to optimize away duplicate state with dynamic, static and unspecified DataVarience. By default
the code now optimizes away duplicate state with either static and unspecied state, previously
it was just handling static state.
implementation of the atomic increment and decrement into a implementation
file.
This way inlining and compiler optimization can no longer happen for these
implementations, but it fixes compilation on win32 msvc targets. I expect
that this is still faster than with with mutexes.
Also the i386 gcc target gets atomic operations with this patch. By using an
implementation file we can guarantee that we have the right compiler flags
available."
atomic related compile failures with msvs2005. Attached changes to make win32
really use the atomic stuff. There are pointer typecast problems and some
historic alignment restrictions that I just took from a previous similar
implementation of mine without looking deep enough. "
From Robert Osfield, made a range of changes to Terry's visitor integrating it into osgUtil::Optimizer and
changing the code to use a style more like the rest of the OSG.
"I have taken the liberty of updating a few files so that there is no longer any derivation from std::vector. I have done this by adding a new file osg/MixinVector and by updating only two others: osg/PrimitiveSet and osg/Array. You will notice that this actually removes what is acknowledged as a \u2018hack\u2019 in osg/PrimitiveSet.
With the original code I did manage to find memory leaks with some compiler options on VC 8 and 9, as well as Intel compiler. I determined the leak existence by instrumenting the destructor code, and by use of a garbage collector as a leak detector (in a similar manner to the Firefox project). Hence in contrast to what I said originally, it is exhibiting symptoms on at least some platforms.
Since I am trying to be a good OSG citizen I got out my editor and started hacking! I have built and tested on Linux (Ubuntu) with GCC 4.x and Windows VC 8 SP1. It appears that nothing is broken, and that I\u2019m using less memory J"
viewport settings in stereo mode. It seems that the SceneView::cull()
method will pass the full size viewport to the left/right
cullvisitors, instead of the modified stereo viewport. I made quite a
few changes to SceneView to fix the issue. The SceneView::cullStage()
method will now receive the viewport as an argument, instead of using
the global viewport. The SceneView::cull() method will pass the
modifed viewport to cullStage when rendering in stereo.
There are 2 new private methods computeLeftEyeViewport() and
computeRightEyeViewport() that will compute the stereo viewports. I
also modified the draw() function so it applies the correct viewport
to the prerender stages. These changes are only necessary for
horizontal/vertical split stereo."
via StateSet::setNestedRenderBin(bool) whether the new RenderBin should be nested
with the existing RenderBin, or be nested with the enclosing RenderStage.
GL_GENERATE_MIPMAP_SGIS is very slow (over half a second for a 720*576
texture). However, glGenerateMipmapEXT() performs well (16ms for the
same texture), so I have modified the attached files to use
Texture::generateMipmap() if glGenerateMipmapEXT is supported, instead
of enabling & disabling GL_GENERATE_MIPMAP_SGIS."
Notes, from Robert Osfield, I've tested the out of the previous path using
GL_GENERATE_MIPMAP_SGIS and non power of two textures on NVidia 7800GT and
Nvidia linux drivers with the image size 720x576 and only get compile times
of 56ms, so the above half second speed looks to be a driver bug. With
Muchael's changes the cost goes done to less than 5ms, so it's certainly
an effective change, even given that Michael's poor expereiences with
GL_GENERATE_MIP_SGIS do look to be a driver bug.
now holds the dirty flag and enables/disables event traversal in response dirty
being set/unset. This allows terrain to be automatically updated in response
to Terrain scale and sample ratio changes.
- Solves issues of loading image data into the texture memory
- Print a warning if images are of different dimensions or have different internal formats (GL specification requires images to be the same)
Patch is tested and seems to work fine. It shouldn't break any other functionality. It should go into include/osg and src/osg
"
multi-threaded paging, where the Pager manages threads of reading local
and http files via seperate threads. This makes it possible to smoothly
browse large databases where parts of the data are locally cached while
others are on a remote server. Previously with this type of dataset
the pager would stall all paging while http requests were being served,
even when parts of the models are still loadable virtue of being in the
local cache.
Also as part of the refactoring the DatabaseRequest are now stored in the
ProxyNode/PagedLOD nodes to facilitate quite updating in the cull traversal,
with the new code avoiding mutex locks and searches. Previous on big
databases the overhead involved in make database requests could accumulate
to a point where it'd cause the cull traversal to break frame. The overhead
now is negligable.
Finally OSG_FILE_CACHE support has been moved from the curl plugin into
the DatabasePager. Eventually this functionality will be moved out into
osgDB for more general usage.
ReaderWriter::ReadResult now has a FILE_REQUEST enum.
ReaderWriter::Options now has a s/getAsynchronousFileReadHint() parameter methods.
libcurl based plugin now detects enabing of the AsynchronousFileReadHint, but
as yet does not handle async requests - handling everything syncronously.
DatabasePager now by default will enable AsynchronousFileReadHint for http
based file requests
avoid threading issues associated with compile running in a parallel with
update/cull on the first frame.
Also added automatic recompile when a new SceneData is applied to a View.
get an excess Tab key report when switching back to an OSG
application (usually FlightGear :-). Although KDE has consumed
the Tab, it's sometimes still in the XKeymapEvent's key_vector,
and followed by a Tab KeyRelease event.
Avoid this artifact by
- asking for a "fresh" keymap (via XQueryKeymap()), rather than
using the unreliable(?) XKeymapEvent's key_vector, and by
- flushing all key events on focus-in (to avoid the KeyRelease)
After Super-press, Tab-press, Super-release, Tab-release (note
the wrong release order!) I still get an extra Tab event. But
this is not surprising and not exactly wrong either. Also it's
hard to avoid, as we can't see what happened to the keyboard
before we regained focus.
Files changed:
src/osgViewer/GraphicsWindowX11.cpp
include/osgViewer/api/X11/GraphicsWindowX11"
It's implemented in the same way that 3D Spherical Display and Panoramic Spherical Display.
You can test it running:
osgviewer --wowvx-20 cow.osg
osgviewer --wowvx-42 cow.osg
depending on the size of your Philips WOWvx display (20" or 42")
Other arguments you can use to control the 3D effect are:
--wow-content <value>
This value defines the kind of content that can be:
0: No depth
1: Signage
2: Movie
3: CGI
4: Still
--wow-factor <value>
Percentage of the display recommended depth value. Default 64, Range [0-255]
--wow-offset <value>
Amount of range behind the screen. Default 128, Range [0-255]
0: Range is shifted in the direction of the viewer.
128: Range is equally divided in front and behind the screen.
255: Range is shifted away from the viewer.
"
from: DEEP_COPY_STATESETS = 8,
to: DEEP_COPY_STATESETS = 1<<3,
showing clearly that this isn't the _value_ 8, but the _bit_ 8. this is an old pattern i see (and like to promulgate) to make code a bit more readable and maintainable.
"
From Robert Osfield, refactored the FrameBufferObejcts::_drawBuffers set up so that its done
within the setAttachment method to avoid potential threading/execution order issues.
hyper keys defined already, but these modifiers were missing in
GUIEventAdapter::ModKeyMask, and the EventQueue ingored them as well.
The attached diff/archive adds the missing parts for Super/Hyper
modifier key support.
I'm aware that this might not be supported on all systems/keyboards
out of the box, but decided to submit it anyway because:
- developers are aware of differences between input devices
(Some mice have scroll wheels, others don't. Some have five or
more buttons, some have only one. Some keyboards don't have
numpads, some have AltGr, some don't etc.)
- even if someone relies on Hyper/Super in distributed software,
this is easy to fix and doesn't create lock-in conditions
- while the names Hyper/Super may only be common on X11, they are
just symbol names and not OS-specific
- even though some systems might not offer these additional modifiers
by default, it's likely that all of them have at least 8 modifier
levels internally, so it should only be a matter of OS configuration
to make them work
- having super/hyper available is useful to offer a user ways
to define local key definitions that are safe from collisions with
predefined "official" key assignments"
Introduced code in BoundgingSphere, BoundingBox, ProxyNode and LOD to utilise the above settings.
Added Matrix::value_type, Plane::value_type, BoundingSphere::value_type and BoundingBox::value_type command line
options that report where the types of floats or doubles.
This change has been done to make it easier for OpenSceneGraph users to check out the svn via https
without any conflicts introduced with a http externals.
as not expiring subgraphs quick enough to enable reasonable load balancing.
New version isn't perfect and will need further work, but does at least reduce
the memory footprint by as much as half on test paths on big databases.
The rewritten method no longer uses the the MaximumNumOfRemovedChildPagedLODs
and MinimumNumOfInactivePagedLODs variables so these and associated methods
for accessing them have been removed.
- /** Set the maximum number of PagedLOD child to remove per frame */
- void setMaximumNumOfRemovedChildPagedLODs(unsigned int number) { _maximumNumOfRemovedChildPagedLODs = number; }
-
- /** Get the maximum number of PagedLOD child to remove per frame */
- unsigned int getMaximumNumOfRemovedChildPagedLODs() const { return _maximumNumOfRemovedChildPagedLODs; }
-
- /** Set the minimum number of inactive PagedLOD child to keep */
- void setMinimumNumOfInactivePagedLODs(unsigned int number) { _minimumNumOfInactivePagedLODs = number; }
-
- /** Get the minimum number of inactive PagedLOD child to keep */
- unsigned int getMinimumNumOfInactivePagedLODs() const { return _minimumNumOfInactivePagedLODs; }
key, but it didn't pick up the initial state. So, if NumLock was on for
the OS at startup (LED on), it was still off for OSG. And the first
keypress turned the LED off, and NumLock on for OSG. The attached fix
picks up the state on every FocusIn, just like it was done in the last
commits for CapsLock. The difference is, that the NumLock mask isn't
standardized (e.g. 0x10 for Linux, and 0x80 for AIX), so we have to do
a reverse lookup (::rescanModifierMapping()).
Note that I could not reproduce the problem on my system, but someone
else confirmed it twice on his, and the patch fixed it for him.
Changed files:
./include/osgViewer/api/X11/GraphicsWindowX11
./src/osgViewer/GraphicsWindowX11.cpp
"
and a new scheme for computing the scaling when using autoscale that introduces smooth
transitions to the scaling of the subgraph so that it looks more natural.
help manage the scaling of particles, whether they should be relative to the
local coordiante frame of the particle system, or be in world coordinates.
It sets osgGA's keymask when restoring keys on FocusIn, according
to the state values of XKeyEvent and XCrossingEvent. (These are
the only source for X11's current capslock state that avoids
pulling in the XKB extension.)
"
osgViewer::StatsHandler and other handlers which allow you to change the
key(s) you would press to get them to do something. Pretty simple change
but useful in our context and possibly in others too."
The win32 pbuffer implementation returned an error unless both the
WGL_ARB_pbuffer and the WGL_ARB_render_texture functions were present.
This was too restrictive, as a pbuffer can usefully be created without
render-to-texture, e.g. for use with glReadPixels. The osg 1.2/Producer
pbuffers worked without RTT, and osgUtil::RenderStage has all the code to
handle both RTT and non-RTT pbuffers, doing a read and copy in the
latter case.
With these changes I have successfully tested the osgprerender example
on a graphics card which supports RTT, and one which doesn't. Plus
tested in my own application.
In order to aid diagnostics I have also added more function status
return checks, and associated error messages. I have included the win32
error text in all error messages output. And there were some errors
with multi-threaded handling of "bind to texture" and a temporary window
context which I have corrected.
These is one (pre-existing) problem with multi-threaded use of pbuffers
in osgViewer & osgprerender, which I have not been able to fix. A win32
device context (HDC) can only be destroyed from the thread that created
it. The pbuffers for pre-render cameras are created in
osgUtil::RenderStage::runCameraSetUp, from the draw thread. But
closeImplementation is normally invoked from the destructor in the main
application thread. With the additional error messages I have added,
osgprerender will now output a couple of warnings from
osgViewer::PixelBufferWin32::closeImplementation() at exit, after
running multi-threaded on windows. I think that is a good thing, to
highlight the problem. I looked into fixing it in osgViewer::Renderer &
osgUtil::RenderStage, but it was too involved for me. My own
application requirements are only single-threaded.
Unrelated fix - an uninitialised variable in
osg::GraphicsThread::FlushDeletedGLObjectsOperation().
"
remain in pressed state after revealing, even if they are no
longer pressed on the keyboard. This can have bad effects,
especially if the stuck keys are modifier keys. One has to
press and release the stuck keys again to reset the wrong state.
The fix keeps track of all key presses and releases. On FocusOut
and UnmapNotify it releases all keys that are in pressed state,
and on KeymapNotify (following a FocusIn), it sets the currently
pressed keys again. To avoid confusion in the OSG-using application
normal keys are always reported released /before/ and pressed
/after/ modifier keys.
As current key states are returned as char[32] keymap by
XQueryKeymap and XKeymapEvent, this format is also used to
recognize modifier keys and for maintaining the current
internal key state. Functions to set/clear/query bits in
such a keymap are added.
The patch was extensively tested with osgkeyboard and
FlightGear under KDE and fvwm2. It was not tested on a
Xinerama setup or with multiple windows, but as _eventDisplay
is used throughout, there should be no problems. The patch also
makes the following changes:
- removes old and obsolete handling of modifier keys in ::adaptKey().
This wasn't only unused, but also wrong (and for that reason commented
out in revision 7066). The modifier states are actually handled
in ./src/osgGA/EventQueue.cpp (EventQueue::keyPress/keyRelease).
- fixes some spelling"
The modifications I made are very small but they are absolutely usefull to use osgIntrospection with visual studio 7.1 or 8 in debug modes.
This should also solve other minor common problems (converter memory leak, virtual destructor for PropertyInfo, etc...).
I choosed two function names : Reflection::uninitialize() and Type::reset(), this can of course be changed if someone has a better idea...
I made the changes against OSG 2.2.0 public release. I tested the result with VS 7.1, VS 7.1 SP1, VS 8.0 SP1 and AQTime 5.0 on Windows XP SP2... All 4 seem to agree : they detected memory leaks before and don't anymore.
Sorry I haven't take the time to test that on linux but the changes are so small I doubt there could be a problem... I let you check that on your side :-).
I hope this will help making OSG an even more wonderfull library."
Attached is a fixed version of OverlayNode.cpp. I fixed CustomPolytope::cut( osg::Plane ) method. Bug was apparent in such scenario:
Let P1 be some random frustum polytope
Let P2 be the polytope that was created from P1 bounding box (P2 contains P1 entirely)
Then ignoring precision errors: P1.cut( P2 ) == P2.cut( P1 ) == P1. But this condition was not always met. Cut failed when some of the polytope reference points happened to lie exactly on some intersecting planes in both P1 & P2 (plane distance was = 0).
I only use CustomPolytope for my shadowing stuff so I did not test how this affects rest of OverlayNode.cpp.
----2----
Also attached is a minor precision improvement for osg::Plane intersect method (double version).
----3----
I have also one observation regarding osg::Plane - There are two intersect vertices methods (float and double flavour):
inline int intersect(const std::vector<Vec3>& vertices) const
inline int intersect(const std::vector<Vec3d>& vertices) const
I guess osg::Plane won't compile when someone changes default vec3 typedef to vec3d. Shouldn't the first method be changed to use vec3f explicitly ? Ie:
inline int intersect(const std::vector<Vec3f>& vertices) const"
pbuffer functions or exactly ask for the extensions we need to call the
apropriate glx extension functions for and around pbuffers extensions.
The glx 1.3 version of this functios are prefered. If this is not pressent we
are looking for the glx extensions and check for them.
Prevously we just used some mix of the glx 1.3 functions or the extension
functions without making sure that this extension is present.
"
carbon-implementation of GraphicsWindow. Now you can use an AGLDrawable
in conjunction with osgViewer/osgCompositeViewer."
Changes from Robert Osfield, changed std::cout to osg::notify(osg::INFO)
creating subclasses of osg::Array that referenced data
stored an application's internal data structures. I took
a stab at implementing that and ran into a couple of
downcasts in Geometry.cpp. Enclosed is my take at fixing
those along with a simple example of how to do this."
in the DatabasePager. The practical effects of these are to greatly reduce startup time
and the time to load an individual scenery tile in FlightGear.
- From my log message:
Minimize the number of StateSets and drawables that are compiled by checking
if they have already been compiled or will be elminated by the
SharedStateManager.
Move the sorting of the dataToCompile queue out of compileGLObjects
into the man pager run function.
Change the SharedStateManager to use maps instead of vectors."
StateSet::removeAssociatedModes(const StateAttribute*)
and a
StateSet::removeAssociatedTextureModes(unsigned, const StateAttribute*)
call. These funktions are just missing for a complete api IMO."
file. The code works well but the SectorPlacer comments stayed in the new
file. I have altered those comments so they now contain valid information for
the BoxPlacer class and the doxygen generated documentation is correct.
"
CullVisitor/SceneView:
*Feature: This version supports multiple clearnodes in the graph, one per renderstage.
Text:
*Feature: Performance Enhancement when calling SetBackdropColor
Material:
*Fix: OpenGL calls are now made according to the OpenGL Standard
"
enhanced version of PolytopeIntersector.
New features of PolytopeIntersector :
* Dimension mask: The user may specify the dimensions of the
primitives to be tested. Checking polytope-triangle and
polytope-quad intersections is rather slow so this can
be turned off.
* Reference plane: The resulting intersections are sorted
by the distance to this plane.
New memebers of PolytopeIntersector::Intersection :
* distance: Distance of localIntersectionPoint to the reference plane
* maxDistance: Maximum distance of all intersectionPoints to the
reference plane.
* intersectionPoints: The points intersecting the planes of the polytope
or points completely inside the polytope.
* localIntersectionPoint: arithmetic mean of all intersection points
* primitiveIndex: Index of the primitive that intersected
I added some more output to the example osgkeyboardmouse."
highlighted problems with Light, ClipPlane and Hint usage in osg::State's usage of cloneType
and reassignment of target/num in StateSet/these StateAttributes.
- Get or set the target number of PagedLOD children to remove per frame.
- Get or set the minimum number of inactive PagedLOD to keep.
Corresponding environment variables have been added too.
The default values reproduce the previous DatabasePager behavior."
- set the resolution of the shadow map; it calls dirty() to
re-initialize at next update
- keep a list of Shader objects to use instead of the default ones, if
the list is empty, the default shaders are used
- explicitly create the Uniform variables, so that subsequent additions
that require more Uniforms can put them in a central place
- set a Light or LightSource to use explicitly for shadow casting,
allows multiple lights in the scene, with one casting shadows
There are two additions that do not ( yet ) function correctly, but in
the present usage they do not interfere with the regular usage of the
techique:
- support for using spotlights, it's using Light.spotCutoff to determine
if it's a spot-light and not point-light,
there is an error in the setup of either the shadow camera or the
texgen, most likely due to the direction of the spotlight, since the
position is being used just like in point or directional lights.
- creation of a debugHUD
the hud is created properly, ( the example included shows it ), but
it displays only white, there has been some discussion of displaying the
shadow map, but I could not find it, the addition of a simple fragment
shader with the appropriate color transform should get this going."
