Hi Robert,
The OpenSceneGraph-3.6 branch compiles and runs properly on my OSX Mojave
(10.14.6) system. I tried osgviewer, several OSG examples, and my own
OSG-based projects.
Thanks,
Ravi
On Mon, Dec 16, 2019 at 12:16 PM Robert Osfield
wrote:
> Hi All,
>
> I have merged the outstanding pull
t; Laurens.
>
> On Wed, Sep 11, 2019 at 9:54 AM Robert Osfield
> wrote:
>
>> Hi Ravi,
>>
>> On Tue, 10 Sep 2019 at 16:34, Ravi Mathur wrote:
>>
>>> I've tried on OSG 3.6.3 and 3.6.4. You can see from the error I posted
>>> that it's because of
10, 2019, 4:14 AM Robert Osfield
wrote:
> Hi Ravi,
>
> On Mon, 9 Sep 2019 at 22:34, Ravi Mathur wrote:
>
>> Has anyone using OSX noticed that the osgslice example does not build?
>> This prevents OSG from building on OSX out of the box. I'm currently using
>> OSX
Hi all,
Has anyone using OSX noticed that the osgslice example does not build? This
prevents OSG from building on OSX out of the box. I'm currently using OSX
10.14.6 & XCode 10.3, but it's been happening for over a year now.
Code:
[100%] Linking CXX executable ../../bin/osgslice
Undefined
Lighting is enabled, which is why the line darkens based on camera
position. Disable lighting on line_geometry. Search osg examples for
GL_LIGHTING to see how.
---
Ravi
Sent from my phone, please excuse my brevity.
On Sun, Dec 23, 2018, 4:46 PM Diego Mancilla Hello,
>
> I'm drawing some
There are plenty of osg examples that show how to create your own event
handler by subclassing from GUIEventHandler. For example look at osgpoints.
Make your own event handler, and change the text visibility (using its node
mask) based on whatever keys you choose.
To update text dynamically, use
If you will be dynamically changing your text, then specify that when you
first create it:
osgText::Text* text = new osgText::Text;
text->setDataVariance(osg::Object::DYNAMIC);
Note that setting the data variance is important for multithreaded Viewers
ng". It's easy for me to pause all active Viewers
while destroying one.
Thanks,
Ravi
On Tue, Dec 18, 2018 at 2:28 PM Ravi Mathur wrote:
> Hi Robert, yes I'm working on a standalone test program now. Multithreaded
> crashes are often hard to reproduce in a vacuum, but as soon as I hav
> Could you create a small test program that reproduces the crash. Details
> on your OS/dev tools/hardware will be useful too.
>
> Cheers.
> Robert.
>
> On Tue, 18 Dec 2018 at 08:50, Ravi Mathur wrote:
>
>> Hello all,
>>
>> I'm seeing a crash on
, smaller mem use
> picter is static version. Note MB usage drop in dynamic version after
> minute or so from the moment updates stopped. I suspect driver compacted
> the memory when it noticed the resources are no longer updated.
>
>
>
> Cheers,
> Hope this helps,
> Woj
Hello all,
I'm running into a strange performance drop issue when using dynamic VBOs
that change frequently. I am measuring performance using framerate with
vsync turned off. I know that framerate isn't always the best performance
measurement, but my example is simple enough and the performance
Answers to all of your questions below.
On Fri, Oct 12, 2018 at 8:49 PM A Z wrote:
> To whom it may concern,
>
> I have a group of basic startup questions about use of the OSG
> C++ 3D graphics library.
>
> *-Is it compatible with Windows, Android, Max, Linux?*
>
Yes.
Hi Steven,
What OS are you on? On OSX, event handling MUST happen from the main
thread, so you have to coordinate the frame loop between threads.
Ravi
On Tue, Sep 11, 2018 at 11:43 AM Robert Osfield
wrote:
> Hi Steven,
>
> I haven't heard of anyone trying to call frame from multiple threads
>
Thanks guys for figuring this out. I ran into the exact same problem
earlier this week with Ubuntu 18.04. I'm using CompositeViewer similar to
the "-3" case of the osgcompositeviewer example (multiple Views in one
window). I call osg::GraphicsContext::createGraphicsContext(traits) with my
custom
Hey Martin,
You are not interpreting the parameters properly. Look at the header for
Matrixd.cpp:
/** Set the position and orientation to be a view matrix,
* using the same convention as gluLookAt.
*/
void makeLookAt(const Vec3d& eye,const Vec3d& center,const Vec3d& up);
It
I've been using Macs and developing on MacOS for over 25 years now, but
still I gotta say I completely understand your frustration Robert. My past
two computers have been Windows laptops, which is also the first time I've
bought non-Mac hardware, mainly because Apple doesn't seem to care about VR
have more
questions.
Thanks!
Ravi Mathur
Sr. Guidance Navigation and Control Engineer
Emergent Space Technologies, Inc.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=73544#73544
___
osg-users mailing list
No worries, thank you for looking into it anyway. :)
To be honest, I don't think that "fixing" this behavior within OSG is worth
it. The code additions to ViewerBase::renderingTraversals() would be
duplicated from Renderer and SceneView, and would only benefit fringe cases
where the user wants
robertosfield wrote:
> Ops, don't send modifications without compiling them first...
>
> The View::LightingMode and SceneView::LightingMode aren't compatible
> enum's so can't be passed as is. I've spotted a better way of doing
> this update, putting into the Renderer::updateSceneView() method
ravidavi wrote:
>
> robertosfield wrote:
> > HI Ravi,
> >
> > I haven't yet looked deeply into the issue as I'm still in post
> > holiday catch up mode. What version of the OSG are you using?
> >
> > Robert.
> >
>
>
> I'm on the OpenSceneGraph-3.5.6 tag.
I just checked, and the same
robertosfield wrote:
> HI Ravi,
>
> I haven't yet looked deeply into the issue as I'm still in post
> holiday catch up mode. What version of the OSG are you using?
>
> Robert.
>
I'm on the OpenSceneGraph-3.5.6 tag.
--
Read this topic online here:
OK I figured it out after digging further into SceneView. In case anyone else
ever runs into this, here is what's going on and how to fix it.
When a new View is created (e.g. osgViewer), it defaults to HEADLIGHT mode and
internally creates the master Camera, along with a Renderer and SceneView
Hi all,
Sorry to bring up yet another question about
osg::View::setLightingMode(NO_LIGHT), but I'm having trouble understanding some
lighting behavior.
Suppose I have a simple scene with no explicitly-specified StateSets ... e.g.
I'm not setting lights or enabling GL_LIGHTING or anything like
Heh I hope I didn't oversell it! :)
OK, I will make a standalone osg example and post it here for people to try. I
will rely on the community to test with shadows and other use cases.
I have some month-end deadlines going on now but will do it in early June.
Ravi
--
Read this
Hi all,
A looong time ago I had created the DepthPartitionNode, which split up an OSG
scene into multiple depth partitions and rendered them with multiple cameras.
This turned into the original osgdepthpartition example. My implementation
created nested cameras, which worked fine at the time,
Aha, you're right I should have said "interleaved". No wonder I wasn't finding
much info, I was using the wrong term! :) Now I see past posts by Robert and
others discussing interleaved arrays. I'll stick with my current approach of
copying the data into separate OSG arrays and deallocate the
Hi all,
I have an array of floats that contains packed vertex data, provided by an
external API, that I want to render. The data represents
[position,normal,texcoord] for each vertex. The array layout is:
Hi Werner,
Instead of modifying the main camera attachments, you might want to create a
slave camera whose sole purpose is to render to FBO. Then you can
enable/disable that camera using slaveCamera->setNodeMask (0x0 to disable,
0x) to enable.
Code:
// Create slave camera and set
OK, after thinking about this, reading MANY other forum posts, and looking into
the osg::Camera code, I've realized that my post title itself was incorrect.
Sharing FBOs is not the appropriate approach, since FBOs are an internal object
to a camera's RenderStage, which itself doesn't exist
Hi all,
I would like to know how to get multiple slave cameras to successively render
to the same texture using multisampling. I've looked at every osg example, but
multisampled FBOs are only discussed in the context of a single rendering
camera.
Without multisampling life is simple. For
Thanks Robert. But still, why does resize() not accept x/y arguments? As it is
now, it cannot be used with CompositeViewer where Cameras are distributed
around the window with various Viewports. Is it just intended for use with the
regular osg::Viewer?
robertosfield wrote:
> Hi Ravi,
>
> The
Hi all,
In osg::Camera, there exist two resize functions: Camera::resize(width, height)
and Camera::setViewport(x, y, width, height). I'm confused conceptually as to
when to use which one. Some examples use resize (osgdistortion), whereas others
use setViewport (osghud).
Looking at the
FYI, my vertex/fragment shaders are using #version 120 because they need to be
compatible with OSX X11 (XQuartz). I have noticed that even on the graphics
cards with the compatibility profile (windows machines with certain ATI/AMD
cards), directly using gl_Vertex instead of osg_Vertex seems to
Hi all,
I see some behavior that I don't understand when using the osg_Vertex shader
attribute across various graphics cards. Hopefully someone can shed some light.
I am enabling OSG uniforms using
GraphicsWindow->getState()->setUseModelViewAndProjectionUniforms(true). Then,
in my vertex
Makes sense. My situation is simple; I have no need to override parts of
shaders, just load a few fragment shaders from files. So I'm going with the
simple approach of an osg::Program per osg::Drawable, with a shared Vertex
osg::Shader and separate Fragment osg::Shaders.
Thanks for clearing
Thank you Robert! Can you speak to the performance of using multiple
osg::Programs vs shader composition?
Another question: suppose I use a separate osg::Program for each osg::Drawable.
Each Program would use the same Vertex osg::Shader object, but have its own
Fragment osg::Shader object.
Hi all,
I have a Geode that contains multiple Drawables. I set a Program containing a
vertex Shader on the Geode's StateSet. This was done because all contained
drawables should use the same vertex shader. Everything works great.
Now, I want to set a fragment shader on just one of the
Fix submitted in Submissions Post
(http://forum.openscenegraph.org/viewtopic.php?t=16036).
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=68244#68244
___
osg-users mailing list
Hi Robert (and any other OSX OSG devs),
Nowadays XCode only includes the OSX 10.11 SDK, which supports deployment
targets back to 10.7. So one can use the 10.11 SDK to target OSX 10.8, which is
what CMake does via CMAKE_OSX_DEPLOYMENT_TARGET.
However, CMake 3.1.0 was the first major release to
Hi all,
I realized that the OpenGL link dependencies for OSG are not quite correct for
OSX/X11. Specifically, if OSG_WINDOWING_SYSTEM is X11 then osgViewer is built
with X11/OpenGL, but osg (base) is still built with Cocoa/OpenGL. This mismatch
of OpenGL implementations can cause very
Fix submitted in Submissions Post
(http://forum.openscenegraph.org/viewtopic.php?p=68188).
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=68190#68190
___
osg-users mailing list
Fix submitted in Submissions Post
(http://forum.openscenegraph.org/viewtopic.php?p=68188).
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=68189#68189
___
osg-users mailing list
Hi Robert,
Excellent, just what I wanted to hear. I'd be happy to provide updates on the
OSX side of the CMake build system. I only have one OS and compiler combo to
code on (10.10.5 with XCode 7.2), but I have a decent amount of experience with
general OSX/CMake development and OSG
Hi all,
I'm surprised nobody has seen this yet, but OSG does not build on OSX when
using XCode 7 and CMake 3.4+. Specifically, the Configure step can be run once,
but cannot ever be run again unless the CMake cache is deleted. The error
through by CMake is the infamous:
Code:
Well I'm dumb. Turns out, I was using CMake to find OpenGL, which linked to the
Cocoa version of GL instead of X11 like I needed. After linking to the correct
OpenGL libraries, I no longer need to call glXMakeCurrent; glViewport does the
job like it's supposed to.
Just thought I'd share for
Hi all,
I have an interesting problem to which I found a strange solution. I don't know
why the solution works, and I'd like to know if anyone here could shed some
light.
I have an OSG-based visualization API that is designed for use from within any
user-provided GUI toolkit. All OpenGL
OK, looks like the fix could be
Code:
IF(APPLE AND NOT (OSG_WINDOWING_SYSTEM STREQUAL "Cocoa"))
RETURN()
ENDIF()
This applies to avfoundation (which only compiles on OSX anyway), as well as
osgmultitouch and osgoscdevice examples, which will continue to compile on
non-apple platforms
Oops the fix I posted is wrong. osgmultitouch and osgoscdevice should continue
to work on non-Apple platforms.
Any ideas on a good fix?
Ravi
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=67769#67769
Hi all,
A few of the newer OSG components explicitly reference
osgViewer::GraphicsWindowCocoa in their code. This causes link errors when
compiling OSG on MacOS with OSG_WINDOWING_SYSTEM set to X11.
The affected components are: avfoundation (osgdb), osgmultitouch (examples),
osgoscdevice
Sure thing Robert.
robertosfield wrote:
> Hi Ravi,
>
>
> Could you post the whole modified file to osg-submissions, I'll review this
> and get it merged.
>
>
> Robert.
>
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=65561#65561
Hi all,
The osgviewerCocoa example gives the following error during build:
Code:
ViewerCocoa.mm:176:9: error:
cast from pointer to smaller type 'NSOpenGLPixelFormatAttribute'
(aka 'unsigned int') loses information
(NSOpenGLPixelFormatAttribute)nil
^~
This occurs on
Hey all,
First off, this is NOT another can we change all vec3 to vec3d
thread. My question is more specific. :)
In osg::CullStack, the EyePointStack is a Vec3 stack, therefore a
Vec3f stack. However, the stack itself is built up using elements
from a osg::Matrix, which could be a
it out on my computer
tomorrow to see if there are any unexpected hangups.
Ravi
On Apr 20, 2009, at 3:55 PM, Robert Osfield wrote:
On Mon, Apr 20, 2009 at 9:37 PM, Ravi Mathur
ravid...@mail.utexas.edu wrote:
So would it be worth it to implement the CullStack::EyePointStack
typedef as
Vec3d
53 matches
Mail list logo