node package manager


basic data flow analyses framework based on esprima


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 (inputlist) {
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`