the OutputStream/InputStream implementations, which was just finished
last weekend with a few tests on Windows and Ubuntu. Hope it could
work and get more feedbacks soon.
I've added a new option "SchemaData" to the osg2 plugin. Developers
may test the new feature with the command line:
# osgconv cow.osg cow.osgb -O SchemaData
It will record all serializer properties used in the scene graph, at
the beginning of the generated file. And when osgviewer and user
applications is going to read the osgb file, the inbuilt data will be
automatically read and applied first, to keep backwards compatibility
partly. This will not affect osgb files generated with older versions.
"
IF(!OSG_COMPILE_FRAMEWORKS) ...
vs.
IF(NOT OSG_COMPILE_FRAMEWORKS) ...
This has the effect of always compiling OpenThreads as a framework under OSX. The CMakeLists.txt that I use to be able to compile the non-framework version of OpenThreads is attached. I simply replaced the '!' replaced by a 'NOT' and added a MESSAGE to notify me when the .framework will be compiled."
in process: a wrong definition of the OSG_FATAL macro, and wrong logic
inside the KeySwitchMatrixManipulator::getDistance() function. I
believe both were slips."
META_OSGMANIPULATOR_Object macro to ensure these classes could work
with their wrappers, and a few naming styles should be changed as
well. Fortunately everything seems to compile fine under Windows and
my new Ubuntu system.
And I finally find the problem of the
serializers/osgTerrain/Terrain.cpp, it just missed an "osg::Group"
before "osg::CoordinateSystemNode" indicator. With the small fix
attached now VPB could generate terrain with osgt/osgb formats."
JIV:"I deleted a line from the ReaderWriterDAE.cpp file that was introduced in rev 11341. I got a crash on this line when the options pointer was NULL."
MP:"Good spot John. The next line is also unnecessary and can be removed as well (attached, plus some minor code beautifying)"
osg::TransferFunction1D* tf(new osg::TransferFunction1D());
tf->allocate(18);
tf->setColor(-10000,osg::Vec4( 1.0, 1.0, 1.0,1.0));
tf->setColor(-1e-6,osg::Vec4( 1.0, 1.0, 1.0,1.0));
tf->setColor( 0,osg::Vec4( 0.0, 0.0, 1.0,1.0));
Remark: The value -1e-6 may be added (falsely) because of a rounding error.
The attached fix prevents assigning NaN values to the image.
"
Texture2DMultismaple as name suggests provides means to directly access subsamples of rendered FBO target. (GLSL 1.5 texelFetch call).
Recently I was working on deferred renderer with OSG, during that I noticed there is no support for multisampled textures (GL_ARB_texture_multisample extension). After consultations with Paul Martz and Wojtek Lewandowski I added Texture2DMultisample class and made few necessary changes around osg::FrameBufferObject, osg::Texture and osgUtil::RenderStage classes."
and from follow email:
"Fixed. According to ARB_texture_multisample extension specification multisample textures don't need TexParameters since they can only be fetched with texelFetch."
- compile errors on windows when compiled with UNICODE flag
- warnings for duplicate WIN32_LEAN_AND_MEAN. I think this should
better fixed by adding WIN32_LEAN_AND_MEAN to vcproj preprocessor
list."
CURL_LIBRARY_DEBUG and ZLIB_LIBRARY_DEBUG. Previously it linked the
debug version against the release libs, which was causing a hang when
running in debug mode on Windows."
modifications of the osgShadow header naming styles as well. The
osgDB::Serializer header is also changed to add new Vec2 serializer
macros because of the needs of osgShadow classes. It should compile
fine on both Windows and Linux. But I have only done a few tests to
generate .osgb, .osgt and .osgx formats with these new wrappers."
Mathieu Marache, he added the last missing piece to this puzzle.
I think it is safe to commit these changes to trunk, as the traditional
way via dylibs should work as before.
Here's some more info how to get frameworks:
With these modifications it is possible to compile frameworks on OS X,
when you set the Cmake-option OSG_COMPILE_FRAMEWORKS to true. If you
want to embed the frameworks in your app-bundle make sure to set
OSG_COMPILE_FRAMEWORKS_INSTALL_NAME_DIR accordingly.
You'll have to build the install-target of the generated xcode-projects
as this sets the install_name_dirs of the frameworks and plugins."
few headers and the osgAnimation sources are also modified to make
everything goes well, including:
A new REGISTER_OBJECT_WRAPPER2 macro to wrap classes like
Skeleton::UpdateSkeleton.
A bug fix in the Seralizer header which avoids setting default values
to objects.
Naming style fixes in osgAnimation headers and sources, also in the
deprecated dotosg wrappers.
A bug fix for the XML support, to write char values correctly.
A small change in the osg::Geometry wrapper to ignore the
InternalGeometry property, which is used by the MorphGeometry and
should not be set by user applications.
The avatar.osg, nathan.osg and robot.osg data files all work fine with
serializers, with some 'unsupported wrapper' warnings when converting.
I'm thinking of removing these warnings by disabling related property
serializers (ComputeBoundingBoxCallback and Drawable::UpdateCallback),
which are seldom recorded by users.
By the way, I still wonder how would we handle the C4121 problem,
discussed some days before. The /Zp compile option is set to 16 in the
attached cmake script file. And is there a better solution now?"
a byte order issue.
The problem is that osg::SwapBytes code has been copied from the old
plugin to the new one, but the latest lib3ds also incorporates code to
handle byte ordering in read & writing. So the net result is that the
swap is done twice.
The solution is simply to remove the custom osg code, and use the
stock lib3ds code. The attached files are against today's revision
11331. I've tested on Sparc & Intel.
"
- Fixed handling of MatrixTransforms (still doesn't support other Transforms types). Fixes things for OSG, DeepExploration, 3DSMax...
- Added support for writing double precision vertices by converting them.
- Added base code for future compatibility option (3rd-party apps that don't read animation data). See "DISABLE_3DS_ANIMATION" compile flag."
running with the OSG_NOTIFY_LEVEL set to INFO.
An iterator was being erased, and then referenced to display a
message. Just changed it to display the message first, then erase."
The plugin can now handle embeded PixelTexture fields in addition to the
already implemented ImageTexture fields.
Fixed a bug with texture repeat being applied to the wrong texture dimension.
Added handling for IndexedLineSet geometries."
now O(n) rather than O(nlogn) where n is the number of requests. The refactoring also cleans up the access of the
request lists so that the code is more readable/maintainable.
Many IDEs will capture the output of processes and highlight stderr messages in red. After upgrading from OSG 1.2, I was noticing that none of the error messages were being highlighted, which I find quite useful."
This was already possible for .frag and .vert so there's no reason why it shouldn't be possible for .geom.
I also changed the alignment of some extension aliases so the file looks a bit better now, but it's up to you if you want to merge that or not.
"
"Here is a small fix in the eventTraversal() function of both viewer
and composite viewer class.
if (getCameraWithFocus())
{
if (getCameraWithFocus()!=getCamera()) // Newly added
{
osg::Viewport* viewport = getCameraWithFocus()->getViewport();
osg::Matrix localCameraVPW =
getCameraWithFocus()->getViewMatrix() *
getCameraWithFocus()->getProjectionMatrix();
if (viewport) localCameraVPW *= viewport->computeWindowMatrix();
osg::Matrix matrix( osg::Matrix::inverse(localCameraVPW) *
masterCameraVPW );
osg::Vec3d new_coord = osg::Vec3d(x,y,0.0) * matrix;
x = new_coord.x();
y = new_coord.y();
}
...
}
I put an additional conditional statement here to ensure that
_cameraWithCamera and _camera are different, otherwise it's no need to
calculate the transition matrix from main camera to focus camera. The
excess calculations of 'matrix' and 'new_coord' may cause
floating-point error and return a slightly wrong result other than an
identity matrix. It seems OK in most cases but will be still pain when
there is little difference between two mouse moving events. "
- Replaced exceptions with assert() or OSG_NOTIFY
- Replaced osg::notify() with OSG_NOTIFY
- Changed braces and tabs to fit OSG coding convention
- Cleaned a few things in code (names, added deallocations upon error)"
currently only Debug appends "d" so the Release and MinSizeWithDebInfo
(and MinSizeRel) all produce the same filenames. This set of changes
lets each build type have a cmake defined string appended, defaulting
to Release none, Debug d, RelWithDebInfo rd, MinSizeRel s. But a user
still can have Release, RelWithDebInfo, and MinSizeRel to produce the
same filenames. It does so by setting the preprocessor define
OSG_LIBRARY_POSTFIX in src/osgDB/CMakeLists.txt to one of the
previously defined cmake variables CMAKE_DEBUG_POSTFIX
CMAKE_RELEASE_POSTFIX CMAKE_RELWITHDEBINFO_POSTFIX
CMAKE_MINSIZEREL_POSTFIX. This method cuts down on the #ifdef _DEBUG
#else preprocessor directives in Registry.cpp as the extension is
always passed in OSG_LIBRARY_POSTFIX. That and __MINGW32__ didn't
have the _DEBUG check which looks like a bug."
And by refactoring a bit of code, I may have fixed some StateSets related bugs (was ignoring StateSets for osg::Groups).
I also added support for Billboard's points, so now "osgconv lz.osg lz.3ds" has an acceptable output. However, there is no rotation depending on billboards' axis, hence the notice "Warning: 3DS writer is incomplete for Billboards (rotation not implemented).". You may want to remove this notice (or lower the notify severity) if you feel 3DS doesn't have to handle such rotations.
The attached archive contains 3 files from 3DS plugin, against rev. 11162.
Please note there is still the textures issue for cow.osg. I guess it's because it's not a "flat, dummy and standard" texture in slot 0... That is to say the only thing the writer can handle at the moment. I guess I won't address this soon.
"
and
"I've detected and fixed another bug in 3DS writer: support for automatic splitting of meshes having >65k faces/points was buggy (was deleting faces).
Here is my four 3DS modified files (in a ZIP), against rev. 11193, including previous fixes AND Stephan's fix about relative filenames."
using osgDB::XmlParser. The extension for XML-formatted scenes is
.osgx, corresponding to .osgb for binary and .osgt for ascii. It could
either be rendered in osgviewer or edited by common web browsers and
xml editors because of a range of changes to fit the XML syntax. For
example, the recorded class names are slight modified, from
'osg::Geode' to 'osg--Geode'.
To quickly get an XML file:
# ./osgconv cow.osg cow.osgx
The StreamOperator header, InputStreram and OutputStream classes are
modified to be more portable for triple ascii/binary/XML formats. I
also fixed a bug in readImage()/writeImage() to share image objects if
needed.
The ReaderWriterOSG2 class now supports all three formats and
reading/writing scene objects (not nodes or images), thanks to
Torben's advice before.
"
The options where not passed on to the image reader plugins when reading the new osg2 format files, so I added the options to the osgDB.:readImageFile function call.
"
absolute filenames for the texture images.
The attached change should fix this by at first looking at the absolute file
name to load a texture and then, if that fails, strip away any paths to try
that again with the bare file name.
The change also fixes a possible exception that could be triggered by an out
of bounds std::string access which is now avoided by using functions from
osgDB/FileUtils.
The change is based on rev 11161."
Old behaviour: "abc.d/filename_no_ext" -> "abc"
New behaviour: "abc.d/filename_no_ext" -> "abc.d/filename_no_ext"
Attached file is against rev. 11158."
- OsgMacroUtils.cmake, SETUP_LINK_LIBRARIES macro : allow linking with debug/release external libraries
- osgQt/CMakeLists.txt : fix the linking to Qt librairies + linking to debug Qt librairies if found
- examples/ qt examples : linking to debug Qt librairies if found"
Attached you will find updates of the files to hopefully solve the warnings (in VS2005 only one warning occured). In addition I fixed a tiny bug that caused a crash with one of my test files."
This change should make it possible to delete PagedLOD's independantly from the DatabasePager, and also prevent issues of
consistency of the pager when subgraphs when are cached elsewhere in the application such as in the Registry filecache.
up vector to setViewMatrixAsLookAt(..) codes in osgShadow. This will addresses previous issues that occured when look vectors
co-incided with the hard coded up vectors.
Fixes in StandardShadowMap.cpp & MinimalShadowMap.cpp were made for spotlight issues. There were cases when further located spotlights were not shadowing properly.
Small tweak in DebugShadowMap & StandardShadowMap.cpp to not limit shadow maps to texture2D (which should also allow texture2D arrays and cube maps). I simply replaced ptr to osg::Texture2D with pointer to osg::Texture. Interpretation of this member could be now changed with change of shaders in derived classes. This may be useful for guys who override LispSM or MinimalBoundsShadowMaps techniques. Could be useful for implementation of PerspectiveCascadedShadowMaps technique for example.
ConvexPolyhedron.cpp & DebugShadowMap.cpp contain debug HUD tweaks.
Change in ConvexPolyhedron.cpp overcomes the regression problem with color per primitive binding which caused that shadow volume outlines stopped to draw. I simply changed PER_PRIMITIVE to PER_PRIMITIVE_SET and it works again.
Other adition is dump method I added to DebugShadowMap which can be used in shadow debugging mode to dump current frame shadow volumes & scene to osg file. It could be then loaded into viewer and freely examined from different angles (which is difficult inside the application if shadow adopts to view and projection). "
options->setPluginStringData("captureVideoDevice", "0");
Lines added in getDevice() are:
int deviceId = atoi(name.c_str());
if(deviceId >= 0 && deviceId < (int)_listDevice.size())
return _listDevice[deviceId];
This makes it easy to use a capture device without knowing it's name. Attached is the whole file against rev 11044"
osgAnimation. It's been tested with the majority of the samples in the
COLLADA test repository and works with all of them either as well as, or
better than, the version of the plugin currently in SVN.
Known issue: vertex animation (AKA morphing) doesn't work at present,
but that's a relatively unpopular method of animating so it's not high
on my priority list."
Follow up email:
"I've been informed that the previous DAE submission didn't build on
unix, so here's the submission again with the fixes. Thanks to Gregory Potdevin and Benjamin Bozou.
Also, my apologies to Roland for not crediting his part in making DAE
animation happen, my work was indeed built on top of his work. Thanks
also to Marius Heise and of course Cedric Pinson."
Changes by Robert Osfield, fixed compile issues when compile without C* automatic conversion enabled in ref_ptr<>
and constructor initialization fixes to address some warnings under gcc.
* Change OcclusionQueryNode::getPassed to take a NodeVisitor rather than the distance from BS center to the eye point. Change where CullVisitor calls this method to use the new parameters.
* getPassed now exits early and returns true to avoid blinking / blink-in of geometry for the first frame or for out-of-range LOD children coming back into view.
* getPassed now considers the distance from the near plane to the bounding sphere (rather than eye point to bounding sphere) when determining if the viewer is "inside" the bounding sphere or not."
The fix is basically as follows:
if( GL3 )
Query indexed extension strings.
else
Query the old way.
The "else" branch is re-indented but otherwise shouldn't contain any changes."
From Robert Osfield, added #if !defined(OSG_GLES1_AVAILABLE) && !defined(OSG_GLES2_AVAILABLE) to new block to prevent it
being compiled under GLES where no glGetStringi exists.
The fix is basically as follows:
if( GL3 )
Query indexed extension strings.
else
Query the old way.
The "else" branch is re-indented but otherwise shouldn't contain any changes."
png_set_expand_gray_1_2_4_to_8() with the 1.2.9 release. This
submission fixes builds of the OSG against versions of libpng < 1.2.9
that don't have the new symbol available. This affects platforms like
Red Hat Enterprise Linux 4 which come with libpng 1.2.7."
Text and Text3D:
1. A new line should be started after a line's last hyphen or before
its last whitespace.
2. If no suitable place to break a line is found, just start new line
after the last character that fits on the line.
3. Whitespace should be removed from the beginning of the new line
(already worked in Text, but not in Text3D).
Line wrapping looks a lot better now with no more lone periods
appearing at the beginning of lines.
Also, right-justified text is more accurate now (slashes would hang
off the end of lines before). With this new code I spotted one
instance where a hyphen stuck out too far, but in general it looks
better. Centered text was not perfect before and still isn't, but I
can't see any significant increase or decrease in quality. The casual
observer would probably never notice a problem.
Also fixed a whitespace problem in Text3D. Not all whitespace was
being removed from the beginning of lines. Now it is all being
removed in the same manner as in Text."
To reproduce, on win32:
-Run osgViewer in a windowed mode, with the cursor off, as such:
osgViewer::Viewer::Windows windows;
viewer.getWindows(windows);
for(osgViewer::Viewer::Windows::iterator itr = windows.begin();
itr != windows.end();
++itr)
{
(*itr)->useCursor( false );
}
-Quickly move the cursor into the window (cursor it should be hidden)
-Resize the window by dragging the border (notice the cursor changes to "resize" cursor)
-Move the cursor back to the inside of the window (notice the cursor is not hidden anymore)
The attached SVN patch will set the cursor to a "NoCursor" during useCursor(false). This correctly stores the no cursor state, so it can be rejuvenated after a future cursor change. This patch also fixes a couple instances where a hidden cursor should show itself, like when it's on the title bar, or the window close button."
of OSG. I modified the osgDB::InputStream and OutputStream and the
PagedLOD wrapper as well. Now all seems to work fine with paged
scenes. I've tested with the puget terrain data and the osgdem
application from VPB:
# osgdem --xx 10 --yy 10 -t ps_texture_4k.tif --xx 10 --yy 10 -d
ps_height_4k.tif -l 8 -v 0.1 -o puget.osgb
As the ive plugin does, The PagedLOD wrapper now automatically add the
latest file path to PagedLODs' databasePath member, to help them find
correct child positions. I also changed the image storage strategy of
the OutputStream class, to store them inline by default. The osgt
extension should also work, in case the image files are also written
to the disk.
"
I create a compositeviewer with two views that share a context. One view is deleted. Texture::TextureObjectSet::discardAllTextureObjects is called, but this does not reset _tail. Now the texture object is again created and addToBack is called from Texture::TextureObjectSet::takeOrGenerate. In addToBack (line 612) _tail is now not 0 (although the list should be empty) and a loop is created from the texture object to itself. Then when the second view is deleted, Texture::TextureObjectSet::deleteAllTextureObjects loops forever. Setting _tail to 0 fixes it for me (see attached)."
which utilises a global recursive mutex that is dedicated to manage Observer and ObserverSet.
The new global mutex for observers avoids problems with deadlocks that were occurring previously when
an osg::Refenced object was being deleted at the same time as on osg::ObserverNodePath.
the extra capability of making it possible for Observers to assume ownership of a object that would otherwsie be deleted.
Added a thread safe ref_ptr<T> observer_ptr<T>::lock() method for robust access to an observed object. This
makes observer_ptr<> more equivilant to boosts weak_ptr.
Introduced the OSG_NOTIFY_DISABLE Cmake variable + include/osg/Config #define to control whether the OpenSceneGraph build
should disable the notification system completely. By setting OSG_NOTIFY_DISABLE to ON in CMake and then rebuilding the
the OSG you can get a slightly smaller (~1%) and more slightly efficient library which can be good for shipping applications,
but with downside of reduced ability to detect runtime problems and their causes.
- Improved identifiers generation in duplicate name handling (was limited to 1000 name collisions, which can be very short for some usages).
- Set all read/write operations use a custom log function that will redirect lib3DS log to osg::notify() (was only used for streams)
- Removed custom code (now uses osgDB::getFilePath())
- Added missing supportsOption() calls
- Cleaned a few minor things"
almost all Linux distributions. Although it is 100% compatible with ld,
by default it gives an error if a library has unresolved symbols at link
time, that is, it has set -Wl,--no-undefined by default. Debian folks
have found that libosg.so and libosgDB.so use some functions belonging
to libdl.so {dlsym,dlopen,dlclose,dlerror} without linking to it.
My changes link those two libraries to libdl.so explicitly in the same
way it is already done for libm.so and librt.so."
SET(COMPRESSION_LIBRARIES ${ZLIB_LIBRARY})
...
LINK_EXTERNAL(${LIB_NAME} ${OSGDB_PLATFORM_SPECIFIC_LIBRARIES})
to
SET(COMPRESSION_LIBRARIES ZLIB_LIBRARY)
...
LINK_EXTERNAL(${LIB_NAME} ${OSGDB_PLATFORM_SPECIFIC_LIBRARIES}
${COMPRESSION_LIBRARIES})
LINK_WITH_VARIABLES(${LIB_NAME} ${COMPRESSION_LIBRARIES})
I notice that the LINK_EXTERNAL macro won't distinguish between debug
and release dependences, which means that osgDB will use the release
version of zlib for all build configurations. Under Win32, this will
cause a manifest problem: all applications using osgDB may complain
"failed to start because msvcr80.dll was not found" or similar
messages.
This change will make it back to normal."
the type writing more explictly tied to the size type, with use of unsigned int as the default size. This approach
ensures that we get the same results under 32 and 64bit builds.
