Ross Barnowski

Gentle ping @kazimierz-256 . Is this issue still valid? There have been some developments on the negative cycle detection front that may be relevant here: see #5089

Thanks for bringing this up @JohnyXia001 . If I'm understanding correctly, the issue is that the current implementation in NetworkX has O(N^2) complexity whereas [references for Kosaraju's algorithm](https://en.wikipedia.org/wiki/Kosaraju's_algorithm#Complexity) note that...

This one has been inactive for a while - the best way to push this forward IMO would be to e.g. try the above proposed update. Would you be interested...

From the discussion here this seems like it'll be a tricky one to get the wording right - any chance you'd be able to wrap up a suggested phrasing in...

Thanks @charmoniumQ - just a quick note about logistics: > More useful type stubs will be coming soon. I think it would be best to have a single PR (i.e....

> did you check what happens if you change the actual networkx .py files in a way that invalidates the .pyi files (e.g. by removing an argument). Does mypy notice...

Given the positive feedback on the changes that had to be made for consistency (thanks @dschult !) I went ahead and did the replacement. Just to recap: this replaces the...

This one's been stale for a while, so I will close it. If the issue persists please feel free to reopen with the latest info.

I took another look at the implementation and the first thing that jumped out as a potential point of variability is the use of `scipy.optimize.linprog` in the `held_karp_ascent` function. By...

Thank you @fmagin for your nice summary, I think your take on the situation largely aligns with my own. I'll go ahead and restate my stance here: I'm personally +1...