[PPL-devel] [Fwd: Re: Question about PPL (missing detail)]

[Roberto Bagnara]

-------- Original Message --------
Subject: Re: Question about PPL (missing detail)
Date: Wed, 08 Oct 2003 19:28:03 +0200
From: Carlos Ureña-Almagro <almagro at ugr.es>
To: Roberto Bagnara <bagnara at cs.unipr.it>
References: <3F83F1DF.9010003 at ugr.es> <3F841488.2020400 at cs.unipr.it>


I understand the problem, I will try to follow your advice.
Thanks a lot for your help.

With respect to my application, at this point I just want to
explore the feasibility of an algortithm for efficiently
compute the first intersection of a line with a huge set
of planar triangles in 3D. This is a common problem in
computer graphics, which has been solved by using an
spatial index (k-d tree or similar) for the triangles.
However, one can map the line and all triangles edges
lines to a 6-dimensional space (using plucker coordinates),
such that the line is mapped to a point, and each triangle
is mapped to the intersection of three half-spaces (one for
each edge). The line intersects the triangle in 3d euclidean
space iif the point is in these three half-spaces in 6d
plucker coordinate space. I want to perform spatial indexing
in this space, thus I need to compute intersections between
6d axis-aligned boxes and those half-spaces obtained from
triangle edges (however, this is just an idea, and PPL is
probably perfect for these purpouses).

Roberto Bagnara wrote:

> Carlos Ureña Almagro wrote:
> 
>>   I'm trying to use PPL library for a project related to Computer
>>   Graphics. First at all, I would like to thank you for your effort
>>   in development and maintenence of that library. I had a trouble
>>   trying to use it, and I would like to know if you have experienced
>>   this and maybe you can point me to the solution. I apologize if
>>   this is a trivial question, or has a trivial solution which I
>>   ignore.
>>
>>   I have suse linux 8.2, with gcc version 3.2, kernel 2.4.19. I
>>   have succesfully installed the binary rpm version of your library.
>>   (I have libstdc++.so.5 in my linux box, rpm missed the corresponding
>>   libstdc++ rpm package, but the file is there, so I ignored the rpm
>>   warning).
>>
>>   After installing, I checked I had the libraries in standard linux lib
>>   folders, and I  wrote a simple C++ program using the PPL. The
>>   program compiled OK, but when I tried to link it, I had an error
>>   message, more concretelly, the linker found references to this
>>   undefined symbol (dots are mine)
>>
>>   standard_alloc_template<...>::_S_force_new
>>
>>   may be the libraries I have are wrong, but every symbol was
>>   correctly resolved except for this, and this is what bothers
>>   me
>>
>>
>>   Any help would be appreciated, thank you very much in advance.
> 
> 
> Dear Carlos,
> 
> the symptom suggests that you are trying to link C++ objects obtained
> with different version of GCC, something that does not work in general.
> Things will improve in the future, but until now the situation has
> been quite discouraging: version 3.0 was not compatible with 3.1
> and the same thing happened in the transition to 3.2 and to 3.3.
> Indeed, even the fact that we distribute RPMs is questionable,
> since they are useless on systems where a different version of GCC
> is installed.  The problem is aggravated by the fact explained
> in our draft FAQ available at http://www.cs.unipr.it/ppl/FAQ.
> The bottom line is that GMP, the PPL and all your C++ code must
> be compiled with the same version of the same compiler (GCC,
> in your case).
> 
> To summarize: you should build everything from sources.
> Start with GMP following the instructions in
> http://www.cs.unipr.it/ppl/Download/requirements
> and then configure, compile and install the PPL.
> Please do not hesitate to contact us again if you need help:
> direct all communication to ppl-devel at cs.unipr.it.
> We would also love to know more about your application.
> All the best,
> 
>      Roberto and the PPL team
> 


-- 
_____________________________________________________________________
Carlos Ureña Almagro (almagro at ugr.es - http://lsi.ugr.es/~curena)
Dpt. L.S.I. - E.T.S. Ingeniería Informática - Universidad de Granada.


-- 
Prof. Roberto Bagnara
Computer Science Group
Department of Mathematics, University of Parma, Italy
http://www.cs.unipr.it/~bagnara/
mailto:bagnara at cs.unipr.it