Revisiting Haskell Network.HTTP

1 · Nelson Elhage · March 20, 2015, 7:03 p.m.
I spent a while talking with Greg Price about the Haskell Network.HTTP issue previously featured here, and he was unconvinced I had the whole story. He spent a while reading some source and thinking about folds, and pointed out today that my analysis was incomplete, and flat-out wrong for the cabal case referenced in the commit. One detail I glossed over in the last post (deliberately for simplicity, but over-zealously!) was the bufOps variable. bufOps is an instance of BufferOps, which provides...