[483] | 1 | dojo.provide("dojox.lang.functional.tailrec"); |
---|
| 2 | |
---|
| 3 | dojo.require("dojox.lang.functional.lambda"); |
---|
| 4 | dojo.require("dojox.lang.functional.util"); |
---|
| 5 | |
---|
| 6 | // This module provides recursion combinators: |
---|
| 7 | // - a tail recursion combinator. |
---|
| 8 | |
---|
| 9 | // Acknowledgements: |
---|
| 10 | // - recursion combinators are inspired by Manfred von Thun's article |
---|
| 11 | // "Recursion Theory and Joy" |
---|
| 12 | // (http://www.latrobe.edu.au/philosophy/phimvt/joy/j05cmp.html) |
---|
| 13 | |
---|
| 14 | // Notes: |
---|
| 15 | // - recursion combinators produce a function, which implements |
---|
| 16 | // their respective recusion patterns. String lambdas are inlined, if possible. |
---|
| 17 | |
---|
| 18 | (function(){ |
---|
| 19 | var df = dojox.lang.functional, inline = df.inlineLambda, _x ="_x"; |
---|
| 20 | |
---|
| 21 | df.tailrec = function( |
---|
| 22 | /*Function|String|Array*/ cond, |
---|
| 23 | /*Function|String|Array*/ then, |
---|
| 24 | /*Function|String|Array*/ before){ |
---|
| 25 | // summary: |
---|
| 26 | // Generates a function for the tail recursion pattern. This is the simplified |
---|
| 27 | // version of the linear recursive combinator without the "after" function, |
---|
| 28 | // and with the modified "before" function. All parameter functions are called |
---|
| 29 | // in the context of "this" object. |
---|
| 30 | // cond: |
---|
| 31 | // The lambda expression, which is used to detect the termination of recursion. |
---|
| 32 | // It accepts the same parameter as the generated recursive function itself. |
---|
| 33 | // This function should return "true", if the recursion should be stopped, |
---|
| 34 | // and the "then" part should be executed. Otherwise the recursion will proceed. |
---|
| 35 | // then: |
---|
| 36 | // The lambda expression, which is called upon termination of the recursion. |
---|
| 37 | // It accepts the same parameters as the generated recursive function itself. |
---|
| 38 | // The returned value will be returned as the value of the generated function. |
---|
| 39 | // before: |
---|
| 40 | // The lambda expression, which is called before the recursive step. |
---|
| 41 | // It accepts the same parameter as the generated recursive function itself, |
---|
| 42 | // and returns an array of arguments for the next recursive call of |
---|
| 43 | // the generated function. |
---|
| 44 | |
---|
| 45 | var c, t, b, cs, ts, bs, dict1 = {}, dict2 = {}, |
---|
| 46 | add2dict = function(x){ dict1[x] = 1; }; |
---|
| 47 | if(typeof cond == "string"){ |
---|
| 48 | cs = inline(cond, _x, add2dict); |
---|
| 49 | }else{ |
---|
| 50 | c = df.lambda(cond); |
---|
| 51 | cs = "_c.apply(this, _x)"; |
---|
| 52 | dict2["_c=_t.c"] = 1; |
---|
| 53 | } |
---|
| 54 | if(typeof then == "string"){ |
---|
| 55 | ts = inline(then, _x, add2dict); |
---|
| 56 | }else{ |
---|
| 57 | t = df.lambda(then); |
---|
| 58 | ts = "_t.t.apply(this, _x)"; |
---|
| 59 | } |
---|
| 60 | if(typeof before == "string"){ |
---|
| 61 | bs = inline(before, _x, add2dict); |
---|
| 62 | }else{ |
---|
| 63 | b = df.lambda(before); |
---|
| 64 | bs = "_b.apply(this, _x)"; |
---|
| 65 | dict2["_b=_t.b"] = 1; |
---|
| 66 | } |
---|
| 67 | var locals1 = df.keys(dict1), locals2 = df.keys(dict2), |
---|
| 68 | f = new Function([], "var _x=arguments,_t=_x.callee,_c=_t.c,_b=_t.b".concat( // Function |
---|
| 69 | locals1.length ? "," + locals1.join(",") : "", |
---|
| 70 | locals2.length ? ",_t=_x.callee," + locals2.join(",") : t ? ",_t=_x.callee" : "", |
---|
| 71 | ";for(;!", |
---|
| 72 | cs, |
---|
| 73 | ";_x=", |
---|
| 74 | bs, |
---|
| 75 | ");return ", |
---|
| 76 | ts |
---|
| 77 | )); |
---|
| 78 | if(c){ f.c = c; } |
---|
| 79 | if(t){ f.t = t; } |
---|
| 80 | if(b){ f.b = b; } |
---|
| 81 | return f; |
---|
| 82 | }; |
---|
| 83 | })(); |
---|
| 84 | |
---|
| 85 | /* |
---|
| 86 | For documentation only: |
---|
| 87 | |
---|
| 88 | 1) The original recursive version: |
---|
| 89 | |
---|
| 90 | var tailrec1 = function(cond, then, before){ |
---|
| 91 | var cond = df.lambda(cond), |
---|
| 92 | then = df.lambda(then), |
---|
| 93 | before = df.lambda(before); |
---|
| 94 | return function(){ |
---|
| 95 | if(cond.apply(this, arguments)){ |
---|
| 96 | return then.apply(this, arguments); |
---|
| 97 | } |
---|
| 98 | var args = before.apply(this, arguments); |
---|
| 99 | return arguments.callee.apply(this, args); |
---|
| 100 | }; |
---|
| 101 | }; |
---|
| 102 | |
---|
| 103 | 2) The original iterative version (before minification and inlining): |
---|
| 104 | |
---|
| 105 | var tailrec2 = function(cond, then, before){ |
---|
| 106 | var cond = df.lambda(cond), |
---|
| 107 | then = df.lambda(then), |
---|
| 108 | before = df.lambda(before); |
---|
| 109 | return function(){ |
---|
| 110 | var args = arguments; |
---|
| 111 | for(; !cond.apply(this, args); args = before.apply(this, args)); |
---|
| 112 | return then.apply(this, args); |
---|
| 113 | }; |
---|
| 114 | }; |
---|
| 115 | |
---|
| 116 | */ |
---|