opensg-users-request,
I download the source from the daily builder, still it can not work for the
same reason.
Is there additional things I should do when I try to compile it in VC7.1?
rgds.
======= 2003-12-10 20:07:00 ��ㄥ�ㄦ�ヤ俊涓�������锛�=======
>Send Opensg-users mailing list submissions to
> [email protected]
>
>To subscribe or unsubscribe via the World Wide Web, visit
> https://lists.sourceforge.net/lists/listinfo/opensg-users
>or, via email, send a message with subject or body 'help' to
> [EMAIL PROTECTED]
>
>You can reach the person managing the list at
> [EMAIL PROTECTED]
>
>When replying, please edit your Subject line so it is more specific
>than "Re: Contents of Opensg-users digest..."
>
>
>Today's Topics:
>
> 1. Re: Compilation error (Polina Kondratieva)
> 2. Re: Memory leak in DrawAction ? (Martin Schmidt)
> 3. Re: Memory leak in DrawAction ? (Dirk Reiners)
> 4. OpenSG Compile problem in VC7.1 (zjj)
> 5. Re: OpenSG Compile problem in VC7.1 (Dirk Reiners)
> 6. clusterserver accross subnets (Jack Gundrum)
> 7. Re: clusterserver accross subnets (Akos Balazs)
>
>--__--__--
>
>Message: 1
>From: Polina Kondratieva <[EMAIL PROTECTED]>
>To: [email protected]
>Subject: Re: [Opensg-users] Compilation error
>Date: Wed, 10 Dec 2003 07:45:48 +0100
>Reply-To: [email protected]
>
>On Wednesday 10 December 2003 00:08, you wrote:
>Hi Dirk,
>
>Unfortunately, this does not work. But the funniest thing is that for one
>NewFolder (it is third NewFolder) test application was successfully
>compilated. What is the problem with two others I have no idea...
>Anyway, thanks a lot for answer.
>
>Regards,
>Polina.
>
>
>> Hi Polina,
>>
>> On Tue, 9 Dec 2003 22:19:51 +0100, Polina Kondratieva <[EMAIL PROTECTED]
>>
>> sb.mpg.de> wrote:
>> > Hi all,
>> >
>> > I've got some strange error while trying to compile example file in the
>> > directory Tutorials.
>> >
>> > I've successfully added two New Folders with two Materials into the
>> > OpenSG project after the following steps:
>> >
>> > 1) put New Folder into the Source/Experimental,
>> > 2) Remove from the all includes 'OpenSG/',
>> > 3) add and correct files common.libs.in, common.mk,
>> > 4) Call configure in the main OpenSG folder
>> > 5) Call make in the Builds/i686-...
>> > 6) Call make in the NewFolderLib
>> > 7) Call make in the NewFolderTest
>> > 8) Call make install in the main OpenSG folder
>> >
>> > and my test program succesfully works in /Builds/i686-pc-linux-gnu-
>> > g++/NewFolderTest
>>
>> Sounds good.
>>
>> > But when I tried to compile it in the Tutorials folder (of course I put
>> > #include<OpenSG/> for all includes) I got the following error message:
>> >
>> > /tmp/ccz7j988.o(.gnu.linkonce.t._ZN3osg21LafortuneMaterialBase12getClassT
>> >ypeEv+0x4)
>> >
>> > : In function `osg::LafortuneMaterialBase::getClassType()':
>> >
>> > /HPS/fopra/polina/Soft/OpenSG/Tutorials/testLafortune.cpp:83: undefined
>> > reference to `osg::LafortuneMaterialBase::_type'
>> > /tmp/ccz7j988.o(.gnu.linkonce.t._ZN3osg5FCPtrINS0_INS0_INS_22AttachmentCo
>> >ntainerPtrENS_8MaterialEEENS_13ChunkMaterialEEENS_17LafortuneMaterialEE5dc
>> >astINS_17FieldContainerPtrEEES7_T_+0x4e)
>> >
>> > : In function
>> >
>> > `osg::FCPtr<osg::FCPtr<osg::FCPtr<osg::AttachmentContainerPtr,
>> > osg::Material>, osg::ChunkMaterial>, osg::LafortuneMaterial>
>> > osg::FCPtr<osg::FCPtr<osg::FCPtr<osg::AttachmentContainerPtr,
>> > osg::Material>, osg::ChunkMaterial>,
>> > osg::LafortuneMaterial>::dcast<osg::FieldContainerPtr>(osg::FieldContaine
>> >rPtr)
>> >
>> >
>> > ':
>> > /HPS/fopra/polina/Soft/OpenSG/Tutorials/testLafortune.cpp:89: undefined
>> > reference to `typeinfo for osg::LafortuneMaterial'
>> > collect2: ld returned 1 exit status
>> > make: *** [testLafortune] Error 1
>> >
>> > Could somebody say me what is the error and how to remove it? (I already
>> > tried to reinstall OpenSG and create Material several times, but error is
>> > still there.)
>>
>> osg-config is currently not flexible enough to support additional libraries
>> automatically. So if you add a library
>> you will have to add it to the link line explicitly, otherwise the compiler
>> will find the headers, because they were installed,
>> but it will not link the lib, resulting in the errors you describe.
>>
>> So add -lOSGNewFolder (or whatever the name is) to the LDFLAGS variable in
>> the Makefile and things should work.
>>
>> Sorry for the confusion
>>
>> Dirk
>
>
>--__--__--
>
>Message: 2
>Date: Wed, 10 Dec 2003 13:22:23 +0100
>From: Martin Schmidt <[EMAIL PROTECTED]>
>To: [email protected]
>Subject: Re: [Opensg-users] Memory leak in DrawAction ?
>Reply-To: [email protected]
>
>Hi Dirk,
>
>Dirk Reiners wrote:
>> [...]
>> I wouldn't say strictly necessary, but if you don't have a good reason
>> to not use
>> the RenderAction I would strongly suggest using it. In all but the simplest
>> cases it will give you a nice speedup, and it is significantly better
>> tested than
>> the DrawAction.
>
>Ok, we should try to switch our code to the RenderAction then.
>
>We are doing stereo with the app and want to use our own calculations
>about where to look at, how left eye and right eye differ and so on (in
>short: calc the model and projection matrices). I was said that the main
>reason for using the DrawAction instead of the RenderAction is, that the
>RenderActions "does too much" and that it is or might be hard to make it
>use _our_ model and projection matrices. I think we have to check if
>this claim is really valid and have a try with the RenderAction ;).
>
>>
>>> Is it possible that we make mistakes in our vrml-loader and create
>>> e.g. geometries in such a way, that they look like they should but
>>> rendering them leaves memory leaks ?
>>
>>
>> I can't think of anything that you could do wrong with a model that
>> would result in constant
>> memory leaks.
>ok
>
>> If you're using 1.2.0 or an older dailybuild one source for a rather
>> constant memory leak is the ChangeList, but usually that's less than
>> 100k/sec. Do you clear the ChangeList somewhere, most appropriately in
>> your display loop? If not, try to add a
>>
>> OSG::Thread::getCurrent()->getChangeList()->clearAll();
>>
>> after the draw and see if that helps.
>
>We are doing this now. I forgot to do it on the cluster slave where
>osg::RemoteAspect::receiveSync is called. But this reminds me to add a
>call to osg::ChangeList::setReadWriteDefault() before osginit once we
>switch to a newer OpenSG version ;).
>
>I am pretty sure it is the DrawAction because the memory leaks also
>depend on the size of the scene and do not occur when we do everything
>but the actual DrawAction::apply(). Since the RenderAction is better in
>many ways (e.g. transparency) there is probably no good reason to use
>the DrawAction anyway.
>
>
>>
>> Hope it helps
>>
>> Dirk
>>
>
>Thank you very much,
> Martin
>
>..............................................
>.. Martin Schmidt
>.. [EMAIL PROTECTED]
>.. http://wwwstud.informatik.uni-rostock.de/~masch
>..............................................
>
>
>
>--__--__--
>
>Message: 3
>To: [email protected]
>Subject: Re: [Opensg-users] Memory leak in DrawAction ?
>From: "Dirk Reiners" <[EMAIL PROTECTED]>
>Organization: Fraunhofer IGD
>Date: Wed, 10 Dec 2003 14:26:27 +0100
>Reply-To: [email protected]
>
>
> Hi Martin,
>
>On Wed, 10 Dec 2003 13:22:23 +0100, Martin Schmidt <[EMAIL PROTECTED]
>rostock.de> wrote:
>>
>> Ok, we should try to switch our code to the RenderAction then.
>>
>> We are doing stereo with the app and want to use our own calculations
>> about where to look at, how left eye and right eye differ and so on (in
>> short: calc the model and projection matrices). I was said that the main
>> reason for using the DrawAction instead of the RenderAction is, that the
>> RenderActions "does too much" and that it is or might be hard to make it
>> use _our_ model and projection matrices. I think we have to check if this
>> claim is really valid and have a try with the RenderAction ;).
>
>If you have your own method to calculate the stereo matrices you might want
>to take a look at the ShearedStereoDecorator, which is our way of doing
>stereo. The
>nice thing about using a Decorator is that you can use the standard mono
>interaction code
>with no change whatsoever.
>
>If that looks too complicated you can use the MatrixCamera, which lets you
>define the matrices directly, with no intervention from OpenSG. We use that
>to render some AR
>szenarios where the matrices are generated by a camera tracker, so we know
>what the
>problem might be. ;)
>
>> We are doing this now. I forgot to do it on the cluster slave where
>> osg::RemoteAspect::receiveSync is called. But this reminds me to add a
>> call to osg::ChangeList::setReadWriteDefault() before osginit once we
>> switch to a newer OpenSG version ;).
>
>Yup. ;)
>
>> I am pretty sure it is the DrawAction because the memory leaks also
>> depend on the size of the scene and do not occur when we do everything
>> but the actual DrawAction::apply(). Since the RenderAction is better in
>> many ways (e.g. transparency) there is probably no good reason to use the
>> DrawAction anyway.
>
>I would think so, too. The MatrixCamera should let you do whatever you
>want,
>so I hope there's no reason to stick to the DrawAction. Is anybody else
>still
>using the DrawAction? If yes, why? I was planning to retire it pretty soon
>as
>I didn't expect it to be used, but this makes me a litlle cautious...
>
>Thanks
>
> Dirk
>
>--
>Dirk Reiners
>
>
>--__--__--
>
>Message: 4
>From: "zjj" <[EMAIL PROTECTED]>
>To: [email protected] <[email protected]>
>Date: Wed, 10 Dec 2003 21:46:30 +0800
>Subject: [Opensg-users] OpenSG Compile problem in VC7.1
>Reply-To: [email protected]
>
>Dear users,
>
> I just download the OpenSG ms_stl version and run the example=
> exe correctly, but when I tried to compile the hello example in=
> VC7.1, I encounter following errors:
>
>
>//***************************************************************=
>************************************
>c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.h(344):=
> error C2146: syntax error : missing ';' before identifier=
> '_IRet'
>c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.h(344):=
> error C2501: 'osg::osgIF<IConditionV,IThenT,IElseT>::_IRet' :=
> missing storage-class or type specifiers
>c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.inl(78):=
> error C2785: 'TypeTraits<FloatTypeT>::RealReturnType=
> osg::osgsqrt(const FloatTypeT)' and 'osg::Real32=
> osg::osgsqrt(const osg::Real32)' have different return types
>c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.inl(78):=
> error C2912: explicit specialization; 'osg::Real32=
> osg::osgsqrt<osg::Real32>(const osg::Real32)' is not a=
> specialization of a function template
>c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.inl(72):=
> fatal error C1903: unable to recover from previous error(s);=
> stopping compilation
>c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.h(344):=
> warning C4346:=
> 'osg::osgIFGen<IConditionV>::osgIFSwitch<T2,T3>::_IRet' :=
> dependent name is not a type
>//***************************************************************=
>************************************
>
>Can somebody help me?
>
>Thanks
>
>
>
>
>
>
>=09
>
>=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1zjj
>[EMAIL PROTECTED]
>=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A1=A12003-12-10
>
>
>
>
>
>--__--__--
>
>Message: 5
>To: [email protected]
>Subject: Re: [Opensg-users] OpenSG Compile problem in VC7.1
>From: "Dirk Reiners" <[EMAIL PROTECTED]>
>Organization: Fraunhofer IGD
>Date: Wed, 10 Dec 2003 14:50:49 +0100
>Reply-To: [email protected]
>
>
> Hi zjj,
>
>On Wed, 10 Dec 2003 21:46:30 +0800, zjj <[EMAIL PROTECTED]> wrote:
>
>> Dear users,
>>
>> I just download the OpenSG ms_stl version and run the example exe
>> correctly, but when I tried to compile the hello example in VC7.1, I
>> encounter following errors:
>>
>>
>> //***************************************************************************************************
>>
>>
>>
>> c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.h(344): error
>> C2146: syntax error : missing ';' before identifier '_IRet'
>> c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.h(344): error
>> C2501: 'osg::osgIF<IConditionV,IThenT,IElseT>::_IRet' : missing storage-
>> class or type specifiers
>> c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.inl(78): error
>> C2785: 'TypeTraits<FloatTypeT>::RealReturnType osg::osgsqrt(const
>> FloatTypeT)' and 'osg::Real32 osg::osgsqrt(const osg::Real32)' have
>> different return types
>> c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.inl(78): error
>> C2912: explicit specialization; 'osg::Real32
>> osg::osgsqrt<osg::Real32>(const osg::Real32)' is not a specialization of
>> a function template
>> c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.inl(72): fatal
>> error C1903: unable to recover from previous error(s); stopping
>> compilation
>> c:\Program Files\OpenSG\include\OpenSG\OSGBaseFunctions.h(344): warning
>> C4346: 'osg::osgIFGen<IConditionV>::osgIFSwitch<T2,T3>::_IRet' :
>> dependent name is not a type
>> //***************************************************************************************************
>>
>> Can somebody help me?
>
>the 1.2 version doesn't support VC 7.1. You can try to use a dailybuild
>snaphot installer from
>http://www.opensg.org/dailybuild_logs .
>
>Hope it helps
>
> Dirk
>
>--
>Dirk Reiners
>
>
>--__--__--
>
>Message: 6
>Date: Wed, 10 Dec 2003 16:45:12 -0500
>From: Jack Gundrum <[EMAIL PROTECTED]>
>To: [email protected]
>Subject: [Opensg-users] clusterserver accross subnets
>Reply-To: [email protected]
>
>We are attempting to use ClusterServer to establish a client/server
>connection between differnet networks.
>Using the client/server tutorial examples we changed the ClusterServer
>to use a IP address and port on the
>machine which runs the client.
>
>The client was changed to use the broadcast address,
>setServiceAddress(BA), of the machine running the server.
>I know very little about networking and was wondering if this is even
>possible? It didnt work.
>
>--
>Jack S.Gundrum
>Research Programmer
>Graduate Education and Research Services
>Academic Services and Emerging Technologies
>Information Technology Services
>Penn State University
>
>
>
>
>--__--__--
>
>Message: 7
>Date: Wed, 10 Dec 2003 23:00:36 +0100 (CET)
>From: Akos Balazs <[EMAIL PROTECTED]>
>To: [email protected]
>Subject: Re: [Opensg-users] clusterserver accross subnets
>Reply-To: [email protected]
>
>Hi Jack,
>
>On Wed, 10 Dec 2003, Jack Gundrum wrote:
>
>> We are attempting to use ClusterServer to establish a client/server
>> connection between differnet networks. Using the client/server tutorial
>> examples we changed the ClusterServer to use a IP address and port on
>> the machine which runs the client.
>>
>> The client was changed to use the broadcast address,
>> setServiceAddress(BA), of the machine running the server. I know very
>> little about networking and was wondering if this is even possible? It
>> didnt work.
>
>I'm not networkguru either, but a few tips:
>
>- your routers have to route multicast packets (this is usually not the
>case for security reasons), plus they have to route them to the correct
>subnets.
>
>- if you have more than one network interface in your client machine, make
>sure you route the multicast packets to the appropriate router (or
>subnet). I've run into this problem in the past. :)
>
>- last but not least I really don't know how secure is routing your OpenSG
>connections over LANs (if you do that), I don't think the protocol was
>designed with security above everything, as it's meant to be used in
>separated cluster systems. And AFAIK the client has to find the servers
>via multicast (at least that's what I _think_) even if you use streamsock
>for the actual data transmission, and I'm not sure how secure that is over
>untrusted networks.
>
>Hope this helps,
> Akos
>
>
>
>
>--__--__--
>
>_______________________________________________
>Opensg-users mailing list
>[email protected]
>https://lists.sourceforge.net/lists/listinfo/opensg-users
>
>
>End of Opensg-users Digest
= = = = = = = = = = = = = = = = = = = =
���������������������������
绀硷��
������������������������zjj
[EMAIL PROTECTED]
������������������������������2003-12-11
-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive? Does it
help you create better code? SHARE THE LOVE, and help us help
YOU! Click Here: http://sourceforge.net/donate/
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
