The data-reify package
data-reify provided the ability to turn recursive structures into explicit graphs. Many (implicitly or explicitly) recursive data structure can be given this ability, via a type class instance. This gives an alternative to using Ref for observable sharing.
Observable sharing in general is unsafe, so we use the IO monad to bound this effect, but can be used safely even with unsafePerformIO if some simple conditions are met. Typically this package will be used to tie the knot with DSL's that depend of observable sharing, like Lava.
Providing an instance for MuRef is the mechanism for allowing a structure to be reified into a graph, and several examples of this are provided.
History: Version 0.1 used unsafe pointer compares. Version 0.2 of data-reify used StableNames, and was much faster. Version 0.3 provided two versions of MuRef, the mono-typed version, for trees of a single type, and the dynamic-typed version, for trees of different types. Version 0.4 used Int as a synonym for Unique rather than Data.Unique for node ids, by popular demand. Version 0.5 merged the mono-typed and dynamic version again, by using DynStableName, an unphantomized version of StableName.
© 2009 Andy Gill; BSD3 license.
|Versions||0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.6.1|
|Dependencies||base (>=3 && <4.2), containers [details]|
|Copyright||(c) 2009 Andy Gill|
|Maintainer||Andy Gill <email@example.com>|
|Category||Language, Data, Parsing, Reflection|
|Uploaded||Wed Sep 9 21:56:20 UTC 2009 by AndyGill|
|Distributions||Fedora:0.6, LTSHaskell:0.6.1, NixOS:0.6.1, Stackage:0.6.1|
|Downloads||4084 total (17 in the last 30 days)|
|Status||Docs uploaded by user
Build status unknown [no reports yet]
For package maintainers and hackage trustees