🚀 go-pugleaf

In article <449i92$ae6@nz12.rz.uni-karlsruhe.de>
ua0y@rzstud1.rz.uni-karlsruhe.de (Frank Muzzulini) writes:

[...] 
>> Bulent Basaran (basaran@prelude.ece.cmu.edu) wrote:
[...] 
>> > 2) Is it possible to derive from a LEDA type?
>> > e.g., 
>> > class C1: public d_array {
>> > ...
>> > }
>> 
>> It is at least possible.  A friend of mine tried to code something
>> like that, but he had some problems in manipulating the object he
>> derived from. Just try it.

C++ permits it, of course, but be very careful. Many LEDA classes have
been written without thinking about possible derivation.

For instance, the `graph' class is the base class for `GRAPH',
`ugraph' and `planar_map'. But a `ugraph' is a graph with its
`undirected' field set to true, thus the behaviour of the undirected
graph is (partly) implemented in the `graph' class. Thus, if you
derived from `graph', in order not to interfere with `ugraph', you
need to know that `graph' was written with `ugraph' in mind... Tricky.

It is one of the coding practice I regret in LEDA. I suppose it was
hardly avoidable.

>>   ___  Frank Muzzulini     ua0y@rz.uni-karlsruhe.de     Muzz@irc

						Olivier L'Heureux

----------------------------------------------------------------------

Laboratoire de microélectronique de l'Université catholique de Louvain

e-mail:	lheureux@dice.ucl.ac.be	               B-1348 Louvain-la-Neuve         
<URL:http://www.dice.ucl.ac.be/~lheureux>                      Belgium
MIME/PGP accepted