Merge pull request #19 from jg-rp/i-regexp

Map JS RegExp to I-Regexp

CHANGELOG.md

@@ -1,5 +1,11 @@
 # JSON P3 Change Log
 
+## Version 1.3.1 (unreleased)
+
+**Fixes**
+
+- Fixed RegExp to I-Regex pattern mapping with the `match` and `search` filter functions. We now correctly match the special `.` character to everything other than `\r` and `\n`.
+
 ## Version 1.3.0
 
 **Fixes**

package.json

@@ -1,6 +1,6 @@
 {
   "name": "json-p3",
-  "version": "1.3.0",
+  "version": "1.3.1",
   "author": "James Prior",
   "license": "MIT",
   "description": "JSONPath, JSON Pointer and JSON Patch",

src/path/functions/match.ts

@@ -58,9 +58,56 @@ export class Match implements FilterFunction {
 
   protected fullMatch(pattern: string): string {
     const parts: string[] = [];
-    if (!pattern.startsWith("^")) parts.push("^");
-    parts.push(pattern);
-    if (!pattern.endsWith("$")) parts.push("$");
+    let nonCaptureGroup = false;
+
+    if (!pattern.startsWith("^") && !pattern.startsWith("^(")) {
+      nonCaptureGroup = true;
+      parts.push("^(?:");
+    }
+    parts.push(this.mapRegexp(pattern));
+
+    if (nonCaptureGroup && !pattern.endsWith("$") && !pattern.endsWith(")$")) {
+      parts.push(")$");
+    }
+
+    return parts.join("");
+  }
+
+  // See https://datatracker.ietf.org/doc/html/rfc9485#name-ecmascript-regexps
+  protected mapRegexp(pattern: string): string {
+    let escaped = false;
+    let charClass = false;
+    const parts: string[] = [];
+    for (const ch of pattern) {
+      switch (ch) {
+        case ".":
+          if (!escaped && !charClass) {
+            parts.push("(?:(?![\r\n])\\P{Cs}|\\p{Cs}\\p{Cs})");
+          } else {
+            parts.push(ch);
+            escaped = false;
+          }
+          break;
+        case "\\":
+          escaped = true;
+          parts.push(ch);
+          break;
+        case "[":
+          charClass = true;
+          escaped = false;
+          parts.push(ch);
+          break;
+        case "]":
+          charClass = false;
+          escaped = false;
+          parts.push(ch);
+          break;
+        default:
+          escaped = false;
+          parts.push(ch);
+          break;
+      }
+    }
     return parts.join("");
   }
 }

src/path/functions/search.ts

@@ -48,12 +48,50 @@ export class Search implements FilterFunction {
     }
 
     try {
-      const re = new RegExp(pattern, "u");
+      const re = new RegExp(this.mapRegexp(pattern), "u");
       if (this.cacheSize > 0) this.#cache.set(pattern, re);
       return !!s.match(re);
     } catch (error) {
       if (this.throwErrors) throw error;
       return false;
     }
   }
+
+  // See https://datatracker.ietf.org/doc/html/rfc9485#name-ecmascript-regexps
+  protected mapRegexp(pattern: string): string {
+    let escaped = false;
+    let charClass = false;
+    const parts: string[] = [];
+    for (const ch of pattern) {
+      switch (ch) {
+        case ".":
+          if (!escaped && !charClass) {
+            parts.push("(?:(?![\r\n])\\P{Cs}|\\p{Cs}\\p{Cs})");
+          } else {
+            parts.push(ch);
+            escaped = false;
+          }
+          break;
+        case "\\":
+          escaped = true;
+          parts.push(ch);
+          break;
+        case "[":
+          charClass = true;
+          escaped = false;
+          parts.push(ch);
+          break;
+        case "]":
+          charClass = false;
+          escaped = false;
+          parts.push(ch);
+          break;
+        default:
+          escaped = false;
+          parts.push(ch);
+          break;
+      }
+    }
+    return parts.join("");
+  }
 }

src/path/lex.ts

@@ -367,6 +367,7 @@ function lexInsideBracketedSelection(l: Lexer): StateFn | null {
     }
 
     if (!l.environment.strict && l.acceptMatchRun(l.environment.keysPattern)) {
+      // FIXME: fall back to legacy behavior if keysPattern is not the default
       switch (l.peek()) {
         case "'":
           l.ignore(); // ~
@@ -377,7 +378,6 @@ function lexInsideBracketedSelection(l: Lexer): StateFn | null {
           l.next();
           return lexDoubleQuoteKeyString(l);
         case "?":
-          l.ignore(); // ~
           l.next();
           l.emit(TokenKind.KEYS_FILTER);
           l.filterLevel += 1;