Commit by vshabanov :: r193 /trunk/camlosg/Make/makedefs: (link)

Added -lpthread to CCLIBS

Commit by vshabanov :: r192 /trunk/camlosg/osgViewer/ (osgViewer_ml.cpp viewer.ml): (link)

render_loop now uses pthread_cond, instead of mutex (which was locked in one thread and unlocke in another).

Commit by vshabanov :: r191 /trunk/camlosg/ (3 files in 2 dirs): (link)

Added some project-specific (non camlosg-specific) machinery for rendering in parallel thread w/o caml_enter/leave_blocking_section.

Commit by vshabanov :: r190 /trunk/camlosg/osgCal/ (osgCal.ml osgCal_ml.cpp): (link)

Added OsgCal.Core_model.load_simple & OsgCal.Model.load_simple

Commit by vshabanov :: r189 /trunk/camlosg/osgGA/osgGA.ml: (link)

OsgGA.Utils renamed to OsgGA.Deprecated_osgGA_utils

Commit by vshabanov :: r188 /trunk/camlosg/ (3 files in 2 dirs): (link)

 * Commented out OsgGA.Gui_event_handler.create (implementation currently doesn't contain CAMLlocal4)
 * Removed caml_enter/leave_blocking_section from OsgCal.Core_model.load

This done temporarily for experimenting with SIGSEGV during game load process.

Commit by vshabanov :: r187 /trunk/camlosg/ (Make/makedefs include/shared_part.h): (link)

allocRawStructWoDtor now uses double array, instead of custom block. alloc_small is faster than alloc_custom and we don't need comparisons on these values.

Commit by vshabanov :: r186 /trunk/camlosg/osg/osg_ml.cpp: (link)

AnimationPath::ControlPoint now also allocated as double array.

Commit by vshabanov :: r185 /trunk/camlosg/osg/ (osg_ml.cpp timer.ml): (link)

Timer internally represented as float, not int64, due to dlow int64 allocations.

Commit by vshabanov :: r184 /trunk/camlosg/ (8 files in 2 dirs): (link)

Vec2/3/4 & Quat are now records with double fields (so their allocation and access to components do not require caml_c_call).

Commit by vshabanov :: r183 /trunk/camlosg/osg/matrix.ml: (link)

Little fix in Osg.Matrix.to_array (it was actually converting to tuple array, not array array)

Commit by vshabanov :: r182 /trunk/camlosg/osg/ (matrix.ml osg_ml.cpp): (link)

Osg.Matrix.translate/scale now donw in caml. Mult only calls C to perform multiplication, not allocation.

Commit by vshabanov :: r181 /trunk/camlosg/ (3 files in 2 dirs): (link)

Osg.Matrix now 16-float element record instead of custom structure.

to_array/to_tuple are now performed directly in caml, translate/scale/rotate are still done in C side, but use alloc_small instead of alloc_custom.

Commit by vshabanov :: r180 /trunk/camlosg/osgViewer/ (osgViewer_ml.cpp viewer.ml): (link)

Added OsgViewer.Viewer.get_camera_manipulator

Commit by vshabanov :: r179 /trunk/camlosg/Make/makerules: (link)

Fixes for make install with NO_BYTECODE option

Commit by vshabanov :: r177 /trunk/camlosg/osgCal/ (osgCal.ml osgCal_ml.cpp): (link)

Added OsgCal.Model.set_auto_update/update

Commit by vshabanov :: r176 /trunk/camlosg/ (9 files in 8 dirs): (link)

Some fixes for static linking OSG (for profiling via gprof)

Commit by vshabanov :: r174 /trunk/camlosg/osgCal/ (osgCal.ml osgCal_ml.cpp): (link)

• Added OsgCal.Mesh_parameters ~no_software_vertex_update
• Added OsgCal.Model.set_time_factor

Commit by vshabanov :: r173 /trunk/camlosg/include/shared_part.h: (link)

Fix for compiling under gcc-4.3

Commit by vshabanov :: r172 /trunk/camlosg/osgText/ (osgText.ml osgText_ml.cpp): (link)

Added osgText.Text.set_kerning_type