- Support for writing unsigned 16-bit images (GL_UNSIGNED_SHORT)
- Code to parse the options string for the following options:
-- Flag to turn off the compression. The PACKBITS compression type causes issues for me with some programs on Windows (Picasa for example).
-- Options to set the XRESOLUTION and YRESOLUTION tags (DPI) in the TIFF file.
Existing behavior (PACKBITS compression, DPI tags left at default) is preserved if the options string is not set.
"
"I checked your solution and found one missing point which makes it still produce the tif error:
The very first seek_set on the empty stream with zero offset."
"This means that the empty stream is seeked again resulting in the fail bit to be set. Your code does not check this case, furthermore you use t_off instead of std::ostream::streampos for the tellp() calls. In this special case (empty stream) tellp() returns -1 which is cast to 0xFFFFFFFFFF as t_off is unsigned. I suggest this addition to your code (within the switch statement)"
Here is a quick list of the modified files:
Archive - getDirectoryContents() no longer pure virtual
Archive.cpp - default getDirectoryContents() implementation
unzip.cpp - modified to fix a bug where the same file will not load twice in a row
ZipArchive.h / ZipArchive.cpp - extends osgDB::Archive and provides support for random access loading within a .zip file
ReaderWriterZip.cpp - modified to use the ZipArchive class"
Here is a quick list of the modified files:
Archive - getDirectoryContents() no longer pure virtual
Archive.cpp - default getDirectoryContents() implementation
unzip.cpp - modified to fix a bug where the same file will not load twice in a row
ZipArchive.h / ZipArchive.cpp - extends osgDB::Archive and provides support for random access loading within a .zip file
ReaderWriterZip.cpp - modified to use the ZipArchive class"
Other changes in this sumbission include the added ability to tessellate polygons in Collada reader, with appropriate options.
- No tessellation
- Tessellate as triangle fan (previous behaviour, kept as default for backward-compatibility)
- Full tessellation
I also put auto_ptr<> for RAII of DAE structure (as discussed), and moved reader options in a structure, as for the writer.
Code also make use of osgDB::ExternalFileWriter I submitted, so please merge this submission after the PluginImageWriter one.
CID 11441: Non-array delete for scalars (DELETE_ARRAY)
"_dwobj::readOpenings(_iobuf *, int)" allocates "obj.fc1". [show details]
"_dwobj::~_dwobj()" uses delete on "obj.fc1" which is an array. It should be deleted with operator delete[]. [show details]
CID 11394: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::IntArray".
Assigning: "this->colorindices" = "new osg::IntArray".
The constructor allocates field "colorindices" of "struct vertexInfo" but there is no destructor.
CID 11395: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::IntArray".
Assigning: "this->coordindices" = "new osg::IntArray".
The constructor allocates field "coordindices" of "struct vertexInfo" but there is no destructor.
CID 11396: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::IntArray".
Assigning: "this->normindices" = "new osg::IntArray".
The constructor allocates field "normindices" of "struct vertexInfo" but there is no destructor.
CID 11397: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::IntArray".
Assigning: "this->txindices" = "new osg::IntArray".
The constructor allocates field "txindices" of "struct vertexInfo" but there is no destructor.
CID 11398: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec2Array".
Assigning: "this->txcoords" = "new osg::Vec2Array".
The constructor allocates field "txcoords" of "struct vertexInfo" but there is no destructor.
CID 11399: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec3Array".
Assigning: "this->norms" = "new osg::Vec3Array".
The constructor allocates field "norms" of "struct vertexInfo" but there is no destructor.
CID 11400: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec3Array".
Assigning: "this->coords" = "new osg::Vec3Array".
The constructor allocates field "coords" of "struct vertexInfo" but there is no destructor.
CID 11401: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec4Array".
Assigning: "this->colors" = "new osg::Vec4Array".
The constructor allocates field "colors" of "struct vertexInfo" but there is no destructor.
CID 11402: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec4Array".
Assigning: "this->polycols" = "new osg::Vec4Array".
The constructor allocates field "polycols" of "struct vertexInfo" but there is no destructor.
CID 11389: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new bsp::VBSPData".
Assigning: "this->bsp_data" = "new bsp::VBSPData".
The constructor allocates field "bsp_data" of "struct bsp::VBSPReader" but the destructor and whatever functions it calls do not free it.
ID 11390: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec3Array".
Assigning: "this->vertices" = "new osg::Vec3Array".
The constructor allocates field "vertices" of "struct prims" but the destructor and whatever functions it calls do not free it.
CID 11391: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec3Array".
Assigning: "this->normals" = "new osg::Vec3Array".
The constructor allocates field "normals" of "struct prims" but the destructor and whatever functions it calls do not free it.
CID 11392: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec3Array".
Assigning: "this->txc" = "new osg::Vec3Array".
The constructor allocates field "txc" of "struct prims" but the destructor and whatever functions it calls do not free it.
CID 11393: Resource leak in object (CTOR_DTOR_LEAK)
Allocating memory by calling "new osg::Vec3Array".
Assigning: "this->txcoords" = "new osg::Vec3Array".
The constructor allocates field "txcoords" of "struct prims" but the destructor and whatever functions it calls do not free it.
CID 11747: Uninitialized pointer field (UNINIT_CTOR)
Non-static class member gset is not initialized in this constructor nor in any functions that it calls.
CID 11851: Unused pointer value (UNUSED_VALUE)
Pointer "gfd" returned by "gr->getField(26)" is never used.
CID 11850: Unused pointer value (UNUSED_VALUE)
Pointer "gfd" returned by "gr->getField(80)" is never used.
CID 11849: Unused pointer value (UNUSED_VALUE)
Pointer "gfd" returned by "gr->getField(3)" is never used.
CID 11848: Unused pointer value (UNUSED_VALUE)
Pointer "gfd" returned by "grec->getField(3)" is never used.
/** Get the file name which represents the archived file.*/
virtual std::string getArchiveFileName() const = 0;
/** return type of file. */
virtual FileType getFileType(const std::string& filename) const = 0;
/** return the contents of a directory.
* returns an empty array on any error.*/
virtual DirectoryContents getDirectoryContents(const std::string& dirName) const = 0;
Added implementations of these new methods into src/osgPlugins/osga/OSGA_Archive.h src/osgPlugins/osga/OSGA_Archive.cpp
The pvr format which can be used as a wrapper for different compressed and uncompressed formats supports this compression algorithm. The original pvr compression uses the pvrtc format. The handling of pvrtc is already implemented in the pvr plugin. PVR provides wrapper functionality for some formats, e.g. etc or even dxt/dds.
Our target system (gles2) is able to use the etc compression format. With minor changes in the submitted files, there is no need to write a separate plugin. However the original pvr texture compression formats are not supported on our target, which is the reason for this extension.
The changes mainly consist in the definition on new enum values in the classes and headers of ReaderWriterPVR,Image and Texture. I also found some locations where the handling of the original pvr textures was not implemented. These are also part of this submission."
of empty node:
if (GetFloat4Param(cot->getParam()->getRef(), f4)) {...}
in the model I have tested cot->getParam()->getRef() return 0x0 and make
it crahes inside GetFloat4Param.
I have added a test before calling GetFloat4Param
if (cot->getParam()->getRef() != 0 &&
GetFloat4Param(cot->getParam()->getRef(), f4)) {...}
"
supportsOption("dds_dxt1_rgb","set the pixel format of DXT1 encoded images to be RGB");
supportsOption("dds_dxt1_rgba","set the pixel format of DXT1 encoded images to be RGBA");
supportsOption("dds_dxt1_detect_rgba","For DXT1 encode images set the pixel format according to presence of transparent pixels.");
And set the default not to modify and detect the RGB or RGBA pixel format for DXT images.
One can override this value via the sampleDensityWhenMoving="value" property in the volume tag, i.e.
<volume sampleDensityWhenMoving="0.01">CardiacCT</volume>
To switch the feature off set the value to 0, i.e.
<volume sampleDensityWhenMoving="0">CardiacCT</volume>
2. Removed unecessary code. Also made sure the images names are not truncated in the middle of an UTF8 character. You'll find there a function called utf8TruncateBytes(), which may be moved in core OSG (osgDB/ConvertUTF I guess). Feel free to do it if you feel it that way.
"
"- In order to build against GLES1 we execute:
$ mkdir build_android_gles1
$ cd build_android_gles1
$ cmake .. -DOSG_BUILD_PLATFORM_ANDROID=ON -DDYNAMIC_OPENTHREADS=OFF
-DDYNAMIC_OPENSCENEGRAPH=OFF -DANDROID_NDK=<path_to_android_ndk>/
-DOSG_GLES1_AVAILABLE=ON -DOSG_GL1_AVAILABLE=OFF
-DOSG_GL2_AVAILABLE=OFF -DOSG_GL_DISPLAYLISTS_AVAILABLE=OFF -DJ=2
-DOSG_CPP_EXCEPTIONS_AVAILABLE=OFF
$ make
If all is correct you will have and static OSG inside:
build_android_gles1/bin/ndk/local/armeabi.
- GLES2 is not tested/proved, but I think it could be possible build
it with the correct cmake flags.
- The flag -DJ=2 is used to pass to the ndk-build the number of
processors to speed up the building.
- make install is not yet supported."
- contributor
- author
- authoring_tool
- comments
- copyright
- source_data
- created
- keywords
- modified
- revision
- subject
- title
- unit
- name
- meter
- up_axis
With this support, users are able to include additional information in their models. Additionally, tools such as sketchup that support asset tags use the values appropriately within their imported models."
1. daeRGeometry.cpp (rev 12132). A tiny sumbission which fixes a quite important bug: a parameter was forgotten in Collada ReaderWriter, and texture coordinates could not be loaded properly. So:
"
Added missing paramter when calling createGeometryData(). Fixes missing texture coordinates (in "bind_vertex_input").
"
2. ReaderWriterDAE.cpp (rev 12132):
"
Added "baseImageDir" as a plugin string data, in order to manually specify base directory to use when relativising image file names. This is used to properly write files, when images are not located in a subdirectory (like "../images" for some software).
"
1. DAE submission:
DAE plugin now correctly writes images URI in Collada file, when images are used twice.
I also greatly improved readability and maintenability of geometry reading (mainly daeRGeometry.cpp), by factorizing code, templatizing it (for double/single precision), and removing ugly macros.
2. osgDB submission:
I updated osgDB::getPathRelative(): it is now far more readable, it handles more cases (especially when you want to relativise "a/c" from "a/b", which results in "../c"), and I added comments to make it clearer to maintain."
DAE plugin was linking ORIGINAL images in the Collada file, using image->getName() as a path (even if images were modified in memory!). As the behaviour was not the one of other plugins (3DS, FBX, and such), I made the plugin relativise images filenames (as those plugins) and write the image which is in memory. However, in order to avoid removing features, I kept the previous behaviour but moved it in an option. Here are the options of the plugin I changed:
- daeForceTexture was unclear in this new context and removed in favor of two new options
- daeLinkOriginalTexturesNoForce: Writes reference to the original image if found, instead of writing the image in memory
- daeLinkOriginalTexturesForce: Writes reference to the original image even if not found, instead of writing the image in memory
Of course, if you specify no option, images are written as for other plugins.
Other thing I changed is the UTF8 support as I told you in a previous conversation. Now there is a simple option, "daeNamesUseCodepage", which makes all names except filenames (materials, animation, geometries...) be considered as encoded using current codepage. If so, they'll be converted to UTF8 when writing; else they are written directly. Of course, filenames follow OSG_USE_UTF8_FILENAME as usual.
I did "
1. It did not support normals applied to individual vertices.
2. It would only support red / green / blue colour triples, but the pmvs models are generated with diffuse colours. (The PLY format, http://local.wasp.uwa.edu.au/~pbourke/dataformats/ply/, lists specular and ambient colour forms as well.)
To partially overcome these limitations, please find attached modified versions of
src/osgPlugins/ply/vertexData.cpp
src/osgPlugins/ply/vertexData.h
The changes I've made are:
1. I have changed the boolean hasColor flag to a vertexField (which is a boolean operation on an enum) to indicate what fields are present in the ply file. (This is required because Turk's ply reader spits out warnings for every line where you try to read fields which do not exist.)
2. I have modified the code to apply valid normals to either triangles or vertices.
3. I have kludged in "support" for the various colour variants. Specifically, all the colour specified can be read from the file. However, they are all applied in the same way (namely as a colour array, bound to each vertex)."
I have attached a patch, against the trunk from 13:30 today, which consists of the following:
1. CMakeModules/FindOpenEXR.cmake: Look for libIlmThread and libIex as well. 2. src/osgPlugins/CMakeList.txt: Only include the exr subdirectory if both the OpenEXR and zip libraries were found. 3. src/osgPlugins/exr/CMakeLists.txt: Add ZIP_LIBRARY to TARGET_EXTERNAL_LIBRARIES."
The DatabasePager now passes the Terrain pointer into the ReaderWriter's via osgDB::Options object,
rather than pushing a NodePath containing the Terrain onto NodeVisitor. This
change means that the DatabasePager nolonger needs to observer the whole NodePath and
will be lighter and quicker for it.
The change also means that ReadFileCallback can now run custom NodeVisitor's on the scene graph without
having to worry about TerrainTile's constructing scene graphs prior to the Terrain being assigned.
Also changed is the NodeVisitor::DatabaseRequestHandler which now requires a NodePath to the node that you wish
to add to rather than just the pointer to the node you wish to add to. This is more robust when handling scenes
with multiple parental paths, whereas previously errors could have occurred due to the default of picking the first
available parental path. This change means that subclasses of DatabasePager will need to be updated to use this new
function entry point.
We should have:
UV for vertex 0, layer 1
UV for vertex 0, layer 2
UV for vertex 1, layer 1
UV for vertex 1, layer 2
...etc...
Instead, we have this:
UV for vertex 0, layer 1
UV for vertex 1, layer 1
...etc...
UV for vertex 0, layer 2
UV for vertex 1, layer 2
...etc..."
* the glsl plugin now supports processing #includes. The file extension sets the shader type.
* the registry releases gl objects of the shared state manager
"
requests to a removed archive file. ReaderWriterTXP::getArchiveName
will create an archive if it doesn't exist. This is causing a crash
if the terrain is removed and there are outstanding database requests
in the database pager. The request will create a new archive, and as
the archive is newly create doesn't have the materials etc to support
the subtile that was just loaded. To fix the problem getArchiveName
will only return existing archives. createArchiveName was added to
create archives.
"
setup the log level to INFO it does not work. It's because of this line:
OSG_INFO << " GetGeoTransform == "<<
dataset->GetGeoTransform(geoTransform)<<"
I have just put the function outside of an stream stuff, and it works.
It's possible we have other code like that in other plugins, I guess
it's a side effect of the MACRO stuff about notification."
- CMakeLists.txt
-- don't look for GL when compiling for iOS (device or simulator), look for OGLES instead
-- use architecture i386 for simulator
-- removed iphoneos-version-min for simulator
- examples/osgviewerIPhone/CMakeLists.txt
-- added build dependencies for osgdb_osg, osgdb_freetype, osgdb_imageio
-- added framework QuartzCore link dependency
- src/osgDB/CMakeLists.txt
-- don't link against Carbon on iOS (device or simulator)
- src/osgPlugins/freetype/CMakeLists.txt
-- don't link against OpenGL on iOS device or simulator
- src/osgViewer/CMakeLists.txt
-- link against OpenGLES on iOS (device or simulator)
- src/osgPlugins/imageio/CMakeLists.txt
-- compile ReaderWriterImageIO_IOS.cpp as Objective-C++
"
hpux. I have skipped irix this time as irix is too dead to keep osg building
there.
As usual, solaris does not like member templates in stl containers.
Some headers missing and link problems due to missing libraries."
attached you'll find the second part of the IOS-submission. It contains
* GraphicsWindowIOS, which supports external and "retina" displays,
multisample-buffers (for IOS > 4.0) and multi-touch-events
* an ios-specific implementation of the imageio-plugin
* an iphone-viewer example
* cMake support for creating a xcode-project
* an updated ReadMe-file describing the necessary steps to get a
working xcode-project-file from CMake
Please credit Thomas Hogarth and Stephan Huber for these changes.
This brings the ios-support in line with the git-fork on github. It
needs some more testing and some more love, the cmake-process is still a
little complicated.
You'll need a special version of the freetype lib compiled for IOS,
there's one bundled in the OpenFrameworks-distribution, which can be used."
Notes, from Robert Osfield, modified CMakeLists.txt files so that the IOS specific paths are within IF(APPLE) blocks.
recently we noticed a little mistake for 3DS files using instances of the same meshs: Every groupnode gets the same name instead of the (correct) instance name of the object. The fix only consists of two additional lines which check whether an instance_name is given for the object and then uses this one instead of the node name."
serialization libraries. My submission mainly includes:
1. Add two new macros USE_DOTOSGWRAPPER_LIBRARY and
USE_SERIALIZER_WRAPPER_LIBRARY. Applications using static OSG must
include corresponding static-link libraries and use these two macros
to predefine native format wrappers. Please see osgstaticviewer and
present3D in the attachment for details.
2. Add a LibraryWrapper.cpp file in each
osgWrappers/deprecated-dotosg/... and osgWrappers/serializers/...
subfolder, which calls all USE_...WRAPPERS macros inside. The
LibraryWrapper file is automatically generated by the
wrapper_includer.cpp (with some slight fixes), which is also attached
for your reference. The deprecated-dotosg/osgAnimation is not included
because it doesn't us REGISTER_DOTOSGWRAPPER to define its wrappers.
3. Modify the ReaderWriterOSG.cpp to prevent calling loadWrappers()
when static build.
4. An uncorrelated fix to Serializer and ObjectWrapper.cpp, which
ensures version variables of serialziers are initialized, and
serializers out-of-version are not written to model files.
"
templated constructor of std::pair not being able to automatically
convert 0 to a pointer. Rather than use preprocessor checks and
#defines, I think a cleaner solution is to cast the std::pair arguments
to the appropriate types to help the compiler out. I attached an updated
version of the file implementing this."
As it works (for our needs at least) I've done and attached a tar.gz file for the VRML2 plugin with a new part to write a VRML2 file from an osg one.
The read part is the same as in osg 2.9.9.
The write part code is in convertToVRML.cpp and .h files. It works for some osg nodes (group, geode, matrixTransform, positionAttitudeTransform and geometry). Textures are converted to jpeg (if not translucent) or png (if translucent).
There are some options that could be given to the writer (with -O switch) :
convertTextures=0 to copy textures without converting them to jpeg or png
convertTextures=-1 do not copy textures, keep them in their original format and location
convertTextures=-2 do not use textures, parse only geometry
convertTextures=-3 (default) convert textures to jpeg or png ones.
textureUnit=X in case of multiple textures, X= texture unit to use (default value=0)
directoryTexture=aPath when texture will be copied, it will be in this directory, not in the current one."
ReadDDSFile warning: couldn't read mipmapData
The issue is caused, when the last block of data is read from the file (less than chunk size of 16384 bytes). The read operation in ReaderWriterGZ::read() then sets the eof and fail bit in the stream and the lines
if (fin.fail())
{
(void)inflateEnd(&strm);
return false;
}
causes the reading to be aborted with the last read data not beeing inflated.
Please find the attached fix for this problem."
video and sound pins by type rather than searching for them by name
since the names of the pins can change based on what kind of file you
are opening. This also removes the need for an explicit check to see
if the file is a .wmv file.
Also changes to the directshow plugin's CMakeLists.txt.
It is not necessary to link against d3dx9 to build the plugin.
"
Now we can use ffmpeg to render webcam video (using the vfwcap device) under Windows:
osgmovie 0 -e ffmpeg -O "format=vfwcap frame_rate=25"
The number 0 just indicates a default device number in ffmpeg.
I think this can work under Linux, too, and should be a bit better than comparing the filename with a '/dev/' string. Just type:
./osgmovie /dev/yourcam -e ffmpeg -O "format=video4linux2 frame_rate=30 size=320x240""
attached. The problem was that the output files were not properly
setting the debug/release libs due to this cmake bug. What occurred was
the release lib was set in all configurations."
changed extensions from .c to .cpp and got compiling as C files as part of the osg core library.
Updated and cleaned up the rest of the OSG to use the new internal GLU.
compatability regression that was introduced almost 2 years ago in
r8834. The IVE version number was bumped to 32 because of the change
in binary layout, but the guard for reading/writing the new field was
checked against 31. Of course this only causes a problem (as for us)
when you've produced IVE files at version 31, which no longer load (or
crash) when loaded by newer OSG/IVE versions."
I also fixed a possible bug in osgDB::XmlParser that doesn't handle control characters (like " to ") when reading node attributes, because the writeWrappedString() and readWrappedString() now depend heavily on control characters. An additional improvement is that osgx now supports comments."
