Control.DeepSeq.Bounded

We provide forcing functions which take a non-strict value of a datatype, and force its evaluation in a pricipled way. As with NFData, one should bear in mind the difference between forcing and demand: in order for your forcing to take effect, demand must be placed on the forcing function itself by the course of execution.

Artificial forcing is most commonly used to suppress space leak, but it has many uses besides. Another is to ensure that exceptions hidden within lazy fields of a data structure don't leak outside the scope of the exception handler; another is to force evaluation of a data structure in one thread, before passing it to another thread (preventing work moving to the wrong threads, a form of time leak).

Unlike DeepSeq, potentially infinite values of coinductive data types are supported by principled bounding of deep evaluation.

Refer to comments in the deepseq package for more information on how artificial forcing can be useful.

Recently, control (static or dynamic) of parallelisation has been added. Control of evaluation order is also supported (via pseq).

Forced evaluation to an arbitrary finite depth

module Control.DeepSeq.Bounded.NFDataN

Forced evaluation over a pattern (arbitrary finite, or dynamic)

module Control.DeepSeq.Bounded.Pattern

module Control.DeepSeq.Bounded.Compile

module Control.DeepSeq.Bounded.PatUtil

module Control.DeepSeq.Bounded.NFDataP

module Control.DeepSeq.Bounded.NFDataPDyn

data Rose a

Constructors

Node a [Rose a]

Instances

Functor Rose

FF Pattern

Eq a => Eq (Rose a)

Show a => Show (Rose a)

NFData a => NFData (Rose a)

Typeable (* -> *) Rose

data F Pattern = FPattern v

Forced evaluation "molecular style"

module Control.DeepSeq.Bounded.Seqable

Generic deriving support via Generics.SOP

module Control.DeepSeq.Bounded.Generic.GSeqable

module Control.DeepSeq.Bounded.Generic.GNFDataN

module Control.DeepSeq.Bounded.Generic.GNFDataP

Re-exported

module Control.DeepSeq

"I can has data family?..."

class FF k where

Associated Types

data F k :: * -> *

Methods

phi :: (Generic v, NFDataN v, NFDataP v) => k -> F k v -> v

The constraints could probably shrink to just Generic with a bit of work... In fact, trying it now -- to do away with NFDataN and NFDataP classes, and just "alias the SOP generic functions", as did in Seqable... [Fail. for now]

Instances

FF Int

FF SeqNode

FF Pattern