fix gcc warning

pretty print

New raster ("bitmap") PS/EPS/PDF output

I forgot to commit these last week: it's just the beginning of some
cleanup of the old 2D algorithm (removed non-functional voronoi
insertion, fixed a couple of memory leaks, added some comments, etc.)

And more airplane work!

- merged STL patch from Nicolas Tardieu. This patch permits to
  create an real elementary surface for each STL face. Until
  we have some code to directly remesh the STL data (which would
  be much nicer and cheaper), this is very useful. We can finally
  use STL representations to create meshable volume models.
  Thanks Nicolas !

- new option to color the mesh by physical entities (if an element
  belongs to multiple physicals, we use the first one)

- additional cleanup of the mesh drawing code

- big cleanup of the geometry creation section in the parser:

  * removed horrible Geo/DataBase.cpp
  * fixed all List-related memory leaks (a lot!)
  * added tests to check if we don't try to recreate existing
    entities (this should prevent many weird bugs due to incorrect
    input files)

- fixed LineLoop/EdgeLoop memory leaks

- many small fixes all over the place

Update copyright notice

Previous commit was completely wrong. Need to think a little more
about this problem.

increment the factor more slowly

Better edge swapping for non-recombined extruded meshes

Removed extra arg for Init_Mesh

Add a couple of 'else if's

new "Delete All;" command + only warn when reloading duplicate geometry
points.

some more cleanups

List_Replace->List_Insert

Handle multiple physicals for single elementary

New MSH file format, version 2.0.

********************************************************
************ Please send me your comments **************
********************************************************

Small changes for skeletton of new mesh format

Added support to visualize mesh partitions. The partition info is currently
set as the physical entity number in the MSH file, but it would be trivial
to change it to something else (e.g., as soon as we generalize the MSH
format to take arbitrary number of tags per element).

Explicitly say that the mesh will be incorrect if the swapping failed

Fixed one more bug for degenerate, non-recombined extrusions, that could
lead to duplicate elements.

Fix seg fault in Read_SMS

Fixed degenrate tets generation (oops!)

- new DecomposeInSimplex plugin
- new "Combine Views" command in the parser
- cleaned up View::tranform and View::smooth (+small bug fix for SLs)
- better check for transparency sorting
- make depend
- new View.AlphaChannel option to change the transparency
  globally for a given view (instead of having to change the colormap
  by hand every time)

With all the above, we can now generate very nice (and correct)
transparent iso-surfaces:

Merge "../tutorial/view3.pos";

minIso = 0;
maxIso = 2;
nbIso = 7;

Plugin(CutMap).iView = 0;

For i In {1:nbIso}
  Plugin(CutMap).A = (maxIso-minIso)/nbIso * i;
  Plugin(CutMap).Run;
EndFor

Delete View[0];
Combine Views;

For i In {1:nbIso}
  Delete View[0];
EndFor

Plugin(DecomposeInSimplex).iView = 0;
Plugin(DecomposeInSimplex).Run;

View[0].AlphaChannel = 0.6;

Fix name

Cleaned up the whole Include/Merge/Open mess (removed the ugly
yyin/yyname tables in the parser, made the Merge routine reentrant
+ uniformized all the INFO/STATUS messages).

This should fix a nasty side effect from this morning's
commit (wrong yynames when doing multiple merges in the same
file).

Fix 2 gcc warnings

- Fix element numbering in Extrude Point and Extrude Line
- Look for config files in $GMSH_HOME before $HOME

- Comment JF's last changes in the parser until we fix the syntax

- Update gl2ps

- Small fix in the configure script to better detect fltk-config

- Fix Parser/Makefile (for older versions of bison)

New native PDF support

Make it possible to change Triangle options

Don't optimize 3D_Mesh.

polish

Remove 2nd order for 3D altogether, since it still produces incorrect
meshes if edges of the surface mesh get swapped

forgot this one

Shout if we try to generate second order meshes in non-implemented
cases.

Some morer work on second order elements