Hanging pointer

fabianpk · January 9, 2006, 07:41

Hello,
I have a problem reconstructing parallel cases, and I do not use any cyclic boundaries. I have a wedge mesh, with wall, wedge, symmetryPlane and patch as boundaries. When I try to reconstructPar I get:

Reconstructing volScalarFields

rho

--> FOAM FATAL ERROR : hanging pointer, cannot dereference

From function PtrList::operator[] const
in file /home/dm2/henry/OpenFOAM/OpenFOAM-1.2/src/OpenFOAM/lnInclude/PtrListI.H at line 76.

This just turned up one day as I was reconstructing a mesh. I use uncompressed, binary, but I don't think this matters because the error even turns up for the 0 time directory.

/Fabian

eugene · January 9, 2006, 08:58

Wedges and cyclics are quite similar to each other. I imagine the bug fix posted in the bug forum will take care of your problem.

fabianpk · January 10, 2006, 02:01

Unfortunately it did not work, the only difference is taht the reference to PtrListI.H refers to the one in my home directory and not the one in /home/dm2/henry

/Fabian

mattijs · January 10, 2006, 04:24

Can you report it as a bug and add a simple testcase?

fabianpk · January 10, 2006, 05:38

Yes, here's a wedge if I decompose this using decomposePar, and then try to reconstruct it, I get the error. So I don't have to create some new data to do it.

http://www.tfd.chalmers.se/~f98faka/...ngPoint.tar.gz

/Fabian

hjasak · January 10, 2006, 05:57

Reproduced it.

You've got a patch called axis with zero faces in it in the original mesh and the code does not handle it. For my taste, you should get rid of the zero-sized patch because you don't need it.

Hrv

#4 0x4053b75c in Foam::error::abort (this=0x40b31720) at error.C:224
#5 0x0806dbfe in Foam::operator<<> (os=@0x8105650, m={fPtr_ = {__pfn = 0x4053b4ca <foam::error::abort()>, __delta = 0}, err_ = @0x40b31720}) at errorManip.H:85
#6 0x08083d32 in Foam::PtrList<foam::fvpatchfield<double> >::operator[] (this=0xbf8dff70, i=5) at PtrListI.H:76
#7 0x0808d631 in GeometricBoundaryField (this=0x9510f64, bmesh=@0xbf8e0a94, field=@0x9510f10, ptfl=@0xbf8dff70) at GeometricBoundaryField.C:200
#8 0x0809ef47 in GeometricField (this=0x9510ed8, io=@0xbf8dfff8, mesh=@0xbf8e01f8, ds=@0x950fc60, iField=@0xbf8dff7c, ptfl=@0xbf8dff70) at GeometricField.C:277
#9 0x0809fbb6 in Foam::geometricFvFieldReconstructor::reconstructFv VolumeField<double> (this=0xbf8e0f14, fieldIoObject=@0x950f648)
at geometricFvFieldReconstructorReconstructFields.C:2 24
#10 0x0809fe6d in Foam::geometricFvFieldReconstructor::reconstructFv VolumeFields<double> (this=0xbf8e0f14, objects=@0xbf8e0ec0)
at geometricFvFieldReconstructorReconstructFields.C:4 64
#11 0x0807936a in main (argc=3, argv=0xbf8e14b4) at reconstructPar.C:241

fabianpk · January 10, 2006, 07:10

You mean that the axis patch of a wedge should not be set and just left empty?

/Fabian

hjasak · January 10, 2006, 07:32

Yes - I've deleted the axis patch from the mesh before the decomposition and it works.

Hrv

sandy · January 28, 2011, 10:29

Hi eugune, in my case I use the cyclic boundaries. Now I meet the same problem "... hanging pointer, cannot dereference ...", you think, what should I do? Please help out.

January 9, 2006, 08:58	Wedges and cyclics are quite s	#2
eugene Senior Member Eugene de Villiers Join Date: Mar 2009 Posts: 725 Rep Power: 21	Wedges and cyclics are quite similar to each other. I imagine the bug fix posted in the bug forum will take care of your problem.

January 10, 2006, 02:01	Unfortunately it did not work,	#3
fabianpk Member Fabian Peng Karrholm Join Date: Mar 2009 Posts: 61 Rep Power: 17	Unfortunately it did not work, the only difference is taht the reference to PtrListI.H refers to the one in my home directory and not the one in /home/dm2/henry /Fabian

January 10, 2006, 05:57	Reproduced it. You've got a	#6
hjasak Senior Member Hrvoje Jasak Join Date: Mar 2009 Location: London, England Posts: 1,905 Rep Power: 33	Reproduced it. You've got a patch called axis with zero faces in it in the original mesh and the code does not handle it. For my taste, you should get rid of the zero-sized patch because you don't need it. Hrv #4 0x4053b75c in Foam::error::abort (this=0x40b31720) at error.C:224 #5 0x0806dbfe in Foam::operator<<> (os=@0x8105650, m={fPtr_ = {__pfn = 0x4053b4ca <foam::error::abort()>, __delta = 0}, err_ = @0x40b31720}) at errorManip.H:85 #6 0x08083d32 in Foam::PtrList<foam::fvpatchfield<double> >::operator[] (this=0xbf8dff70, i=5) at PtrListI.H:76 #7 0x0808d631 in GeometricBoundaryField (this=0x9510f64, bmesh=@0xbf8e0a94, field=@0x9510f10, ptfl=@0xbf8dff70) at GeometricBoundaryField.C:200 #8 0x0809ef47 in GeometricField (this=0x9510ed8, io=@0xbf8dfff8, mesh=@0xbf8e01f8, ds=@0x950fc60, iField=@0xbf8dff7c, ptfl=@0xbf8dff70) at GeometricField.C:277 #9 0x0809fbb6 in Foam::geometricFvFieldReconstructor::reconstructFv VolumeField<double> (this=0xbf8e0f14, fieldIoObject=@0x950f648) at geometricFvFieldReconstructorReconstructFields.C:2 24 #10 0x0809fe6d in Foam::geometricFvFieldReconstructor::reconstructFv VolumeFields<double> (this=0xbf8e0f14, objects=@0xbf8e0ec0) at geometricFvFieldReconstructorReconstructFields.C:4 64 #11 0x0807936a in main (argc=3, argv=0xbf8e14b4) at reconstructPar.C:241 __________________ Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk

January 10, 2006, 07:32	Yes - I've deleted the axis pa	#8
hjasak Senior Member Hrvoje Jasak Join Date: Mar 2009 Location: London, England Posts: 1,905 Rep Power: 33	Yes - I've deleted the axis patch from the mesh before the decomposition and it works. Hrv __________________ Hrvoje Jasak Providing commercial FOAM/OpenFOAM and CFD Consulting: http://wikki.co.uk

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
FATAL ERROR hanging pointer cannot dereference	tsencic	OpenFOAM Running, Solving & CFD	8	January 7, 2021 03:10
Hanging Elements	Dipl.-Ing. Jayadi Lukito	CFX	0	January 29, 2008 03:18
postprocessing keeps hanging	leo	FLUENT	1	October 22, 2007 09:16
Accessing a pointer away from current pointer	shephali shrimali	FLUENT	0	May 10, 2007 07:54
hanging fan	Mavinakere.	FLUENT	3	August 3, 2003 22:34

January 9, 2006, 07:41	Hello, I have a problem reco	#1
fabianpk Member Fabian Peng Karrholm Join Date: Mar 2009 Posts: 61 Rep Power: 17	Hello, I have a problem reconstructing parallel cases, and I do not use any cyclic boundaries. I have a wedge mesh, with wall, wedge, symmetryPlane and patch as boundaries. When I try to reconstructPar I get: Reconstructing volScalarFields rho --> FOAM FATAL ERROR : hanging pointer, cannot dereference From function PtrList::operator[] const in file /home/dm2/henry/OpenFOAM/OpenFOAM-1.2/src/OpenFOAM/lnInclude/PtrListI.H at line 76. This just turned up one day as I was reconstructing a mesh. I use uncompressed, binary, but I don't think this matters because the error even turns up for the 0 time directory. /Fabian

January 10, 2006, 04:24	Can you report it as a bug and	#4
mattijs Senior Member Mattijs Janssens Join Date: Mar 2009 Posts: 1,419 Rep Power: 26	Can you report it as a bug and add a simple testcase?

January 10, 2006, 05:38	Yes, here's a wedge if I decom	#5
fabianpk Member Fabian Peng Karrholm Join Date: Mar 2009 Posts: 61 Rep Power: 17	Yes, here's a wedge if I decompose this using decomposePar, and then try to reconstruct it, I get the error. So I don't have to create some new data to do it. http://www.tfd.chalmers.se/~f98faka/...ngPoint.tar.gz /Fabian

January 10, 2006, 07:10	You mean that the axis patch o	#7
fabianpk Member Fabian Peng Karrholm Join Date: Mar 2009 Posts: 61 Rep Power: 17	You mean that the axis patch of a wedge should not be set and just left empty? /Fabian

January 28, 2011, 10:29		#9
sandy Senior Member Sandy Lee Join Date: Mar 2009 Posts: 213 Rep Power: 18	Hi eugune, in my case I use the cyclic boundaries. Now I meet the same problem "... hanging pointer, cannot dereference ...", you think, what should I do? Please help out.