Hi André,
Thanks for the exhustive testing. I don't quite have the same set up
here, but a bit similiar I have a X2 3800, 2GB RAM, Linux, 7800GT x 2,
2 screens (one per card). All the examples you have problems run
great on my system, nothing is slow, no crashes, no artifacts on any
of the
Hi robert,
so with the latest CVS version and the missing fix i did myself for my
system
i started the runexample.bat:
crash: at the end, need ctrl+c to break
osgscalarbar
osgvertexprogram
all other worked very nice, without dead lock, crash ,
nice work you did last night !
many thanks,
Hi Adrian,
On 2/15/07, Adrian Egli [EMAIL PROTECTED] wrote:
so with the latest CVS version and the missing fix i did myself for my
system
What fix was this? The setFrameStamp?
i started the runexample.bat:
crash: at the end, need ctrl+c to break
osgscalarbar
osgvertexprogram
Were these
2007/2/15, Robert Osfield [EMAIL PROTECTED]:
Hi Adrian,
On 2/15/07, Adrian Egli [EMAIL PROTECTED] wrote:
so with the latest CVS version and the missing fix i did myself for my
system
What fix was this? The setFrameStamp?
YES the one you did
i started the runexample.bat:
crash: at
HI Adrain,
On 2/15/07, Adrian Egli [EMAIL PROTECTED] wrote:
at the end, yes no access violation, but didn't terminate them self
Sounds like a dead lock on exit. Could you run these apps in a
debugger and try to exit, then ctrl-c it to find out where the
deadlock is occurring. Note there
Hi Robert,
osgprecipitation makes use of GLSL heavily, and while it vertex
and pixel fill requirments
won't stress a 7900GT, I wouldn't be surprised if inbuilt graphics
will struggle. Could it
be the once of screens is being driven by the NVIDIA 6150?
Yes. Performance in
Hi Andre,
On 2/15/07, André Garneau [EMAIL PROTECTED] wrote:
Yes. Performance in single-threaded mode is excellent though. Anyway since
this is not a typical configuration, I'll redo the tests without the 6150.
Normally one would expect multi-threaded multi-context to run faster
than single
Hi Adrian,
Thanks for the changed files, the problem is that these files are
autogenerated so the next time I update the wrappers these changes
will be lost. In this instances its the I_BaseType macro defined in
include/osgIntrospection/ReflectionMacros that will where the mods
need to happen.
Hi robert,
:-( too bad, ok but if you autogenerate the files, means that i have to do a
work around in the marco.
you can replace the current macro by following macro:
#define I_BaseType(x) \
{ \
typedef x BASE_TYPE_X;\
addBaseType(typeof( BASE_TYPE_X )); \
BaseTypeConverters
Hi Adrian,
On 2/15/07, Adrian Egli [EMAIL PROTECTED] wrote:
Hi robert,
:-( too bad, ok but if you autogenerate the files, means that i have to do a
work around in the marco.
you can replace the current macro by following macro:
#define I_BaseType(x) \
{ \
typedef x BASE_TYPE_X;\
AM
To: osg users
Subject: Re: [osg-users] Please test CVS version of OpenSceneGraph
Hi robert,
:-( too bad, ok but if you autogenerate the files, means that i have to
do a work around in the marco.
you can replace the current macro by following macro:
#define I_BaseType(x
Hi Win32's,
In light of the continuing threading problems under Win32 I think I'll
default osgViewer's threading model to SingleThreaded and check this
into CVS. This will take the pressure off solving all the various
combinations of threading right away.
Robert.
On 2/15/07, Mike Wittman [EMAIL PROTECTED] wrote:
The template-argument-list errors look to be due to a lack of spaces in the
typeof macro. Here's a new Reflection header which should fix those
problems. For consistency I think that's the best solution.
Thanks Mike, merged and submitted to
On 2/15/07, Robert Osfield [EMAIL PROTECTED] wrote:
Hi Win32's,
In light of the continuing threading problems under Win32 I think I'll
default osgViewer's threading model to SingleThreaded and check this
into CVS. This will take the pressure off solving all the various
combinations of
On 2/15/07, Adrian Egli [EMAIL PROTECTED] wrote:
in the attachment the marco file:
C:\dev\OpenSceneGraphCVS\OpenSceneGraph\include\osgIntrospection\ReflectionMacros
Thanks, I've just reviewed the change and it looks safe. But..
hopefully redundant because Mke's fix to the typeof macro that
Robert Osfield wrote:
Work on setting up the new OpenSceneGraph server is underway, one of
the items on the list for migration is migrating from CVS to SVN.
Before we do this migration I would like to make sure that we leave
the CVS versions in a good state i.e. compiling on all the major
Hi Robert,
good news:
(1) build under VC2003 without any error !
(2) no crash when running runexample.osg
/adegli
2007/2/15, Martin Spott [EMAIL PROTECTED]:
Robert Osfield wrote:
Work on setting up the new OpenSceneGraph server is underway, one of
the items on the list for migration is
Hi Martin,
On 2/15/07, Martin Spott [EMAIL PROTECTED] wrote:
Robert Osfield wrote:
Work on setting up the new OpenSceneGraph server is underway, one of
the items on the list for migration is migrating from CVS to SVN.
Before we do this migration I would like to make sure that we leave
the
Hi Robert,
Robert Osfield wrote:
What viewer codes does FlightGear currently use?
Mathias told me that osgViewer ist _not_ being used - apparently
there's sort of a generic viewer in one of these libraries:
quickstep: 18:12:05 /usr/local/src/FlightGear ldd /opt/FlightGear/bin/fgfs |
grep -i
Hi J-S,
osgmanipuilator is a new example, and I haven't yet had a chance to
port it across to osgViewer. Work in this direction has started, but
requires mods to further mods to osgManipulator to make it a bit more
flexible w.r.t how its set up. Could you send a complete modified
Hi Adrain,
Thanks for screenshots, it does like a thread issue of some kind.
Again, could please answer the following questions:
What hardware do you have ?
1.a) Graphics hardware
1.b) CPU type, number of cores, are they enabled?
1.c) Number of screens and their configuration
It is
CPU:
Intel Dual Core : T7600 2.33 GHz
2 GB RAM
GPU:
ATI Mobility Radeon X1600 : Driver ATI Version 8.24
Windows:
Windows XP : Version 2002, Service Pack 2, all securtiy updates installed /
updated today !
Screen:
1
** TEST with SingleThreaded
C:\Dokumente und
Hi Adrian,
On 2/14/07, Adrian Egli [EMAIL PROTECTED] wrote:
CPU:
Intel Dual Core : T7600 2.33 GHz
2 GB RAM
GPU:
ATI Mobility Radeon X1600 : Driver ATI Version 8.24
Windows:
Windows XP : Version 2002, Service Pack 2, all securtiy updates installed /
updated today !
Screen:
1
Thanks, this is
i did following test, and crash !
** TEST with SingleThreaded
C:\Dokumente und Einstellungen\adegliset OSG_THREADING=SingleThreaded
C:\Dokumente und Einstellungen\adegliosgshaders
GraphicsWindowWin32::setPixelFormat() - Found a matching pixel format but
without the
Robert Osfield schrieb:
So please please test the CVS version of the OSG,
On OS X osgviewer crashes:
show the stats via s and then show the help via h - bang. Here are
some stacktraces for the various threading-modes:
If you show first the help and then the stats it works, even if you hide
Hi Adrian,
On 2/14/07, Adrian Egli [EMAIL PROTECTED] wrote:
i did following test, and crash !
** TEST with SingleThreaded
C:\Dokumente und Einstellungen\adegliset OSG_THREADING=SingleThreaded
C:\Dokumente und Einstellungen\adegliosgshaders
GraphicsWindowWin32::setPixelFormat()
HI Stephan,
On 2/14/07, Stephan Maximilian Huber [EMAIL PROTECTED] wrote:
Robert Osfield schrieb:
So please please test the CVS version of the OSG,
On OS X osgviewer crashes:
show the stats via s and then show the help via h - bang. Here are
some stacktraces for the various threading-modes:
Hi Robert,
another crash on my side:
osgprerender --window cow.osg will crash here: (regardless of the
threading model)
#00x0029b421 in osg::GraphicsContext::releaseContext at
GraphicsContext.cpp:277
#10x0127a0f1 in osgUtil::RenderStage::draw at RenderStage.cpp:847
#20x01275c8b
HI Stephan,
On 2/14/07, Stephan Maximilian Huber [EMAIL PROTECTED] wrote:
I am switching the threading-model via 'm'. Regardless, what
threading-model I chose, showing the help after stats will crash. But I
have to show full stats, showing only the fps will work.
I did a quick test and set the
Hi Stephan,
Thanks for the excellent testing. I have just tried:
osgprerender --window cow.osg
And it crashes here too, which is good news as it means I have a good
chance of tracking the problem down and fixing it and knowing that it
is really fixed. This one is probably related to the
Hello Robert, good morning (in my timezone at least),
Could you send a complete modified osgmanipulator.dsp?
Here you are. (Hope it comes through ok, I don't use Thunderbird but still...)
The problem was a missing space between one libpath option and the next, which
prevented _both_ from being
Robert Osfield schrieb:
Could you try:
osgviewer file.osg --SingleThreaded
Then see if the stats fails.
Here are the results for osgviewer, pressing 2 x s, press 1 x h:
osgviewer cow.osg --SingleThreaded OK
osgviewer cow.osg --CullDrawThreadPerContext OK
osgviewer cow.osg
Hi Stephan,
On 2/14/07, Stephan Maximilian Huber [EMAIL PROTECTED] wrote:
Here are the results for osgviewer, pressing 2 x s, press 1 x h:
osgviewer cow.osg --SingleThreaded OK
osgviewer cow.osg --CullDrawThreadPerContext OK
osgviewer cow.osg --CullThreadPerCameraDrawThreadPerContext
Thanks J-S, fix now merged and submitted to CVS.
On 2/14/07, Jean-Sebastien Guay [EMAIL PROTECTED] wrote:
Hello Robert, good morning (in my timezone at least),
Could you send a complete modified osgmanipulator.dsp?
Here you are. (Hope it comes through ok, I don't use Thunderbird but
-
[EMAIL PROTECTED] On Behalf Of Robert Osfield
Sent: Tuesday, February 13, 2007 4:50 PM
To: osg users
Subject: Re: [osg-users] Please test CVS version of OpenSceneGraph
HI Adrian,
On 2/13/07, Adrian Egli [EMAIL PROTECTED] wrote:
thanks robert, i sould upgrade the compiler vc 2003 to vc 2005
, February 13, 2007 4:50 PM
To: osg users
Subject: Re: [osg-users] Please test CVS version of OpenSceneGraph
HI Adrian,
On 2/13/07, Adrian Egli [EMAIL PROTECTED] wrote:
thanks robert, i sould upgrade the compiler vc 2003 to vc 2005? but
i
don't
like doing this at the moment because
Robert Osfield schrieb:
I have just reviewed the code (src/osgViewer/Viewer.cpp) that checks
the OSG_THREADING and it looks OK to me, I did test it when I added
the code at worked then. Any chance you could put some debug info in
the Viewer::suggestBestThreadingModel() so to see why it might be
Hi Adrian,
I have introduced various extra measures to ensure that the
DataVariance is set correctly, and in the case of osgshaders this
changes what was a STATIC into a DYNAMIC setting for the StateSet
holding the updating uniforms. I have also changed osgpick to set the
DataVariance of the
Hi Adrian,
On 2/14/07, Adrian Egli [EMAIL PROTECTED] wrote:
i did also the test
osgviewer cow.osg --SingleThreaded OK
osgviewer cow.osg --CullDrawThreadPerContext OK
osgviewer cow.osg
--CullThreadPerCameraDrawThreadPerContext CRASH
osgviewer cow.osg
Hi Robert,
GraphicsWindowWin32::setPixelFormat() - Found a matching pixel format but
without the WGL_SWAP_METHOD_ARB specification for screen #0
The ATI driver probably doesn't swap the WGL_SWAP_METHOD_ARB control,
this may or may not be an issue, perhaps Andreas can comment. If its
not
Hi André,
On 2/14/07, André Garneau [EMAIL PROTECTED] wrote:
Did not read the full thread yet, but this message just really means that
the WGL_SWAP_METHOD_ARB specification in the graphic traits is being ignored
because the OpenGL driver does not support it.
It is harmless and just a warning
Like J-S, I also get this compile error from osgManipulator. This is from
CVS about 12 hours old using VS8. The code appears to be trying to
initialize an iterator with an int.
Paul Martz
Skew Matrix Software LLC
http://www.skew-matrix.com
303 859 9466
-
Project Core
Hi robert,
i updated my CVS and will redo the test.
the crash happends any time. i can't say when. just it happends :-(
/adegli
2007/2/14, Robert Osfield [EMAIL PROTECTED]:
Hi Adrian,
On 2/14/07, Adrian Egli [EMAIL PROTECTED] wrote:
i did also the test
Hi Robert,
Thanks for the clarification. So it should be safe to change this
warning to just an info message then?
Yes.
André
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
Thanks for the data point Paul. I've just done a clean build of the
osgManipulator wrapper and it builds cleanly with gcc 4.02. The error
is pretty hard to decode as there is templates and macro's flying
around. I don't get the error so I can't even tweak and test.
On 2/14/07, Paul Martz
Hi Adrian,
On 2/14/07, Adrian Egli [EMAIL PROTECTED] wrote:
Hi robert,
i updated my CVS and will redo the test.
the crash happends any time. i can't say when. just it happends :-(
Once you have an updated version it'll be interesting to see how
things are going. I will at least closed off
On 2/14/07, Paul Martz [EMAIL PROTECTED] wrote:
Hi Robert -- This is not a problem with the wrapper. It's in core
osgManipulator. I'll play with it this evening if I have time and see if I
can create a fix.
Ahh now it makes more sense, I've been off on a wild goose chase
looking at the
Hello Robert,
Thanks for the data point Paul. I've just done a clean build of the
osgManipulator wrapper and it builds cleanly with gcc 4.02. The error
is pretty hard to decode as there is templates and macro's flying
around. I don't get the error so I can't even tweak and test.
It's
Hi Stephan,
On 2/14/07, Stephan Maximilian Huber [EMAIL PROTECTED] wrote:
Robert Osfield schrieb:
Could you try:
osgviewer file.osg --SingleThreaded
Then see if the stats fails.
Here are the results for osgviewer, pressing 2 x s, press 1 x h:
osgviewer cow.osg --SingleThreaded OK
Hi Stephan,
I've now fixed the crash when running osgprerender --window cow.osg,
but the copy texture doesn't yet work so we get a black window. I
believe this later problem is down to makeContextCurrentImplementation
not being implemented on the various GraphicsWindow implementations.
I'll do
Robert Osfield schrieb:
I have now added a Mutex to osg::Stats and use of it in the various
get/set methods.
I have tested it out by running with DrawThreadPerContext and
CullThreadPerCameraDrawThreadPerContext and holding down the 's' key
and so far so good, no crashes. I'm not noticing a
Hi Stephan,
On 2/14/07, Stephan Maximilian Huber [EMAIL PROTECTED] wrote:
Could you do a cvs update and see if the above now work?
I did a cvs-update and made a clean compile -- I am sorry, your fix does
not work
osgviewer cow.osg --DrawThreadPerContext will crash after pressing 2 x
's' and
PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Robert Osfield
Sent: February-14-07 2:20 PM
To: osg users
Subject: Re: [osg-users] Please test CVS version of OpenSceneGraph
Hi Stephan,
On 2/14/07, Stephan Maximilian Huber [EMAIL PROTECTED] wrote:
Could you do a cvs update and see if the above now
Hi André,
On 2/15/07, André Garneau [EMAIL PROTECTED] wrote:
Have this crash on Windows as well.
The problem is that the stats object gets deleted while the 'h' command is
processed and there is still a pending reference to this object in the
BlockDrawCallback object. The next time this one
Hi Robert,
Most excellent, yet this all makes sense now, the only thing that
doesn't make sense is why Linux was working so robustly... when
toggling help and stats and running multi-threaded.
There two solutions I've opted for (now checked into CVS) is that the
callbacks in
Hi Robert,
Once you have an updated version it'll be interesting to see how
things are going. I will at least closed off some avenues for
failure.
It might be worth forcing osgViewer::Viewer down on to using just
SingleThreaded. Setting OSG_THREADING to SingleThreaded should be
example.
André
-Original Message-
From: André Garneau [mailto:[EMAIL PROTECTED]
Sent: February-15-07 12:28 AM
To: 'osg users'
Subject: RE: [osg-users] Please test CVS version of OpenSceneGraph
Hi Robert,
Once you have an updated version it'll be interesting to see how
things are going
Robert Osfield wrote:
Hi All,
Work on setting up the new OpenSceneGraph server is underway, one of
the items on the list for migration is migrating from CVS to SVN.
Before we do this migration I would like to make sure that we leave
the CVS versions in a good state i.e. compiling on all the
HI Gordon,
On 2/12/07, Gordon Tomlinson [EMAIL PROTECTED] wrote:
Just a thought but would it better to wait for the VCS to SVN move until say
after the 1.3 release and tag happens
Seems a natural clean time to do such a big change rather than in the middle
of the release, just a thought
SVN
Hi J-S,
I have just done a search through OpenThreads and OpenSceneGraph and
can't find any places when the kernel2.lib and user32.lib are butted
together. Could you do a clean check out of OpenThreads and see how
you get on.
Robert.
On 2/13/07, Jean-Sebastien Guay [EMAIL PROTECTED] wrote:
Hi OM,
Sorry about this missed check-in, the missing files should now be checked in.
Robert.
On 2/13/07, Ole-Morten Duesund [EMAIL PROTECTED] wrote:
Robert Osfield wrote:
Hi All,
Work on setting up the new OpenSceneGraph server is underway, one of
the items on the list for migration is
some work some not
2007/2/13, Robert Osfield [EMAIL PROTECTED]:
Hi Adrian,
Do any of the examples work fine?
I've been able to create a osgblendfunction crash, although my toggle
the blend function using key right, key left. I have solved this
crash by change the DataVariance of the
Hi Adrian,
On 2/13/07, Adrian Egli [EMAIL PROTECTED] wrote:
some work some not
Does the message:
GraphicsWindowWin32::setPixelFormat() - Found a matching pixel format
but without the WGL_SWAP_METHOD_ARB specification for screen #0
Appear for all examples? Do any of these examples work?
Hello Robert,
I have just done a search through OpenThreads and OpenSceneGraph and
can't find any places when the kernel2.lib and user32.lib are butted
together.
Sorry about that, it was my fault. I reinstalled Visual C++ Express Edition
lately, and adding user32.lib (and other libs) to the
Yes, it does appear only with this message.
GraphicsWindowWin32::setPixelFormat() - Found a matching pixel format but
without the WGL_SWAP_METHOD_ARB specification for screen #0
Warning: deleting still referenced object 015872E0 of type 'class
osg::Referenced *'
the final reference count
HI J-S,
On 2/13/07, Jean-Sebastien Guay [EMAIL PROTECTED] wrote:
All is going well now, OSG is building. I'll give a status report when I come
back from work. But I can see for now that there are missing introspection
files as someone else mentioned.
I checked in the missing osgIntrospection
Compiling Error:
\dev\OpenSceneGraphCVS\OpenSceneGraph\src\osgWrappers\osg\PrimitiveSet.cpp(875)
: error C2947: expecting '' to terminate template-argument-list, found ''
BEGIN_OBJECT_REFLECTOR(osg::VectorGLsizei)
I_BaseType(std::vectorGLsizei);
I_Constructor0(VectorGLsizei,
Hi Adrian,
I have just do a build on \src\osgWrappers\osg\PrimitiveSet.cpp and it
compiles fine, so I suspect a compiler bug, what compiler are you
using?
Have you successfully compiled osgWrappers before?
Robert.
On 2/13/07, Adrian Egli [EMAIL PROTECTED] wrote:
Compiling Error:
i am using VC 2003, 7.1.3088
2007/2/13, Robert Osfield [EMAIL PROTECTED]:
Hi Adrian,
I have just do a build on \src\osgWrappers\osg\PrimitiveSet.cpp and it
compiles fine, so I suspect a compiler bug, what compiler are you
using?
Have you successfully compiled osgWrappers before?
Robert.
On 2/13/07, Adrian Egli [EMAIL PROTECTED] wrote:
i am using VC 2003, 7.1.3088
Have you been able to compile the osgWrappers before? Is this error a new one?
Could you also clarify some of my earlier questions about what
hardware you have, and whether the warning reported to the console has
may i first time watched this bug. may it was still there i don't know.
2007/2/13, Robert Osfield [EMAIL PROTECTED]:
On 2/13/07, Adrian Egli [EMAIL PROTECTED] wrote:
i am using VC 2003, 7.1.3088
Have you been able to compile the osgWrappers before? Is this error a new
one?
Could you also
complete error list:
+
osgPlugin COLLADA dae warning LNK4098: defaultlib 'MSVCRT' conflicts with
use of other libs; use /NODEFAULTLIB:library
osgPlugin COLLADA dae warning LNK4099: PDB 'libcollada_stderrplugin.pdb' was
not found with
Hi Adrain,
All the non osgWrappers errors are down to not having all the
dependencies installed. If you don't need these features then just
disable their build.
The osgWrapper issue looks like a compiler bug. Any chance you could upgrade?
Robert.
On 2/13/07, Adrian Egli [EMAIL PROTECTED]
thanks robert, i sould upgrade the compiler vc 2003 to vc 2005? but i don't
like doing this at the moment because everything i did, is based on vc2003,
so
i am not shure that the other projects would well working. is there no way
to fix it ?
same issue like runexamples.bat
/adegli
2007/2/13,
Hi All,
I have merged and number of build and execution fixes today, so CVS
should now be in a better position under Windows and X11. XCode
projects will still be out of sync though.
Please keep testing :-)
Robert.
___
osg-users mailing list
HI Adrian,
On 2/13/07, Adrian Egli [EMAIL PROTECTED] wrote:
thanks robert, i sould upgrade the compiler vc 2003 to vc 2005? but i don't
like doing this at the moment because everything i did, is based on vc2003,
so i am not shure that the other projects would well working. is there no way
to
Hello Robert,
The osgWrapper issue looks like a compiler bug. Any chance you could
upgrade?
To me it just seems like you need a space when you have a template class as a
parameter to another template class... In this case, this is a snippet of the
specific code which generates the error:
Hello Robert,
I have merged and number of build and execution fixes today, so CVS
should now be in a better position under Windows and X11. XCode
projects will still be out of sync though.
Please keep testing :-)
Hi Robert
Just a thought but would it better to wait for the VCS to SVN move until say
after the 1.3 release and tag happens
Seems a natural clean time to do such a big change rather than in the middle
of the release, just a thought
Best Regards
Gordon
What's the point of waiting?
On Mon, 12 Feb 2007 22:27:02 +0100, Gordon Tomlinson
[EMAIL PROTECTED] wrote:
Hi Robert
Just a thought but would it better to wait for the VCS to SVN move until
say
after the 1.3 release and tag happens
Seems a natural clean time to do such a big change
Hi robert,
i checkout last CVS version and compiled it: when i am running on a windows
XP machine the runexamples.bat i get some really strange
error:
C:\dev\OpenSceneGraphCVS\OpenSceneGraphecho osgversion
osgversion
C:\dev\OpenSceneGraphCVS\OpenSceneGraphosgversion
OpenSceneGraph Library 1.2
It starts for example osgblendequation cessnafire.osg, then the viewer gets
blocked and returns after a few seconds with an access violation error
2007/2/12, Adrian Egli [EMAIL PROTECTED]:
Hi robert,
i checkout last CVS version and compiled it: when i am running on a
windows XP machine the
Hello Robert,
So please please test the CVS version of the OSG,
Starting to test on Windows, I run into a problem early on, compiling
OpenThreads (Release DLL nd Debug DLL builds give this error, but static builds
don't):
1LINK : fatal error LNK1181: cannot open input file
83 matches
Mail list logo
