Basic data flow analyses framework based on esprima
$ npm install analyses
analyses depends on a globally available Map
, which can either be provided by
running with --harmony
or by preloading a shim before using the library.
analyses itself explicitly does not depend on a shim itself, to be more forward
compatible and to give you more choice.
Since the collections, specifically Set
in --harmony
do not support any method to iterate over them,
analyses provides a minimal analyses.Set
shim with some useful set operations.
This will change once native --harmony
Sets are actually useful.
var cfg = esgraph(esprima.parse(''));
var output = analyses(cfg, function (input, list) {
this; // the cfg node
input; // the input set
list; // the worklist `.push()` nodes to it.
return new analyses.Set(); // either return a new output Set
// or return an output Set and `enqueue: false` so the worklist algorithm does
// not check and enqueue successors itself.
return {output: new analyses.Set(), enqueue: false}
}, {
// direction:
// forward or backward; defaults to forward
direction: 'forward',
// custom merge function:
// typically union or intersect; defaults to union
// merge function takes an array of inputs
// `analyses.merge()` wraps a function which takes a pair `a, b` of inputs
merge: analyses.merge(analyses.Set.union),
// custom equals function:
// this is used to determine if the output of a node still changes and to
// not enqueue any more successors and stop the iteration; defaults to
// Set.equals
equals: analyses.Set.equals
});
// the output is a `Map` from CFG nodes to a result set of type `analyses.Set`
output.get(cfg[0]).values()
LGPLv3