Ted

Apologies if this sounds naive but in the meantime (while someone constructs an optimal algorithm), can we just treat the list of edges as an array and use something simple...

Hmmm I was even more naive than I thought. Binary search requires order preservation between the elements of the array and the natural numbers. This is doable with ordered pairs...

Could we perhaps implement edge removal in a way similar to `graph-tool`? This is the approach (referenced [here](http://graph-tool.skewed.de/static/doc/quickstart.html)): Removing a vertex is an O(N) operation. The vertices are internally stored...

This may explain why the build is failing: ``` julia julia> g = simple_graph(5, is_directed=false) Undirected Graph (5 vertices, 0 edges) julia> add_edge!(g, 1, 2) edge [1]: 1 -- 2...

No worries. I need a break from this too. :) For future reference: If `e in edges(g) == true`, shouldn't it follow that `revedge(e) in edges(g) == true` for undirected...

@pozorvlak Any headway regarding edge indexing?

@lindahua From the user's perspective, the ability to remove or exclude edges/vertices is an essential part of graph construction. It would be strange for the default graph object to not...

@pozorvlak Thanks for accommodating me (I'm sure it's evident I'm new to all this). Is there an expedient way to do tests in Julia locally without relying on Travis? Right...

I'll edit this PR later to just remove incident edges when running `remove_vertex!`. Also, I added `remove_edge!` and `remove_vertex!` in `src/Graphs.jl` but I'm getting the same output. Do I need...

Same here. Just getting blank tweets.