[Paraview] dlopen question

Takuya OSHIMA oshima at eng.niigata-u.ac.jp
Tue Sep 23 23:36:14 EDT 2008


Hi Mattijs,

Not sure if this helps, but interestingly a VisIt guy had a very
similar problem around a week ago. You can follow the discussion by
searching for "help with dlopen" in
https://email.ornl.gov/pipermail/visit-developers/2008-September/thread.html .

The caveat is that he says RTLD_LAZY|RTLD_GLOBAL solved the problem
while apparently not in this case...

Takuya OSHIMA, Ph.D.
Faculty of Engineering, Niigata University
8050 Ikarashi-Ninocho, Nishi-ku, Niigata, 950-2181, JAPAN


From: Mattijs Janssens <m.janssens at opencfd.co.uk>
Subject: [Paraview] dlopen question
Date: Mon, 22 Sep 2008 10:22:49 +0100

> In our code we use dlopen with RTLD_LAZY|RTLD_GLOBAL to optionally load 
> additional libraries.
> 
> We noticed that this gives problems in the paraview reader. The same code 
> which happily loads when running from the command line now gives undefined 
> symbols when used through the reader.
> 
> E.g. the additional library libAdditional uses a symbol aaa from library 
> libmain which is already a dependency of my application. (so 'ldd 
> application' shows up libmain). My application does a dlopen of libAdditional 
> and it all works fine.
> 
> Now when invoked in a reader (the reader itself I assume is loaded using 
> dlopen into Paraview?) the dlopen of libAdditional complains about unresolved 
> symbol aaa. Funnily enough it works if we first do a dlopen of libmain and 
> then of libAdditional.
> 
> Is this related to the flags passed into dlopen? To complicate the matter, 
> from what we heard, the problem does not appear on OSX.
> 
> Regards,
> 
> Mattijs
> 
> -- 
> 
> Mattijs Janssens
> 
> OpenCFD Ltd.
> 9 Albert Road,
> Caversham,
> Reading RG4 7AN.
> Tel: +44 (0)118 9471030
> Email: M.Janssens at OpenCFD.co.uk
> URL: http://www.OpenCFD.co.uk
> _______________________________________________
> ParaView mailing list
> ParaView at paraview.org
> http://www.paraview.org/mailman/listinfo/paraview


More information about the ParaView mailing list