(*>) and (<*) could be used to apply recognizing parsers for the discarded
half. This makes a huge gain for uu-parsinglib. uu-parsinglib's P_m monad
could be extended just as it has done with P_f and P_h to also wrap its
existing R monad, which would let it apply the parser as a recognizer
efficiently.
And for parsimony it allows me to treat that side of the alternative grammar
as a right seminearring ignoring the argument, this increases sharing
opportunities for my grammar fragments, because pure nodes in recognizers
can be treated as epsilons in the grammar and safely elided.
-Edward Kmett
On Wed, Aug 19, 2009 at 11:40 AM, Ross Paterson
On Mon, Jun 29, 2009 at 01:45:45PM +0100, Ross Paterson wrote:
The proposal is that the following functions
(<$) :: Functor f => a -> f b -> f a (*>) :: Applicative f => f a -> f b -> f b (<*) :: Applicative f => f a -> f b -> f a some :: Alternative f => f a -> f [a] many :: Alternative f => f a -> f [a]
are moved into the corresponding classes, with the existing implementations as default definitions.
Henning asked for concrete examples (i.e. code) where this would give a substantial performance win while still defining the same function, including termination properties. (But it seems reasonable not to rely on RULES to improve the asymptotic complexity.)
I gave an example for (<$). Does anyone have concrete examples for the others? _______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries