Hi Janek, yes to both -- in a way. See Section 5.3 here for lists: http://dl.acm.org/citation.cfm?id=2543736 For my usual work, I use stream fusion and manually 'flatten' everything in all of ADPfusion and rather large bunch of other work building on top of that. ;-) Giegerich's original ADP is full or list comprehensions -- every single function uses them, and does not require a lot of additional machinery to run. http://bibiserv.techfak.uni-bielefeld.de/adp/ Note that if you want to introduce deep optimizations, it'll be a larger project. See also Coutts' phd thesis ([2] in our paper), and the original stream fusion paper [3]. Gruss, Christian === [2] D. Coutts. Stream Fusion: Practical Shortcut Fusion for Coinductive Sequence Types. PhD thesis, University of Oxford, 2010. [3] D. Coutts, R. Leshchinskiy, and D. Stewart. Stream fusion: From lists to streams to nothing at all. In Proceedings of the 12th ACM SIGPLAN International Conference on Functional Programming, pages 315– 326, Freiburg, Germany, 2007. ACM. * Jan Stolarek [23.07.2014 13:57]:

Haskellers,

recently I've been looking into the possibility of creating some new optimisations for GHC. These would be mostly aimed at list comprehensions. Here's where I need your help:

1. Do you have complex list comprehensions usage examples from real code? By complex I mean nested list comprehensions, reading from more than one list ([ ...| x <- xs, y <- ys ... ]) etc.

2. Do you have list comprehensions code that you had to optimize by hand because GHC was unable to make them fast enough?

Janek _______________________________________________ Glasgow-haskell-users mailing list Glasgow-haskell-users@haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-users