Menu
▾
▴
csp-commits
|
From: <mk...@us...> - 2003-05-03 09:47:57
|
Update of /cvsroot/csp/APPLICATIONS/CSPSim
In directory sc8-pr-cvs1:/tmp/cvs-serv8675
Modified Files:
CHANGES.current CHANGES
Log Message:
see CHANGES.current
Index: CHANGES.current
===================================================================
RCS file: /cvsroot/csp/APPLICATIONS/CSPSim/CHANGES.current,v
retrieving revision 1.27
retrieving revision 1.28
diff -C2 -d -r1.27 -r1.28
*** CHANGES.current 23 Apr 2003 18:47:29 -0000 1.27
--- CHANGES.current 3 May 2003 09:47:52 -0000 1.28
***************
*** 1,178 ****
! Version 0.3.2 (in progress)
===========================
! 2003-04-23: delta
! Added m_CharacterSize to ScreenInfo to control size of character
! (m_FontSIze controls now font resolution).
!
! Cleaned a bit this code.
!
! Added average fps.
!
! 2003-04-23: onsight
! Upgraded configure.in to check for OSG 0.9.4. The full test,
! which was broken in 0.9.3, works again. The test for osgGLUT
! has been removed since it is no longer needed.
!
! Removed osgGLUT from the Linux makefile.
!
! 2003-04-22: onsight
! Port to OSG 0.9.4, mostly changes to text display code. The new
! osgText code seems a bit inefficient, causing a frame rate hit
! for the onscreen status display. Text under Linux appears to be
! working again though.
!
! Changed classes that had been inheriting from simdata::Object
! just for reference counting to use simdata::Referenced instead.
!
! All non-XML simdata references now use simdata::Ref<>, and the
! XML refences use simdata::Link<>. simdata::Pointer<>, which had
! been used for both purposes, is no longer available. Ref is the
! basic smart-pointer class with reference counting. Link is
! similar but meant only for simdata::Objects that are creating
! automatically as part of the data archive loading process (ie.
! members that have SIMDATA_XML(...) lines in the interface.
!
! Some general header cleanups, which managaed to eliminate some
! unnecessary interdependency.
!
! 2003-04-21: delta
! Removed non necessaray calls to DegreesToRadians and
! RadiansToDegrees. Now angles must be specified in radians. If it
! is needed to display angle in degree call
! simdata::RadiansToDegrees at this stage.
!
! 2003-04-19: delta
! Got ride of more than one hundred of warnings (complaining about
! casting from double to float) in ColorSpace.cpp while compiling
! with vc++.
!
! 2003-04-19: onsight
! Changed AircraftObject:: m_heading, m_pitch, m_roll to
! m_Heading, m_Pitch, m_Roll. These are now in radians, and
! converted to degrees only if necessary (for display).
! setAttitude(x, y, z) still takes degrees.
!
! 2003-04-18: delta
! Added Runge-Kutta Cask-Karp numerical method.doSimStep2() is now
! calling it.
!
! Added a note on initialization of 'pause'.
!
! Added a simdata::Pointer<DynamicalObject> parameter in
! ObjectStats methods.
!
! Added a check to non nullitty of m_GameScreen in
! CSPSim::setActiveObject.
!
! Fixed a getStats non const method in AircraftObject class.
!
! Checked Onsight changes: a couple of forward class declarations
! causing compilation problems in vc++ have been fixed thanks to
! Onsight.
!
! Updated .NET project.
! 2003-04-18: onsight
! Lots of changes to many parts of CSPSim. First and foremost, the
! scene graph is now in VirtualScene. VirtualBattlefield no longer
! contains scene graph code. Much of VB still needs to be replaced
! though.
!
! The scene graph is reorganized slightly, in anticipation of
! adding more static and dynamic objects:
!
! +------------ ROOT -------------+
! | |
! BACKGROUND +--------OBJECTS-----+
! | | | |
! EYE_TRANSFORM TERRAIN STATIC DYNAMIC
! | |
! SKY SKYLIGHTS
! / \
! SUNLIGHT MOONLIGHT
!
! Fixed an initialization error in AeroDynamics that could cause
! m_GE to be NAN.
!
! Fixed the initialization of the gear retract state.
!
! Moved the first object update out of the main simloop so that
! the pause (several seconds) occurs before time starts counting.
! Limited the FrameTime to 0.2 seconds (was 1.0), which prevents
! the FM from exploding. Removed the initial 'paused' state.
!
! Changed SimObject to use quaternions for attitude, no 3x3
! matrices anymore.
!
! Added a SceneModel class to separate the model transform and
! switch from SimObject. The SceneModel member is only constructed
! if the object is added the scene. The ObjectModel is always
! loaded, so that bounds checking, collision tests, and model
! related data are available independent of the scene graph.
! SceneModel now uses osg::PositionAttitudeTransform instead of
! MatrixTransform.
!
! Set a default joystick environment variable for SDL if none is
! specified. (Linux only)
!
! Fixed AircraftObject::setAttitude(pitch, roll, heading)
! parameter order and units. This order is now correct, and the
! units are degrees. Previously the parameters were mislabelled
! and doubly converted from degrees to radians (resulting in very
! small angles).
!
! Moved access to the core components (Battlefield, Scene, etc)
! out of various classes that had been storing copies. Most access
! is now through CSPSim::theSim.
!
! Switched CSPSim to use simdata::DataManager instead of
! simdata::DataArchive.
!
! Major rewrite of the smoke trail code. The design is much
! cleaner now, but more work remains to be done.
!
! Various method name changes to conform to the standard case
! convensions (mixed with lower first letter).
!
! Removed "m_Army" from SimObject. This will eventually be
! replaced with a full object class, not just an integer.
!
! Made several bools in SimObject into bit flags.
!
! Greater use of smart-pointers (simdata::Pointer and
! osg::ref_ptr)
!
! Corrected handling of Terrain object pointers (now all smart
! pointers). Changed demeter to use osg::Referenced instead of
! it's own home-built work- alike.
! Moved the eye position in view <1> to the correct cockpit
! position. The 3D model remains on in the interior view now. The
! offset to the pilot's head is specified in the ObjectModel XML
! data. (There's a temporary contact point at the same location to
! make it easy to see where this is).
! Fixed ThrottleInput so that the throttle starts at zero for
! keyboard-only input, and has the correct sense ('=' increases,
! '-' decreases).
! Changed the log sync and closing to ensure that the log is
! flushed properly under all circumstances and that it doesn't
! segfault when the program quits. I still don't fully understand
! all the issues, and it may not work perfectly under Windows as
! is.
! Reordered various ctor initializers to silence gcc warnings.
! Added Bin/TestObjects.py to allow objects to be added to the sim
! for testing purposes. Currently, only the active object (created
! in CSPSim::init at startup) has physics. All other aircraft
! objects are frozen (this is set in AircraftObject---the physics
! code is commented out for non-human controlled a/c). To switch
! the active object, hit <spacebar>.
! Started a Falcon 4 aircraft data file converter in Tools/XML. It
! is currently can be used to convert the thrust tables to
! simdata's XML format.
- *********** VC USERS: please add VirtualScene.cpp to the VC project files
--- 1,44 ----
! Version 0.3.3 (in progress)
===========================
! 2003-05-02: onsight
! ==========> VC USERS: add Theater.cpp to the project.
! I've been making many changes and additions, but have been too
! lax in updating the changelog. Here's what I can remember and
! I will try to fill in the missing pieces as I come across them.
! Jitter fix. The scene graph has been restructured so that the
! eyepoint is at the origin. Dynamic objects, feature cells
! (more on that below), and the terrain transform are updated
! every frame to compensate for motion of the camera. Terrain
! lattice boundaries (the old local coordinate system) only affect
! terrain, so other objects can now move/exist continuously in
! adjacent terrain blocks. Particle effects are now in a global
! coordinate system which causes the to jitter somewhat. This
! could still be improves somewhat by moving the global origin
! to the center of the theater.
!
! Some fixes to CSPSim::cleanup() to avoid segfaults at program
! exit.
! Added much infrastructure to VirtualBattlefield. The battlefield
! now consists of a grid of cells that track static and dynamic
! objects, manage bubbles and aggregation/deaggregation, and
! dynamically add and remove objects from the scene graph.
! Added theater infrastructure to load static objects into the
! battlefield from XML. The terrain object is now part of the
! theater. There are some new files in Data/XML/theater to
! provide some very basic groups of static features for testing.
! Added aggregate(), deaggregate(), and setVisible() methods to
! DynamicObject. These are called automatically by the battlefield.
! They don't do anything in DynamicObject, but can be extended
! in subclasses to implement aggregation logic.
! Made DynamicObject::setHuman() private, and setAir() and
! setGrounded() protected. Changes of the 'human' state (when the
! active object changes) must now go through VirtualBattlefield's
! setHuman() method so that the bubbles are updated correctly.
Index: CHANGES
===================================================================
RCS file: /cvsroot/csp/APPLICATIONS/CSPSim/CHANGES,v
retrieving revision 1.4
retrieving revision 1.5
diff -C2 -d -r1.4 -r1.5
*** CHANGES 18 Apr 2003 12:07:33 -0000 1.4
--- CHANGES 3 May 2003 09:47:52 -0000 1.5
***************
*** 2,5 ****
--- 2,184 ----
to CVS.
+ Version 0.3.2 (2003-05-02)
+ ===========================
+
+ 2003-04-23: delta
+ Added m_CharacterSize to ScreenInfo to control size of character
+ (m_FontSIze controls now font resolution).
+
+ Cleaned a bit this code.
+
+ Added average fps.
+
+ 2003-04-23: onsight
+ Upgraded configure.in to check for OSG 0.9.4. The full test,
+ which was broken in 0.9.3, works again. The test for osgGLUT
+ has been removed since it is no longer needed.
+
+ Removed osgGLUT from the Linux makefile.
+
+ 2003-04-22: onsight
+ Port to OSG 0.9.4, mostly changes to text display code. The new
+ osgText code seems a bit inefficient, causing a frame rate hit
+ for the onscreen status display. Text under Linux appears to be
+ working again though.
+
+ Changed classes that had been inheriting from simdata::Object
+ just for reference counting to use simdata::Referenced instead.
+
+ All non-XML simdata references now use simdata::Ref<>, and the
+ XML refences use simdata::Link<>. simdata::Pointer<>, which had
+ been used for both purposes, is no longer available. Ref is the
+ basic smart-pointer class with reference counting. Link is
+ similar but meant only for simdata::Objects that are creating
+ automatically as part of the data archive loading process (ie.
+ members that have SIMDATA_XML(...) lines in the interface.
+
+ Some general header cleanups, which managaed to eliminate some
+ unnecessary interdependency.
+
+ 2003-04-21: delta
+ Removed non necessaray calls to DegreesToRadians and
+ RadiansToDegrees. Now angles must be specified in radians. If it
+ is needed to display angle in degree call
+ simdata::RadiansToDegrees at this stage.
+
+ 2003-04-19: delta
+ Got ride of more than one hundred of warnings (complaining about
+ casting from double to float) in ColorSpace.cpp while compiling
+ with vc++.
+
+ 2003-04-19: onsight
+ Changed AircraftObject:: m_heading, m_pitch, m_roll to
+ m_Heading, m_Pitch, m_Roll. These are now in radians, and
+ converted to degrees only if necessary (for display).
+ setAttitude(x, y, z) still takes degrees.
+
+ 2003-04-18: delta
+ Added Runge-Kutta Cask-Karp numerical method.doSimStep2() is now
+ calling it.
+
+ Added a note on initialization of 'pause'.
+
+ Added a simdata::Pointer<DynamicalObject> parameter in
+ ObjectStats methods.
+
+ Added a check to non nullitty of m_GameScreen in
+ CSPSim::setActiveObject.
+
+ Fixed a getStats non const method in AircraftObject class.
+
+ Checked Onsight changes: a couple of forward class declarations
+ causing compilation problems in vc++ have been fixed thanks to
+ Onsight.
+
+ Updated .NET project.
+ 2003-04-18: onsight
+ Lots of changes to many parts of CSPSim. First and foremost, the
+ scene graph is now in VirtualScene. VirtualBattlefield no longer
+ contains scene graph code. Much of VB still needs to be replaced
+ though.
+
+ The scene graph is reorganized slightly, in anticipation of
+ adding more static and dynamic objects:
+
+ +------------ ROOT -------------+
+ | |
+ BACKGROUND +--------OBJECTS-----+
+ | | | |
+ EYE_TRANSFORM TERRAIN STATIC DYNAMIC
+ | |
+ SKY SKYLIGHTS
+ / \
+ SUNLIGHT MOONLIGHT
+
+ Fixed an initialization error in AeroDynamics that could cause
+ m_GE to be NAN.
+
+ Fixed the initialization of the gear retract state.
+
+ Moved the first object update out of the main simloop so that
+ the pause (several seconds) occurs before time starts counting.
+ Limited the FrameTime to 0.2 seconds (was 1.0), which prevents
+ the FM from exploding. Removed the initial 'paused' state.
+
+ Changed SimObject to use quaternions for attitude, no 3x3
+ matrices anymore.
+
+ Added a SceneModel class to separate the model transform and
+ switch from SimObject. The SceneModel member is only constructed
+ if the object is added the scene. The ObjectModel is always
+ loaded, so that bounds checking, collision tests, and model
+ related data are available independent of the scene graph.
+ SceneModel now uses osg::PositionAttitudeTransform instead of
+ MatrixTransform.
+
+ Set a default joystick environment variable for SDL if none is
+ specified. (Linux only)
+
+ Fixed AircraftObject::setAttitude(pitch, roll, heading)
+ parameter order and units. This order is now correct, and the
+ units are degrees. Previously the parameters were mislabelled
+ and doubly converted from degrees to radians (resulting in very
+ small angles).
+
+ Moved access to the core components (Battlefield, Scene, etc)
+ out of various classes that had been storing copies. Most access
+ is now through CSPSim::theSim.
+
+ Switched CSPSim to use simdata::DataManager instead of
+ simdata::DataArchive.
+
+ Major rewrite of the smoke trail code. The design is much
+ cleaner now, but more work remains to be done.
+
+ Various method name changes to conform to the standard case
+ convensions (mixed with lower first letter).
+
+ Removed "m_Army" from SimObject. This will eventually be
+ replaced with a full object class, not just an integer.
+
+ Made several bools in SimObject into bit flags.
+
+ Greater use of smart-pointers (simdata::Pointer and
+ osg::ref_ptr)
+
+ Corrected handling of Terrain object pointers (now all smart
+ pointers). Changed demeter to use osg::Referenced instead of
+ it's own home-built work- alike.
+
+ Moved the eye position in view <1> to the correct cockpit
+ position. The 3D model remains on in the interior view now. The
+ offset to the pilot's head is specified in the ObjectModel XML
+ data. (There's a temporary contact point at the same location to
+ make it easy to see where this is).
+
+ Fixed ThrottleInput so that the throttle starts at zero for
+ keyboard-only input, and has the correct sense ('=' increases,
+ '-' decreases).
+
+ Changed the log sync and closing to ensure that the log is
+ flushed properly under all circumstances and that it doesn't
+ segfault when the program quits. I still don't fully understand
+ all the issues, and it may not work perfectly under Windows as
+ is.
+
+ Reordered various ctor initializers to silence gcc warnings.
+
+ Added Bin/TestObjects.py to allow objects to be added to the sim
+ for testing purposes. Currently, only the active object (created
+ in CSPSim::init at startup) has physics. All other aircraft
+ objects are frozen (this is set in AircraftObject---the physics
+ code is commented out for non-human controlled a/c). To switch
+ the active object, hit <spacebar>.
+
+ Started a Falcon 4 aircraft data file converter in Tools/XML. It
+ is currently can be used to convert the thrust tables to
+ simdata's XML format.
+
+ *********** VC USERS: please add VirtualScene.cpp to the VC project files
+
Version 0.3.1 (2003-04-18)
===========================
|
Thanks for helping keep SourceForge clean.
X