re2js 2.6.0 → 2.7.0

package/build/index.umd.js

@@ -2,7 +2,7 @@
  * re2js
  * RE2JS is the JavaScript port of RE2, a regular expression engine that provides linear time matching
  *
- * @version v2.6.0
+ * @version v2.7.0
  * @author Oleksii Vasyliev
  * @homepage https://github.com/le0pard/re2js#readme
  * @repository github:le0pard/re2js
@@ -2159,9 +2159,9 @@
         case Inst.NOP:
           return `nop -> ${this.out}`;
         case Inst.LB_WRITE:
-          return `lbwrite ${this.lb} -> ${this.out}`;
+          return `lbwrite ${this.arg} -> ${this.out}`;
         case Inst.LB_CHECK:
-          return `lbcheck ${this.lb} -> ${this.out}, ${this.arg}`;
+          return `lbcheck ${this.arg} -> ${this.out}`;
         case Inst.RUNE:
           if (this.runes === null) {
             return 'rune <null>';
@@ -2600,17 +2600,17 @@
               continue;
             }
           case Inst.LB_WRITE:
-            this.lbTable[Math.abs(inst.lb)] = pos;
+            this.lbTable[Math.abs(inst.arg)] = pos;
             pc = inst.out;
             continue;
           case Inst.LB_CHECK:
-            if (inst.lb > 0) {
+            if (inst.arg > 0) {
               // Positive Lookbehind
-              if (this.lbTable[inst.lb] === pos) {
+              if (this.lbTable[inst.arg] === pos) {
                 pc = inst.out; // Flattened tail recursion
                 continue;
               }
-            } else if (this.lbTable[-inst.lb] !== pos) {
+            } else if (this.lbTable[-inst.arg] !== pos) {
               // Negative Lookbehind
               pc = inst.out; // Flattened tail recursion
               continue;
@@ -4690,7 +4690,7 @@
     }
     lookBehind(a, lb) {
       const id = this.newInst(Inst.LB_WRITE);
-      this.prog.getInst(id.i).lb = lb;
+      this.prog.getInst(id.i).arg = lb;
 
       // Create the prefix wildcard `.*` for the lookbehind automaton
       const any = this.rune(Compiler.ANY_RUNE(), 0);
@@ -4698,7 +4698,7 @@
       const lbAutomaton = this.cat(dotStar, a);
       this.prog.patch(lbAutomaton.out, id.i);
       const checkId = this.newInst(Inst.LB_CHECK);
-      this.prog.getInst(checkId.i).lb = lb;
+      this.prog.getInst(checkId.i).arg = lb;
 
       // Save the starting point of this lookbehind automaton
       this.prog.lbStarts.push(lbAutomaton.i);
@@ -5473,6 +5473,7 @@
     static ERR_UNEXPECTED_PAREN = 'unexpected )';
     static ERR_NESTING_DEPTH = 'expression nests too deeply';
     static ERR_LARGE = 'expression too large';
+    static ERR_INVALID_CAPTURE_IN_LOOKBEHIND = 'invalid capture in lookbehind';
 
     // maxHeight is the maximum height of a regexp parse tree.
     // It is somewhat arbitrarily chosen, but the idea is to be large enough
@@ -5876,6 +5877,18 @@
       }
       return x;
     }
+
+    // recursively check for captures
+    static hasCapture(re) {
+      if (re === null) return false;
+      if (re.op === Regexp.Op.CAPTURE) return true;
+      if (re.subs) {
+        for (let sub of re.subs) {
+          if (Parser.hasCapture(sub)) return true;
+        }
+      }
+      return false;
+    }
     constructor(wholeRegexp, flags = 0) {
       this.wholeRegexp = wholeRegexp;
       // Flags control the behavior of the parser and record information about
@@ -6559,7 +6572,7 @@
             case 1:
               // Impossible but handle.
               re.op = Regexp.Op.EMPTY_MATCH;
-              re.subs = null;
+              re.subs = Regexp.emptySubs();
               break;
             case 2:
               {
@@ -7010,6 +7023,9 @@
 
       // Handle lookbehinds
       if (re2.lb !== 0) {
+        if (Parser.hasCapture(re1)) {
+          throw new RE2JSSyntaxException(Parser.ERR_INVALID_CAPTURE_IN_LOOKBEHIND, this.wholeRegexp);
+        }
         if (re2.lb > 0) {
           re2.op = Regexp.Op.PLB;
         } else {
@@ -8667,7 +8683,6 @@
      */
     *matchAll(input) {
       const m = this.matcher(input);
-      const inputStr = typeof input === 'string' ? input : m.matcherInput.asCharSequence();
       while (m.find()) {
         // Build the match array starting with the full match
         const result = [m.group(0)];
@@ -8680,7 +8695,7 @@
 
         // Attach native RegExp match properties
         result.index = m.start(0);
-        result.input = inputStr;
+        result.input = input; // Retains original String or Uint8Array so index aligns perfectly
 
         // Attach named capture groups if they exist
         const namedGroups = this.namedGroups();