Ondrej Kokes

There is a fix in https://github.com/kokes/nbviewer.js/commit/138eca77fdf558e4e067dc3fd5d0de373a0330e5, but I'd like to see if we've fixed this completely. Plus I'd like to know if we need to treat the URL any differently...

Can you benchmark the imports alone? I presume much of your benchmark is spent import pandas, which is quite a heavy dependency (takes 0.7s, warm, on my machine). Plus there's...

`yield from` can be replaced by ```python data = DBF(...) for el in data: yield el ```

> Any more ideas of how we could get this closer to 74 seconds? It comes down to what the limits are. I tried running the three obvious pieces of...

> In the loop you use to save the CSV, you are instantiating a new DataFrame object at every iteration, this definetly adds up in wasted time. You don't need...

Remember that the index needs to be updated upon each mutation of its parent DataFrame/column.

That's the idea, I explicitly close it from both sides, but upon further investigation, there are other issues at play. I further simplified the code to just this (+ logging)...

But even in the first case (when no data got exchanged) did the peer learn that - when the server closed it. So proper accounting should lead to the fact...

I finally see where the confusion on my part comes from. The `Close()` method on the client side is to close the `io.Writer` as it's from the SendStream interface, **it...

Only looked at it briefly just now, but already ran into issues with smaller `MaxDatagramFrameSize` (protocol violation upon reception of 1200-byte long datagrams). But at least got a partial repro....