now, with a 'resize' routine for the engine we can/should also recreate some
buffers that are created only at init time (e.g., 2d filters, dome fbos, ...).
This bug was always present in Blender (since 2.49 at least).
Bugfix supported by NF-UBC Nereus Program as part of the development
of OceanViz/NereusViz
You can see a screenshot of the funcionality here:
http://wiki.blender.org/index.php/Doc:2.6/manual/Game_Engine/Performance/Display/Framerate_and_Profile
This patch creates a bar-like graph to quickly allow the game dev to see the performance changes.
Also it changes the font to monospace (too allow ' ' padding) and reduced shadow border to
match the blenderplayer one.
Patch finalized and commited at Global Game Jam Vancouver (last one, time to sleep forever now)
# Print a message when an async LibLoad is done
import bge
def finished_cb(status):
print("Library (%s) loaded in %.2fms." % (status.libraryName, status.timeTaken))
bge.logic.LibLoad('myblend.blend', 'Scene', async=True).onFinish = finished_cb
LibLoad() now returns a KX_LibLoadStatus object for information on the library loading. LibNew() and LibFree() are unaffected by this commit. In other words, the async option only works for LibLoad(). Furthermore it only works for Scenes, not Actions or Meshes.
* Cleaning up the conversion code to avoid a per-face material conversion. Materials are now stored in buckets and only converted if a new material is found. This replaces some of Campbell's earlier work on the subject. His work wasn't as thorough, but it was much safer for a release.
* Shaders are only compiled for LibLoaded materials once. Before they could be compiled twice, which could really slow things down.
* Refactoring the rasterizer code to use a strategy design pattern to handle different geometry rendering methods such as immediate mode, vertex arrays and vertex buffer objects. VBOs are added, but they will be disabled in a following commit since they are still slower than vertex arrays with display lists. However, VBOs are still useful for mobile, so it's good to keep them around.
* Better multi-uv support. The BGE should now be able to handle more than two UV layers, which should help it better match the viewport.
viewport. This helps to eliminate OpenGL calls in weird places like the physics
code and to reduce glGet calls, which are expensive.
There should be no functional changes (except maybe a very slight speed improvement).
The problem is motion blur was being treated as a per scene operation, but all scenes were trying to use the same accumulation buffer. Now motion blur is done in EndFrame() instead of PostRenderScene().
---------------------------------------------------------------
This was a test drive to see how painful the merge will be.
Next batches are:
- use desktop option for fullscreen
- multisampling option
- bullet collision mask
- python
- storage (vbo, dl, ...)
- lighting
[lighting still needs review]
[python could use review, although it should be straightforward]
[storage should be tested more I think]
Merged /branches/soc-2011-cucumber:r
36991,37059,37157,37416,37497-37499,37501,37522,39036,40593
36991:
==UI==
* Made some options available in Blender Game that were only available in Blender Render (camera resolution, animation fps)
* Created a panel for the embedded player
* Renamed the FPS option for the standalone player to Refresh Rate
* Moved framing options to display
* Made a button to launch the blender player from within blender (only tested on windows for now)
37059:
==UI==
* Added the option to change the exit key for the BGE. The UI currently just sets a number, and this feature most likely does not work for blenderplayer yet. More work on this to come.
* Removed the physics settings from the scene panel for the BGE.
* Added an Add menu in the logic brick header.
37157:
Making the bake options available in Blender Game
37416:
Making the exit key UI element accept key presses instead of numbers. It still does not work for the Blenderplayer, and it does not limit the input to key presses (other events don't work for exiting)
37497:
Some more work on getting the exit key to work in the Blenderplayer.
Input is now restricted to keyboard events only for the exit key UI.
37498:
Some clean up from the last commit.
The exit key setting affects the Blenderplayer now.
37499:
Cleaning up some duplicate code. Now the reverseTranslateTable for converting blender key codes to ketsji key codes is only defined in BL_BlenderDataConverter.
37501:
Centralizing the exit key methods to the keyboard devices. This should make it easier to get exit key control to the python API.
[37517: committed previously]
37522:
Moved control of the exit key away from the keyboard devices, and moved it to ketsjiengine.
Added setExitKey and getExitKey to the python API
39036:
A couple of the doversions were in the wrong spot. This should fix some issues with the exit key not being set.
[not committed entirely, see below]]
40552: space_logic.py (* fixed an error in space_logic.py *)
40593:
launch blenderplayer from ui not working in OSX fix - by Daniel Stokes and me
########################################################
code left behind (to be included in next commit):
########################################################
{
/* Initialize default values for collision masks */
Object *ob;
for(ob=main->object.first; ob; ob=ob->id.next)
ob->col_group = ob->col_mask = 1;
}
- Added support of variable size sensor width and height.
- Added presets for most common cameras, also new presets can be defined by user.
- Added option to control which dimension (vertical or horizontal) of sensor
size defines FOV. Old behavior of automatic FOV calculation is also kept.
- Renderer, viewport, game engine and collada importer/exporter should
deal fine with this changes. Other exporters would be updated soon.
geez, who coded the font object? or even worse, who did the dome code?
Don't coders talk? tsc tsc ...
Now seriously, KX_KetsjiEngine::RenderFonts() could be moved to inside the KX_Scene class. It probably should (so I could call it from inside KX_Dome::RenderDomeFrame()). Not critical, so not changing it for now.
http://projects.blender.org/tracker/?func=detail&aid=28167&group_id=9&atid=306
Game Actuator (restart or load a new file) will not keep some settings alive (as
we had in 2.49).
In 2.49 the solution used was to use Blender globals (G.fileflags) to get/set
those settings. That was causing the blender file to change if you change the
material mode from the game.
In 2.5 this never worked, and the implementation was buggy (it's relying in the
scene settings, which get reset ever time we restart/load a new file).
My idea for fixing this is to create a new struct (GlobalSettings) where we
store any setting to be preserver during the course of the game. This is
specially important for options that require the game to restart/load new file
(graphic ones). But it later can be expanded to support other things such as
audio settings (e.g. volume), ...
I'm also planning to expand it for stereo and dome settings, but I prefer to
first get this committed and then build a new patch on top of that.
I had some problems in finding a correct way for build/link the blenderplayer
changes, so although it's working I'm not sure this is the best code (e.g. I
couldn't make forward declaration to work in GPG_Application.h for the struct
GlobalSettings so I ended up including KX_KetsjiEngine.h)
[note: I talked with Brecht and he find this is an ok solution. He implemented
it originally so it's good to have his go. However I still think there must be a way to make forward declaration to work. I will see with other devs if there is a better solution]
[also I'm likely renaming glsl to flags later if there are more settings stored in the flags to be used. But for now we are only handling glsl flags]
Final GSoC commit.
* Bugfix: Negative frames crashed
* Bugfix: JOS sample buffer size prediction error (wasted memory)
* Optimisation: for JOS upsampling (around 12 % difference measured here)
* Optimisation: Better filter for JOS resampling
* Bugfix: Error in relative 3D audio code.
* Removed Attenuation
* Bugfix: Multiple scenes in BGE lead to errors, BGE audio now all relative, to support multiple scenes.
High quality resampling on mixdown, linear for playback.
* Lots of improvements and fixes for the JOS resampler, now it works fine!
* High quality filter coefficients for the JOS resampler (sorry for the 5 MB source file).
* Fix for GE orientation bug. Note: moto uses x,y,z,w quaternion storage, while rest of blender uses w,x,y,z.
* Minor changes/fixes.
* Pepper depends on ffmpeg 0.7.1 or higher now, windows and mac build systems set to ffmpeg-0.8
* Fixed orientation retrieval in OpenAL device code.
* Added stopAll() method to AUD_IDevice (also for Python) and call it on BGE exit
* Changed BGE to use audaspace via native C++ instead over the C API.
* Made AUD_SequencerFactory and AUD_SequencerEntry thread safe.
* Changed sound caching into a flag which fixes problems on file loading, especially with undo.
* Removed unused parameter from sound_mute_scene_sound
* Fixed bug: changing FPS didn't update the sequencer sound positions.
* Fixed bug: Properties of sequencer strips weren't set correctly.
* Minor warning fixes.
This was never highly tested, that's why I never committed (my patch for this was from September 2010).
But once again I got a report that this bug was a deal-break and the patch seems to work for this artist.
I believe it's working, but I will keep my eyes open for this.