Re: [jgrapht-users] Bug in removeAllEdges?

[Trevor H. <tr...@vo...>]

On Nov 25, 2006, at 1:24 PM, John V. Sichi wrote:

> Trevor Harmon wrote:
>> The attached program throws a ConcurrentModificationException, and  
>> I don't know why. Even stranger, if outgoingEdgesOf is changed to  
>> edgesOf, the exception is not thrown. Doesn't make sense. Is this  
>> a bug?
>
> Not a bug, unless you consider the program below to indicate a bug  
> in the JDK :)

Let me rephrase my question then. The following JDK program works as  
expected:

List<String> list = new ArrayList<String>();
list.add("arm");
list.add("leg");
List<String> sublist = list.subList(0,1);
sublist.clear();

But the following JGraphT program throws a  
ConcurrentModificationException:

DirectedGraph<String, DefaultEdge> g =
	new DefaultDirectedGraph<String, DefaultEdge>(DefaultEdge.class);
String s1 = "s1";
String s2 = "s2";
g.addVertex(s1);
g.addVertex(s2);
g.addEdge(s1, s2);
g.edgesOf(s1).clear();

Regardless of this problem, how is one supposed to remove all  
outgoing edges from a vertex? That's all I'm trying to do. With the  
large number of methods available in JGraphT, I would have expected  
the API to provide a straightforward way of doing this (without  
having to iterate over the edges myself).

> The outgoingEdgesOf vs edgesOf behavior is just a matter of getting  
> lucky with one.

Well, that part I would still say is a bug. I could understand it if  
outgoingEdgesOf, incomingEdgesOf, and edgesOf all ended up causing  
the ConcurrentModificationException -- at least that would be  
consistent. But that's not the case here. The behavior of an API  
shouldn't be left to chance. :)

Trevor