Avoid double extraction of substrings in various MATCH_ bytecodes

peggyjs · Aug 21, 2023 · fe01e6f · fe01e6f
1 parent 736d6cc
commit fe01e6f
Show file tree

Hide file tree

Showing 2 changed files with 67 additions and 22 deletions.
diff --git a/lib/compiler/passes/generate-js.js b/lib/compiler/passes/generate-js.js
@@ -313,6 +313,7 @@ function generateJS(ast, options) {
   function generateRuleFunction(rule) {
     const parts = [];
     const stack = new Stack(rule.name, "s", "var", rule.bytecode);
+    let skipAccept = false;
 
     function compile(bc) {
       let ip = 0;
@@ -351,6 +352,50 @@ function generateJS(ast, options) {
         parts.push("}");
       }
 
+      /*
+        MATCH_* opcodes typically do something like
+
+          if (<test>(input.substr(peg$currPos, length))) {
+            sN = input.substr(peg$currPos, length);
+            ...
+          } else {
+            sN = peg$FAILED;
+            ...
+          }
+
+        compileInputChunkCondition will convert that to
+
+          sN = input.substr(peg$currPos, length);
+          if (<test>(sN)) {
+            ...
+          } else {
+            sN = peg$FAILED;
+            ...
+          }
+
+          and avoid extracting the sub string twice.
+      */
+      function compileInputChunkCondition(
+        cond, argCount, inputChunkLength
+      ) {
+        const baseLength = argCount + 3;
+        let inputChunk = inputChunkLength === 1
+          ? "input.charAt(peg$currPos)"
+          : "input.substr(peg$currPos, " + inputChunkLength + ")";
+        if (bc[ip + baseLength] === op.ACCEPT_N
+              && bc[ip + baseLength + 1] === inputChunkLength) {
+          skipAccept = true;
+          parts.push(stack.push(inputChunk));
+          inputChunk = stack.pop();
+        } else if (inputChunkLength === 1) {
+          cond = cond.replace("$$$.charCodeAt(0)", "input.charCodeAt(peg$currPos)");
+        }
+        compileCondition(
+          cond.replace("$$$", inputChunk),
+          argCount
+        );
+      }
+
       function compileLoop(cond) {
         const baseLength = 2;
         const bodyLength = bc[ip + baseLength - 1];
@@ -505,41 +550,41 @@ function generateJS(ast, options) {
             break;
 
           case op.MATCH_STRING:       // MATCH_STRING s, a, f, ...
-            compileCondition(
+            compileInputChunkCondition(
               ast.literals[bc[ip + 1]].length > 1
-                ? "input.substr(peg$currPos, "
-                    + ast.literals[bc[ip + 1]].length
-                    + ") === "
-                    + l(bc[ip + 1])
-                : "input.charCodeAt(peg$currPos) === "
+                ? "$$$ === " + l(bc[ip + 1])
+                : "$$$.charCodeAt(0) === "
                     + ast.literals[bc[ip + 1]].charCodeAt(0),
-              1
+              1,
+              ast.literals[bc[ip + 1]].length
             );
             break;
 
           case op.MATCH_STRING_IC:    // MATCH_STRING_IC s, a, f, ...
-            compileCondition(
-              "input.substr(peg$currPos, "
-                + ast.literals[bc[ip + 1]].length
-                + ").toLowerCase() === "
-                + l(bc[ip + 1]),
-              1
+            compileInputChunkCondition(
+              "$$$.toLowerCase() === " + l(bc[ip + 1]),
+              1,
+              ast.literals[bc[ip + 1]].length
             );
             break;
 
           case op.MATCH_CHAR_CLASS:   // MATCH_CHAR_CLASS c, a, f, ...
-            compileCondition(
-              r(bc[ip + 1]) + ".test(input.charAt(peg$currPos))",
-              1
+            compileInputChunkCondition(
+              r(bc[ip + 1]) + ".test($$$)", 1, 1
             );
             break;
 
           case op.ACCEPT_N:           // ACCEPT_N n
-            parts.push(stack.push(
-              bc[ip + 1] > 1
-                ? "input.substr(peg$currPos, " + bc[ip + 1] + ")"
-                : "input.charAt(peg$currPos)"
-            ));
+            if (skipAccept) {
+              stack.sp++;
+              skipAccept = false;
+            } else {
+              parts.push(stack.push(
+                bc[ip + 1] > 1
+                  ? "input.substr(peg$currPos, " + bc[ip + 1] + ")"
+                  : "input.charAt(peg$currPos)"
+              ));
+            }
             parts.push(
               bc[ip + 1] > 1
                 ? "peg$currPos += " + bc[ip + 1] + ";"

diff --git a/test/api/pegjs-api.spec.js b/test/api/pegjs-api.spec.js
@@ -356,7 +356,7 @@ describe("Peggy API", () => {
           it("labelled rule name", () => check("RULE_2 'named'", source, "RULE_2", "peg$parseRULE_2() {"));
           it("literal expression", () => check("'a'", source, null, "input.charCodeAt(peg$currPos) === 97"));
           it("multichar literal", () => check("'def'", source, null, "input.substr(peg$currPos, 3) === peg$c3"));
-          it("chars expression", () => check("[abc]", source, null, "peg$r0.test(input.charAt(peg$currPos))"));
+          it("chars expression", () => check("[abc]", source, null, /s\d = input\.charAt\(peg\$currPos\);\s*if \(peg\$r0\.test\(s\d\)\)/));
           it("rule expression", () => check("RULE_2", source, null, "peg$parseRULE_2();"));
           it("choice expression", () => check(
             "RULE_1 / RULE_2",