package fix
Install
dune-project
Dependency
Authors
Maintainers
Sources
md5=af7099d1cf24654bd379953b8e3d063f
sha512=147bde60a96191f5f0090842df37a12509d78cf6e9dda482fbd000f61215c3d12115854fc331ae39009f26948d6d527187adade160e7ab27a6c4f0d8b324490a
doc/fix/Fix/DataFlow/index.html
Module Fix.DataFlowSource
This module performs a forward data flow analysis over a (possibly cyclic) directed graph. Like Fix.Fix, it computes the least function of type variable -> property that satisfies a fixed point equation. It is less widely applicable than Fix.Fix, but, when it is applicable, it can be both easier to use and more efficient. It does not perform dynamic dependency discovery. The submodule Fix.DataFlow.ForCustomMaps is particularly tuned for performance.
Run requires a type variable that is equipped with an implementation of imperative maps, a type property that is equipped with leq and join functions, and a data flow graph whose edges describe the propagation of properties. It performs a forward data flow analysis and returns its result.
ForOrderedType is a special case of Run where it suffices to pass an ordered type T as an argument. A reference to a persistent map is used to hold the memoization table.
module ForHashedType
(T : Hashtbl.HashedType)
(P : sig ... end)
(G : sig ... end) :
sig ... endForHashedType is a special case of Run where it suffices to pass a hashed type T as an argument. A hash table is used to hold the memoization table.
ForIntSegment is a special case of Run where the type of variables is the integer segment [0..n). An array is used to hold the table.
module ForCustomMaps
(P : sig ... end)
(G : sig ... end)
(V : sig ... end)
(B : sig ... end) :
sig ... endForCustomMaps is a forward data flow analysis that is tuned for greater performance. It internally relies on CompactQueue, instead of Queue. Furthermore, instead of relying on a full-fledged implementation of maps as described by MINIMAL_IMPERATIVE_MAPS, it expects the user to create and initialize two maps V and B that satisfy the signature ARRAY. This typically allows the user to choose an efficient, specialized data representation.