Re: FoldrW/buildW issues

Hi, Am Mittwoch, den 24.09.2014, 16:37 -0400 schrieb David Feuer:

On Sep 12, 2014 2:35 PM, "Joachim Breitner" wrote:

...

I once experimented with a magic "oneShot :: (a -> b) -> (a -> b)" function, semantically the identity, but tell the compiler not to share the result of the computation. Using that in the definition of foldl-as-foldr, one can get the same effect as Call Arity, but a bit more reliable. I need to investigate if that solves the sumConcatInits problem.

One nice thing about this idea (which sounds like it must be related to the "state hack", but is more explicit) is that it presumably applies also to similar situations in the State and ST monads, when a state transformer is only used once. Could you explain, perhaps, what compiler transformation this enables, and how you implemented it?

I guess it is used in various places, and I woudn’t know all of them. The one that I was aiming for was is this one case x of True -> z False -> \s(one-shot). e to \s(one-shot) . case x of True -> z s False -> e explained in Note [Combining case branches] at https://github.com/ghc/ghc/blob/master/compiler/coreSyn/CoreArity.lhs#L671 The implementation is attached to my previous mail.

It would be nice if the compiler could figure this out for itself,

Well, that’s what I thought, and CallArity is what I came up with :-) Greetings, Joachim -- Joachim “nomeata” Breitner mail@joachim-breitner.de • http://www.joachim-breitner.de/ Jabber: nomeata@joachim-breitner.de • GPG-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org