GroteGnoom

You don't seem to use the visited vector to check for cycles. Maybe you can add a test case for for a 2-edge cycle, which I think is detected correctly,...

Hmm, all right, then it's definitely not easy to fix the general case, but checking for NULLs or empty strings would at least solve one particular version of the problem.

This is a list of examples that are more than 100 lines after #2175: ``` (smaller_examples)> find . -type f -name "*.c" | xargs wc -l | sort | awk...

No, please work on master, you should be able to work independently from #2175. #2175 is already way too big, and I shouldn't have done that. If you work on...

We've merged #2175. @Jeroenvh99 have you started working on this? Do you need any help, for example in picking a file to work on?

Yes! I've assigned you to it. If you have any questions, please don't hesitate to ask, including about quickly building and testing your code. An initial build should be simple...

I'm not sure I understand your question completely. igraph is written in C, so you can't overload functions, so we'll have to think of a new name. Maybe something like...

We now have ``` igraph_error_t igraph_neighbors(const igraph_t *graph, igraph_vector_int_t *neis, igraph_integer_t pnode, igraph_neimode_t mode) { if (!igraph_is_directed(graph) || mode == IGRAPH_ALL) { return igraph_i_neighbors(graph, neis, pnode, mode, IGRAPH_LOOPS_TWICE, IGRAPH_MULTIPLE); }...

> eigenvector_centrality computes the left eigenvector of the adjacency matrix. In order to have more control over its behaviour, the directed argument should be replaced with a mode argument. So...

> Yes. But I think this is very low priority. It's one of the 0.11 milestones, so I thought this would have at least higher priority than non-milestoned issues.