Really fix fft2 regression. #9740
authorDavid Feuer <David.Feuer@gmail.com>
Wed, 29 Oct 2014 14:47:57 +0000 (15:47 +0100)
committerJoachim Breitner <mail@joachim-breitner.de>
Wed, 29 Oct 2014 14:48:04 +0000 (15:48 +0100)
commit64d0a198be05c7baff36e43ab96928a402f00a19
tree5a9488cef1f4ce7d25200cf7f90d920b21bfc510
parentf1090855d9b8d33b3194364dcca0683d89049f03
Really fix fft2 regression. #9740

Rewrite `take` more aggressively for fusion. Add some more explicit
strictness to `unsafeTake` and `unsafeDrop` that seems to help code size and
allocation just a drop in some nofib tests. They were not previously
strict in their numerical arguments, but always called in contexts where
those had been forced; it didn't make a difference in simple test cases,
but made a small difference for nofib. See #9740.

Differential Revision: https://phabricator.haskell.org/D394
libraries/base/GHC/List.lhs