Szabolcs Horvát

> Letting square lattice labels stay as "row,col" would be useful. That's not how it is now, but that's exactly how it will be in the future.

Originally, I did not notice `CompleteGraphQ` ... `IGCompleteQ` differs in that it ignores multi-edges and self-loops. ``` In[25]:= glist = {Graph[{1 2}], Graph[{1 2, 1 1}], Graph[{1 2, 1 2}]};...

@jplauri Is connectedness usually required for block graphs? The current `IGCactusQ` implementation does require it.

Wikipedia and MathWorld include connectedness in their cactus definitions. Wikipedia also refers to cactus graphs as *Husimi trees*, where "tree" suggests connectedness. It mentions that "Husimi tree" is sometimes used...

These are all good suggestions. Thanks! I'll also make a few changes to make it easier to get started with development, and will send you an email during next week.

I think this result is correct. Can you explain more why you think it is not? Deleting these vertex sets disconnects the graph. They are indeed the smallest possible sets...

Milestone removed as there is a workaround for this in Mma code. Will keep the bug open until the issue is fixed in the C code upstream.

It'll be a while before I can find the time to look at this (Windows is not my main system), but this is very likely because of a bad compiler...

> and then build a project via CMake I don't think that will make a difference here. The issue is still to use the correct command line flags, linking the...

I can't comment much on this today (deadlines...) but please try using the MinGW driver, not the GenericCompiler driver. I know that's not what the documentation says, but the documentation...