(new-codegen) Use common code for matching literals, classes and any

lib/compiler/opcodes.js

@@ -36,10 +36,10 @@ let opcodes = {
 
   // Matching
 
-  MATCH_ANY:         17,   // MATCH_ANY a, f, ...
-  MATCH_STRING:      18,   // MATCH_STRING s, a, f, ...
-  MATCH_STRING_IC:   19,   // MATCH_STRING_IC s, a, f, ...
-  MATCH_CLASS:       20,   // MATCH_CLASS c, a, f, ...
+  MATCH_ANY:         17,   // MATCH_ANY
+  MATCH_LITERAL:     18,   // MATCH_LITERAL l
+  MATCH_LITERAL_IC:  19,   // MATCH_LITERAL_IC l
+  MATCH_CLASS:       20,   // MATCH_CLASS c
   ACCEPT_N:          21,   // ACCEPT_N n
   ACCEPT_STRING:     22,   // ACCEPT_STRING s
   EXPECT:            23,   // EXPECT e

lib/compiler/passes/generate-bytecode.js

@@ -153,37 +153,21 @@ let visitor = require("../visitor");
 // Matching
 // --------
 //
-// [17] MATCH_ANY a, f, ...
+// [17] MATCH_ANY
 //
-//        if (input.length > currPos) {
-//          interpret(ip + 3, ip + 3 + a);
-//        } else {
-//          interpret(ip + 3 + a, ip + 3 + a + f);
-//        }
+//        state.matchAny()
 //
-// [18] MATCH_STRING s, a, f, ...
+// [18] MATCH_LITERAL l
 //
-//        if (input.substr(currPos, literals[s].length) === literals[s]) {
-//          interpret(ip + 4, ip + 4 + a);
-//        } else {
-//          interpret(ip + 4 + a, ip + 4 + a + f);
-//        }
+//        state.matchLiteral(l)
 //
-// [19] MATCH_STRING_IC s, a, f, ...
+// [19] MATCH_LITERAL_IC l
 //
-//        if (input.substr(currPos, literals[s].length).toLowerCase() === literals[s]) {
-//          interpret(ip + 4, ip + 4 + a);
-//        } else {
-//          interpret(ip + 4 + a, ip + 4 + a + f);
-//        }
+//        state.matchLiteralIC(l)
 //
-// [20] MATCH_CLASS c, a, f, ...
+// [20] MATCH_CLASS c
 //
-//        if (classes[c].test(input.charAt(currPos))) {
-//          interpret(ip + 4, ip + 4 + a);
-//        } else {
-//          interpret(ip + 4 + a, ip + 4 + a + f);
-//        }
+//        state.matchClass(c)
 //
 // [21] ACCEPT_N n
 //
@@ -302,6 +286,13 @@ function generateBytecode(ast) {
     );
   }
 
+  function buildTryCondition(match, commonCode, thenCode, elseCode) {
+    if (match > 0) { return thenCode; }
+    if (match < 0) { return elseCode; }
+
+    return commonCode;
+  }
+
   function buildBreakCondition(match, condCode, n, thenCode, elseCode) {
     thenCode = buildSequence(thenCode, [op.BREAK, n]);
 
@@ -947,11 +938,11 @@ function generateBytecode(ast) {
         // save one |substr| call that would be needed if we used |ACCEPT_N|.
         return buildSequence(
           context.reportFailures ? [op.EXPECT, expectedIndex] : [],
-          buildCondition(
+          buildTryCondition(
             match,
             node.ignoreCase
-              ? [op.MATCH_STRING_IC, stringIndex]
-              : [op.MATCH_STRING, stringIndex],
+              ? [op.MATCH_LITERAL_IC, stringIndex]
+              : [op.MATCH_LITERAL, stringIndex],
             node.ignoreCase
               ? [op.ACCEPT_N, node.value.length]
               : [op.ACCEPT_STRING, stringIndex],
@@ -976,7 +967,7 @@ function generateBytecode(ast) {
 
       return buildSequence(
         context.reportFailures ? [op.EXPECT, expectedIndex] : [],
-        buildCondition(
+        buildTryCondition(
           match,
           [op.MATCH_CLASS, classIndex],
           [op.ACCEPT_N, 1],
@@ -991,7 +982,7 @@ function generateBytecode(ast) {
 
       return buildSequence(
         context.reportFailures ? [op.EXPECT, expectedIndex] : [],
-        buildCondition(
+        buildTryCondition(
           node.match|0,
           [op.MATCH_ANY],
           [op.ACCEPT_N, 1],

lib/compiler/passes/generate-js.js

@@ -427,38 +427,33 @@ function generateJS(ast, options) {
             compileLoop();
             break;
 
-          case op.MATCH_ANY:          // MATCH_ANY a, f, ...
-            compileCondition("input.length > peg$state._offset", 0);
+          case op.MATCH_ANY:          // MATCH_ANY
+            parts.push(resStack.push("peg$state.matchAny()"));
+            ip++;
             break;
 
-          case op.MATCH_STRING:       // MATCH_STRING s, a, f, ...
-            compileCondition(
-              ast.literals[bc[ip + 1]].length > 1
-                ? "input.substr(peg$state._offset, "
-                    + ast.literals[bc[ip + 1]].length
-                    + ") === "
-                    + l(bc[ip + 1])
-                : "input.charCodeAt(peg$state._offset) === "
-                    + ast.literals[bc[ip + 1]].charCodeAt(0),
-              1
-            );
+          case op.MATCH_LITERAL: {    // MATCH_LITERAL l
+            let str = ast.literals[bc[ip + 1]];
+            parts.push(resStack.push(
+              str.length > 1
+                ? "peg$state.matchLiteral(" + l(bc[ip + 1]) + ")"
+                : "peg$state.matchChar(" + l(bc[ip + 1]) + ", " + str.charCodeAt(0) + ")"
+            ));
+            ip += 2;
             break;
-
-          case op.MATCH_STRING_IC:    // MATCH_STRING_IC s, a, f, ...
-            compileCondition(
-              "input.substr(peg$state._offset, "
-                + ast.literals[bc[ip + 1]].length
-                + ").toLowerCase() === "
-                + l(bc[ip + 1]),
-              1
-            );
+          }
+          case op.MATCH_LITERAL_IC:    // MATCH_LITERAL_IC l
+            parts.push(resStack.push(
+              "peg$state.matchLiteralIC(" + l(bc[ip + 1]) + ")"
+            ));
+            ip += 2;
             break;
 
-          case op.MATCH_CLASS:        // MATCH_CLASS c, a, f, ...
-            compileCondition(
-              r(bc[ip + 1]) + ".test(input.charAt(peg$state._offset))",
-              1
-            );
+          case op.MATCH_CLASS:        // MATCH_CLASS c
+            parts.push(resStack.push(
+              "peg$state.matchClass(" + r(bc[ip + 1]) + ")"
+            ));
+            ip += 2;
             break;
 
           case op.ACCEPT_N:           // ACCEPT_N n