As pointed out in #43, geometry shaders are officially supported by OpenGL since 3.2.

Reference: http://open.gl/geometry

Could be an optional feature based on renderer version (OGL 2 or 3.2)

There's no project version indication in current root CMakeLists.txt.

It should have one.

I have Been recently Using horde and with EVERY program I compile run except for the examples gives me this:

invalid ...... handle in .......

with every function that includes nodes
here is one of my report logs:

0.000
Initializing GL2 backend using OpenGL driver '3.3.0' by 'NVIDIA Corporation' on 'GeForce 8400 GS/PCIe/SSE2'
0.039
Invalid resource handle in h3dLoadResource
0.041
Invalid resource handle in h3dLoadResource
0.042
Loading resource '$Tex2D'
0.043
Loading resource '$Tex3D'
0.044
Invalid resource handle in h3dAddNodes
0.046
Invalid node handle in h3dSetupModelAnimStage
0.049
Invalid resource handle in h3dAddCameraNode
0.052
Invalid node handle in h3dSetNodeParamI
0.053
Invalid node handle in h3dSetNodeParamI
0.055
Invalid node handle in h3dSetNodeParamI
0.057
Invalid node handle in h3dSetNodeParamI
0.058
Invalid node handle in h3dSetupCameraView
0.060
Invalid resource handle in h3dResizePipelineBuffers
0.061
Invalid node handle in h3dSetModelAnimParams
0.063
Invalid node handle in h3dSetNodeTransform
0.065
Invalid node handle in h3dRender
0.089
Invalid node handle in h3dSetModelAnimParams
0.091
Invalid node handle in h3dSetNodeTransform
0.093
Invalid node handle in h3dRender
0.094
Invalid node handle in h3dSetModelAnimParams
0.095
Invalid node handle in h3dSetNodeTransform
0.097
Invalid node handle in h3dRender
0.102
Invalid node handle in h3dSetModelAnimParams
0.103
Invalid node handle in h3dSetNodeTransform
0.105
Invalid node handle in h3dRender
0.106
Invalid node handle in h3dSetModelAnimParams
0.107
Invalid node handle in h3dSetNodeTransform
0.110
Invalid node handle in h3dRender
0.111
Invalid node handle in h3dSetModelAnimParams
0.113
Invalid node handle in h3dSetNodeTransform
0.115
Invalid node handle in h3dRender
0.118
Invalid node handle in h3dSetModelAnimParams
0.119
Invalid node handle in h3dSetNodeTransform
0.121
Invalid node handle in h3dRender
0.123
Invalid node handle in h3dSetModelAnimParams
0.174
Invalid node handle in h3dSetNodeTransform
0.176
Invalid node handle in h3dRender
0.177
Invalid node handle in h3dSetModelAnimParams
0.178
Invalid node handle in h3dSetNodeTransform
0.180
Invalid node handle in h3dRender
0.183
Invalid node handle in h3dSetModelAnimParams
0.184
Invalid node handle in h3dSetNodeTransform
0.185
Invalid node handle in h3dRender
0.187
Invalid node handle in h3dSetModelAnimParams

and so on..

any response would be usefull

I think a project need that people at less know how to use the same tools you are using (enve if they choose to use other later for their own setup).

So I suggest to delete the root dir CMake for

• Can be confusing to "new" people not used to cmake, they could think it is a part of cmake and not even look inside. Or because there is already files for VS you would not get why these are there.
• cmake . -G "Xcode" is the correct one to run on OSX as today.
• you could include in the readme the reference to cmake . that I think will try to run the default for the system (anyone can say if it is true or not), and perhaps include a CMAKE_USAGE.md or something that list all the possible outputs like:
  • For Linux or default developer tools of system cmake .
  • For XCode cmake . -G "Xcode"
  • For VS 2005 cmake -G "Visual Studio 8 2005"
  • For VS 2008 cmake -G "Visual Studio 9 2008"

And there could be also the indication in how to enable/disable terrain extension. So people get more feedback about cmake build chain and purpose.

Both 1.0 and current revisions fail:

make[1]: don't know how to make install/strip. Stop

I'm pretty sure something's wrong with your cmake makefiles.

On FreeBSD.

I've got a compiler warning in following function. The compiler is complaining that the function is not returning anything. Should the function signature be changed to void?

bool TextureResourceEx::importTexGL(uint32 texGL, int width, int height )
{
    _loaded = true;
    _width = width;
    _height = height;

    if( _texObject == defTex2DObject )
    {
        m_imported = false;
        RenderDeviceInterface *rdi = Modules::renderer().getRenderDevice();
        _texObject = rdi->createTexture( _texType, _width, _height, _depth == 0 ? 1 : _depth, _texFormat,
                                         false, false, false, _sRGB );
    }
    replaceTexObj( texGL );
    m_imported = true;
}

Hi!

Just cloned the master head when revisiting an older test case, getting some deprecation warnings.

This is just a warning, but considering everything else in Horde3D is pedantic perfect, you'd want to fix this too perhaps. It's an easy fix:
http://stackoverflow.com/questions/9838385/replace-of-xkeycodetokeysym

I could even provide a patch, although it looks like a case of copy paste from the already found sollution.

When I make Horde3D in mingw32, I got the error :

`strncpy_s' was not declared in this scope

I found there is a macro in https://github.com/horde3d/Horde3D/blob/master/Horde3D/Source/Shared/utPlatform.h#L77

#if (!defined( _MSC_VER ) || (defined( _MSC_VER ) && (_MSC_VER < 1400))) && !defined( __MINGW32__ )
#   define strncpy_s( dst, dstSize, src, count ) strncpy( dst, src, count < dstSize ? count : dstSize )
#endif

Why !define(__MINGW32__) here ?

Add an OpenGL 4-based renderer to better support modern hardware capabilities (comparing to D3D11).

Useful links:

http://www.horde3d.org/forums/viewtopic.php?f=1&t=1904
http://rastergrid.com/blog/2011/10/opengl-vs-directx-the-war-is-far-from-over/

I have Horde3D as a submodule for one of my projects and it complains that the submodule was modified when Mac OS X binaries were built. Here's a fix:

https://gist.github.com/anonymous/8217677

I am getting following error when i try to execute:

cmake -DHORDE3D_FORCE_DOWNLOAD_GLFW=ON ../Tools/Horde3DEditor

CMake Warning at ../../CMakeLists.txt:19 (find_package):
  By not providing "FindGLFW.cmake" in CMAKE_MODULE_PATH this project has
  asked CMake to find a package configuration file provided by "GLFW", but
  CMake did not find one.

  Could not find a package configuration file provided by "GLFW" with any of
  the following names:

    GLFWConfig.cmake
    glfw-config.cmake

  Add the installation prefix of "GLFW" to CMAKE_PREFIX_PATH or set
  "GLFW_DIR" to a directory containing one of the above files.  If "GLFW"
  provides a separate development package or SDK, be sure it has been
  installed.


-- Found Qt4: /usr/bin/qmake-qt4 (found version "4.8.6") 
-- Found Qt4: /usr/bin/qmake-qt4 (found version "4.8.6") 
CMake Error at CMakeLists.txt:28 (SET_PROPERTY):
  set_property could not find TARGET Terrain.  Perhaps it has not yet been
  created.


CMake Error at CMakeLists.txt:39 (SET_PROPERTY):
  set_property could not find TARGET Chicago.  Perhaps it has not yet been
  created.

Errors:

/usr/ports/devel/horde3d/work/Horde3D-57678aa/Horde3D/Source/ColladaConverter/converter.cpp:915:20: error: cannot take the address of an rvalue of type 'float'
                fwrite_le<float>(&Matrix4f().x[j], 1, f);
                                 ^~~~~~~~~~~~~~~~
1 error generated.

/usr/ports/devel/horde3d/work/Horde3D-57678aa/Horde3D/Source/Horde3DEngine/egRenderer.cpp:2301:78: error: cannot take the address of an rvalue of type 'float'
        _renderDevice->setShaderConst( _defColorShader.uni_worldMat, CONST_FLOAT44, &Matrix4f().x[0] );
                                                                                    ^~~~~~~~~~~~~~~~

and warnings

/usr/ports/devel/horde3d/work/Horde3D-57678aa/Horde3DEditor/src/QPropertyEditor/Property.cpp:78:7: warning: case value not in enumerated type 'QVariant::Type' [-Wswitch]
        case QMetaType::Float:  
             ^
/usr/ports/devel/horde3d/work/Horde3D-57678aa/Horde3DEditor/src/QPropertyEditor/Property.cpp:103:7: warning: case value not in enumerated type 'QVariant::Type' [-Wswitch]
        case QMetaType::Float:  
             ^
/usr/ports/devel/horde3d/work/Horde3D-57678aa/Horde3DEditor/src/QPropertyEditor/Property.cpp:123:7: warning: case value not in enumerated type 'QVariant::Type' [-Wswitch]
        case QMetaType::Float:  
             ^

When building the project with a makefile generated by CMake, the samples don't work.

Xcode generates the executable in Horde3D/Binaries/Darwin/Debug/or Horde3D/Binaries/Darwin/Release/.

The makefile generates the executable in Horde3D/Binaries/Darwin/(not in a subfolder). Therefore, the resources in Content can not be found.

Marciano proposed on the forum:

We are currently thinking about taking out the scene tree (transformation hierarchy) from the Horde core. Why?

I begin to think more and more that is not a good idea if the 3d engine (and even less the renderer) is modifying transformations. Without a scene tree, all horde scene objects would just be added directly into a flat list. Besides making the game engine integration strategy clearer, this would also help to further simplify the horde code and would definitely be beneficial for performance, as the scene tree does not have to be updated every frame (an operation which causes a lot of cache misses).

The change would have the introduction of model resources as dependency. Along with it, we would probably give more control over the model update as well. It needs to be possible to explicitly and immediately apply the skeletal animation to a model, so that the joint matrices can be read back to manually attach an object to a joint. A simple function updateModel could take care of that and control as well when morphers and software skinning should be applied. BTW, if anyone really needs it, it would be possible to have a scene tree in the utility library.

Discussion on official thread:
http://horde3d.org/forums/viewtopic.php?f=8&t=1117

I think .NET related stuff should not live in the same source-tree of default C/C++ version, so I propose to move .NET samples to Bindings/C#/Source/Samples/.

Hello,
Do you know about Conan?
Conan is modern dependency manager for C++. And will be great if your library will be available via package manager for other developers.

Here you can find example, how you can create package for the library.

If you have any questions, just ask :-)

Hi guys,

I propose to add some new API functions relative to animation area.

For example, currently it looks impossible to retrieve info about the current weight and elapsed time of an animation.

There's a "h3dSetModelAnimParams(...)" function but not a "h3dGetModelAnimParams()" one.

The proposal is discussed in the related thread on the official forum:

http://www.horde3d.org/forums/viewtopic.php?f=8&t=1899&p=9496#p9496

Current loader assumes binary data endianness conforms architecture.

This http://www.horde3d.org/forums/viewtopic.php?f=1&t=1130 would be very useful to have. It would probably be a good idea to also provide a user pointer.

The reason we need this is that we have a virtual machine that runs glsl-like code that we would like to use to simulate wind and other forces acting on the particles.

Marciano proposed on the forum:

The idea is to introduce model resources. Model scene nodes will still exist but they just reference a Model resource instead of having child nodes for their data. Since the model resources will solely contain meshes and joints, it is a lot easier to optimize them.

Discussion threads:
http://horde3d.org/forums/viewtopic.php?f=8&t=733
http://horde3d.org/forums/viewtopic.php?f=8&t=1896

As reported by @attilaz in its fork:

Shipping GLFW with Horde3D SDK makes compiling and running samples easier by newbies, but I don't think it is a good idea to ship an old version of GLFW 2.7 missing a lot of bug fixes or going in trouble of updating bundled GLFW with every release with GLFW 3 being on the edge.

So my suggestion is to remove bundled GLFW from Horde3D SDK or atleast from the git repository and dynamically link against the installed version on target system.

Built-in samples could just link dynamically the lib.

A requirement check should be added to CMakeLists.txt.

Occasionally you can see extra triangles in the terrain.

These are most noticable if you go "under" the terrain and can see that every other triangle is exactly vertical

from the top of the terrain you see all the triangles and some extra ones you shouldn't

Hi! Not sure if this is a bug or if I'm screwing something up, I just recently started tinkering with Horde3D master repos.

While trying to adapt the SDL example to SDL2 I've come across a weird problem, I try to load the pipelines as resources:

    H3DRes test = h3dAddResource(H3DResTypes::Pipeline, "pipelines/forward.pipeline.xml", 0);
    if (!h3dutLoadResourcesFromDisk("<full and relative path to Horde3D/Horde3D/Binaries/Content tried>")) {
        if (!h3dutDumpMessages()) {
            fprintf(stderr, "ERROR: unable to log?\n");
        }
        fprintf(stderr, "ERROR: unable to load resources\n");
    }

h3dutLoadResourcesFromDisk always returns false, I've tried relative paths, and "." (copying all the content to my local path).

I tried dumping messages to a log to maybe find out why, but there is no log file available and no output to stdout/stderr. Subsequent call to h3dGetError returns false and tried setting MaxLogLevel to both 0 and 1000, neither has given any output afaik.

First, I'm wondering what I'm doing wrong since I'm unable to load any resources (the samples work fine) and secondly I'm wondering how to get information about WHAT the actual error is. Error parsing XML? Error finding the file? So far I've seen no output from Horde3D at all regardless of MaxLogLevel, is there no way to find out what error actually occured when calling a function?

Thanks for the help!

It looks like the entire engine's codebase doesn't follow the C++ const-correctness (in particular for getter functions).

Is it a precise design choice?

Some cents on the topic (in particular why C++ const-ness tends to be useless nowadays):

http://www.cprogramming.com/tutorial/const_correctness.html
http://duramecho.com/ComputerInformation/WhyHowCppConst.html

The samples (Chicago, Knight, etc.) doesn't seem to work on Linux. All I get is a black screen.
Tried two different nvidia machines with x64 Arch Linux.

This issue might be caused by the hack that made it compile in the first place (#66).

Edit: If I hit F7 (debug mode?) I can see some khaki colored plane when I move my mouse around.

Edit2: I'm very sorry. I didn't see the error log html file. It didn't find any resources because I ran the binaries from the cmake build directory.

This is more like a question than an issue. I tried to convert some animated character models(dae format) to geo format that H3D can process. I encountered a warning message saying my character has more than 75 joints. Once the model is loaded into the engine, the joint's index (in abc.scene.xml file) which higher than 75 are missing and so does the meshes. Is there a restriction with H3D that prevent loading models that have more than 75 joints?

Great work on the new version thou. Thanks a ton.

It would be nice to have proper directional lighting instead of workarounds or quick hacks, so this is more of a feature request or enhancement, as discussed or related to this thread here and from older posts

http://www.horde3d.org/forums/viewtopic.php?f=2&t=2304

The request was posted on the forum:

http://www.horde3d.org/forums/viewtopic.php?f=1&t=1904

This could be useful to improve the appeal of Horde3D in the "professional" market (i.e.: sim industry, 3D visualization, etc.)

Add an Direct3D 11-based renderer to better support Windows ecosystem (WinPhone8/WinRT/Xbox).

Work by attila:
https://github.com/attilaz/horde3d-x/tree/gles2_d3d11

Discussion on threads:
http://horde3d.org/forums/viewtopic.php?f=8&t=1886

Add an OpenGL ES 2-based renderer to support mobile platforms like iOS and Android.

Work by attila:
https://github.com/attilaz/horde3d-x/tree/gles2_d3d11

Discussion on threads:
http://horde3d.org/forums/viewtopic.php?f=8&t=1886
http://horde3d.org/forums/viewtopic.php?f=8&t=1036

Tested with Void Linux on (nvidia) gt640 gpu and also on an intel HD5500 gpu, tried adding content dir for in place running without effect (but still no error)

Documentation is currently out-of-date. Should be re-generated before releasing.

Currently, in samples code, we reinvent the wheel every time with a lot of duplicated code.

It makes the sample-specific code bloating (hiding the interesting logic with a lot of boring configuration code), hard to manage (fixes should be duplicated too for each sample) and hard to extend (adding more samples).

I propose to create a base (very minimal) sample framework which handles:

• Default window creation, configuration and handling.
• Default application logic (Exit on ESC, Fullscreen toggle on F1, etc.).
• Hooks for sample-specific logic.

It should be implemented using an abstract BaseSampleApplication class, shared by all samples and inherited by sample concrete application class. The concrete class should be written in a header file with a default name (i.e: app.h) included by the shared main source file.

RenderDevice::_activeVertexAttribsMask seems to be uninitialized before it's first read on line 1219.
Edit: Seems to be fixed on dev branch.

Redundancy should be avoided on source repository: if a file could be generated by others, it should be generated if it useful and when it is useful.

All pre-generated Visual Studio solutions should be removed by the source tree and generated by CMake when they are really needed (i.e. included in a Windows-specific binary SDK package).

It allows a better separation between build and source files and avoid sync issues between CMake file and VS solutions (new files should be added to CMakeLists.txt and NOT to VS project files!).

I think we should change a little bit the way the Render Device, Renderer and Modules classes are in relationship. At the moment is a little bit confusing.

The idea is to give ownership of low-level Renderer class to RenderDevice instead of high-level Modules singleton container, this also allows to remove the gRDI global pointer.

Discussion on this thread:

http://www.horde3d.org/forums/viewtopic.php?f=8&t=1908

H3D_DEBUGGER_OUTPUT: Send log messages to the output channel of the attached debugger

This is only working on windows. It's really useful to have that while developing, it saves you to check the html log file. On most other platform and stdout or stderr is shown on the debugger, so the fix is simple.

Currently the terrain extension uses the same vertex buffer two times in it's geometry. Calling destroyGeometry destroys the bound buffers by default and releases the same buffer multiple times because it does not check if a buffer bound previously has been already destroyed.
Probably the best would be to reference count the _buffer entries and only destroy buffers in destroyGeometry if no one else is using it anymore. Similar to how it is done with the resources. Maybe _buffers should be handled as Resources anyway. This way one vertex buffer may be used in different Geometries if needed.

I can not build the project on MacOSX 10.9 with XCode 6.0.1, CMake 3.0.2, I got this error:
CMake Error at CMakeLists.txt:42 (include):
include could not find load file:
Extensions/Extensions.txt

Then I created an empty Extensions.txt file in Extensions folder, and got this Error:
Horde3D/Source/Horde3DEngine/egModules.cpp:29:11: fatal error:
'egExtensions_auto_include.h' file not found

include "egExtensions_auto_include.h"

I had updated all Horde3D samples to latest version of GLFW 3. Everything works and compiles without issues, but there is problem with cursor position when mouse is moved and FPS camera rotates / resets itself to Z-axis instead of beginning movement at start/center position of window. Camera movement code is same:

void cursor_position_callback(GLFWwindow* window, double xpos, double ypos)
{
    if( !running )
    {
        mx0 = xpos; my0 = ypos;
        return;
    }
    app->mouseMoveEvent( (double)(xpos - mx0 ), (double)(my0 - ypos) );
    mx0 = xpos; my0 = ypos;
}

I think it has something to do with GLFW 3 (cursor position which start at X0 and Y0 of window, instead of center), instead of Horde3D but Im posting this issue here because I was not able to find any solution on Horde3D forums.

Planning on using soft bodies on Bullet, and I couldn't find any public function that would help me on this.

Build of Horde3D master fails on Visual Studio 2012.

error C2039: 'max' : is not a member of 'std' ..\horde3d\horde3d\source\colladaconverter\daeLibAnimations.h 141 1 ColladaConv

The cause seems to be a reorganization of STL in last VS libs.

The fix should be to add the following line to daeLibAnimations.h:

#include <algorithm>

I propose to move all the Overlays logic to a separated extension.

Why?

Well, basically because:

1. In real use cases, I think, a stand-alone GUI solution is used instead of basic Overlays.
2. Overlays are just a restricted use case of, more general, geometry primitives drawing (see: #24)
3. Split the core from extension concepts is usually a winning idea.

Discussion on this thread:

http://www.horde3d.org/forums/viewtopic.php?f=8&t=2162

The idea is to provide a simple API (maybe in the Utility library) to render native debug primitives like cubes, spheres, cones, cylinders, quads, triangles, etc.

The API should be simple enough, like the Overlay's one. The system should be able to batch for material all the primitives.

Discussion on this thread:

http://www.horde3d.org/forums/viewtopic.php?f=8&t=1906

I'm not sure if it should be considered an Horde3D bug (maybe not), but I post it just for reference.

The building of Horde3D fails with GCC 4.7.X due to a bug in the included rapidxml library. The bug is related to the change in the GCC behavior about symbol name lookups.

The build error is the following (this is not exactly from my build, but it almost the same):

[,,,]
../Shared/rapidxml_print.h: In instantiation of 'OutIt rapidxml::internal::print_node(OutIt, const rapidxml::xml_node<Ch>*, int, int) [with OutIt = std::ostream_iterator<char, char, std::char_traits<char> >; Ch = char]':
../Shared/rapidxml_print.h:390:57:   required from 'OutIt rapidxml::print(OutIt, const rapidxml::xml_node<Ch>&, int) [with OutIt = std::ostream_iterator<char, char, std::char_traits<char> >; Ch = char]'
../Shared/rapidxml_print.h:403:9:   required from 'std::basic_ostream<Ch>& rapidxml::print(std::basic_ostream<Ch>&, const rapidxml::xml_node<Ch>&, int) [with Ch = char]'
../Shared/rapidxml_print.h:414:31:   required from 'std::basic_ostream<Ch>& rapidxml::operator<<(std::basic_ostream<Ch>&, const rapidxml::xml_node<Ch>&) [with Ch = char]'
../Shared/rapidxml_print.h:115:17: error: 'print_children' was not declared in this scope, and no declarations were found by argument-dependent lookup at the point of instantiation [-fpermissive]
[,,,]

The bug was already reported on the Horde3D forum at:

http://www.horde3d.org/forums/viewtopic.php?f=3&t=1735

More info here (with a valid workaround):

http://stackoverflow.com/questions/14113923/rapidxml-print-header-has-undefined-methods
http://gcc.gnu.org/gcc-4.7/porting_to.html

You need to install it along with the .desktop file and icons.

I think the output folder should be moved on root directory, making the build process more clear.

Master is missing Extensions/Extensions.txt, and cmake won't build without it.
Seems like everything builds ok if I checkout the file from develop.

Oh, and it's actually the same issue as in #65.

You currently install into /usr, which isn't the destination on BSDs.