CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

Mohamed Yousef wrote:
> space isn't a problem
> but the problem is that , with deletion of faces/edges/vertices , the 
> two vectors will be immediately invalidated
> and you will have to reconstruct them again from beginning
>
> what is more problematic is that , even if you decided to monitor all 
> deletion/addition and try to make partial reconstruction of
> vectors , you will have to mimic and analyze all inner/low level half 
> edge operations to update your vectors accordingly (may be by a custom 
> half-edge that does this extra work for every operation )
>
> Regards,
> Mohamed Yousef
Maybe I was not clear.
You can use the list based HDS to have support for removal
and in parallel maintain your vector of handles.
Then I might not get your problem but from what I understand, I see no 
"inner/low level half-edge operations" except maintaining the vector of 
handles.

If I am wrong, can you please state what you expect to be done using the
polyhedron class (your requirements).


S.

> On 27 April 2010 15:25, Sebastien Loriot (GeometryFactory) <sloriot.ml 
> <http://sloriot.ml>@gmail.com <http://gmail.com>> wrote:
>
>     Mohamed Yousef wrote:
>
>         The real problem with solutions like this , is that it's built
>         by hand
>         and every time you update P , you will have to go through the
>         very lengthy process of updating id's
>
>         the only real solution is to use Vector based Polyhedron's
>         instead of List ones , but those are limited and more importantly
>         doesn't support deleting nodes
>
>         this problem is a real weak point of cgal , facing applications
>         that require regular , random update of Polyhedron and need
>         random access / indices
>
>     If space is not a problem, maybe you can use an additional vector of
>     handles together with the polyhedron so that you can have random access
>     as well as direct access. Maintaining in parallel the vector should not
>     be too expensive (depending on your application).
>
>     S.
>
>
>
>
>         Regards,
>         Mohamed Yousef
>
>
>         On 27 April 2010 14:45, 
> <
>  
> <mailto:>
>         
> <mailto:
>  
> <mailto:>>>
>  wrote:
>
>            Thank you.
>
>            Le 27/04 - 14:02, Sebastien Loriot (GeometryFactory) a écrit :
>             > 
>
>  
> <mailto:>
>  
> <mailto:
>         
> <mailto:>>
>  wrote:
>             > >It's what I want.
>             > >But it seems that by default the vertex class has no
>         member id.
>             > >Is there an option ?
>             >
>             > Have a look at this item class
>             > CGAL/Polyhedron_items_with_id_3.h
>             >
>             > Facets, halfedges and vertices will have an id.
>             >
>             > S.
>             >
>             > >
>             > >Le 27/04 - 11:49, Sebastien Loriot (GeometryFactory) a
>         écrit :
>             > 
> >>
>  
> <mailto:>
>         
> <mailto:
>  
> <mailto:>>
>  wrote:
>             > >>>I have a polyhedron P.
>             > >>>And I want to associate to each face an index : the
>         distance from
>             > >>>P->facets_begin() to the iterator pointing to the face (for
>            example).
>             > >>>
>             > >>>Now, given a halfedge h, I would like to know the number of
>            h.facets().
>             > >>>I tried std::distance(P.facets_begin(), h.facets()) but it
>            doesn't work
>             > >>>since one is an iterator and the other a handle.
>             > >>>
>             > >>>Is there a simple way to do that ?
>             > >>>
>             > >>Is something like this a correct solution for your problem?
>             > >>
>             > >>unsigned i=0
>             > >>for (Polyhedron::Facet_iterator
>             > >>it=P.facets_begin();it!=P.facets_end();++it){
>             > >>  it->id()=i++;
>             > >>}
>             > >>
>             > >>S.
>             > >
>             >
>             >
>             > --
>             > You are currently subscribed to cgal-discuss.
>             > To unsubscribe or access the archives, go to
>             > https://lists-sop.inria.fr/wws/info/cgal-discuss
>
>            --
>            You are currently subscribed to cgal-discuss.
>            To unsubscribe or access the archives, go to
>            https://lists-sop.inria.fr/wws/info/cgal-discuss
>
>
>
>
>     -- 
>     You are currently subscribed to cgal-discuss.
>     To unsubscribe or access the archives, go to
>     https://lists-sop.inria.fr/wws/info/cgal-discuss