Tim Bernhard

The issue seems to have been the directness of the graph and its consequences on the adjacency list. My tests pass, and I would consider them complex enough to assume...

I cannot reproduce the "Non-matching number of IGRAPH_FINALLY and IGRAPH_FINALLY_CLEAN." locally. Any tips, dear @szhorvat or @ntamas ?

Thank you for these useful insights and pointing to the right CMake variable.

It might be that I misunderstand something, but isn't `igraph_fundamental_cycles()` merged with #1957 exactly this feature (at least, if called as often as many cycles are searched)?

I see, then I misunderstood your code, thanks for the response. I therefore implemented a somewhat naïve approach to the problem in C++ and attached the resulting class. If you...

Thank you for this comprehensive answer and the hints and tips! > - Which algorithm does this implement? I have not put any research into algorithms as I needed this...

Thank you for your patience. You might be right, my understanding of a simple cylce also includes that it cannot be broken down into smaller cycles, as shown in the...

Thank you very much for these useful insights and examples. I will go back to the drawing board for a more generally interesting implementation then, starting with the literature. I...

The naïve way to generalize my approach to all simple cycles is by replacing the step "find shortest path from the neighbour back to junction (without the direct edge)" with...

I would conclude my literature study confirming that Johnson's algorithm is indeed the way to go. I will refer to https://github.com/igraph/igraph/blob/master/CONTRIBUTING.md and https://github.com/igraph/igraph/wiki/Tips-on-writing-igraph-code for a first overview of what to...