioning
  
  i.e. the "-p" option of gerris. Incorrect object casting between GtsWGnode and
  GfsBox were causing segfaults for some configurations.

2009-06-29 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Divergence redistribution between merged cells for moving boundaries

2009-06-25 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Fix for problem with small cells and moving solid interface in second 
  order method

2009-06-16 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Bug fix for dividing by zero in moving code

2009-07-19 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 2009-07-17

2009-07-18 Stephane Popinet <popinet at users.sf.net>
  * GfsEventBalance flows can transfer more than one box at a time

2009-07-17 Stephane Popinet <popinet at users.sf.net>
  * Various small fixes

2009-07-16 Stephane Popinet <popinet at users.sf.net>
  * gfs_domain_locate() uses a Cartesian array for fast indexing
  
  The previous systematic traversal of all the GfsBoxes was very slow 
  (when the number of boxes was large e.g. due to "splitting"). 
  This was particularly noticeable for the VOF algorithm (which does
  many locate() calls to reconstruct local Cartesian stencils).

2009-07-15 Stephane Popinet <popinet at users.sf.net>
  * New option '--event' for gerris2D/3D

2009-07-15 Stephane Popinet <popinet at users.sf.net>
  * Fixed GfsVariableCurvature timers

2009-07-15 Stephane Popinet <popinet at users.sf.net>
  * New '--pid' option for gerris2D/3D

2009-07-14 Stephane Popinet <popinet at users.sf.net>
  * A few fixes for OutputTiming

2009-07-14 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Output timing now outputs sorted timings

2009-07-14 Stephane Popinet <popinet at users.sf.net>
  * Entirely solid boxes can be used
  
  They will be automatically removed from the simulation rather than 
  making the simulation abort with an error message.

2009-07-14 Stephane Popinet <popinet at users.sf.net>
  * Dynamic parallel load-balancing (EventBalance)

2009-07-13 Stephane Popinet <popinet at users.sf.net>
  * daily script only generates launchpad snapshots

2009-07-13 Stephane Popinet <popinet at users.sf.net>
  * Fixed header dependency for domain.c

2009-07-12 Stephane Popinet <popinet at users.sf.net>
  * Added openmpi-bin as a dependency for debian package
  
  This is a workaround for the broken libopenmpi1 debian package.

2009-06-28 Stephane Popinet <popinet at users.sf.net>
  * GfsBoundaryMpi is included by default (even in the serial version)
  
  This simplifies the code and also means that the serial version can be used
  to join and load simulation files generated by parallel runs (e.g. a serial
  version of GfsView ran on a different system will now be able to visualise
  the results of a parallel run).

2009-06-22 Stephane Popinet <popinet at users.sf.net>
  * Fixed Debian version numbering

2009-06-22 Stephane Popinet <popinet at users.sf.net>
  * Automatically include aclocal m4 macros

2009-05-13 Stephane Popinet <popinet at users.sf.net>
  * Debian packages are built with MPI support

2009-05-13 Stephane Popinet <popinet at users.sf.net>
  * Added 'launchpad-snapshot' Makefile target

2009-07-02 Stephane Popinet <popinet at users.sf.net>
  * GfsOutputDropletSums and GfsRemoveDroplets should now work in parallel
  
  and also for periodic boundary conditions in serial and in parallel.

2009-07-09 Stephane Popinet <popinet at users.sf.net>
  * EventBalance implements dynamic load balancing on 2 PEs only

2009-07-08 Stephane Popinet <popinet at users.sf.net>
  * OutputTime reports average CPU time for parallel runs

2009-07-08 Stephane Popinet <popinet at users.sf.net>
  * OutputBalance ignores box pids for serial runs

2009-07-06 Stephane Popinet <popinet at users.sf.net>
  * Box positions set in the parameter file are preserved

2009-07-01 Stephane Popinet <popinet at users.sf.net>
  * GfsInitFraction should now work also for parallel simulations
  
  It used to fail for domains containing disconnected GfsBoxes (which can
  happen for some domain decompositions in parallel).

2009-06-28 Stephane Popinet <popinet at users.sf.net>
  * Added version info to read()/write() methods of GfsDomain

2009-06-28 Stephane Popinet <popinet at users.sf.net>
  * New functions gfs_send_boxes() and gfs_receive_boxes()
  
  To move GfsBoxes between processors in parallel.

2009-06-25 Stephane Popinet <popinet at users.sf.net>
  * Major bug fix for GfsBoundaryPeriodic
  
  Amazingly the implementation of the match() method seems to have been wrong
  right from the initial implementation (then part of BoundaryMpi). Yet, despite
  being completely inconsistent, this implementation managed to "work"
  undetected for several years... The bug also affected parallel simulations
  through inheritance to BoundaryMpi.

2009-06-23 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for file access test in parallel

2009-06-23 Stephane Popinet <popinet at users.sf.net>
  * Improved error messages in parallel (includes processor name)

2009-06-22 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged release_21_06_09

2009-06-21 Stephane Popinet <popinet at users.sf.net>
  * Added configuration options to "--version" info

2009-06-21 Stephane Popinet <popinet at users.sf.net>
  * GfsOutputSimulation can write "joined" parallel simulation files
  
  Only for GFS and text formats.

2009-06-21 Stephane Popinet <popinet at users.sf.net>
  * GfsEventScript is only called for rank 0 PE in parallel

2009-06-20 Stephane Popinet <popinet at users.sf.net>
  * Added MPI_Finalize() as required by MPI standard

2009-06-19 Stephane Popinet <popinet at users.sf.net>
  * Added man pages for gfsjoin2D, gfsjoin3D

2009-06-19 Stephane Popinet <popinet at users.sf.net>
  * New 'keep' option for gfsjoin2D

2009-06-19 Stephane Popinet <popinet at users.sf.net>
  * GfsDomain automatically converts GfsBoundaryMPI into edges on read

2009-06-19 Stephane Popinet <popinet at users.sf.net>
  * Updated test case references

2009-06-19 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for GfsSolidMoving/GfsSurface write() method

2009-06-18 Stephane Popinet <popinet at users.sf.net>
  * New gfsjoin2D, gfsjoin3D tools
  
  These are replacing the gfsjoin script. Note however that they can only be used
  to join parallel files created using this and later versions of Gerris. Older
  parallel simulation files should still use the gfsjoin script.

2009-06-18 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for write() method of GfsSurface

2009-05-14 Stephane Popinet <popinet at users.sf.net>
  * GfsVariableTerrain also reconstructs H and P of GfsRiver

2009-05-13 Stephane Popinet <popinet at users.sf.net>
  * Adaptivity cannot coarsen by more than one level per timestep

2009-05-13 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for incorrect generation of dependencies in tests docs

2009-05-12 Stephane Popinet <popinet at users.sf.net>
  * Check that SolidMoving is used within SimulationMoving

2009-05-12 Stephane Popinet <popinet at users.sf.net>
  * Renamed GfsMovingSimulation to GfsSimulationMoving

2009-05-12 Stephane Popinet <popinet at users.sf.net>
  * Updated moving hexagon test case

2009-05-12 Stephane Popinet <popinet at users.sf.net>
  * Split 2nd-order moving in moving2.c

2009-05-12 Stephane Popinet <popinet at users.sf.net>
  * Fix for remeshing of moving solids

2009-05-12 Stephane Popinet <popinet at users.sf.net>
  * Diffusion now works with moving boundaries

2009-05-06 Stephane Popinet <popinet at users.sf.net>
  * Fix for SurfaceBc implementation of moving boundaries

2009-05-05 Stephane Popinet <popinet at users.sf.net>
  * Solid motion is now specified through SurfaceBc

2009-05-05 Stephane Popinet <popinet at users.sf.net>
  * Simplification of redistribute_destroyed_cells_content()

2009-05-05 Stephane Popinet <popinet at users.sf.net>
  * Simplified moving_cell_fine_init and removed old_solid_coarse_fine

2009-05-05 Stephane Popinet <popinet at users.sf.net>
  * Simplified moving_simulation_set_timestep()

2009-05-04 Stephane Popinet <popinet at users.sf.net>
  * Fixes for memory leaks in initial moving boundary implementation

2009-05-03 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Translating hexagon test case

2009-05-03 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Initial moving solid boundary implementation

2009-05-12 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for corner refinement bug fix

2009-05-11 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for violation of 'corner refinement criterion'

2009-05-09 Stephane Popinet <popinet at users.sf.net>
  * autogen.sh copies missing files

2009-05-08 Stephane Popinet <popinet at users.sf.net>
  * New 'time_order' parameter for GfsRiver

2009-05-04 Stephane Popinet <popinet at users.sf.net>
  * More generic valgrind suppression file (ld.so.supp)

2009-05-04 Stephane Popinet <popinet at users.sf.net>
  * Default boundary conditions can be set for each variable

2009-05-02 Stephane Popinet <popinet at users.sf.net>
  * New 'cleanup' method for GfsVariable

2009-04-30 Stephane Popinet <popinet at users.sf.net>
  * Changed meaning of gfs_simulation_get_solids() (!!)

2009-05-02 Stephane Popinet <popinet at users.sf.net>
  * Debian package must depend on libgts-snapshot-dev

2009-05-02 Stephane Popinet <popinet at users.sf.net>
  * Bug fixes for 'daily' script

2009-05-01 Stephane Popinet <popinet at users.sf.net>
  * daily script properly signs debian repositories

2009-05-01 Stephane Popinet <popinet at users.sf.net>
  * Divide-by-zero exceptions use <fenv.h>

2009-05-02 Stephane Popinet <popinet at users.sf.net>
  * Man pages need to be distributed

2009-05-01 Stephane Popinet <popinet at users.sf.net>
  * Updated debian package build (based on Drew Parson's)

2009-05-01 Stephane Popinet <popinet at users.sf.net>
  * Added man pages from debian (thanks Ruben and Drew!)

2009-04-30 Stephane Popinet <popinet at users.sf.net>
  * Advection merging updates can be parameterised

2009-04-30 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_surface_transformation()

2009-04-30 Stephane Popinet <popinet at users.sf.net>
  * gfs_cell_traverse_cut() can traverse "destroyed" cells

2009-04-30 Stephane Popinet <popinet at users.sf.net>
  * Projections can use an optional initial divergence field

2009-04-30 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_advance_tracers()

2009-04-30 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_cell_is_small()

2009-04-21 Stephane Popinet <popinet at users.sf.net>
  * New 'ppm2video' script
  
  Uses ffmpeg to convert PPM to (almost) any video format.

2009-04-21 Stephane Popinet <popinet at users.sf.net>
  * Updated reference for 'reynolds' test case

2009-04-19 Stephane Popinet <popinet at users.sf.net>
  * Fix for dimensioning of GfsSurfaceBc

2009-04-03 Stephane Popinet <popinet at users.sf.net>
  * Fix for dependencies of examples

2009-04-02 Stephane Popinet <popinet at users.sf.net>
  * Fix for negative clock return values (thanks to G. McBain)

2009-03-27 Stephane Popinet <popinet at users.sf.net>
  * gerris.dic is installed in $prefix/share/gerris/gerris.dic

2009-03-16 Daniel Fuster <dfuster at gmail.com>
  * vim dictionary

2009-03-18 Stephane Popinet <popinet at users.sf.net>
  * Fix for compilation on Mac OSX 10.4, 10.5

2009-03-16 Stephane Popinet <popinet at users.sf.net>
  * OutputDropletSums sorts droplets by volume

2009-03-15 Stephane Popinet <popinet at users.sf.net>
  * Fixed dimensioning of GfsVariableCurvature

2009-03-13 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for surface tension coefficient and dimensions

2009-02-27 Stephane Popinet <popinet at users.sf.net>
  * Fix (again!) for module dependencies

2009-02-26 Stephane Popinet <popinet at users.sf.net>
  * Fixed obsolete lseek() call in RStarTree

2009-02-26 Stephane Popinet <popinet at users.sf.net>
  * Fix for hardcoded path in wavewatch module

2009-02-25 Stephane Popinet <popinet at users.sf.net>
  * Fix for incorrect 'dist' dependency in wavewatch module

2009-02-24 Stephane Popinet <popinet at users.sf.net>
  * Fix for 'listing' option of wavewatch

2009-02-24 Stephane Popinet <popinet at users.sf.net>
  * Fixed changelog in gerris.spec

2009-02-23 Stephane Popinet <popinet at users.sf.net>
  * Fix for wavewatch module makefile

2009-02-23 Stephane Popinet <popinet at users.sf.net>
  * Automatic type checking for wavewatch C/Fortran linking

2009-02-23 Stephane Popinet <popinet at users.sf.net>
  * Wavewatch initialisation files are created in the background

2009-02-23 Stephane Popinet <popinet at users.sf.net>
  * New 'wavewatch' module

2009-02-20 Stephane Popinet <popinet at users.sf.net>
  * Fix for energy calculation in GfsWave

2009-02-03 Stephane Popinet <popinet at users.sf.net>
  * Applied patch by Laurent Roblou to FES2004 library

2009-02-02 Stephane Popinet <popinet at users.sf.net>
  * New object GfsVariableFunction
  
  Used to efficiently refine adaptively a variable defined by a GfsFunction.

2009-01-29 Stephane Popinet <popinet at users.sf.net>
  * Small optimisation for dry cells in GfsRiver

2009-01-16 Stephane Popinet <popinet at users.sf.net>
  UNDO: New GfsBcValve BC for GfsRiver

2009-01-29 Stephane Popinet <popinet at users.sf.net>
  * Cron script darcs pull from the correct repositories

2009-01-29 Stephane Popinet <popinet at users.sf.net>
  * Added fortran dependency for new stokes module

2009-01-28 Stephane Popinet <popinet at users.sf.net>
  * Small optimisation for gfs_domain_traverse_merged()

2009-01-28 Stephane Popinet <popinet at users.sf.net>
  * Less strict check for parabola test case

2009-01-27 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated GfsView parameter files for updated GfsGlLinear object

2009-01-21 Stephane Popinet <popinet at users.sf.net>
  * Fix for hardcoding of path in Terrain module

2009-01-16 Stephane Popinet <popinet at users.sf.net>
  * New GfsVariableTerrain for adaptive refinement of GfsRiver

2009-01-16 Stephane Popinet <popinet at users.sf.net>
  * New GfsBcValve BC for GfsRiver

2008-08-22 Stephane Popinet <popinet at users.sf.net>
  * New object GfsInitStokesWave

2008-07-28 Stephane Popinet <popinet at users.sf.net>
  * Iterative Stokes wave solver of John Chaplin

2008-05-06 Stephane Popinet <popinet at users.sf.net>
  * Added GFS_FES2004 environment variable

2007-09-18 Stephane Popinet <popinet at users.sf.net>
  * New EventBalance object for parallel load-balancing

2007-09-02 Stephane Popinet <popinet at users.sf.net>
  * New VariableDiagonal for Poisson solver diagonal dominance evaluation

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * Updated GfsRiver examples

2009-01-10 Stephane Popinet <popinet at users.sf.net>
  * New 'hump' GfsRiver example

2009-01-09 Stephane Popinet <popinet at users.sf.net>
  * New dam break example

2009-01-22 Stephane Popinet <popinet at users.sf.net>
  * Updated parabola test case

2009-01-22 Stephane Popinet <popinet at users.sf.net>
  * SourceCoriolis properly takes friction term into account

2009-01-22 Stephane Popinet <popinet at users.sf.net>
  * Friction for the River model is implemented using SourceCoriolis

2009-01-22 Stephane Popinet <popinet at users.sf.net>
  * Fix for terrain module when refining an existing mesh

2009-01-22 Stephane Popinet <popinet at users.sf.net>
  * Log timing statistics for events

2009-01-16 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for ppmcombine

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * Updated parabola test case

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for incorrect write() method of GfsRefineTerrain

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * stderr and stdout are muted in parallel

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for gfs_all_reduce()

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * Consistent descriptions and derived variables for GfsRiver

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * Removed Riemann Roe solver from GfsRiver

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for overflow in gfscompare

2009-01-15 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for GfsInit erasing variables' descriptions

2009-01-14 Stephane Popinet <popinet at users.sf.net>
  * Added timers to GfsRiver

2009-01-11 Stephane Popinet <popinet at users.sf.net>
  * Bathymetry reconstruction ensures hydrostatic balance

2009-01-10 Stephane Popinet <popinet at users.sf.net>
  * New 'parabola' test case for GfsRiver

2009-01-10 Stephane Popinet <popinet at users.sf.net>
  * Cleaned up GfsRiver code

2009-01-09 Stephane Popinet <popinet at users.sf.net>
  * Topography source terms for GfsRiver are balanced also with adaptivity

2009-01-09 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for gfs_remove_ponds()

2008-12-17 Stephane Popinet <popinet at users.sf.net>
  * Uses the standard definitions for H and Zb

2008-12-10 Stephane Popinet <popinet at users.sf.net>
  * Implementation of the Audusse et al scheme

2008-12-10 Stephane Popinet <popinet at users.sf.net>
  * Wetting-drying for GfsRiver

2008-12-10 Stephane Popinet <popinet at users.sf.net>
  * Topographic source terms for River model

2008-12-10 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for HLLC riemann solver

2008-12-09 Stephane Popinet <popinet at users.sf.net>
  * 2nd-order non-linear shallow-water solver

2008-12-09 Stephane Popinet <popinet at users.sf.net>
  * BcSubcritical boundary condition

2008-12-09 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_center_minmod_gradient

2008-11-14 Stephane Popinet <popinet at users.sf.net>
  * Added "none" option for gradient of AdvectionParams
  
  This means that the advection schemes can be made first-order in space.

2008-11-14 Stephane Popinet <popinet at users.sf.net>
  * Initial nonlinear shallow-water solver
  
  1D, first-order in space and time.

2009-01-09 Stephane Popinet <popinet at users.sf.net>
  * Fix for hevea bug in gfs2doc

2008-12-18 Stephane Popinet <popinet at users.sf.net>
  * Parallel outputs are only done for files with '%d' format

2008-12-17 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for OutputScalarSum in parallel

2009-01-08 Stephane Popinet <popinet at users.sf.net>
  * OutputTime also outputs real elapsed time

2008-12-10 Stephane Popinet <popinet at users.sf.net>
  * OutputSimulation was not talking dimensioning into account for text format

2008-12-02 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for BC of diffusion equation

2008-11-28 Stephane Popinet <popinet at users.sf.net>
  * Gnuplot output from gfs2oogl takes mapping into account

2008-11-28 Stephane Popinet <popinet at users.sf.net>
  * OutputParticle now derives from OutputLocation
  
  Note also that the syntax has changed.

2008-11-07 Stephane Popinet <popinet at users.sf.net>
  * ASCII float conversion format option for OutputSimulation

2008-11-07 Stephane Popinet <popinet at users.sf.net>
  * Improved robustness of VOF volume computation

2008-11-04 Stephane Popinet <popinet at users.sf.net>
  * Slip length in BcNavier is now a GfsFunction

2008-11-04 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for GfsRemoveDroplets (thanks to G. Tomar)

2008-11-04 Stephane Popinet <popinet at users.sf.net>
  * Fixed dimensioning of OutputSolidForce.

2008-10-30 Stephane Popinet <popinet at users.sf.net>
  * Dimensioning for AdaptGradient

2008-10-28 Stephane Popinet <popinet at users.sf.net>
  * Dimensioning of Flather boundary conditions

2008-10-24 Stephane Popinet <popinet at users.sf.net>
  * New 'ppm2theora' script

2008-10-24 Stephane Popinet <popinet at users.sf.net>
  * 'tides' example uses new 3D ocean code and scaling system

2008-10-24 Stephane Popinet <popinet at users.sf.net>
  * rsurfacedraw outputs aspect ratio statistics

2008-10-23 Stephane Popinet <popinet at users.sf.net>
  * Fixed dimensioning of GfsEventHarmonic

2008-10-22 Stephane Popinet <popinet at users.sf.net>
  * GTS and CGD files take mapping into account

2008-10-22 Stephane Popinet <popinet at users.sf.net>
  * Terrain databases were incompatible between 32 and 64 bits
  
  This is fixed but you will need to regenerate the databases...

2008-10-16 Stephane Popinet <popinet at users.sf.net>
  * Dimensioning for 3D ocean model

2008-10-21 Stephane Popinet <popinet at users.sf.net>
  * Increased significant digits of text export

2008-10-16 Stephane Popinet <popinet at users.sf.net>
  * New "dL" derived variable

2008-10-16 Stephane Popinet <popinet at users.sf.net>
  * Consistent z-scaling of terrain

2008-10-16 Stephane Popinet <popinet at users.sf.net>
  * Terrain module works only with relative bilinear coordinates

2008-10-16 Stephane Popinet <popinet at users.sf.net>
  * Fixed access modes for Terrain module

2008-10-09 Stephane Popinet <popinet at users.sf.net>
  * New Terrain module uses relative sums
  
  This fixes round-off problems in the previous implementation of 
  the new version. Note that the terrain databases need to be regenerated.

2008-10-09 Stephane Popinet <popinet at users.sf.net>
  * New 'rsurfacedraw' command

2008-10-06 Stephane Popinet <popinet at users.sf.net>
  * Added new "min" and "max" derived Terrain variables

2008-10-02 Stephane Popinet <popinet at users.sf.net>
  * New version of the terrain module using region_sum queries
  
  This patch contains both the old and the new version, for debugging purposes.

2008-10-02 Stephane Popinet <popinet at users.sf.net>
  * New function r_surface_query_region_sum()
  
  Efficiently computes the statistics for a given region using the hierarchical
  statistics stored in the R*-tree.

2008-10-05 Stephane Popinet <popinet at users.sf.net>
  * Fixed headers for compilation on Mac OSX

2008-10-05 Stephane Popinet <popinet at users.sf.net>
  * Fix for incorrect versioning of debian packages

2008-09-29 Stephane Popinet <popinet at users.sf.net>
  * darcs2dist uses sftp rather than ssh

2008-09-23 Stephane Popinet <popinet at users.sf.net>
  * Updated URL for sourceforge server upgrade

2008-09-18 Stephane Popinet <popinet at users.sf.net>
  * Fix for dimensioning of SourceTension when using reduced gravity

2008-09-11 Stephane Popinet <popinet at users.sf.net>
  * Added missing required files for test cases

2008-09-09 Stephane Popinet <popinet at users.sf.net>
  * Axisymmetric spurious currents test case

2008-09-09 Stephane Popinet <popinet at users.sf.net>
  * Axisymmetric surface tension implementation

2008-09-09 Stephane Popinet <popinet at users.sf.net>
  * Replaced GFS_VARIABLE() with GFS_VALUE() in tension.c, vof.c

2008-09-09 Stephane Popinet <popinet at users.sf.net>
  * Updated test case references

2008-09-08 Stephane Popinet <popinet at users.sf.net>
  * Fixes for missing dimensioning
  
  VariableCurvature, VariableDistance, VariableFiltered, SourceTension.

2008-09-06 Stephane Popinet <popinet at users.sf.net>
  * Implicit viscosity now works for axisymmetric domains
  
  The explicit viscosity scheme has also been simplified.

2008-09-06 Stephane Popinet <popinet at users.sf.net>
  * Axisymmetric viscous flow past a sphere test case

2008-08-06 Stephane Popinet <popinet at users.sf.net>
  * New object GfsAdaptError
  
  Uses an error estimate based on the norm of the Hessian matrix.

2008-08-01 Stephane Popinet <popinet at users.sf.net>
  * Reimplementation of SourceVicosityExplicit
  
  Also works for axisymmetric flows (but for 3D flows yet).

2008-08-01 Stephane Popinet <popinet at users.sf.net>
  * Explicit viscosity lid-driven cavity test case

2008-09-04 Stephane Popinet <popinet at users.sf.net>
  * gfs2oogl outputs dimensional values in mixed cells

2008-09-03 Stephane Popinet <popinet at users.sf.net>
  * VariableDistance now works also in 3D

2008-08-21 Stephane Popinet <popinet at users.sf.net>
  * New user function "correctness()" for complexity of VOF interfaces

2008-08-22 Stephane Popinet <popinet at users.sf.net>
  * Spatial functions can now be time-dependent

2008-08-19 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases references

2008-08-10 Stephane Popinet <popinet at users.sf.net>
  * check.py is distributed in the tarball

2008-08-10 Stephane Popinet <popinet at users.sf.net>
  * Computation of maximum interface curvature is more robust

2008-08-07 Stephane Popinet <popinet at users.sf.net>
  * VariableFiltered is initialised
  
  This is important if a filtered variable is used for the initial approximate
  projection (e.g. for the density).

2008-07-29 Stephane Popinet <popinet at users.sf.net>
  * Removed left-over debugging message

2008-07-28 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases references

2008-07-26 Stephane Popinet <popinet at users.sf.net>
  * gfs_init_solid_fractions() always resets the old solids first

2008-06-20 Stephane Popinet <popinet at users.sf.net>
  * Quick fix for surfaces exactly coincident with mesh
  
  i.e. it should not be necessary anymore to shift implicit surfaces by epsilon 
  to get the expected results.

2008-05-01 Stephane Popinet <popinet at users.sf.net>
  * Robust treatment of border cases for gfs_line_center, gfs_plane_center

2008-07-23 Stephane Popinet <popinet at users.sf.net>
  * Cleaner axi test case parameter file

2008-07-26 Stephane Popinet <popinet at users.sf.net>
  * Restructured correct_normal_velocity()
  
  This also fixes a bug introduced by the new axisymmetric formulation.

2008-07-26 Stephane Popinet <popinet at users.sf.net>
  * Fix for "corner refinement" topology violation

2008-07-25 Stephane Popinet <popinet at users.sf.net>
  * Fixed incorrect L-dimensioning of force output

2008-07-22 Stephane Popinet <popinet at users.sf.net>
  * Axisymmetric advection

2008-07-22 Stephane Popinet <popinet at users.sf.net>
  * Removed obsolete charset option for html doc generation

2008-07-20 Stephane Popinet <s.popinet at niwa.co.nz>
  * Clean target for examples

2008-07-19 Stephane Popinet <popinet at users.sf.net>
  * Axisymmetric Euler solver

2008-07-18 Stephane Popinet <popinet at users.sf.net>
  * Axisymmetric potential flow around a sphere test case

2008-07-19 Stephane Popinet <popinet at users.sf.net>
  * L-dimensioning for OutputScalarSum, VTK and Tecplot

2008-07-19 Stephane Popinet <popinet at users.sf.net>
  * New function "area" for interfacial area density

2008-06-09 Stephane Popinet <popinet at users.sf.net>
  * Wave model uses user-defined length units

2008-06-08 Stephane Popinet <popinet at users.sf.net>
  * Solid flux correction for wave model

2008-06-08 Stephane Popinet <popinet at users.sf.net>
  * New 'garden sprinkler effect' example for wave model

2008-07-18 Stephane Popinet <popinet at users.sf.net>
  * Several important fixes for L-dimensioning

2008-07-16 Stephane Popinet <popinet at users.sf.net>
  * gfs_domain_cfl() should now take source terms into account

2008-07-16 Stephane Popinet <popinet at users.sf.net>
  * Minor fixed to RPM spec file (I. Vari)

2008-06-07 Stephane Popinet <popinet at users.sf.net>
  * Maximum timestep is not limited

2008-07-10 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for cgd files and face boundary conditions

2008-06-08 Stephane Popinet <popinet at users.sf.net>
  * New "Energy" derived variable for wave model

2008-06-07 Stephane Popinet <popinet at users.sf.net>
  * New object GfsInitWave

2008-06-07 Stephane Popinet <popinet at users.sf.net>
  * nk and ntheta are parameters for wave model

2008-06-07 Stephane Popinet <popinet at users.sf.net>
  * Fixed scaling of energy for wave model

2008-05-11 Stephane Popinet <popinet at users.sf.net>
  * Initial sub-cycling implementation for wave model

2008-05-09 Stephane Popinet <popinet at users.sf.net>
  * Initial wave model

2008-06-13 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for broken BcNavier

2008-06-09 Stephane Popinet <popinet at users.sf.net>
  * Dimensional size of the unit box can be set

2008-07-08 Stephane Popinet <popinet at users.sf.net>
  * Multigrid "tricks" for diffusion solver

2008-07-07 Stephane Popinet <popinet at users.sf.net>
  * darcs2dist puts tarballs in a tarballs/ directory

2008-07-07 Stephane Popinet <popinet at users.sf.net>
  * Fixed typo in RefineHeight (thanks to G. McBain)

2008-07-07 Stephane Popinet <popinet at users.sf.net>
  * Added missing gfsxref to tarball

2008-06-20 Stephane Popinet <popinet at users.sf.net>
  * More accurate streamlines
  
  Uses new function gfs_mixed_cell_interpolate().

2008-07-04 Stephane Popinet <popinet at users.sf.net>
  * Scripts for automatic cross-reference generation
  
  For automatically linking keyword wiki documentation with the examples and 
  test cases.

2008-07-02 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases references

2008-06-20 Stephane Popinet <popinet at users.sf.net>
  * New object GfsBoundaryGradient

2008-07-01 Stephane Popinet <popinet at users.sf.net>
  * RPM spec update for Fedora (I. Vari)

2008-07-01 Stephane Popinet <popinet at users.sf.net>
  * Several fixes for GfsCartesianGrid object

2008-06-20 Stephane Popinet <popinet at users.sf.net>
  * darcs2dist makes tarballs with consistent version numbers

2008-07-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for doc generation with new gfs-highlight script

2008-06-20 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for adaptive refinement (maxlevel was not always respected)

2008-06-19 Stephane Popinet <popinet at users.sf.net>
  * VariableFiltered defines values on all levels
  
  This should fix bugs when using a filtered variable within alpha (alpha needs 
  to be defined on all levels for the viscous terms).

2008-06-19 Stephane Popinet <popinet at users.sf.net>
  * Updated spurious currents test case reference

2008-06-13 Stephane Popinet <popinet at users.sf.net>
  * Fix for linking utility libraries with libtool

2008-06-11 Stephane Popinet <popinet at users.sf.net>
  * Fix for OutputSimulation silly bug

2008-06-11 Stephane Popinet <popinet at users.sf.net>
  * Fix for VTK/Tecplot output in parallel

2008-06-11 Stephane Popinet <popinet at users.sf.net>
  * New gfs-highlight script
  
  Uses GNU source-highlight for syntax highlighting of simulation files.

2008-06-09 Stephane Popinet <popinet at users.sf.net>
  * GfsAdaptGradient has been generalised to functions

2008-05-09 Stephane Popinet <popinet at users.sf.net>
  UNDO: gfs_domain_add_variable() takes a class parameter

2008-06-07 Stephane Popinet <popinet at users.sf.net>
  * OutputSimulation really writes all the variables by default

2008-05-11 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_event_redo()

2008-05-09 Stephane Popinet <popinet at users.sf.net>
  * gfs_domain_add_variable() takes a class parameter

2008-05-09 Stephane Popinet <popinet at users.sf.net>
  * Generalised gfs_matrix_new() to non-square matrices

2008-06-08 Stephane Popinet <popinet at users.sf.net>
  * Checks that density is not negative

2008-06-06 Stephane Popinet <popinet at users.sf.net>
  * Fix for inconsistent treatment of maxlevel adaptive parameter
  
  This patch should ensure that cells are never more refined than maxlevel.

2008-06-06 Stephane Popinet <popinet at users.sf.net>
  * Fix for automake/libtool object issue

2008-04-23 Stephane Popinet <popinet at users.sf.net>
  UNDO: Use dolt instead of standard libtool

2008-05-13 Stephane Popinet <popinet at users.sf.net>
  * Improved options for xyz2rsurface

2008-05-13 Stephane Popinet <popinet at users.sf.net>
  * Improved path system for Terrain module

2008-05-13 Stephane Popinet <popinet at users.sf.net>
  * Fix for hard-coded module names in parameter files

2008-05-12 Stephane Popinet <popinet at users.sf.net>
  * New dependencies for RPM packages (I. Vari)

2008-05-10 Stephane Popinet <popinet at users.sf.net>
  * cron script uses new "devel" branches

2008-05-10 Stephane Popinet <popinet at users.sf.net>
  * Added cron job script in tools/

2008-05-09 Stephane Popinet <popinet at users.sf.net>
  * Fixed missing dependency for Terrain module

2008-05-06 Stephane Popinet <popinet at users.sf.net>
  * Added debian dependencies for Tide and Map modules

2008-05-08 Stephane Popinet <popinet at users.sf.net>
  * GFS_MODULES_DIR is defined in Makefile
  
  This should fix problems when using the --libdir option of configure

2008-05-07 Stephane Popinet <popinet at users.sf.net>
  * Poiseuille flow test case

2008-05-06 Stephane Popinet <popinet at users.sf.net>
  * Integrated FES2004 build within primary build

2008-05-06 Stephane Popinet <popinet at users.sf.net>
  * Initial import of FES2004_distribution_soft-1.0.9
    
  A GPL-ed library to read FES2004 tidal atlas data.

2008-05-06 Stephane Popinet <popinet at users.sf.net>
  * New derived variables rx, ry and rz

2008-05-01 Stephane Popinet <popinet at users.sf.net>
  * Improved robustness of solid fraction checks

2008-04-24 Stephane Popinet <popinet at users.sf.net>
  * New abstract class GfsGenericSurface

2008-04-29 Stephane Popinet <popinet at users.sf.net>
  * Map module also scales z coordinate

2008-04-29 Stephane Popinet <popinet at users.sf.net>
  * Solid boundary generation changes for Terrain module

2008-04-29 Stephane Popinet <popinet at users.sf.net>
  * Extension of GfsRefineTerrain to 3D

2008-04-29 Stephane Popinet <popinet at users.sf.net>
  * Removed "Writing new objects" section of tutorial
  
  This has been superseded by the wiki version in the Gerris programming course.

2008-04-29 Stephane Popinet <popinet at users.sf.net>
  * Added missing awk script in debian package

2008-04-24 Stephane Popinet <popinet at users.sf.net>
  * Thin solid cells are flagged and "fixed" before checks

2008-04-24 Stephane Popinet <popinet at users.sf.net>
  * New GfsTerrain class

2008-04-23 Stephane Popinet <popinet at users.sf.net>
  * Terrain module uses R*-tree

2008-04-23 Stephane Popinet <popinet at users.sf.net>
  * Use dolt instead of standard libtool

2008-04-23 Stephane Popinet <popinet at users.sf.net>
  * Minor changes to RStarTree (updated header files etc...)

2008-04-10 Stephane Popinet <popinet at users.sf.net>
  * Large file (> 2GiB) support for RStarTree

2008-04-10 Stephane Popinet <popinet at users.sf.net>
  * Upgraded/integrated Makefile.am for RStarTree

2008-04-10 Stephane Popinet <popinet at users.sf.net>
  * R*-tree implementation of Norbert Beckmann
  
  This is the original implementation as unpacked from http://www.rtreeportal.org.

2008-04-08 Stephane Popinet <popinet at users.sf.net>
  * New terrain module for height-field databases

2008-04-01 Stephane Popinet <popinet at users.sf.net>
  * New module for FES2004 tidal boundary conditions

2008-04-01 Stephane Popinet <popinet at users.sf.net>
  * GModule can take parameters

2008-03-28 Stephane Popinet <popinet at users.sf.net>
  * New Map module for cartographic projections

2008-03-25 Stephane Popinet <popinet at users.sf.net>
  * Support for geometric "mapping" of domain coordinates

2008-01-30 Stephane Popinet <popinet at users.sf.net>
  * Single-layer ocean model can now use "3D" code rather than "2D3"

2008-04-18 Stephane Popinet <popinet at users.sf.net>
  * Stricter checking when compiling user-defined GfsFunctions

2008-04-18 Stephane Popinet <popinet at users.sf.net>
  * Need to make sure that all PEs use the same timestep!
  
  Important bug fix for the parallel version.

2008-04-05 Stephane Popinet <popinet at users.sf.net>
  * Check for unclosed statement in parameter files

2008-03-25 Stephane Popinet <popinet at users.sf.net>
  * Checks that variable names are not reserved keywords

2008-03-18 Stephane Popinet <popinet at users.sf.net>
  * Examples files are distributed only if they are version-controlled

2008-03-13 Stephane Popinet <popinet at users.sf.net>
  * Removed MRS acceleration
  
  While this accelerated convergence somewhat in some cases, it seemed to prevent
  convergence in other cases. This patch is not an exact "rollback" of the
  initial MRS implementation because it does not rollback other simple but 
  important changes to the Poisson solver which seem to really improve robustness.

2008-03-13 Stephane Popinet <popinet at users.sf.net>
  * Fix for MRS implementation in parallel

2008-03-11 Stephane Popinet <popinet at users.sf.net>
  * Minimal Residual Smoothing implementation
  
  This guarantees that the Poisson solver does not diverge (but does not
  guarantee convergence). It can also accelerate convergence depending on the
  problem. Coupled with other smaller changes included in this patch ('minlevel'
  tuning in particular) this improves the robustness of the solver.

2008-03-12 Stephane Popinet <popinet at users.sf.net>
  * Series 60 ship waves example

2008-03-10 Stephane Popinet <popinet at users.sf.net>
  * SourceControl was broken by momentum source term patch

2008-03-07 Stephane Popinet <popinet at users.sf.net>
  * 'omega' relaxation parameter is written only if different from default
  
  This means that simulation files are backward-compatible with older versions of
  gerris/gfsview. 

2008-03-06 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for sign error in new momentum source term implementation

2008-03-06 Stephane Popinet <popinet at users.sf.net>
  * GfsOutput pipes now know the pre-defined Gerris shell variables

2008-03-06 Stephane Popinet <popinet at users.sf.net>
  * gerris.spec version number fix

2008-03-05 Stephane Popinet <popinet at users.sf.net>
  * Optimised temporary variables usage for "gc" option

2008-03-05 Stephane Popinet <popinet at users.sf.net>
  * Cosmetics

2008-03-05 Stephane Popinet <popinet at users.sf.net>
  * Removed generation of Postscript docs

2008-03-04 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for VOF patch introducing "_x", "_y" and "_alpha" variables

2008-02-29 Stephane Popinet <popinet at users.sf.net>
  * VOF interface normals and alpha values are named variables
  
  e.g. for a VOF tracer called "T" in 2D, three new variables are automatically 
  created "T_x", "T_y" and "T_alpha".

2008-03-04 Stephane Popinet <popinet at users.sf.net>
  * GfsSource terms for velocity components are consistent with projection
  
  i.e. they are computed in a manner consistent with the pressure gradient 
  discretisation (in a manner similar to the surface tension). This is necessary
  in particular to guarantee exact hydrostatic balance with non-linear pressure
  distributions (e.g. the 'hydrostatic/quadratic' test case).

2007-10-18 Stephane Popinet <popinet at users.sf.net>
  * Added stability conditions for SourceDiffusionExplicit and SourceViscosityExplicit

2008-03-04 Stephane Popinet <popinet at users.sf.net>
  * New quadratic hydrostatic pressure test case

2008-02-29 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for OutputDropletSums when using functions as volume fraction

2008-02-27 Stephane Popinet <popinet at users.sf.net>
  * New air/water capillary wave test case

2008-02-27 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases for surface tension/advection bug fix

2008-02-27 Stephane Popinet <popinet at users.sf.net>
  * Pressure-correction in the Crank-Nicholson scheme is turned off by default
  
  This was degrading the results for a number of test cases (capillary waves in
  particular). The reasons why are still unclear.
  
  Note that this does not mean that this version of the code is the same as the
  version prior to the "Pressure term is included in RHS of viscosity solve"
  patch, as this version also includes an important bug fix for the advection
  terms when surface tension is present.
  
  The pressure-correction Crank-Nicholson scheme can be turned off explicitly 
  when required (e.g. "hydrostatic" test case).

2008-02-22 Stephane Popinet <popinet at users.sf.net>
  * Cleanup of GfsEventScript temporary files handling

2008-02-22 Stephane Popinet <popinet at users.sf.net>
  * Pressure term is included in RHS of viscosity solve
  
  This is an important change to the timestepping which should improve
  things significantly in particular when large source terms are included
  (e.g gravity + hydrostatic pressure).

2008-02-22 Stephane Popinet <popinet at users.sf.net>
  * Fix for initial timestep in Boussinesq example

2008-01-22 Stephane Popinet <popinet at users.sf.net>
  * New 'hydrostatic' test case

2008-02-21 Stephane Popinet <popinet at users.sf.net>
  * Bug fix when restarting simulations using SurfaceBc

2008-02-21 Stephane Popinet <popinet at users.sf.net>
  * Bug fixes for VTK/Tecplot output

2008-01-30 Stephane Popinet <popinet at users.sf.net>
  * Added 'omega' over-relaxation parameter

2008-01-25 Stephane Popinet <popinet at users.sf.net>
  * Removed gfs_get_from_below_extensive

2008-01-16 Stephane Popinet <popinet at users.sf.net>
  * GfsInit becomes a standard GfsEvent when either step or istep is specified 

2008-02-11 Stephane Popinet <popinet at users.sf.net>
  * VTK/Tecplot export

2008-01-29 Stephane Popinet <popinet at users.sf.net>
  * Improved robustness of checks for 'beta' and 'waves' test cases

2008-01-16 Stephane Popinet <popinet at users.sf.net>
  * Bug fix when using multiple GfsTime objects in parameter files

2008-01-16 Stephane Popinet <popinet at users.sf.net>
  * Various 'fixme' comments

2008-01-16 Stephane Popinet <popinet at users.sf.net>
  * New "dV" intrinsic domain variable

2008-01-16 Stephane Popinet <popinet at users.sf.net>
  * Renamed GfsOutputDropletStats to GfsOutputDropletSums
  
  Note that the syntax has changed, please check the doc.

2008-01-16 Stephane Popinet <popinet at users.sf.net>
  * Updated gfs_domain_remove_droplets() to use new GFS_VALUE macro
  
  Eventually this macro will replace GFS_VARIABLE.

2008-01-16 Stephane Popinet <popinet at users.sf.net>
  * RemoveDroplets takes an extra optional argument
  
  Which allows flexible selection of the "droplet" criterion.

2008-01-15 Stephane Popinet <popinet at users.sf.net>
  * Force reference of domain in Variable and DerivedVariable

2008-01-14 Stephane Popinet <popinet at users.sf.net>
  * Fix for missing variable descriptions

2008-01-13 Stephane Popinet <popinet at users.sf.net>
  * Deferred GfsFunction compilation
  
  Optionally, GfsFunctions are compiled only when they are used for the first
  time. This greatly improves the speed of GfsView when reading input piped from
  Gerris (when the simulation contains one or several GfsFunctions).

2008-01-13 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases references

2008-01-12 Stephane Popinet <popinet at users.sf.net>
  * Timestep is set properly to guarantee CFL condition
  
  This should fix CFL problems when using the VOF scheme.

2008-01-11 Stephane Popinet <popinet at users.sf.net>
  * New object GfsOutputDropletStats

2008-01-11 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_domain_tag_droplets()
  
  Used by GfsRemoveDroplets. This is more generic than the previous 
  implementation.

2008-01-11 Stephane Popinet <popinet at users.sf.net>
  * New intersection, union and difference predefined macros

2008-01-10 Stephane Popinet <popinet at users.sf.net>
  * 'error.gfs' dump files are indexed by PE number

2008-01-10 Stephane Popinet <popinet at users.sf.net>
  * Stricter syntax checking for read() method of Domain

2008-01-08 Stephane Popinet <popinet at users.sf.net>
  * RPM build fix (I. Vari)
  
  Removed %{?_smp_mflags} from make due to intermittent
  build errors on some SMP systems.

2007-12-21 Stephane Popinet <popinet at users.sf.net>
  * gfs_domain_cfl takes MAC velocities into account

2007-12-20 Daniel Fuster <dfuster at gmail.com>
  * gfsjoin201207
  Gfsjoin bug: fixed

2007-12-19 Stephane Popinet <popinet at users.sf.net>
  * New Cook Strait tides example

2007-12-18 Stephane Popinet <popinet at users.sf.net>
  * Simplified regular expression in m4.awk
  
  Hopefully this will fix issues with POSIX regex etc...

2007-12-06 Stephane Popinet <popinet at users.sf.net>
  * Updated debian package dependencies
  
  This should fix problems with recent Ubuntu versions (>= 7.10).

2007-12-05 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for reading VariablePosition reference position

2007-11-15 Stephane Popinet <popinet at users.sf.net>
  * In VOF advection GfsAdvectionParams needs to be initialised properly

2007-11-12 Stephane Popinet <popinet at users.sf.net>
  * VOF advection should now work also with embedded solid boundaries

2007-11-12 Stephane Popinet <popinet at users.sf.net>
  * Fixed package (install) dependencies

2007-11-12 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for refinement of coarse VOF cells in parallel
  
  Thanks to Daniel Fuster for reporting the problem.

2007-11-09 dfuster at gmail.com
  * gfsjoin1108
  Some modifications are introduced to fix some problems which could appear with the older version
  Checking options are added
  Minor format changes

2007-11-09 Stephane Popinet <popinet at users.sf.net>
  * cell_init() is now a method of GfsDomain
  
  This simplifies applying a consistent initialisation when creating
  children cells, in particular for the complicated initialisation
  occuring when refining interface cells between direction-sweeps during
  VOF advection.

2007-11-06 Stephane Popinet <popinet at users.sf.net>
  * New script gfsjoin (written by Daniel Fuster)

2007-11-01 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for "classes" internal utility with MPI

2007-10-30 Stephane Popinet <popinet at users.sf.net>
  * Added spec file for RPM packaging (thanks to Ivan Vari)

2007-10-30 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for MPI version of gfs_domain_stats_balance()

2007-10-24 Stephane Popinet <popinet at users.sf.net>
  * Boundary conditions are applied within Init events
  
  This is necessary e.g. when subsequent variables are derived from the gradients
  of previously initialised variables.

2007-10-23 Stephane Popinet <popinet at users.sf.net>
  * Minimised round-off errors in face_fractions()

2007-10-15 Stephane Popinet <popinet at users.sf.net>
  * Updated waves test case references

2007-10-12 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for gfs_domain_boundary_locate()
  
  This is important only for VOF boundary conditions.

2007-10-12 Stephane Popinet <popinet at users.sf.net>
  * Improved secant-bisection root-finding for implicit surfaces

2007-10-10 Stephane Popinet <popinet at users.sf.net>
  * Refines cells which are too coarse for VOF advection
  
  Cells are too coarse when one of their neighboring cells is finer and
  contains and interface which will be advected in them at the next
  timestep.

2007-10-11 Stephane Popinet <popinet at users.sf.net>
  * New pre-defined implicit surface 'cube'

2007-10-11 Stephane Popinet <popinet at users.sf.net>
  * Both GfsRefineSolid and GfsRefineSurface now work with implicit surfaces

2007-10-10 Stephane Popinet <popinet at users.sf.net>
  * New 'cfactor' parameter for GfsAdapt

2007-10-09 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_domain_reshape()
  
  Quoting the doc: 
  "Force the grading of the tree hierarchy of domain, matches the
  boundaries, recomputes merged cells and applies the boundary
  conditions for all variables."

2007-09-28 Stephane Popinet <popinet at users.sf.net>
  * gfs_cell_is_cut() is more clever when dealing with implicit surfaces
  
  This means in particular that GfsRefineSurface now also works with implicit 
  surfaces.

2007-10-05 Stephane Popinet <popinet at users.sf.net>
  * Fix for compatibility with glib-1.2 series

2007-10-03 Stephane Popinet <popinet at users.sf.net>
  * Miscellaneous doc/comments fixes

2007-09-28 Stephane Popinet <popinet at users.sf.net>
  * GfsNorm ignores values with a weight of zero for max norm

2007-09-28 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for text macros within comments

2007-09-28 Stephane Popinet <popinet at users.sf.net>
  * New "w" option for GfsOutputErrorNorm

2007-09-20 Stephane Popinet <popinet at users.sf.net>
  * Fixed incorrect label for capillary wave test case

2007-09-18 Stephane Popinet <popinet at users.sf.net>
  * Added figure to capillary wave test case

2007-09-18 Stephane Popinet <popinet at users.sf.net>
  * Updated Poisson test cases references
  
  The errors have been increased somewhat by using the more robust
  gradient_fine_coarse() implementation.

2007-09-18 Stephane Popinet <popinet at users.sf.net>
  * Improved formatting of test suite doc header

2007-09-18 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for gfs_domain_stats_balance() on serial parameter files

2007-09-18 Stephane Popinet <popinet at users.sf.net>
  * Removed obsolete installation instructions from the tutorial

2007-09-14 Stephane Popinet <popinet at users.sf.net>
  * New option 'data' for gerris (when splitting or partitioning)

2007-09-14 Stephane Popinet <popinet at users.sf.net>
  * OutputBalance takes "pid" into account even for non-parallel simulations

2007-09-14 Stephane Popinet <popinet at users.sf.net>
  * Fixed gfsview parameters for Boussinesq example

2007-09-14 Stephane Popinet <popinet at users.sf.net>
  * Fix for domain splitting with periodic boundary conditions

2007-09-14 Stephane Popinet <popinet at users.sf.net>
  * Domain traversal should now be deterministic
  
  ...as well as the rest of Gerris. This was not the case before because of the
  "randomization" effect of using hash tables indexed on pointers to store the
  graph nodes (i.e. the GfsBoxes). This was a pain for debugging and also caused
  boxes to "jump around" when visualising periodic domains with several boxes.

2007-09-13 Stephane Popinet <popinet at users.sf.net>
  * New "Id" and "Pid" derived variables

2007-09-13 Stephane Popinet <popinet at users.sf.net>
  * Added clock start/stop to avoid harmless warnings

2007-09-13 Stephane Popinet <popinet at users.sf.net>
  * Error message when trying to partition a domain composed of too few boxes

2007-09-12 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for bad comment indentation in parameter files views

2007-09-11 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for SourceControl divide by zero

2007-09-07 Stephane Popinet <popinet at users.sf.net>
  * Minor maintenance (comments, valgrind stuff etc...)

2007-09-09 Stephane Popinet <popinet at users.sf.net>
  * Automake 1.6 is OK (I guess 1.4 is not)

2007-09-08 Stephane Popinet <popinet at users.sf.net>
  * Fixed boundary conditions for boussinesq example
  
  As pointed out by Phil Rubini, the previous boundary conditions were 
  misleading: they were equivalent to flow in a closed box rather than the
  expected vertical open channel.

2007-09-08 Stephane Popinet <popinet at users.sf.net>
  * Fixed typo in oscillation.gfs

2007-09-07 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases references

2007-09-07 Stephane Popinet <popinet at users.sf.net>
  * Test cases are run in the order they appear in the final report

2007-09-07 Stephane Popinet <popinet at users.sf.net>
  * Added 'Define' to known classes

2007-09-07 Stephane Popinet <popinet at users.sf.net>
  * Fixed formatting of gfs2tex parameter file view

2007-09-07 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for uninitialised normal z-component in parabola_fit_init()
  
  ..or how 4 characters can lead to no end of trouble...

2007-09-06 Stephane Popinet <popinet at users.sf.net>
  * Macro support needs to be turned on explicitly
  
  To avoid m4 problems when trying to apply macros to 
  binary simulation files.

2007-09-06 Stephane Popinet <popinet at users.sf.net>
  * Fix for plotting style of reynolds test case

2007-09-06 Stephane Popinet <popinet at users.sf.net>
  * New inviscid droplet oscillation test case

2007-09-05 Stephane Popinet <popinet at users.sf.net>
  * Updated spurious and capwave test cases

2007-09-05 Stephane Popinet <popinet at users.sf.net>
  * 'kinetic' test case uses a mollified volume fraction

2007-09-05 Stephane Popinet <popinet at users.sf.net>
  * Macro support in parameter files

2007-09-05 Stephane Popinet <popinet at users.sf.net>
  * Checks that automake version is >= 1.8

2007-09-05 Stephane Popinet <popinet at users.sf.net>
  * Added assertions alpha > 0.
  
  Alpha (1/rho) is user defined and could take negative values if care is not
  taken.

2007-08-30 Stephane Popinet <popinet at users.sf.net>
  * gradient_fine_coarse() uses only directly neighbouring cells
  
  Using a third cell in the opposite direction, although nominally second-order
  in space could cause instabilities. The reduced spatial order of this new
  implementation does not seem to affect the convergence rates for the test cases
  in the test suite.

2007-09-03 Stephane Popinet <popinet at users.sf.net>
  * Dependency fix for order of variables when writing simulation files
  
  Physical parameters can depend on defined variables. The incorrect order
  could prevent from restarting a simulation.

2007-08-31 Stephane Popinet <popinet at users.sf.net>
  * Initial projection is skipped when restarting a simulation

2007-08-31 Stephane Popinet <popinet at users.sf.net>
  * Constants can be expressions (not for all objects yet)

2007-08-19 Stephane Popinet <popinet at users.sf.net>
  * VariableFiltered was not taking the number of iterations into account

2007-08-30 Stephane Popinet <popinet at users.sf.net>
  * Updated CSS stylesheet for consistency with wiki

2007-08-29 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for adaptation with 'maxcells' set

2007-08-29 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for memory corruption when trying to do "twisted" streamlines in 2D

2007-08-22 Stephane Popinet <popinet at users.sf.net>
  * Major bug fix for VOF advection scheme
  
  The previous implementation did not take into account properly the
  compression/expansion of cell volumes at each step of the split
  scheme. As a result the overall scheme had very poor mass
  conservation.

2007-07-24 Stephane Popinet <popinet at users.sf.net>
  * New test case for kinetic energy conservation in multiphase flows

2007-05-24 Stephane Popinet <popinet at users.sf.net>
  UNDO: Refines cells which are too coarse for VOF advection
  
  Cells are too coarse when one of their neighboring cells is finer and
  contains and interface which will be advected in them at the next
  timestep.

2007-05-24 Stephane Popinet <popinet at users.sf.net>
  * Refines cells which are too coarse for VOF advection
  
  Cells are too coarse when one of their neighboring cells is finer and
  contains and interface which will be advected in them at the next
  timestep.

2007-08-22 Stephane Popinet <popinet at users.sf.net>
  * Updated spurious currents test case

2007-08-22 Stephane Popinet <popinet at users.sf.net>
  * Updated 'shear' VOF advection test case
  
  Including test for mass conservation.

2007-08-22 Stephane Popinet <popinet at users.sf.net>
  * New VOF advection test case with adaptive refinement

2007-08-22 Stephane Popinet <popinet at users.sf.net>
  * Ignore white spaces at the start of a parameter file

2007-05-30 Stephane Popinet <popinet at users.sf.net>
  * Capillary wave test case uses a larger domain aspect ratio
  
  This improves the convergence significantly and the results
  compare well with those of Gerlach et al (added in the biblio).

2007-05-23 Stephane Popinet <popinet at users.sf.net>
  * Face curvature does not use vof-weighted averaging anymore

2007-08-15 Stephane Popinet <popinet at users.sf.net>
  * Automated darcs versioning ignores changes in the doc/ directory

2007-05-25 Stephane Popinet <popinet at users.sf.net>
  * Kmax values for "nearly empty" interfacial cells are filtered

2007-05-23 Stephane Popinet <popinet at users.sf.net>
  * VariableCurvature optionally computes the maximum curvature

2007-05-22 Stephane Popinet <popinet at users.sf.net>
  * Cleanup of HF curvature calculation

2007-05-22 Stephane Popinet <popinet at users.sf.net>
  * Several improvements to height-function curvature calculation
  
  - Interface positions are taken into account only if they are far enough apart.
  - Nearest neighbour interpolation is used before resorting to facet-fitting.
  - Various options for circle-fitting and paraboloids of different degrees etc...

2007-05-10 Stephane Popinet <popinet at users.sf.net>
  * Paraboloid fitting for curvature calculation in 3D

2007-05-05 Stephane Popinet <popinet at users.sf.net>
  * Parabola fitting uses normal direction and local interface position

2007-05-03 Stephane Popinet <popinet at users.sf.net>
  * Parabola fitting for gfs_height_curvature()

2007-04-30 Stephane Popinet <popinet at users.sf.net>
  * More robust implementation of gfs_height_curvature()
  
  Tries all directions of integration (based on normal orientation) rather than
  only the first guess.

2007-08-12 Stephane Popinet <popinet at users.sf.net>
  * Fix for ffmpeg options and mktemp in ppm2mpeg
  
  Newer ffmpeg versions seem to have stricter syntax for the bitrate.
  
  Also mktemp does not seem to accept the "-t" option on all systems.

2007-07-28 Stephane Popinet <popinet at users.sf.net>
  * BcNavier for Navier slip condition

2007-07-26 Stephane Popinet <popinet at users.sf.net>
  * Explicit message when trying to use implicit surfaces with RefineSolid

2007-08-07 Stephane Popinet <popinet at users.sf.net>
  * Fixed comment formatting for logo example

2007-08-07 Stephane Popinet <popinet at users.sf.net>
  * Testing framework now depends on bash
  
  So that we can use the "set -o pipefail" option to be sure to capture all
  errors when running a test command which is made of independent piped 
  commands (e.g. "gerris2D logo.gfs | gfsview2D logo.gfv").

2007-08-02 Stephane Popinet <popinet at users.sf.net>
  * Improved error checking for ppm2mpeg

2007-07-06 Stephane Popinet <popinet at users.sf.net>
  * Fix for corner case in face_bilinear()

2007-06-09 Stephane Popinet <popinet at users.sf.net>
  * New object GfsSourceViscosityExplicit

2007-05-23 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for marginal case in myc (3D version)

2007-05-23 Stephane Popinet <popinet at users.sf.net>
  * GfsFunction does not use variable index anymore
  
  The index of a variable can vary during the simulation. Use the
  variable's address instead.

2007-05-22 Stephane Popinet <popinet at users.sf.net>
  * Fix for degenerate case in gfs_domain_remove_droplets

2007-05-20 Stephane Popinet <popinet at users.sf.net>
  * G_MAXDOUBLE values are ignored by gfs_norm_add() and gts_range_add_value()

2007-04-27 Stephane Popinet <popinet at users.sf.net>
  * Fix for inconsistent momentum diffusion
  
  When using SourceDiffusion for variable density flows
  (thanks to Shahriar Afkhami for reporting this).

2007-04-27 Stephane Popinet <popinet at users.sf.net>
  * New exported function gfs_multilevel_params_stats_write()

2007-04-26 Stephane Popinet <popinet at users.sf.net>
  * Reformatted bibtex reference

2007-04-25 Stephane Popinet <popinet at users.sf.net>
  * More detailed versioning of test cases summary

2007-04-25 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for GfsSurface transformations

2007-04-24 Stephane Popinet <popinet at users.sf.net>
  * Moved 'shear' example to test cases

2007-04-24 Stephane Popinet <popinet at users.sf.net>
  * Fix for gfs2doc when used on test cases

2007-04-24 Stephane Popinet <popinet at users.sf.net>
  * Updated examples to use implicit surfaces

2007-04-24 Stephane Popinet <popinet at users.sf.net>
  * Configure checks for linker flags required for module compilation
  
  This should fix issues on the various versions of Mac OSX.

2007-04-24 Stephane Popinet <popinet at users.sf.net>
  * Rotations for GfsSurface are defined only via principal axis

2007-04-24 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for Fedora tarball

2007-04-23 Stephane Popinet <popinet at users.sf.net>
  * Minor bug fixes

2007-04-20 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for expression parsing

2007-04-20 Stephane Popinet <popinet at users.sf.net>
  * Updated wave tests to use implicit surfaces

2007-04-20 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_surface_segment_normal()

2007-04-20 Stephane Popinet <popinet at users.sf.net>
  * New predefined spatial function 'sphere()'

2007-04-20 Stephane Popinet <popinet at users.sf.net>
  * Updated tests with new implicit surfaces

2007-04-20 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for "flattened" surfaces in RefineSurface

2007-04-20 Stephane Popinet <popinet at users.sf.net>
  * Secant/bisection root-finding for implicit surfaces

2007-04-18 Stephane Popinet <popinet at users.sf.net>
  * Solid boundaries can be defined using implicit surfaces
  
  # a cylinder of radius 0.0625
  Solid (x*x + y*y - 0.0625*0.0625)
  
  This can also be used to define volume fractions (using GfsInitFraction).

2007-04-20 Stephane Popinet <popinet at users.sf.net>
  * Docs are not included in the distribution tarball anymore

2007-04-17 Stephane Popinet <popinet at users.sf.net>
  * All surfaces are now defined using a new GfsSurface object
  
  Note that this is a new object, the previous "GfsSurface" has been replaced with
  "GfsSolid".

2007-04-16 Stephane Popinet <popinet at users.sf.net>
  * Detailed versioning based on darcs

2007-04-16 Stephane Popinet <popinet at users.sf.net>
  * darcs2dist checks that the tarball builds properly

2007-04-16 Stephane Popinet <popinet at users.sf.net>
  * Added missing desktop files to tarball distribution

2007-04-13 Stephane Popinet <popinet at users.sf.net>
  * GfsSurface allows transformations of the input GTS surface

2007-04-13 Stephane Popinet <popinet at users.sf.net>
  * New GfsSurface class replaces GtsSurfaceFile and GtsSurface
  
  The GtsSurfaceFile keyword is still supported, however the GtsSurface keyword is
  not and should be replaced by "GfsSurface {}" in old simulation files.

2007-04-13 Stephane Popinet <popinet at users.sf.net>
  * New tool gfs2gfs to convert old simulations to the current format

2007-04-13 Stephane Popinet <popinet at users.sf.net>
  * First parameter block of GfsEvent is now optional
  
  Provided it is not followed by another parameter block.

2007-04-12 Stephane Popinet <popinet at users.sf.net>
  * curvature_coarse_fine() did not work when using a VariableDistance

2007-04-12 Stephane Popinet <popinet at users.sf.net>
  * Simplified layered (2D3) GfsOcean implementation
  
  A first step towards a full 3D (non-layered) ocean model.
  
  Note that variables "HU" and "HV" are no longer defined. "U" and "V"
  should be used instead (particularly for Flather BCs in parameter
  files).

2007-04-12 Stephane Popinet <popinet at users.sf.net>
  * Updated nz test case

2007-04-12 Stephane Popinet <popinet at users.sf.net>
  * Updated circle/refined test case for new fine/coarse gradient calculation

2007-04-11 Stephane Popinet <popinet at users.sf.net>
  * Changed fine_coarse gradient calculation
  
  Fine cells on the left-hand-side of the fine-coarse boundary are taken
  into account only if they are not mixed. This is to ensure the
  stability of the GfsOcean model (the nz test case).

2007-04-11 Stephane Popinet <popinet at users.sf.net>
  * New adaptive coastally-trapped wave test case

2007-04-11 Stephane Popinet <popinet at users.sf.net>
  * New gravity waves in Cook strait test case

2007-04-05 Stephane Popinet <popinet at users.sf.net>
  * Simplified Poisson coefficients calculation

2007-03-29 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for gfs_domain_traverse_cut_2D()

2007-03-29 Stephane Popinet <popinet at users.sf.net>
  * Merged advection fluxes redistribution is done differently
  
  This mitigates the appearance of negative values for positive tracers near
  solid boundaries.

2007-03-27 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for non-initialised fractions in GfsOcean

2007-03-27 Stephane Popinet <popinet at users.sf.net>
  * New derived variables for surface fractions

2007-03-27 Stephane Popinet <popinet at users.sf.net>
  * Adaptivity along solid boundaries should now work (but not for 2D3 yet)

2007-03-23 Stephane Popinet <popinet at users.sf.net>
  * Added check for positiveness of solid->s[0]

2007-03-23 Stephane Popinet <popinet at users.sf.net>
  * ftt_cell_locate() works for 2D3

2007-03-20 Stephane Popinet <popinet at users.sf.net>
  UNDO: Source terms for non-advected variables are taken into account

2007-03-20 Stephane Popinet <popinet at users.sf.net>
  * Source terms for non-advected variables are taken into account

2007-03-20 Stephane Popinet <popinet at users.sf.net>
  * Improved error messages for Cartesian grid interpolation

2007-03-19 Cedric Penard <cedric.penard at ifremer.fr>
  * Cartesian functions

2007-03-19 Stephane Popinet <popinet at users.sf.net>
  * gfs_plane_center() works with any normal
  
  Not just with normals in the positive quadrant.

2007-03-19 Stephane Popinet <popinet at users.sf.net>
  * New functions gfs_vof_plane_facet() and gfs_vof_plane_center()

2007-03-18 Stephane Popinet <popinet at users.sf.net>
  * Simplified circular wave test case (using GfsGlobal)

2007-03-16 Stephane Popinet <popinet at users.sf.net>
  * "Permanent" cells cannot be coarsened
  
  At the moment this is used to prevent coarsening of the initial
  embedded solid surface: a first step towards adaptivity of embedded
  solid surfaces.

2007-03-16 Stephane Popinet <popinet at users.sf.net>
  * Special 2D3 case for interpolate_1D1 upwinding

2007-03-16 Stephane Popinet <popinet at users.sf.net>
  * Initial timestep is set to zero (not one)

2007-03-16 Stephane Popinet <popinet at users.sf.net>
  * Small fix for GfsCartesianGrid

2007-03-15 Stephane Popinet <popinet at users.sf.net>
  * New "twod" option for GfsRefineSurface

2007-03-15 Stephane Popinet <popinet at users.sf.net>
  * Merged cells work for 2D3

2007-03-15 Stephane Popinet <popinet at users.sf.net>
  * Small cells cannot be close to boundaries

2007-03-15 Stephane Popinet <popinet at users.sf.net>
  * New derived variable "dt" (timestep)

2007-03-14 Stephane Popinet <popinet at users.sf.net>
  * GfsOcean uses new variables "HU" and "HV" to store the barotropic velocity field
  
  This simplifies the implementation of Flather boundary conditions.

2007-03-13 Stephane Popinet <popinet at users.sf.net>
  * Cleanup of GfsCartesianGrid

2007-03-13 Cedric Penard <cedric.penard at ifremer.fr>
  * Fix for merge with stable branch

2007-03-13 C[_\c3_][_\a9_]dric P[_\c3_][_\a9_]nard <cedric.penard at ifremer.fr>
  * New object GfsCartesianGrid

2007-03-13 Stephane Popinet <popinet at users.sf.net>
  * Updated couette.gfs
  
  Comments in functions must now use the C format ('/*' not '#').

2007-03-09 Stephane Popinet <popinet at users.sf.net>
  * Further bug fixing for comments in GfsFunction expressions

2007-03-09 Stephane Popinet <popinet at users.sf.net>
  * Updated examples doc

2007-03-08 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for Dirichlet BC for VOF tracers

2007-03-08 Stephane Popinet <popinet at users.sf.net>
  * GfsVariablePosition takes an optional 'ref' argument

2007-03-05 Stephane Popinet <popinet at users.sf.net>
  * Updated Rayleigh-Taylor example

2007-03-04 Stephane Popinet <popinet at users.sf.net>
  * Fix for reseting 'status' when computing solid fractions

2007-03-04 Stephane Popinet <popinet at users.sf.net>
  * Init events are turned into standard events when specifying 'step'

2007-03-04 Stephane Popinet <popinet at users.sf.net>
  * Updated shear example

2006-12-02 Stephane Popinet <popinet at users.sf.net>
  UNDO: Init events are also realised at half timesteps

2007-03-02 Stephane Popinet <popinet at users.sf.net>
  * Fixed dependencies for gfsclasses.py

2007-03-02 Stephane Popinet <popinet at users.sf.net>
  * Cleanup of Couette flow test case

2007-03-02 Stephane Popinet <popinet at users.sf.net>
  * Fix for face interpolation of diffusion coefficients

2007-03-02 Stephane Popinet <popinet at users.sf.net>
  * Unified script/code treatment for GfsGlobal, GfsEventScript and GfsOutput
  
  This fixes a bug in GfsOutput and GfsEventScript where "#!" were
  treated as comments.

2007-03-02 Stephane Popinet <popinet at users.sf.net>
  * Debian package recommends various bits and bobs

2007-03-02 Stephane Popinet <popinet at users.sf.net>
  * Updated surface tension test cases

2007-03-02 Stephane Popinet <popinet at users.sf.net>
  * Added Gerris logo example

2007-03-01 Stephane Popinet <popinet at users.sf.net>
  * Fixed face_bc() for GfsBcDirichlet

2007-03-01 Stephane Popinet <popinet at users.sf.net>
  * VOF normal is computed using "myc" in 2D (thanks to Ruben Scardovelli)

2007-02-28 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases

2007-02-28 Stephane Popinet <popinet at users.sf.net>
  * Temporary bug fix for GfsEventStop
  
  For more "complex" variables (e.g. VariableTracerVOF) EventStop will
  not work properly if adaptivity is used. There is no simple solution
  to this problem.

2007-02-28 Stephane Popinet <popinet at users.sf.net>
  * Added icons for mime types

2007-02-27 Stephane Popinet <popinet at users.sf.net>
  * Fix for merge with CSF branch

2007-02-27 Stephane Popinet <popinet at users.sf.net>
  * Added tolerance to avoid "missed" synchronised events

2007-02-27 Stephane Popinet <popinet at users.sf.net>
  * Fix for overflow exception

2007-02-22 Stephane Popinet <popinet at users.sf.net>
  * Updated gfs_classes()

2007-02-12 Stephane Popinet <popinet at users.sf.net>
  * Mixed Youngs-Centered VOF normal calculation

2007-02-09 Stephane Popinet <popinet at users.sf.net>
  * Symmetry boundary conditions for VOF
  
  These are approximate only, in particular the boundary conditions for
  the normal should be anti-symmetric not symmetric (as implemented in
  this patch). The effect of this approximation should be small however.
  
  Also, not sure what the symmetry conditions actually do when combined
  with complex solid boundaries.

2007-02-09 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for periodic boundary conditions

2007-02-07 Stephane Popinet <popinet at users.sf.net>
  * Fix for small VOF fragments in vof_fine_coarse()

2007-01-31 Stephane Popinet <popinet at users.sf.net>
  * Simplified and more robust Youngs stencil routine

2007-01-30 Stephane Popinet <popinet at users.sf.net>
  * Removed obsolete AdaptNotBox stuff

2007-01-28 Stephane Popinet <popinet at users.sf.net>
  * VOF Tracers are now defined using GfsVariableTracerVOF
  
  Normal directions and alpha are pre-computed and stored together with the VOF 
  fraction.

2007-01-27 Stephane Popinet <popinet at users.sf.net>
  * Stability criterion for gravity waves only takes into account interface cells

2006-12-11 Stephane Popinet <popinet at users.sf.net>
  * VariableCurvature uses either height-function method or levelset
  
  According to the arguments (i.e. the second argument specifies a 
  VariableTracer or a VariableDistance).

2006-12-06 Stephane Popinet <popinet at users.sf.net>
  * New object GfsVariablePosition
  
  For implementation of "reduced gravity".

2006-08-13 Stephane Popinet <popinet at users.sf.net>
  * Periodic boundary conditions do not use connectivity
  
  This made calculating geometric relationships difficult (e.g. distance between
  cells etc...). Periodic links are now automatically replaced by a new type of
  boundary condition (GfsBoundaryPeriodic). This is also used as a base class 
  for a simplified GfsBoundaryMPI implementation.
  
  This should fix the bug when using the diffusion solver with embedded solid 
  boundaries combined with periodic boundary conditions.

2006-08-14 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_domain_filter()

2006-04-22 Stephane Popinet <popinet at users.sf.net>
  * New object GfsEventSumDirection

2007-02-22 Stephane Popinet <popinet at users.sf.net>
  * gfs_simulation_adapt() does not take a "stats" parameter anymore

2007-02-13 Stephane Popinet <popinet at users.sf.net>
  * Removed warning for interpolation of GTS files in functions

2007-02-01 Stephane Popinet <popinet at users.sf.net>
  * Merged advection fluxes were not correctly weighted
  
  For solid boundaries with a variable resolution.

2007-01-28 Stephane Popinet <popinet at users.sf.net>
  * Updated SourceTensionCSS to reflect new staggered timestepping of tracers

2007-01-28 Stephane Popinet <popinet at users.sf.net>
  * Initialisation of simulation is done using gfs_simulation_init()

2007-01-21 Stephane Popinet <popinet at users.sf.net>
  * Optimised implementation of ftt_cell_locate()

2006-12-02 Stephane Popinet <popinet at users.sf.net>
  * Init events are also realised at half timesteps

2007-01-19 Stephane Popinet <popinet at users.sf.net>
  * VOF works (as in "does not crash") with embedded solid boundaries

2007-01-19 Stephane Popinet <popinet at users.sf.net>
  * New gravity wave test case

2007-01-19 Stephane Popinet <popinet at users.sf.net>
  * Updated capillary wave test cases

2007-01-19 Stephane Popinet <popinet at users.sf.net>
  * Tracers are now defined at time n+1/2
  
  The resulting "leap-frog" time integration allows to recover second-order 
  accuracy for variable-density and surface-tension-driven flows.

2007-01-18 Stephane Popinet <popinet at users.sf.net>
  * Fix for face coordinates calculation in GfsFunction

2006-12-22 Stephane Popinet <popinet at users.sf.net>
  * New test case for capillary wave with fluids of different densities

2006-12-19 Stephane Popinet <popinet at users.sf.net>
  * gfs_vof_facet() could fail for degenerate cases

2006-12-19 Stephane Popinet <popinet at users.sf.net>
  * Changed the format used by gfs_write_mac_velocity() to gnuplot

2006-12-19 Stephane Popinet <popinet at users.sf.net>
  * Default CFL for VOF advection has been set to 0.45
  
  To have some margin. Also a warning has been added in case the true
  CFL is larger than 0.5 (when doing the VOF advection proper).

2006-12-14 Stephane Popinet <popinet at users.sf.net>
  * Density (and viscosity) is computed using face-interpolated volume fraction values
  
  This uses the new gfs_function_face_value(). Large density ratio flows
  can now be computed correctly. Needs further testing.

2006-12-14 Stephane Popinet <popinet at users.sf.net>
  * Updated Rayleigh-Taylor example
  
  Now it really works, thanks to the new density calculation. The results
  compare well to other simulations (Surfer and Marker code results in my
  thesis).

2006-12-14 Stephane Popinet <popinet at users.sf.net>
  * gfs_function_face_value() uses the face-interpolated values of variables

2006-12-11 Stephane Popinet <popinet at users.sf.net>
  * Added description for VariableDistance

2006-12-11 Stephane Popinet <popinet at users.sf.net>
  * Fixed constant for stability constraint of surface tension

2006-12-06 Stephane Popinet <popinet at users.sf.net>
  * Functions to compute cell/plane intersections have been moved from GfsView
  
  They are used to compute 3D VOF facets.
  Note also that the interface of gfs_vof_facet() has changed.

2006-12-05 Stephane Popinet <popinet at users.sf.net>
  * "Non-hydrostatic pressure" implementation
  
  This is useful for two-phase flows with a (flattish) interface, variable 
  density and gravity. The hydrostatic pressure is substracted off and the 
  remaining pressure jump across the interface due to gravity is discretised 
  using the same technique as for surface tension. This guarantees an accurate 
  solution free of spurious (gravity) currents.

2006-12-01 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for VOF boundary conditions

2006-11-28 Stephane Popinet <popinet at users.sf.net>
  * VOF plane is defined in full cells
  
  This prevents floating point exceptions when assigning undefined values to VOF
  plane variables.

2006-11-28 Stephane Popinet <popinet at users.sf.net>
  * Code cleanup for tension coefficients

2006-11-28 Stephane Popinet <popinet at users.sf.net>
  * Generic boundary conditions cannot be applied for VOF fractions
  
  A special case treatment has been added for symmetry and Dirichlet conditions.

2006-11-24 Stephane Popinet <popinet at users.sf.net>
  * Multiple SourceTension terms are permitted
  
  For example a combination of surface tension and "reduced gravity".

2006-11-23 Stephane Popinet <popinet at users.sf.net>
  * SourceTension can be used to implement "reduced gravity"

2006-11-23 Stephane Popinet <popinet at users.sf.net>
  * Maximum CFL is 0.5 for VOF

2006-11-07 Stephane Popinet <popinet at users.sf.net>
  * More robust algorithm for computation of local interface height

2006-11-03 Stephane Popinet <popinet at users.sf.net>
  * Height-Function curvature calculation should now work in 3D
  
  and also on 3D adaptive grids.

2006-11-03 Stephane Popinet <popinet at users.sf.net>
  * Limit on the size of the HF-curvature calculation "stencil"

2006-11-02 Stephane Popinet <popinet at users.sf.net>
  * 3D implementation of gfs_youngs_normal()

2006-11-02 Stephane Popinet <popinet at users.sf.net>
  * "Exact" implementation of gfs_plane_alpha() in 3D
  
  This replaces the Newton iterations version. It is much more accurate and
  hence ensures better volume conservation.

2006-11-02 Stephane Popinet <popinet at users.sf.net>
  * Singular 3D VOF volume calculations use the 2D algorithm
  
  This guarantees exact volume representation.

2006-11-01 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for VOF advection with periodic boundary conditions

2006-10-23 Stephane Popinet <popinet at users.sf.net>
  * Added missing variable descriptions

2006-10-22 Stephane Popinet <popinet at users.sf.net>
  * Important bug fix for boundary conditions on alpha

2006-10-21 Stephane Popinet <popinet at users.sf.net>
  * Fixed VOF-plane calculation for isolated interface fragments
  
  The zero-norm of the singular interface normal in this case was causing NaN
  problems.

2006-10-20 Stephane Popinet <popinet at users.sf.net>
  * Fixed VOF advection boundary conditions

2006-10-20 Stephane Popinet <popinet at users.sf.net>
  * Capillary wave test case took too long to run with 7 levels

2006-10-20 Stephane Popinet <popinet at users.sf.net>
  * Figure for PLIC height calculation

2006-10-20 Stephane Popinet <popinet at users.sf.net>
  * Updated spurious currents test case

2006-10-19 Stephane Popinet <popinet at users.sf.net>
  * Added warnings when HF-curvature fails

2006-10-19 Stephane Popinet <popinet at users.sf.net>
  * SourceTension and SourceTensionCSS are derived from a new SourceTensionGeneric object
  
  Also SourceTensionCSS "works" again thanks to the new function 
  gfs_youngs_gradient (actually a renamed version of the old 
  gfs_youngs_normal implementation).

2006-10-18 Stephane Popinet <popinet at users.sf.net>
  * GfsSourceTension takes an extra 'sigma' argument

2006-10-18 Stephane Popinet <popinet at users.sf.net>
  * Updated capillary wave test case

2006-10-18 Stephane Popinet <popinet at users.sf.net>
  * Cleanup of HF-curvature implementation

2006-10-18 Stephane Popinet <popinet at users.sf.net>
  * HF curvature calculation works on adaptive meshes!
  
  But in 2D only for the moment. This patch also contains various (unused) 
  routines implementing other ways of computing the curvature (based on VOF 
  interface reconstructions, mean-square parabola and circle fit  etc...)

2006-10-08 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_vof_interpolate()

2006-10-06 Stephane Popinet <popinet at users.sf.net>
  * New functions gfs_fit_curvature() and gfs_shahriar_curvature()

2006-10-06 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_domain_boundary_locate()

2006-08-15 Stephane Popinet <popinet at users.sf.net>
  * VariableCurvature uses height method rather than levelset

2006-08-15 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_height_curvature()
  
  Only works in 2D and on non-refined grids for the moment.

2006-08-15 Stephane Popinet <popinet at users.sf.net>
  * Normal computation with the "column" method

2006-08-15 Stephane Popinet <popinet at users.sf.net>
  * Volume fraction weighting of surface tension term

2006-08-15 Stephane Popinet <popinet at users.sf.net>
  * Density is taken into account for surface tension term

2006-08-14 Stephane Popinet <popinet at users.sf.net>
  * New VOF advection implementation
  
  Works with variable interface resolution but not with solid boundaries yet.
  Uses "Eulerian" rather than "Lagrangian" PLIC advection.

2006-08-14 Stephane Popinet <popinet at users.sf.net>
  * GfsVariableCurvature has moved to levelset.c

2006-08-14 Stephane Popinet <popinet at users.sf.net>
  * Filtering of curvature along the interface
  
  For levelset curvature only and disabled by default.

2006-08-14 Stephane Popinet <popinet at users.sf.net>
  * New implementation of Youngs gradient
  
  Works on adaptive meshes but only in 2D for the moment.

2006-08-14 Stephane Popinet <popinet at users.sf.net>
  * Updated stability criterion for surface tension

2006-08-14 Stephane Popinet <popinet at users.sf.net>
  * Tracer advection does not save "half" values anymore

2006-05-23 Stephane Popinet <popinet at users.sf.net>
  * gfs_line_alpha() uses explicit formula
  
  ... rather than Newton iterations.

2006-02-22 Stephane Popinet <popinet at users.sf.net>
  * VOF should work across refinement levels
  
  Also gfs_plane_volume() and gfs_plane_alpha() now do their own
  symmetries (i.e. do not require m.x, m.y and m.z to be positive
  anymore).

2006-02-21 Stephane Popinet <popinet at users.sf.net>
  * Simplification of gfs_plane_volume()

2006-02-06 Stephane Popinet <popinet at users.sf.net>
  * CSF uses volume-fraction-weighted tension terms

2006-02-05 Stephane Popinet <popinet at users.sf.net>
  * VariableCurvature must be interpolated with care on adaptive mesh

2006-02-05 Stephane Popinet <popinet at users.sf.net>
  * Sigma and theta have nothing to do in VariableCurvature
  
  ...they will be back in a latter version of SourceTension.

2006-02-05 Stephane Popinet <popinet at users.sf.net>
  * Updated spurious current test case

2006-02-05 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for VariableCurvature computation near boundaries

2006-02-04 Stephane Popinet <popinet at users.sf.net>
  * VariableDistance computes the distance function only locally
  
  i.e. in a neighbourhood of the interface just large enough for
  VariableCurvature.

2006-02-04 Stephane Popinet <popinet at users.sf.net>
  * New functions gfs_center_gradient_stencil() and gfs_interpolate_stencil()

2006-02-03 Stephane Popinet <popinet at users.sf.net>
  * For curvature estimation, normals are computed using simple gradients
  
  ...rather than Youngs gradients. This stabilises the surface tension,
  improves the spurious currents and it is simpler.

2006-02-03 Stephane Popinet <popinet at users.sf.net>
  * Renamed VariableLevelSet as VariableDistance

2006-02-03 Stephane Popinet <popinet at users.sf.net>
  * Back to semi-implicit surface tension for capillary wave test case
  
  Although the explicit scheme appeared to be more accurate it was just
  by chance (i.e. the error increased when the timestep decreased).

2006-02-03 Stephane Popinet <popinet at users.sf.net>
  * VariableCurvature is defined only on the interface
  
  It now is defined as the curvature of the piece of interface cutting
  the cell. If the cell is not cut by the interface it is set to
  G_MAXDOUBLE.

2006-02-03 Stephane Popinet <popinet at users.sf.net>
  * VOF-based levelset function computation is back
  
  It works well with the new curvature calculation.

2006-02-02 Stephane Popinet <popinet at users.sf.net>
  UNDO: Levelset function is computed using the VOF-reconstructed interface

2006-02-02 Stephane Popinet <popinet at users.sf.net>
  * Youngs-averaging of normal divergence was unstable in some cases

2006-02-02 Stephane Popinet <popinet at users.sf.net>
  * The distance function value was not always correct for cells cut by the isoline

2006-02-02 Stephane Popinet <popinet at users.sf.net>
  * Control of "implicitness" of VariableCurvature

2006-02-01 Stephane Popinet <popinet at users.sf.net>
  * "Semi-implicit" scheme for VariableCurvature

2006-01-30 Stephane Popinet <popinet at users.sf.net>
  * VariableCurvature uses Youngs divergence
  
  Rather than gfs_center_gradient(). This improves the spurious currents slightly.

2006-02-02 Stephane Popinet <popinet at users.sf.net>
  * Stability condition for surface tension was still too restrictive

2006-02-02 Stephane Popinet <popinet at users.sf.net>
  * Levelset function is computed using the VOF-reconstructed interface

2006-02-02 Stephane Popinet <popinet at users.sf.net>
  * Updated spurious and capwave test cases

2006-02-02 Stephane Popinet <popinet at users.sf.net>
  * Stability constraint for surface tension was too restrictive

2006-01-30 Stephane Popinet <popinet at users.sf.net>
  * K8 smoothing kernel doesn't work
  
  Bubbles appear to be unstable for all smoothing lengths larger than twice the
  grid size. For these lengths K8 does not improve over the default vertex filter.

2006-01-30 Stephane Popinet <popinet at users.sf.net>
  * VariableFiltered uses K8 smoothing kernel

2006-01-30 Stephane Popinet <popinet at users.sf.net>
  * Levelset sign was not computed properly

2006-01-30 Stephane Popinet <popinet at users.sf.net>
  * VariableLevelSet computes the levelset function

2006-01-29 Stephane Popinet <popinet at users.sf.net>
  * Stability criterion for CSF surface tension

2006-01-29 Stephane Popinet <popinet at users.sf.net>
  * Source terms have a new generic stability() method

2006-01-29 Stephane Popinet <popinet at users.sf.net>
  * CSF surface tension implementation
  
  This is the default. The old CSS implementation (2D only) is still available as
  SourceTensionCSS.

2007-02-27 Stephane Popinet <popinet at users.sf.net>
  * Added missing html doc titles

2007-02-22 Stephane Popinet <popinet at users.sf.net>
  * Added mime-types

2007-02-13 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for mixed-cells output in gfs2oogl

2007-02-21 Stephane Popinet <popinet at users.sf.net>
  * darcs2dist creates 'versioned' snapshots

2007-02-21 Stephane Popinet <popinet at users.sf.net>
  * MacOSX does not like nested functions
  
  They were used to implement dx(), dy(), dz() in GfsFunction. They have been
  replaced by using ugly (but portable) global variables.

2007-02-21 Stephane Popinet <popinet at users.sf.net>
  * Fixed labels in tutorial

2007-02-21 Stephane Popinet <popinet at users.sf.net>
  * Fix for darcs2dist

2007-02-20 Stephane Popinet <popinet at users.sf.net>
  * Bigger fonts and thicker lines for lid test figures

2007-02-20 Stephane Popinet <popinet at users.sf.net>
  * Added anchor points for some sections of the tutorial

2007-02-20 Stephane Popinet <popinet at users.sf.net>
  * Workaround for character encoding bug in hevea 1.08

2007-02-21 Stephane Popinet <popinet at users.sf.net>
  * Updated source doc templates

2007-02-20 Stephane Popinet <popinet at users.sf.net>
  * FAQ is now in the wiki not in the source doc

2007-02-20 Stephane Popinet <popinet at users.sf.net>
  * GfsOutputAdaptStats works again

2007-02-19 Stephane Popinet <popinet at users.sf.net>
  * Option for color names in tests documentation

2007-02-19 Stephane Popinet <popinet at users.sf.net>
  * Parameter files in docs now link to "Object Hierarchy" wiki

2007-02-16 Stephane Popinet <popinet at users.sf.net>
  * Uses hevea rather than latex2html to generate HTML docs

2007-02-15 Stephane Popinet <popinet at users.sf.net>
  * Poisson test cases were not reliable due to race conditions in command pipes

2007-02-15 Stephane Popinet <popinet at users.sf.net>
  * Much simplified GfsOutput command pipe implementation

2007-02-14 Stephane Popinet <popinet at users.sf.net>
  * Fix for variables in POSIX shells (for test suite)

2007-02-14 Stephane Popinet <popinet at users.sf.net>
  * Fix for getopt_long and unknown arguments

2007-02-14 Stephane Popinet <popinet at users.sf.net>
  * Updated Rayleigh-Taylor example

2007-02-14 Stephane Popinet <popinet at users.sf.net>
  * New ppm2mpeg helper script (uses ffmpeg)
  
  ffmpeg is better than mjpegtools and simpler to install (debian packages are
  readily available).

2007-01-31 Stephane Popinet <popinet at users.sf.net>
  * Python test script works with dash

2007-01-30 Stephane Popinet <popinet at users.sf.net>
  * New GfsGlobal object defines "global" functions

2007-01-19 Stephane Popinet <popinet at users.sf.net>
  * Improved gfs_domain_remove_droplets()

2007-01-18 Stephane Popinet <popinet at users.sf.net>
  * GfsAdaptNotBox is deprecated

2006-12-19 Stephane Popinet <popinet at users.sf.net>
  * Thin cells filled with fluid could create an incorrect topology

2006-12-19 Stephane Popinet <popinet at users.sf.net>
  * Replaced assertion with warning in solid fractions calculation

2006-12-11 Stephane Popinet <popinet at users.sf.net>
  * "levelmin" is taken into account only for coarsening the mesh

2006-12-06 Stephane Popinet <popinet at users.sf.net>
  * Fixed Tangaroa example doc

2006-12-06 Stephane Popinet <popinet at users.sf.net>
  * Default boundary conditions for velocity diffusion were not applied properly

2006-12-05 Stephane Popinet <popinet at users.sf.net>
  * Added Tangaroa example

2006-11-23 Stephane Popinet <popinet at users.sf.net>
  * Aborts if the Poisson solver fails to converge

2006-11-24 Stephane Popinet <popinet at users.sf.net>
  * Fixed function inlining mess

2006-11-24 Stephane Popinet <popinet at users.sf.net>
  * glib.h is not needed if gts.h is already included

2006-11-24 Stephane Popinet <popinet at users.sf.net>
  * Updated merging test case for new adaptive algorithm

2006-11-23 Stephane Popinet <popinet at users.sf.net>
  * Added "hysteresis" to adaptive algorithm
  
  This prevents cells being created and destroyed in quick succession.

2006-11-23 Stephane Popinet <popinet at users.sf.net>
  * Standard error should be used rather than standard output in the examples

2006-11-23 Stephane Popinet <popinet at users.sf.net>
  * GfsVariableResidual wasn't scaled properly

2006-11-22 Stephane Popinet <popinet at users.sf.net>
  * Adaptive algorithm uses a simpler technique by default
  
  The "global optimisation" technique is overkill when "maxcells" is not set.
  Also, it does not allow for the simple "boolean" combination of several 
  criteria which is possible with the simple algorithm.

2006-11-04 Stephane Popinet <popinet at users.sf.net>
  * Improved error messages for non-closed solid surfaces

2006-11-02 Stephane Popinet <popinet at users.sf.net>
  * Creeping Couette flow does not need advection terms

2006-11-02 Stephane Popinet <popinet at users.sf.net>
  * bat2gts sets higher output precision

2006-11-02 Stephane Popinet <popinet at users.sf.net>
  * gfs2oogl can output all variables in mixed cells
  
  Not just P and Vorticity.

2006-10-25 Stephane Popinet <popinet at users.sf.net>
  * Thin cells are replaced either with fluid cells or solid cells
  
  They were replaced blindly by solid cells which was a very bad
  approximation in some cases.

2006-10-25 Stephane Popinet <popinet at users.sf.net>
  * Removed non-standard C++-style comments

2006-10-21 Stephane Popinet <popinet at users.sf.net>
  * Gerris abort generates an error.gfs dump (for post-mortem debugging)

2006-10-19 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for GfsEventStop
  
  EventStop was using a temporary variable to store the previous timestep values.
  This variable was not initialised when adapting the mesh (temporary variables 
  are not initialised by default).
  
  This has been fixed by using a "real" variable instead but with a NULL name.
  
  This meant allowing variables with a NULL name i.e. hidden permanent variables.

2006-09-27 Stephane Popinet <popinet at users.sf.net>
  * Inlined functions are defined once in a separate file

2006-10-19 Stephane Popinet <popinet at users.sf.net>
  * Simplified Debian package dependencies (again)

2006-10-18 Stephane Popinet <popinet at users.sf.net>
  * Simplified Debian package dependencies
  
  To make them compatible with Ubuntu. It's a bit messy though as I am not sure
  what the issues are with binary compatibility of Ubuntu/Debian libraries etc...

2006-10-14 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for OutputPPM for 2D GfsOcean

2006-08-13 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for error handling in gfs_domain_read()

2006-08-12 Stephane Popinet <popinet at users.sf.net>
  * Removed DX support (use GfsView instead!)

2006-10-12 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases references for new timestep calculation

2006-10-12 Stephane Popinet <popinet at users.sf.net>
  * Fixed timestep calculation for 'infinite' events

2006-10-11 Stephane Popinet <popinet at users.sf.net>
  * Fixed stupid timestep calculation (thanks to Chris Johnson for the bug report)

2006-10-10 Stephane Popinet <popinet at users.sf.net>
  * Increased tolerance for capillary wave test case

2006-10-09 Stephane Popinet <popinet at users.sf.net>
  * Updated test cases references

2006-10-09 Stephane Popinet <popinet at users.sf.net>
  * Z-moments need to be computed in 2D for solid force

2006-10-06 Stephane Popinet <popinet at users.sf.net>
  * Fixed autogen.sh for Mac OSX

2006-10-06 Stephane Popinet <popinet at users.sf.net>
  * Fixed dependencies for Debian snapshots

2006-08-08 Stephane Popinet <popinet at users.sf.net>
  * Set the default minimum of iterations for multilevel solvers to one

2006-08-03 Stephane Popinet <popinet at users.sf.net>
  * Fixed obsolete simulation file in FAQ

2006-07-19 Stephane Popinet <popinet at users.sf.net>
  * Configure was not working properly with empty "modules" directory
  
  This was a problem only for systems not supporting dynamic modules.

2006-07-18 Stephane Popinet <popinet at users.sf.net>
  * Advection update using averaged values for momentum in mixed cells was ignored
  
  This caused the channel test case to fail. Ultimately however this average value
  hack should not be used at all.

2006-05-12 Stephane Popinet <popinet at users.sf.net>
  * Quirk's merged fluxes are used for tracer advection rather than averaging
  
  They should be used also for momentum advection but this needs testing.

2006-07-03 Stephane Popinet <popinet at users.sf.net>
  * Check for ensuring that solid fractions are positive
  
  Round-off errors could cause the solid fractions to be a small negative value. This could cause problems for
  algorithms assuming (correctly) that volume fraction is always positive.

2006-06-18 Stephane Popinet <popinet at users.sf.net>
  * Unbuffered python for test cases (to force "realtime" status)

2006-06-11 Stephane Popinet <popinet at users.sf.net>
  * Variables have a new "description" field

2006-06-11 Stephane Popinet <popinet at users.sf.net>
  * DerivedVariable is now a proper object
  
  ... and also has a new "description" field.

2006-05-28 Stephane Popinet <popinet at users.sf.net>
  * More explicit message for systems which do not support functions

2006-05-12 Stephane Popinet <popinet at users.sf.net>
  * Coarse cells on the Poisson hierarchy with only one neighbor are "turned off"
  
  these cells could cause poor convergence of the Poisson solver (see the dumbell
  test case).

2006-04-25 Stephane Popinet <popinet at users.sf.net>
  * Added "dumbbell" Poisson test case

2006-04-22 Stephane Popinet <popinet at users.sf.net>
  * GfsFunction understands dx(), dy() and dz()

2006-06-12 Stephane Popinet <popinet at users.sf.net>
  * Missing script in lid test case

2006-04-03 Stephane Popinet <popinet at users.sf.net>
  tagged release_0_9_2

2006-04-03 Stephane Popinet <popinet at users.sf.net>
  * Updated docs

2006-03-31 Stephane Popinet <popinet at users.sf.net>
  * Removed obsolete pdflatex doc stuff

2006-03-30 Stephane Popinet <popinet at users.sf.net>
  * Cells with solid volume fraction of 0 are removed

2006-03-24 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for relaxation on lowest level

2006-03-24 Stephane Popinet <popinet at users.sf.net>
  * Implicit Coriolis should now work with the Navier-Stokes solver

2006-03-24 Stephane Popinet <popinet at users.sf.net>
  * RefineDistance uses a lower-bound for the distance...
  
  ... rather than the distance from the center of the cell.

2006-03-15 Stephane Popinet <popinet at users.sf.net>
  * Thin cells are removed
  
  "Thin" cells are topologically complex cut cells which lead to
  inaccurate volume-of-fluid representation of solid surfaces.  In some
  cases the presence of these cells could lead to instabilities in the
  projection. In all cases they would lead to inaccurate velocities.

2006-03-15 Stephane Popinet <popinet at users.sf.net>
  * Updated tests for modified initial approx projection

2006-03-14 Stephane Popinet <popinet at users.sf.net>
  * RefineSolid defines a new derived variable "SolidCurvature"

2006-03-14 Stephane Popinet <popinet at users.sf.net>
  * New function gfs_solid_is_thin()

2006-03-14 Stephane Popinet <popinet at users.sf.net>
  * New interface for gfs_domain_add_derived_variable()

2006-03-14 Stephane Popinet <popinet at users.sf.net>
  * Fix for gradient computation at coarse/fine solid boundaries
  
  The previous version could use information from the wrong side of the
  solid surface when constructing interpolants for cells close to a
  solid boundary i.e. information was "leaking through" the solid
  surface. New weighting and checks with solid surface fractions should
  now avoid this.

2006-03-14 Stephane Popinet <popinet at users.sf.net>
  * Merged cells can be formed using children cells at coarse/fine interfaces

2006-03-10 Stephane Popinet <popinet at users.sf.net>
  * Time step is set before the first projection
  
  Previously the scaling for the tolerance on the divergence of the
  first projection was based on a timestep of unity (i.e. usually much
  larger than the subsequent timesteps).

2006-03-10 Stephane Popinet <popinet at users.sf.net>
  * Debian snapshot packages were not correctly versioned

2006-03-08 Stephane Popinet <popinet at users.sf.net>
  * Leave "specks" alone

2006-03-08 Stephane Popinet <popinet at users.sf.net>
  * Updated boundaries and channel tests for weighted-centered pressure-gradient

2006-03-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Approximate projection uses fraction-weighted average pressure gradients
  
  The previous unweighted averages were unstable for the "thin plate"
  test case which has been added to the test suite.

2006-03-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Topology was not computed correctly for solid fractions computation in 3D

2006-03-05 Stephane Popinet <popinet at users.sf.net>
  * Fix for special cases when computing solid fractions
  
  Solid fractions of cells cut more than once by the surface were not 
  consistently computed by the approximate "VOF" technique. This was causing
  convergence problems in some cases.
  
  These cells (rare) are now treated using a simple approximation.

2006-03-08 Stephane Popinet <popinet at users.sf.net>
  * Removed obsolete "GtsInterface" stuff

2006-03-03 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated test cases for changes in projection

2006-03-03 Stephane Popinet <s.popinet at niwa.co.nz>
  * Examples tarball did not include animations

2006-02-14 Stephane Popinet <popinet at users.sf.net>
  * gfs_domain_norm_residual() was not computing a dimensionless divergence
  
  It now returns the error in volume conservation per timestep relative
  to the volume of the cell.

2006-02-09 Stephane Popinet <popinet at users.sf.net>
  * OutputSimulation has an optional "text" format

2006-02-09 Stephane Popinet <popinet at users.sf.net>
  * Updated Debian dependencies

2006-01-30 Stephane Popinet <popinet at users.sf.net>
  * New functions gfs_vof_plane() and gfs_vof_facet()

2006-02-05 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for inconsistent calls for divergence calculation

2006-02-05 Stephane Popinet <popinet at users.sf.net>
  * OutputTiming also reports the maximum number of variables allocated

2006-02-05 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for "Pmac"
  
  Pmac was allocated as a temporary variable and thus was not
  interpolated from one timestep to the next when using adaptive
  refinement (causing trouble with the convergence rate of the Poisson
  solver). It is now a default variable.

2006-02-05 Stephane Popinet <popinet at users.sf.net>
  * GfsAdapts should be treated as standard events
  
  ... in particular to avoid problems due to a changed order of events
  in output files.

2006-02-03 Stephane Popinet <popinet at users.sf.net>
  * New "nitermin" param for multilevel solvers
  
  This is optional but useful in some cases to prevent noise building up in a 
  stationary solution (spurious currents test case for example).

2006-02-01 Stephane Popinet <popinet at users.sf.net>
  * OutputTiming computes the total running time
  
  i.e. does not exclude the time taken by Outputs and Events.

2006-01-30 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for GfsVariable half-events
  
  Half-events were never processed for GfsVariables.

2006-01-30 Stephane Popinet <popinet at users.sf.net>
  * VariableFiltered works with odd numbers of iterations

2006-02-03 Stephane Popinet <popinet at users.sf.net>
  * Updated macros for automake 1.8

2006-01-28 Stephane Popinet <popinet at users.sf.net>
  * Generalised gfs_correct_normal_velocities() for CSF surface tension

2006-01-28 Stephane Popinet <popinet at users.sf.net>
  * New object GfsVariableCurvature

2006-01-28 Stephane Popinet <popinet at users.sf.net>
  * MAC pressure is saved
  
  Both the MAC and approximate projections thus have better initial guesses. When
  looking for stationary solutions this makes a big difference.

2006-01-28 Stephane Popinet <popinet at users.sf.net>
  * Simplification of Boussinesq example

2006-01-28 Stephane Popinet <popinet at users.sf.net>
  * Fixed typo in spurious currents test case description

2006-01-28 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for GfsSourceDiffusionExplicit

2006-01-28 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for generic coarse_fine() method for GfsVariable

2006-01-27 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for special case of solid boundary
  
  When children mixed cells are all empty, a divide-by-zero would occur when
  computing the center of "mass" of the parent cell.

2006-01-27 Stephane Popinet <popinet at users.sf.net>
  * Better checks for construction of merged cells lists

2006-01-21 Stephane Popinet <popinet at users.sf.net>
  * VOF tracers are adaptively refined consistently

2006-01-21 Stephane Popinet <popinet at users.sf.net>
  * Generic coarse_fine() method for GfsVariable

2006-01-19 Stephane Popinet <popinet at users.sf.net>
  * New criterion for "small" cells
  
  Each direction is tested independently and the presence or absence of 
  neighbors is taken into account.

2006-01-17 Stephane Popinet <popinet at users.sf.net>
  * S now returns normalised surface fraction

2006-01-17 Stephane Popinet <popinet at users.sf.net>
  * New object GfsOutputScalarMaxima

2005-12-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed typo in tutorial

2005-12-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated test references
  
  GCC version 4 seems to produce slightly different results.

2005-12-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated debian packaging and docs

2005-12-02 Stephane Popinet <s.popinet at niwa.co.nz>
  * Debian package suggests gfsview

2005-12-02 Stephane Popinet <s.popinet at niwa.co.nz>
  * cvs2dist user proper versioning

2005-12-02 Stephane Popinet <s.popinet at niwa.co.nz>
  * Debian package snapshots

2005-11-30 Stephane Popinet <popinet at users.sf.net>
  * Debian package stuff (from Marcelo's official debian archive)

2005-11-29 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for use of uninitialised heap info in adaptive refinement

2005-11-29 Stephane Popinet <s.popinet at niwa.co.nz>
  * Workaround for bug in latex2html

2005-11-28 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated release instructions

2005-10-17 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 0.8.0

2005-10-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated docs for release 0.8.0

2005-10-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputSolidForce also computes moments

2005-10-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added missing dependencies for building docs

2005-10-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * Couette test case uses RMS-norm rather than maximum

2005-10-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs_mixed_cell_gradient uses Dirichlet conditions if set

2005-10-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * Thin-wall and waves test fixes

2005-10-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * GfsFunction aborts on user declaration of reserved variable names

2005-10-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated channel test reference

2005-10-11 Stephane Popinet <s.popinet at niwa.co.nz>
  * Couette flow test case

2005-10-11 Stephane Popinet <s.popinet at niwa.co.nz>
  * New functions to compute 2nd invariant of shear strain rate tensor

2005-10-10 Stephane Popinet <s.popinet at niwa.co.nz>
  * Multilevel diffusion params are defined as part of GfsDiffusion

2005-10-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * Variable viscosity needs to be defined on all levels

2005-10-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * New parameter "beta" controls the implicitness of the diffusion solver

2005-10-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * Makefile.deps are cleaned (to force them to be updated)

2005-10-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Variable viscosity coupled with variable density should now work

2005-10-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Geometry is generated rather than stored in rt example

2005-10-04 Stephane Popinet <popinet at users.sf.net>
  * Surface tension should work with variable density

2005-10-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * Moved lid-driven cavity example to test suite

2005-10-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added examples.tar.gz target

2005-10-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for upwinding of MAC velocities
  
  The "lid-driven cavity" test case was broken by patch "Face boundary
  conditions modified for consistent upwinding". When Dirichlet
  conditions are imposed on the normal velocities, strict upwinding can
  cause a violation of the solvability condition for the Poisson
  equation. This patch fixes this problem by imposing the boundary
  condition but only for normal MAC velocities (in effect reverting to
  the old formulation in this case only).

2005-10-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for type cast bothering gcc

2005-10-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Restructured test/ directory

2005-10-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed obsolete tests

2005-10-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Simple advection test case

2005-10-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Geostrophic adjustment test cases

2005-10-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * New Rayleigh-Taylor example

2005-09-29 Stephane Popinet <popinet at users.sf.net>
  * SourceControl does not need a timescale

2005-09-29 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated capillary wave test

2005-09-28 Stephane Popinet <popinet at users.sf.net>
  * Capillary wave test case

2005-09-28 Stephane Popinet <s.popinet at niwa.co.nz>
  * Solid fractions are matched at coarse/fine boundaries

2005-09-28 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated channel test

2005-09-27 Stephane Popinet <s.popinet at niwa.co.nz>
  * Variable half-events for ocean models

2005-09-26 Stephane Popinet <popinet at users.sf.net>
  * New GfsVariableFiltered class

2005-09-26 Stephane Popinet <popinet at users.sf.net>
  * Variable density has been fixed

2005-09-26 Stephane Popinet <s.popinet at niwa.co.nz>
  * Some derived variables can be used in GfsFunction even with cell = NULL

2005-09-26 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputHistogram can take an optional dependent variable

2005-09-23 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated tests and examples for obsolete gfsview option

2005-09-23 Stephane Popinet <s.popinet at niwa.co.nz>
  * EventHarmonic can optionally compute the mean-squares error

2005-09-21 Stephane Popinet <s.popinet at niwa.co.nz>
  * Unified implementation of 2D and 2D3 ocean models
  
  The 2D3 "baroclinic" model only works with a single layer for the moment.
  The results should be exactly identical to the 2D ocean model.

2005-09-20 Stephane Popinet <popinet at users.sf.net>
  * Pressure is centered in coastally-trapped waves test case

2005-09-20 Stephane Popinet <s.popinet at niwa.co.nz>
  * New function gfs_set_2D_solid_fractions_from_surface

2005-09-20 Stephane Popinet <s.popinet at niwa.co.nz>
  * New function gfs_cell_traverse_cut_2D

2005-09-20 Stephane Popinet <s.popinet at niwa.co.nz>
  * RefineHeight uses same interpolation routine as GfsFunction

2005-09-19 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated channel test reference

2005-09-19 Stephane Popinet <s.popinet at niwa.co.nz>
  * Weaker criteria for Poisson test cases

2005-09-16 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed obsolete gerris_dir stuff

2005-09-16 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputEnergy has been removed (it can be done better using functions)

2005-09-16 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed old Poisson tests

2005-09-16 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated Poisson test cases

2005-09-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated channel and boundaries tests for "old" centered pressure gradient calculation

2005-09-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * New circular waves shallow-water test case

2005-09-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Weighted centered pressure gradient is now an option used only by the ocean models

2005-09-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Cosmetics

2005-09-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * EventFilter takes a time scale as extra argument
  
  Exponential filtering with the given timescale is used rather than
  discrete full filtering.

2005-09-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated references for Poisson tests

2005-09-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated tests for default erelax of one

2005-09-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * Coriolis source term takes an optional linear drag parameter

2005-09-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * Update of 'channel' and 'boundaries' test cases for centered error estimation

2005-09-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * 'erelax' is one by default (rather than 2)
  
  In most cases this is more efficient.

2005-09-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * Bug fix for OutputCorrelation

2005-09-06 Stephane Popinet <s.popinet at niwa.co.nz>
  *  at  is replaced with # before compiling functions
  
  Lines starting with # are treated as comments and ignored, which is a
  problem if precompiler directives are used in the function. Using  at 
  instead of # allows the use of precompiler directives.

2005-09-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * Channel and boundaries test cases compare non-centered variables

2005-09-02 Stephane Popinet <s.popinet at niwa.co.nz>
  * Variables are NOT centered by default
  
  All variables were set to centered by default. Not sure how this
  crept into the code but this could have potentially important follow
  up effects. Need to check on this.

2005-08-30 Stephane Popinet <s.popinet at niwa.co.nz>
  * Cleanup of ocean divergence calculation

2005-08-30 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for missing params following changes in gfs_poisson_cycle

2005-08-29 Stephane Popinet <s.popinet at niwa.co.nz>
  * CFL is applicable only if some advection scheme is used

2005-08-29 Stephane Popinet <s.popinet at niwa.co.nz>
  * Simplified center-of-mass checks

2005-08-26 Stephane Popinet <s.popinet at niwa.co.nz>
  * Small fixes for pickier gcc-4.0 warnings

2005-08-26 Stephane Popinet <s.popinet at niwa.co.nz>
  * "Exponential" relaxation is controlable ("erelax" parameter)

2005-08-26 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed typos in FAQ

2005-08-26 Stephane Popinet <s.popinet at niwa.co.nz>
  * Compilation of functions on MacOSX

2005-08-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for stupid bug in gfs_init_write

2005-08-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * Extrusion of closed profiles is done correctly (no duplicate vertices)

2005-08-16 Stephane Popinet <s.popinet at niwa.co.nz>
  * Bi-directional streamlines had duplicate vertices

2005-08-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Bug fix for 2nd order streamline integration
  
  Streamline interval length is not constrained by cell size anymore.

2005-08-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * New functions gfs_domain_cell_traverse_condition, gfs_domain_cell_point_distance2
  
  And also ftt_cell_box. gfs_domain_cell_traverse_condition is used
  as a generic replacement for gfs_domain_cell_traverse_box.

2005-08-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated references for test cases (for the improved Poisson solver)

2005-08-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * Change to GFV files for new format

2005-08-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * Generalisation of streamline creation functions

2005-08-10 Stephane Popinet <popinet at users.sf.net>
  * Bug fix for environment variables in gfs2tex

2005-08-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for incorrect merging test case reference

2005-08-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * New and updated Poisson test cases
  
  Rather than plotting residual vs multigrid cycles, we plot residual vs
  CPU time which is much more meaningful. A new thin plate test is
  representative of pathological cases which did not work well with the
  former version of the multigrid Poisson solver.

2005-08-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * Improvements to the Poisson solver
  
  After a full review of the Poisson solver: convergence tests with pre
  and post relaxations, FMG implementation etc... the following
  relatively minor changes were made: (1) the prolongation operator uses
  "second-order" gradient-based interpolation rather than straight
  injection, (2) the number of relaxations increases exponentially for
  coarser levels. None of the more complex changes seemed to improve
  convergence.

2005-08-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * New GfsClock object replaces GTimer
  
  Times are now given as user CPU time rather than wall-clock time.

2005-08-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Changes to GFV files for new GfsView file format

2005-08-04 Sebastien Delaux <s.delaux at niwa.co.nz>
  * cell->data is set to NULL by gfs_cell_cleanup

2005-08-04 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Fix for incorrect error message in FTT I/O

2005-08-04 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Bug fix for stupid use of destroyed event

2005-08-03 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs_face_ca was failing in particular cases on 3D meshes

2005-08-02 Stephane Popinet <s.popinet at niwa.co.nz>
  * Introduced a new GFS_FLAG_GRADIENT_BOUNDARY cell flag to fix Flather ocean bc
  
  The Flather ocean bc do not like the centered pressure gradient
  interpolation used to correct the centered velocities. To revert to
  the old "upwind" gradient interpolation a new flag has been
  introduced. This flag is set automatically by the Flather bc.

2005-08-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * ca.z is initialised to zero in 2D

2005-08-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated poisson test cases with solid boundaries

2005-07-21 Stephane Popinet <s.popinet at niwa.co.nz>
  * Changed the way gfs2tex sets environment variables

2005-07-21 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added new-style Poisson test cases

2005-07-20 Stephane Popinet <popinet at users.sf.net>
  * New GfsPoisson solver

2005-07-20 Stephane Popinet <s.popinet at niwa.co.nz>
  * `shapes' has been moved to tools and is installed

2005-07-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs_function_value() can take NULL as cell argument

2005-07-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Cosmetics

2005-07-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Face function evaluation knows about mixed cells
  
  This means that boundary conditions dependent on spatial coordinates
  will correctly account for cell faces on the boundaries of the domain
  being cut by an embdedded surface.

2005-07-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed broken FAQ link

2005-07-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * Cleaned up darcs2dist script

2005-07-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added ChangeLog target

2005-07-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added tarball target for tests docs

2005-07-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * Changed computation of centered pressure gradients
  
  This influences only the pressure gradient in mixed cells. It is now
  computed as the surface-fraction-weighted averaged rather than just
  the average. This is intuitively more consistent as this provides a
  continuous interpolation when the face fractions vanish on one side of
  the cell. More importantly this greatly stabilises the ocean model
  near sharp headlands. It has a minimal influence on the convergence
  rates of the test cases with boundaries (slightly improves the V
  component and slightly degrades the U component).

2005-07-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * Version set to 0.8.0 prior to release

2005-07-12 Stephane Popinet <s.popinet at niwa.co.nz>
  * Variable "P" is centered by default

2005-07-12 Stephane Popinet <s.popinet at niwa.co.nz>
  tagged 0.6.4

2005-07-11 Stephane Popinet <s.popinet at niwa.co.nz>
  * Merged Ocean1 model from 'ocean' branch
  
  This does not include the "fraction-weighted pressure correction" of
  centered velocities which means that some coastlines configurations
  will not be stable.

2005-07-11 Stephane Popinet <s.popinet at niwa.co.nz>
  * Bug fix for VOF plane center calculation
  
  Assertions were too limiting when cells were completely full or empty.

2005-07-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Restructuring of source terms class hierarchy
  
  Former "vector" source terms (coriolis, surface tension, viscosity...)
  are now derived from a common "velocity" source class which
  automatically checks for U,V,W variables. This means that the syntax
  in parameter files for these source terms has changed (no need to
  specify a variable anymore).

2005-07-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Spurious currents test updated for new syntax

2005-07-09 Sebastien Delaux <s.delaux at niwa.co.nz>
  * Fix for locale-independent parsing of parameter files

2005-07-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Minor layout changes for merging test

2005-07-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated channel test reference

2005-07-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * New function gfs_function_expression()

2005-07-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed obsolete "derived" variable stuff

2005-07-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for floating-point bug in calculation of 2D center of mass
  
  Also added checks for consistency of the center of mass and center of
  area positions.

2005-07-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs2tex test() method uses gfsview-batch instead of gfsview (tests are non-interactive)

2005-07-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated reference for spurious current test

2005-07-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for merging conflict of speck removal stuff

2005-07-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for deprecated casts in gfs_object_simulation()

2005-07-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Upgrade of EventFilter to new memory management

2005-03-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Specks are removed automatically from the simulation
  
  "Specks" are tiny (less than one cell volume) solid domains.

2005-03-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * New ocean test energy.sh, temporary modification of waves.sh
  
  waves.sh uses the Ocean1 2D model.

2005-02-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * New GfsEventFilter class
  
  Uses averages of corner values as a filter. This is very efficient to filter
  out grid-scale oscillations of crappy schemes...

2005-07-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated boundaries test

2005-07-06 Stephane Popinet <s.popinet at niwa.co.nz>
  * Changed BC for boussinesq example

2005-07-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * New memory management changes for ocean model

2005-07-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * EventSum takes a function rather than a variable, EventSum2 has been removed

2005-07-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * Replaced SourceViscosity with SourceDiffusion in examples for now

2005-07-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * Cleanup of test script for doc/examples

2005-07-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * Spurious currents convergence is computed for finer grids

2005-07-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * VOF scheme was not allocating temporary variable

2005-07-05 Stephane Popinet <s.popinet at niwa.co.nz>
  * GfsFunction identifier-finding routine was too naive

2005-07-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated tutorial

2005-07-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated reference doc

2005-07-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for non-advected tracers with sources

2005-06-30 Stephane Popinet <popinet at users.sf.net>
  * Allows for "tracers" only governed by source terms

2005-06-30 Stephane Popinet <popinet at users.sf.net>
  * Updated reference manual

2005-07-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * boussinesq example updated for new gfsview option

2005-07-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for expression parsing in GfsFunction

2005-07-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for I/O of GfsDomain

2005-06-30 Stephane Popinet <s.popinet at niwa.co.nz>
  * Generalised parsing of GfsFunction

2005-06-30 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed bug in dynamic variable allocation

2005-06-29 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs_function_description() can return long or short descriptions

2005-06-29 Stephane Popinet <s.popinet at niwa.co.nz>
  * Non-repetitive outputs are always overwritten

2005-06-29 Stephane Popinet <s.popinet at niwa.co.nz>
  * Derived variables are associated with GfsDomain

2005-06-28 Stephane Popinet <s.popinet at niwa.co.nz>
  * GfsInit creates unknown variables and preserves the order of initialisation

2005-06-28 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for incorrect parsing of GfsFunction

2005-06-28 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed obsolete MPI tests

2005-06-28 Stephane Popinet <s.popinet at niwa.co.nz>
  * Restructured GfsVariable implementation
  
  All the variables are now dynamically allocated. Temporary variables are
  allocated when needed. There is no limit (other than memory) on the maximum
  number of variables per cell.
  
  This also uncovered a serious bug: when using diffusive tracers, the solution of
  the diffusion equation would overwrite the stored values of the pressure
  gradients used to correct the advective terms for the velocity. With the new
  temporary variable allocation scheme messes like that should be avoided.

2005-06-27 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed obsolete gfs2other and unsupported gfs2vtk

2005-06-23 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed Makefile for old-style surface tension test

2005-06-22 Stephane Popinet <s.popinet at niwa.co.nz>
  * New-style spurious current test case

2005-06-21 Stephane Popinet <s.popinet at niwa.co.nz>
  * New-style channel test, old tests have been removed

2005-06-21 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputScalarHistogram correctly uses new variable definition and computes instantaneous histograms

2005-06-21 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfscompare uses gfs_cell_interpolate() for error calculation of centred variables

2005-06-21 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed FAQ broken link

2005-06-21 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added 'boundaries' test case

2005-06-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * merging test uses batch-mode gfsview for figure generation

2005-06-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * New 4-way vortex merging test case

2005-06-16 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added simple periodic test case

2005-06-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added biblio to tests doc and a few bug fixes

2005-06-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Automated generation of summary for test docs

2005-06-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated reference for 'reynolds/box' test case, removal of corresponding old tests

2005-06-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added PPM animate question to FAQ

2005-06-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Auto-documented Euler test cases replace old-style tests
  
  Only the equivalent test cases for reference2D/stationary1.xmgr and
  reference2D/stationary4box1.xmgr have been implemented at this point.

2005-06-13 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * New 'Reynolds' test case, replacement for 'stationary'
  
  Using auto-documenting with gfs2doc.

2005-06-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * cast fix for gfs2oogl

2005-06-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * New macro ftt_vector_norm()

2005-06-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * Numbers are allowed in identifiers of macros and variables

2005-06-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * "S" is a derived macro for solid area

2005-06-13 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputScalarHistogram takes weight as argument

2005-06-10 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputScalar takes the new GfsFunction as argument instead of a GfsVariable

2005-06-10 Stephane Popinet <s.popinet at niwa.co.nz>
  * GfsFunction now knows about "derived variables"
  
  This is the first step toward removing old-style derived variables
  altogether.

2005-06-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputScalarHistogram starts when specified and properly reopens static files

2005-06-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * GfsFunction recognises domain variables directly

2005-06-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Face boundary conditions modified for consistent upwinding
  
  The former face boundaries were not "upwind aware", this was causing
  problems when, for example, a negative normal velocity was set as inflow
  condition on the left of the domain and at the same time a Dirichlet BC on a
  tracer.

2005-06-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Bug fix for tracer advection (thanks to David Sterling)

2005-06-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * New object GfsOutputScalarHistogram

2005-06-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * New macro gfs_cell_volume()

2005-06-02 Stephane Popinet <s.popinet at niwa.co.nz>
  * New gfs_solid_normal() and gfs_pressure_force() functions

2005-06-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed FAQ icons (part 2)

2005-06-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed FAQ icons

2005-06-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Restructuring of checks for interpolation in face_bilinear

2005-06-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added FAQ and updated links and style sheets

2005-05-31 Stephane Popinet <s.popinet at niwa.co.nz>
  * New SourceControl object

2005-05-31 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputSimulation uses binary format by default

2005-05-27 Stephane Popinet <s.popinet at niwa.co.nz>
  * Preliminary restructuring of Euler test cases

2005-05-27 Stephane Popinet <s.popinet at niwa.co.nz>
  * Cosmetics for advection tests

2005-05-27 Stephane Popinet <s.popinet at niwa.co.nz>
  * Restructured advection order test cases

2005-05-26 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated installation instructions

2005-05-26 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for GfsEventScript unused handling of error output

2005-05-25 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added boussinesq.gfv gfsview parameter file

2005-04-24 Stephane Popinet <s.popinet at niwa.co.nz>
  * InitVorticity takes vorticity function as argument
  
  Test cases have been updated and obsolete modules have been removed.

2005-04-20 Stephane Popinet <s.popinet at niwa.co.nz>
  * Bug fix for InitVorticity

2005-04-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * Update for xmgr -> xmgrace transition

2005-04-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * VOF scheme has a CFL limit of one (not 0.5)

2005-04-14 Stephane Popinet <s.popinet at niwa.co.nz>
  * gerris can take parameter file on standard input

2005-04-10 Stephane Popinet <s.popinet at niwa.co.nz>
  * Minor bugfix for gfs2tex

2005-04-10 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated boussinesq example

2005-04-10 Stephane Popinet <s.popinet at niwa.co.nz>
  * Boussinesq example now uses scripting mode of GfsView

2005-04-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated examples Makefile.am dependencies

2005-04-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added Boussinesq example to doc

2005-04-06 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * New Boussinesq example

2005-04-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * More robust (hopefully) FPU exceptions control

2005-04-07 Stephane Popinet <s.popinet at niwa.co.nz>
  * Decent autogen.sh script

2005-04-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Automatic dependency of examples doc on files generated by examples

2005-04-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Shear example updated for new C expressions in GfsFunction

2005-04-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed problems with C expressions in GfsFunction
  
  Also added check for pkg-config in configure. pkg-config is required for
  inlined functions in parameter files.

2005-04-03 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * GfsInit and GfsOutputErrorNorm can use functions of variables

2005-04-03 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * Fixed lid example typos

2005-04-03 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * New time-reversed VOF advection example

2005-04-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * New lid-driven cavity example

2005-04-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * Force flushing of OutputLocation

2005-04-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * OutputLocation can now work with a list of several locations

2005-04-01 Stephane Popinet <s.popinet at niwa.co.nz>
  * EventStop can now store the difference used as criterion

2005-03-23 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added missing signal header in utils.c (for BSD)

2005-03-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * doc/examples/Makefile.deps is distributed to avoid a python dependency

2005-03-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * bat2gts does not need to construct closed geometry anymore

2005-03-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed obsolete CVS usermap file

2005-03-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added automated testing of examples collection

2005-03-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * autogen.sh must not be distributed in tarballs

2005-03-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added simple autogen.sh script

2005-03-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * Automatic distribution of examples + improved gfs2tex code

2005-03-17 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs2doc checks that the example runs

2005-03-16 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs2doc script for automatically generated example documentation
  
  Also improved cross-links for colorised parameter files.

2005-03-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Initial implementation of examples scripts

2005-02-22 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added scripts in tools to distribution

2005-02-22 Stephane Popinet <s.popinet at niwa.co.nz>
  * Bug fix for boundary matching of solid fractions

2005-02-22 Stephane Popinet <s.popinet at niwa.co.nz>
  * Mixed cell refinement fixes in 3D
  
  The advection fix could be improved (it does not currently take into account
  the solid fractions, unlike the 2D version of the fix).

2004-11-09 Stephane Popinet <popinet at users.sourceforge.net>
  * Removed some of the "mixed cell refinement" constraints (gerris--ocean--0.7--patch-2)
  gerris--ocean--0.7--patch-2
  Keywords: 
  
  Probably needs more careful review.
  

2005-02-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * GfsFunction can use C statements (instead of full functions)
  
  but modules are not supported anymore by GfsFunction (that was never used
  anyway).
  
  This patch means that things like:
  
    Init {} { U = sqrt(x) }
  
  are now legal.

2005-02-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * GfsFunction can take a GTS surface as argument
  
  and interpolates the z-coordinate at location (x,y) to get the value of the
  function.

2005-02-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added ChangeLog to distribution target

2005-02-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Removed arch obsolete garbage

2005-02-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Spurious current test with adaptivity

2005-02-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs_object_simulation() does not work for GfsBc objects

2005-02-15 Stephane Popinet <s.popinet at niwa.co.nz>
  * Changed interface to gfs_function_read()

2005-02-11 Stephane Popinet <s.popinet at niwa.co.nz>
  * New GfsAdaptFunction class
  
  Which uses the new GfsFunction to allow users to specify adaptive criteria
  based on the values of local variables.
  
  This can be used for example to refine around an interface depending on the local
  value of the surface tension.

2005-02-11 Stephane Popinet <s.popinet at niwa.co.nz>
  * Cosmetics

2005-02-11 Stephane Popinet <s.popinet at niwa.co.nz>
  * GfsFunction can now use model variables
  
  This can be used for example to define source terms dependent on other
  variables.

2005-02-11 Stephane Popinet <s.popinet at niwa.co.nz>
  * Cosmetics

2005-02-10 Stephane Popinet <s.popinet at niwa.co.nz>
  * Capillary wave test

2005-02-10 Stephane Popinet <s.popinet at niwa.co.nz>
  * Surface tension only works in 2D for now

2005-02-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * Renamed _gfs_source_tension_ to _T

2005-02-09 Stephane Popinet <s.popinet at niwa.co.nz>
  * gfs_cell_corner_interpolator now works for periodic BCs
  
  But only when the periodic BCs are not cut by a solid boundary.

2005-02-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added surface tension tests to configure

2005-02-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Minimum timestep is 1e-9
  
  Several functions implicitly rely on the timestep being strictly positive.

2005-02-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Spurious currents test

2005-02-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Replaced gfs_youngs_gradient with gfs_youngs_normal
  
  Which computes the three components in one operation and should be about
  three times as fast as three calls to gfs_youngs_gradient.

2005-02-08 Stephane Popinet <s.popinet at niwa.co.nz>
  * Proper scaling of surface tension terms
  
  The surface tension term also uses the volume fraction at time n+1 which
  improves the stability significantly.

2005-02-04 Stephane Popinet <s.popinet at niwa.co.nz>
  * Replaced "marker" surface tension implementation with tensorial-CSF formulation

2005-01-30 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * Implementation of Young's normal calculation for VOF

2005-01-30 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * Default adaptive maxlevel set to 5
  
  The previous default was infinite which was not too good for memory use...

2005-01-30 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * VOF advection tests

2005-01-30 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * Timestep is computed using minimum CFL of all variables

2005-01-30 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * New GfsAdvection simulation class

2005-01-30 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * Tracers now take advection/diffusion parameters

2005-01-30 Stephane Popinet <stephane.popinet at paradise.net.nz>
  * Graphic advection test restructured
  
  Uses gerris/gfsview rather than advection.c/gts2oogl/geomview etc...

2005-01-28 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fix for cygwin built

2005-01-19 Stephane Popinet <s.popinet at niwa.co.nz>
  * darcs2dist and bat2gts are installed

2005-01-19 Stephane Popinet <s.popinet at niwa.co.nz>
  * darcs2dist takes URL argument

2005-01-19 Stephane Popinet <s.popinet at niwa.co.nz>
  * Clearer error message for boundaries outside fluid domain

2005-01-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Automatic ChangeLog generation and darcs2dist snapshot script

2005-01-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Added GTS dependency to pkgconfig files

2005-01-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed test/mpi

2005-01-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed test/advection
  
  The results needs checking.

2005-01-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Fixed test/poisson
  
  test/poisson/order still gives some inconsistent results.

2005-01-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Updated test/poisson/convergence/reference2D

2004-12-09 Stephane Popinet <popinet at users.sourceforge.net>
  * Matching of surface fractions for refined mixed cells (gerris--ocean--0.7--patch-27)
  gerris--ocean--0.7--patch-27
  Keywords: 
  
  Also removal of obsolete gfs_refine_mixed() function and improvement
  in error reporting for check_area_fractions().
  

2004-12-07 Stephane Popinet <popinet at users.sourceforge.net>
  * Important bug fix for GfsEventHarmonic (gerris--ocean--0.7--patch-26)
  gerris--ocean--0.7--patch-26
  Keywords: 
  
  The right-hand-side for the least-mean-square of the offset was not
  properly computed.
  

2004-12-07 Stephane Popinet <popinet at users.sourceforge.net>
  * isocube.h is added to distribution (gerris--ocean--0.7--patch-25)
  gerris--ocean--0.7--patch-25
  Keywords: 
  
  

2004-12-07 Stephane Popinet <popinet at users.sourceforge.net>
  * gfs2oogl profile output now works without -c option (gerris--ocean--0.7--patch-24)
  gerris--ocean--0.7--patch-24
  Keywords: 
  
  

2004-12-01 Stephane Popinet <popinet at users.sourceforge.net>
  * Changed the way "special cases" are handled in VOF (gerris--ocean--0.7--patch-23)
  gerris--ocean--0.7--patch-23
  Keywords: 
  
  This is simpler and fixes a serious bug in the new solid fraction
  algorithm.
  

2004-12-01 Stephane Popinet <popinet at users.sourceforge.net>
  * New function gfs_domain_combine_traverse() (gerris--ocean--0.7--patch-22)
  gerris--ocean--0.7--patch-22
  Keywords: 
  
  For traversing two GfsDomain "simultaneously".
  

2004-11-29 Stephane Popinet <popinet at users.sourceforge.net>
  * Made gfs_divergence() interface consistent with gfs_velocity_norm() etc... (gerris--ocean--0.7--patch-21)
  gerris--ocean--0.7--patch-21
  Keywords: 
  
  i.e. it now takes an extra GfsVariable argument.
  

2004-11-29 Stephane Popinet <popinet at users.sourceforge.net>
  * Replaced static cell_is_cut function with exported gfs_cell_is_cut (gerris--ocean--0.7--patch-20)
  gerris--ocean--0.7--patch-20
  Keywords: 
  
  Used for new gfsview solid entity.
  

2004-11-25 Stephane Popinet <popinet at users.sourceforge.net>
  * Added isocube.h to sources (gerris--ocean--0.7--patch-19)
  gerris--ocean--0.7--patch-19
  Keywords: 
  
  

2004-11-22 Stephane Popinet <popinet at users.sourceforge.net>
  * Refinements interpolate existing values (gerris--ocean--0.7--patch-18)
  gerris--ocean--0.7--patch-18
  Keywords: 
  
  This allows to restart a simulation while changing the refinements of
  some parts.
  

2004-11-17 Stephane Popinet <popinet at users.sourceforge.net>
  * In 2D3 ftt_cell_child_corner() now works for all directions (gerris--ocean--0.7--patch-17)
  gerris--ocean--0.7--patch-17
  Keywords: 
  
  

2004-11-17 Stephane Popinet <popinet at users.sourceforge.net>
  * New GfsEventHarmonic class (gerris--ocean--0.7--patch-16)
  gerris--ocean--0.7--patch-16
  Keywords: 
  
  Does on-the-fly harmonic analysis of a variable.
  

2004-11-16 Stephane Popinet <popinet at users.sourceforge.net>
  * New utility function gfs_matrix_inverse() (gerris--ocean--0.7--patch-14)
  gerris--ocean--0.7--patch-14
  Keywords: 
  
  

2004-11-16 Stephane Popinet <popinet at users.sourceforge.net>
  * Option not to output solid surface in GfsOutputSimulation (gerris--ocean--0.7--patch-13)
  gerris--ocean--0.7--patch-13
  Keywords: 
  
  

2004-11-16 Stephane Popinet <popinet at users.sourceforge.net>
  * Weaker solid fraction consistency checks to accomodate mixed cell refinement (gerris--ocean--0.7--patch-12)
  gerris--ocean--0.7--patch-12
  Keywords: 
  
  

2004-11-16 Stephane Popinet <popinet at users.sourceforge.net>
  * Added tolerance on timestep calculation (gerris--ocean--0.7--patch-11)
  gerris--ocean--0.7--patch-11
  Keywords: 
  
  To try to avoid very small timesteps (< 1e-9) due to round-off errors
  on times of multiple simultaneous events.
  

2004-11-16 Stephane Popinet <popinet at users.sourceforge.net>
  * Removed mixed cell refinement constraint for face traversal (gerris--ocean--0.7--patch-10)
  gerris--ocean--0.7--patch-10
  Keywords: 
  
  

2004-11-16 Stephane Popinet <popinet at users.sourceforge.net>
  * bat2gts needs units specified (gerris--ocean--0.7--patch-7)
  gerris--ocean--0.7--patch-7
  Keywords: 
  
  The default length unit for GMT is inches, bat2gts works only if units
  are cm. I have added options to the mapproject commands to override
  the defaults.
  

2004-11-09 Stephane Popinet <popinet at users.sourceforge.net>
  * More robust GfsRefineHeight algorithm (gerris--ocean--0.7--patch-4)
  gerris--ocean--0.7--patch-4
  Keywords: 
  
  Uses the four cell corner height values rather than just the central
  one.
  

2004-11-09 Stephane Popinet <popinet at users.sourceforge.net>
  * Uses g[0] rather than res for temporary storage in adaptive algorithm (gerris--ocean--0.7--patch-1)
  gerris--ocean--0.7--patch-1
  Keywords: 
  
  We want to keep res in case its value is required for monitoring the
  poisson solver.
  

2004-11-04 Stephane Popinet <popinet at users.sourceforge.net>
  * New class GfsRefineHeight (gerris--mainline--0.7--patch-29)
  gerris--mainline--0.7--patch-29
  Keywords: 
  
  For height/depth dependent refinement in ocean model.
  

2004-11-04 Stephane Popinet <popinet at users.sourceforge.net>
  * gfs_domain_remove_ponds() now matches boundaries (gerris--mainline--0.7--patch-28)
  gerris--mainline--0.7--patch-28
  Keywords: 
  
  Only if ponds have been removed.
  

2004-11-04 Stephane Popinet <popinet at users.sourceforge.net>
  * Added targets libdx2D3 and gerris2D3.pc (gerris--mainline--0.7--patch-27)
  gerris--mainline--0.7--patch-27
  Keywords: 
  
  

2004-11-03 Stephane Popinet <popinet at users.sourceforge.net>
  * New function gfs_domain_remove_specks() (gerris--mainline--0.7--patch-26)
  gerris--mainline--0.7--patch-26
  Keywords: 
  
  "Specks" are solid islands not containing any empty cell.
  
  They cause stability problems for the free-surface solver of the ocean
  model.
  
  They cannot be considered resolved anyway (by definition their maximum
  thickness is smaller than one grid cell).
  

2004-11-03 Stephane Popinet <popinet at users.sourceforge.net>
  * New class GfsRemovePonds (gerris--mainline--0.7--patch-25)
  gerris--mainline--0.7--patch-25
  Keywords: 
  
  Removes small "ponds" from the domain.
  

2004-11-03 Stephane Popinet <popinet at users.sourceforge.net>
  * GfsVariable are now events (gerris--mainline--0.7--patch-24)
  gerris--mainline--0.7--patch-24
  Keywords: 
  
  These events are activated at the start of the timestep loop in run()
  methods.
  
  This means that the syntax for variables in parameter files has been
  changed.
  
  This involved a substantial cleanup and restructuring of the events
  and their links with simulations.
  
  The first application of this new mechanism is the new
  GfsVariableResidual.
  

2004-11-03 Stephane Popinet <popinet at users.sourceforge.net>
  * Split GfsVariable in new variable.{ch} file (gerris--mainline--0.7--patch-23)
  gerris--mainline--0.7--patch-23
  Keywords: 
  
  It was getting a bit too big and dispersed all over the place.
  

2004-11-02 Stephane Popinet <popinet at users.sourceforge.net>
  * Fixes to solid fraction algorithm for 2D3 (gerris--mainline--0.7--patch-22)
  gerris--mainline--0.7--patch-22
  Keywords: 
  
  Because the aspect ratios of the 2D3 cells are not constant.
  

2004-11-02 Stephane Popinet <popinet at users.sourceforge.net>
  * New GfsRefineSurface class (gerris--mainline--0.7--patch-21)
  gerris--mainline--0.7--patch-21
  Keywords: 
  
  The GfsRefineSolid and GfsRefineDistance classes have also been
  reworked.
  

2004-11-02 Stephane Popinet <popinet at users.sourceforge.net>
  * Fractions initialisation uses new solid fraction algorithm (gerris--mainline--0.7--patch-20)
  gerris--mainline--0.7--patch-20
  Keywords: 
  
  The gfs_cell_init_fraction() function has been replaced by
  gfs_domain_init_fraction() which now calls
  gfs_domain_init_solid_fractions(). 
  
  The GfsInitFraction event has been changed accordingly.
  

2004-11-02 Stephane Popinet <popinet at users.sourceforge.net>
  * New calculation of 3D center of mass of solid fractions (gerris--mainline--0.7--patch-19)
  gerris--mainline--0.7--patch-19
  Keywords: 
  
  Uses the new 3D VOF gfs_plane_center() function.
  

2004-11-01 Stephane Popinet <popinet at users.sourceforge.net>
  * Fixed bug when computing fractions of degenerate intersections (gerris--mainline--0.7--patch-18)
  gerris--mainline--0.7--patch-18
  Keywords: 
  
  We don't want to use VOF when the normal is undefined (|n| = 0),
  rather use the face fractions directly (since the flux contribution of
  the "normal" part is zero) to compute the fraction.
  

2004-11-01 Stephane Popinet <popinet at users.sourceforge.net>
  * Added VOF calculation of 3D solid fraction (gerris--mainline--0.7--patch-17)
  gerris--mainline--0.7--patch-17
  Keywords: 
  
  The center of mass calculation is not done yet.
  

2004-11-01 Stephane Popinet <popinet at users.sourceforge.net>
  * Changed 2D bounding boxes for cut cells traversal (gerris--mainline--0.7--patch-16)
  gerris--mainline--0.7--patch-16
  Keywords: 
  
  Only the faces crossing the z=0 plane should be considered.
  

2004-11-01 Stephane Popinet <popinet at users.sourceforge.net>
  * Removed upstream self-intersection checks (gerris--mainline--0.7--patch-15)
  gerris--mainline--0.7--patch-15
  Keywords: 
  
  This can cause the solid fractions algorithm to fail due to
  consistency checks being violated by self-intersecting surfaces.
  
  The messages will be a bit obscure. We'll fix that as we go.
  
  The advantage is that in possibly a large number of cases with "minor"
  self-intersections (e.g. nearly collocated vertices etc...) the
  algorithm will proceed and should produce valid solid fractions.
  

2004-11-01 Stephane Popinet <popinet at users.sourceforge.net>
  * Domain traversal for painting algorithm (gerris--mainline--0.7--patch-14)
  gerris--mainline--0.7--patch-14
  Keywords: 
  
  The previous "box" traversal did not work for the painting algorithm
  (which does not know box boundaries).
  
  Traversing the whole domain rather than each box fixes the problem.
  
  Also, the painting algorithm used to also traverse boundary cells. A
  check has been added to avoid that.
  

2004-11-01 Stephane Popinet <popinet at users.sourceforge.net>
  * preliminary version of new 3D solid fraction computation (gerris--mainline--0.7--patch-13)
  gerris--mainline--0.7--patch-13
  Keywords: 
  
  Only computes the face fractions at this point (the painting algorithm
  works fine unchanged from 2D).
  

2004-11-01 Stephane Popinet <popinet at users.sourceforge.net>
  * Changed "temporary" to "permanent" for GfsVariable (gerris--mainline--0.7--patch-12)
  gerris--mainline--0.7--patch-12
  Keywords: temporary
  
  permanent is a pointer to the permanent variable stored in this
  temporary variable.
  

2004-10-28 Stephane Popinet <popinet at users.sourceforge.net>
  * New algorithm for 2D solid fractions computation (gerris--mainline--0.7--patch-11)
  gerris--mainline--0.7--patch-11
  Keywords: 
  
  Does not use the gts_surface_inter etc... functions but a simple
  computation of the intersections of the sides of the cell with the
  surface, combined with a linear approximation of the piece of the
  surface contained in the cell.
  
  This is *much* faster, simpler and should be robust. It is also much
  less picky about the degeneracies of the surfaces it can deal with.
  
  It does not work yet for a varying level of refinement along the
  surface.
  

2004-10-28 Stephane Popinet <popinet at users.sourceforge.net>
  * Check for diffusion terms before warning (gerris--mainline--0.7--patch-10)
  gerris--mainline--0.7--patch-10
  Keywords: 
  
  The "soid surface cuts boundary cells" warning is only relevant if
  diffusion terms are present.
  

2004-10-28 Stephane Popinet <popinet at users.sourceforge.net>
  * Improved solid fractions computation (gerris--mainline--0.7--patch-9)
  gerris--mainline--0.7--patch-9
  Keywords: 
  
  Both solid fractions computation and GfsRefineSolid use the new
  gfs_cell_traverse_cut function. This has several advantages: no need
  for pre-computation of a bounding-box tree, faster (O(log(N))).
  
  GfsRefineSolid is *much* faster (at least one order of magnitude).
  
  A paiting algorithm is used to set fractions for cells which are not
  cut by the solid boundary. This is simpler than the ray-casting
  technique (no need for bounding-box tree).
  

2004-10-28 Stephane Popinet <popinet at users.sourceforge.net>
  * Added timers for initial refinement and solid fractions (gerris--mainline--0.7--patch-8)
  gerris--mainline--0.7--patch-8
  Keywords: 
  
  

2004-10-27 Stephane Popinet <popinet at users.sourceforge.net>
  * New function gfs_cell_traverse_cut (gerris--mainline--0.7--patch-7)
  gerris--mainline--0.7--patch-7
  Keywords: 
  
  Fast traversal of cells cut by a surface. The cost of traversal scales
  roughly like log(N) where N is the total number of cells traversed.
  

2004-10-27 Stephane Popinet <popinet at users.sourceforge.net>
  * Added bat2gts script (gerris--mainline--0.7--patch-6)
  gerris--mainline--0.7--patch-6
  Keywords: 
  
  For generating ocean bathymetry files.
  

2005-01-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Bumped micro version

2004-10-21 Stephane Popinet <popinet at users.sourceforge.net>
  * Automatically adds PS and Div variables to ocean model (gerris--mainline--0.7--patch-5)
  gerris--mainline--0.7--patch-5
  Keywords: 
  
  Using a overloading of the read() method of GfsOcean.
  

2004-10-20 Stephane Popinet <popinet at users.sourceforge.net>
  * Replaced GfsOutputStreakline with GfsOutputParticle (gerris--mainline--0.7--patch-4)
  gerris--mainline--0.7--patch-4
  Keywords: 
  
  GfsOutputStreakline was weird. GfsOutputParticle is a lagrangian
  tracer integrator.
  

2004-10-20 Stephane Popinet <popinet at users.sourceforge.net>
  * Fixed cloning of GfsVariable. (gerris--mainline--0.7--patch-3)
  gerris--mainline--0.7--patch-3
  Keywords: 
  
  The (non)-cloning of the surface_bc field of GfsVariable was crashing
  the code when destroying cloned variables. This has been solved by not
  cloning surface_bc and source terms.
  

2004-10-20 Stephane Popinet <popinet at users.sourceforge.net>
  * Fixed DX modules (gerris--mainline--0.7--patch-2)
  gerris--mainline--0.7--patch-2
  Keywords: 
  
  

2004-10-19 Stephane Popinet <popinet at users.sourceforge.net>
  * Updates for release 0.6.0 (gerris--mainline--0.5--patch-8)
  gerris--mainline--0.5--patch-8
  Keywords: 0.6.0
  
  Tutorial and reference manual have been updated.
  A fix for MacOSX libtool has been included.
  Fix for incorrect call to gfs_interpolate in gfs2oogl.
  Fix for branch "stable" in cvs2cl.pl
  

2004-10-18 Stephane Popinet <popinet at users.sourceforge.net>
  * GTS solid files can also be written in binary format. (gerris--mainline--0.5--patch-7)
  gerris--mainline--0.5--patch-7
  Keywords: 
  
  

2004-10-14 Stephane Popinet <popinet at users.sourceforge.net>
  * gfs_interpolate uses proper bi(tri)linear interpolation (gerris--mainline--0.5--patch-6)
  gerris--mainline--0.5--patch-6
  Keywords: 
  
  Corner values for the cell are computed using gfs_cell_corner_value
  and then used to do the bi(tri)linear interpolation. This has been
  tested succinctly and provides true continuous (C1)
  interpolation. This was not the case before.
  
  An important note is that the value at the center of the cell obtained
  by (bi)trilinear interpolation is NOT equal to the variable value at
  the center (it is equal to the mean of the corner values).
  

2004-10-14 Stephane Popinet <popinet at users.sourceforge.net>
  * Support for binary simulation files (gerris--mainline--0.5--patch-5)
  gerris--mainline--0.5--patch-5
  Keywords: 
  
  Binary files speed up reading time by a factor of ~5 for big simulations.
  There is no consideration of endianess at this point.
  

2004-10-13 Stephane Popinet <popinet at users.sourceforge.net>
  * Checks for periodic bc in interpolations (gerris--mainline--0.5--patch-4)
  gerris--mainline--0.5--patch-4
  Keywords: 
  
  Periodic boundary conditions do not work yet with interpolations for
  solid boundary fluxes and for corner values.
  

2004-10-11 Stephane Popinet <popinet at users.sourceforge.net>
  * Control of diffusion bc on immersed boundaries (gerris--mainline--0.5--patch-3)
  gerris--mainline--0.5--patch-3
  Keywords: 
  
  Boundary conditions on immersed boundaries for the diffusion equation
  can now be set in parameter files.
  

2004-10-08 Stephane Popinet <popinet at users.sourceforge.net>
  * Conversion of .cvsignore to .arch-inventory (gerris--mainline--0.5--patch-1)
  gerris--mainline--0.5--patch-1
  Keywords: 
  
  

2005-01-18 Stephane Popinet <s.popinet at niwa.co.nz>
  * Initial commit (s.popinet at niwa.co.nz--2004/gerris--mainline--0.5)