Space leaks in the real world

Work in progress

• maximumBy in base

  https://gitlab.haskell.org/ghc/ghc/-/issues/10830

• A leak in Shake

  https://neilmitchell.blogspot.com/2013/02/chasing-space-leak-in-shake.html

• Two leaks in Hoogle and one in Shake

  https://neilmitchell.blogspot.com/2015/09/three-space-leaks.html

• A simple space leak in a recursive accumulator

  https://stackoverflow.com/questions/48034553/how-to-avoid-haskell-space-leak

• Two leaks in Shake, one in HashMap

  https://neilmitchell.blogspot.com/2020/05/fixing-space-leaks-in-ghcide.html

• discogs2pg

  I’m not sure I really consider this a space leak. It’s just bad design that getTables takes a list that it doesn’t need. On the other hand it’s probably also bad design that store takes a [a] rather than some sort of stream.

  https://clrnd.com.ar/posts/2015-08-21-profiling-a-haskell-space-leak.html

• Space leak in GHCi

  https://simonmar.github.io/posts/2018-06-20-Finding-fixing-space-leaks.html

• Space leak in nested monadic loops

  Quite different from most space leaks, and really, really bad

  https://ro-che.info/articles/2017-01-10-nested-loop-space-leak

  Discussion on GHC issue

• Space leak that should have been fixed with strict fields

  https://sriramsami.com/haskell-optimization/

Taxonomy

• Edward Yang’s Space Leak Zoo

Problems with unnecessary strictness

• A strictness slowdown in uniplate

  https://neilmitchell.blogspot.com/2013/08/destroying-performance-with-strictness.html