Tamás Nepusz

> What could be done is to switch to completely different method of generating the documentation, avoid DocBook Note that we have recently accepted a PR in the C core...

So it looks like the hourglass algorithm only brings some (not significant) improvement compared to the tree algorithm if the graph is relatively large (i.e. vcount=100 is not enough to)...

No, we don't have an API for that yet. I was thinking that we could solve this with `permute_vertices()` and `permute_edges()` but apparently these currently assume that the target graph...

> I believe you are right. Yes, you are correct. This seems to be a bug indeed. Great job!

Strictly speaking, the input is not even valid GraphML so we should have rejected it in the first place. I believe that there are two bugs to be fixed here:...

Duplicate attribute detection was broken in the GraphML reader; this is now fixed. There's still one bug: `attr.type` is prefix-matched so `attr.type="s"` matches `"string"` and is thus treated as a...

Fixed both in `master` and in `develop`. It was a bug in the trie data structure that allowed an insertion of an empty string but then reported the empty string...

Commit 57cd9eb07 works for me on my machine, both with the minimized and the unminimized test case. Can you post the test case here that fails for you?

Okay, so strictly speaking this should be a parse error as well, but it's harder to detect than the previous cases. Right now we have a trie that maps from...

Let's try to fix the issue of the default argument only for the time being. Would it suffice if we changed the behaviour in 1.0 such that any parametrization that...