issue pegjs#30: begin implementing ranges support. Changes in grammar…

… and handling in passes.
Mingun · Jan 1, 2014 · 252b43e · 252b43e
1 parent 93d3c6a
commit 252b43e
Show file tree

Hide file tree

Showing 6 changed files with 94 additions and 4 deletions.
diff --git a/lib/compiler/opcodes.js b/lib/compiler/opcodes.js
@@ -42,5 +42,10 @@ module.exports = {
   /* Failure Reporting */
 
   SILENT_FAILS_ON:  24,   // SILENT_FAILS_ON
-  SILENT_FAILS_OFF: 25    // SILENT_FAILS_FF
+  SILENT_FAILS_OFF: 25,   // SILENT_FAILS_FF
+
+  /* Checking array length */
+
+  CHECK_MIN_LEN:    27,   // CHECK_MIN_LEN len
+  CHECK_MAX_LEN:    28    // CHECK_MAX_LEN len
 };
diff --git a/lib/compiler/passes/generate-bytecode.js b/lib/compiler/passes/generate-bytecode.js
@@ -168,10 +168,21 @@ var utils = require("../../utils"),
  * [25] SILENT_FAILS_OFF
  *
  *        silentFails--;
+ *
+ * Checking array length
+ * ---------------------
+ *
+ * [27] CHECK_MIN_LEN len
+ *
+ *        stack.push(len !== undefined && stack.top().length < len)
+ *
+ * [28] CHECK_MAX_LEN len
+ *
+ *        stack.push(len !== undefined && stack.top().length >=len)
  */
 module.exports = function(ast) {
   var consts = [];
-
+  /// Add new constant to constant list, if it not exist, and return they index in that list.
   function addConst(value) {
     var index = utils.indexOf(consts, function(c) { return c === value; });
 
@@ -512,6 +523,30 @@ module.exports = function(ast) {
       );
     },
 
+    range: function(node, context) {
+      var emptyArrayIndex = addConst('[]');
+          expressionCode  = generate(node.expression, {
+            sp:     context.sp + 1,
+            env:    { },
+            action: null
+          });
+
+      return buildSequence(
+        [op.PUSH, emptyArrayIndex], // var result = []
+        expressionCode,             // var elem = expr();
+        buildLoop(
+          [op.WHILE_NOT_ERROR],     // while (elem !== null) {
+          buildSequence(
+            [op.APPEND],            //   result.push(elem);
+            /*TODO: new opcode*/    //   if (max !== undefined && result.length >= max) break;
+            expressionCode          //   elem = expr();
+          )
+        ),                          // }
+        /*TODO: new opcode*/        // if (min !== undefined && result.length < min) result = null;
+        [op.POP]                    // return result;
+      );
+    },
+
     rule_ref: function(node) {
       return [op.RULE, utils.indexOfRuleByName(ast, node.name)];
     },

diff --git a/lib/compiler/passes/generate-javascript.js b/lib/compiler/passes/generate-javascript.js
@@ -773,7 +773,7 @@ module.exports = function(ast, options) {
   }
 
   parts.push([
-    '',
+    '    //{ Helper functions',
     '    function text() {',
     '      return input.substring(peg$reportedPos, peg$currPos);',
     '    }',
@@ -936,17 +936,19 @@ module.exports = function(ast, options) {
     '        posDetails.column',
     '      );',
     '    }',
-    ''
+    '    //}'
   ].join('\n'));
 
   if (options.optimize === "size") {
     parts.push(indent4(generateInterpreter()));
     parts.push('');
   } else {
+    parts.push('    //{ Parse rule functions')
     utils.each(ast.rules, function(rule) {
       parts.push(indent4(generateRuleFunction(rule)));
       parts.push('');
     });
+    parts.push('    //}')
   }
 
   if (ast.initializer) {

diff --git a/lib/compiler/passes/report-left-recursion.js b/lib/compiler/passes/report-left-recursion.js
@@ -45,6 +45,7 @@ module.exports = function(ast) {
     optional:     checkExpression,
     zero_or_more: checkExpression,
     one_or_more:  checkExpression,
+    range:        checkExpression,
 
     rule_ref:
       function(node, appliedRules) {

diff --git a/lib/compiler/passes/report-missing-rules.js b/lib/compiler/passes/report-missing-rules.js
@@ -27,6 +27,7 @@ module.exports = function(ast) {
     optional:     checkExpression,
     zero_or_more: checkExpression,
     one_or_more:  checkExpression,
+    range:        checkExpression,
 
     rule_ref:
       function(node) {

diff --git a/src/parser.pegjs b/src/parser.pegjs
@@ -138,6 +138,38 @@ suffixed
         expression: expression
       };
     }
+  / expression:primary r:range {
+      if (!(r.delimiter !== undefined)) {
+        if (!(r.max !== undefined)) {// unbounded
+          if (r.min === 0) {
+            return {
+              type:       "zero_or_more",
+              expression: expression
+            };
+          } else
+          if (r.min === 1) {
+            return {
+              type:       "one_or_more",
+              expression: expression
+            };
+          }
+        } else
+        if (r.max === 1) {
+          if (r.min === 0) {
+            return {
+              type:       "optional",
+              expression: expression
+            };
+          } else
+          if (r.min === 1) {
+            return expression
+          }
+        }
+      }
+      r.type = "range";
+      r.expression = expression;
+      return r;
+    }
   / primary
 
 primary
@@ -152,6 +184,16 @@ primary
   / dot { return { type: "any" }; }
   / lparen expression:expression rparen { return expression; }
 
+range
+  = range_open r:range2 delimiter:(comma primary)? range_close {
+    r.delimiter = delimiter !== "" ? delimiter[1] : undefined;
+    return r;
+  }
+range2
+  = min:int? dots max:int? {return {min:min!==""?min:0, max:max!==""?max:undefined};}
+  / val:int {return {min:val, max:val};}
+int = n:digit+ __ {return parseInt(n.join(''),10);}
+
 /* "Lexical" elements */
 
 action "action"
@@ -179,6 +221,10 @@ plus      = "+" __ { return "+"; }
 lparen    = "(" __ { return "("; }
 rparen    = ")" __ { return ")"; }
 dot       = "." __ { return "."; }
+comma     = "," __ { return ","; }
+dots      = ".." __{ return ".."; }
+range_open= "|" __ { return "|"; }
+range_close="|" __ { return "|"; }
 
 /*
  * Modeled after ECMA-262, 5th ed., 7.6, but much simplified: