jsonrepair - npm Package Compare versions

Comparing version 3.5.1 to 3.6.0

lib/cjs/regular/jsonrepair.js

@@ -319,5 +319,57 @@ "use strict";
       i++;
-      const isEndOfString = stopAtDelimiter ? i => (0, _stringUtils.isDelimiter)(text[i]) : i => isEndQuote(text.charCodeAt(i));
-      while (i < text.length && !isEndOfString(i)) {
-        if (text.charCodeAt(i) === _stringUtils.codeBackslash) {
+      while (true) {
+        if (i >= text.length) {
+          // end of text, we are missing an end quote
+          if (!stopAtDelimiter) {
+            // retry parsing the string, stopping at the first next delimiter
+            i = iBefore;
+            output = output.substring(0, oBefore);
+            return parseString(true);
+          }
+
+          // repair missing quote
+          str = (0, _stringUtils.insertBeforeLastWhitespace)(str, '"');
+          output += str;
+          return true;
+        } else if (isEndQuote(text.charCodeAt(i))) {
+          // end quote
+          // let us check what is before and after the quote to verify whether this is a legit end quote
+          const iQuote = i;
+          const oQuote = str.length;
+          str += '"';
+          i++;
+          output += str;
+          parseWhitespaceAndSkipComments();
+          if (stopAtDelimiter || i >= text.length || (0, _stringUtils.isDelimiter)(text.charAt(i)) || (0, _stringUtils.isQuote)(text.charCodeAt(i))) {
+            // The quote is followed by a delimiter or the end of the text,
+            // so the quote is indeed the end of the string
+            parseConcatenatedString();
+            return true;
+          }
+          if ((0, _stringUtils.isDelimiter)(text.charAt(prevNonWhitespaceIndex(iQuote - 1)))) {
+            // This is not the right end quote: it is preceded by a delimiter,
+            // and NOT followed by a delimiter. So, there is an end quote missing
+            // parse the string again and then stop at the first next delimiter
+            i = iBefore;
+            output = output.substring(0, oBefore);
+            return parseString(true);
+          }
+
+          // revert to right after the quote but before any whitespace, and continue parsing the string
+          output = output.substring(0, oBefore);
+          i = iQuote + 1;
+
+          // repair unescaped quote
+          str = str.substring(0, oQuote) + '\\' + str.substring(oQuote);
+        } else if (stopAtDelimiter && (0, _stringUtils.isDelimiter)(text[i])) {
+          // we're in the mode to stop the string at the first delimiter
+          // because there is an end quote missing
+
+          // repair missing quote
+          str = (0, _stringUtils.insertBeforeLastWhitespace)(str, '"');
+          output += str;
+          parseConcatenatedString();
+          return true;
+        } else if (text.charCodeAt(i) === _stringUtils.codeBackslash) {
+          // handle escaped content like \n or \u2605
           const char = text.charAt(i + 1);
@@ -349,2 +401,3 @@ const escapeChar = escapeCharacters[char];
         } else {
+          // handle regular characters
           const char = text.charAt(i);
@@ -369,33 +422,6 @@ const code = text.charCodeAt(i);
         if (skipEscapeChars) {
-          const processed = skipEscapeCharacter();
-          if (processed) {
-            // repair: skipped escape character (nothing to do)
-          }
+          // repair: skipped escape character (nothing to do)
+          skipEscapeCharacter();
         }
       }
-      const hasEndQuote = (0, _stringUtils.isQuote)(text.charCodeAt(i));
-      if (hasEndQuote) {
-        str += '"';
-        i++;
-      } else {
-        // repair missing quote
-        str = (0, _stringUtils.insertBeforeLastWhitespace)(str, '"');
-      }
-      output += str;
-      parseWhitespaceAndSkipComments();
-
-      // See whether we have:
-      // (a) An end quote which is not followed by a valid delimiter
-      // (b) No end quote and reached the end of the input
-      // If so, revert parsing this string and try again, running in a more
-      // conservative mode, stopping at the first next delimiter
-      const isAtEnd = i >= text.length;
-      const nextIsDelimiter = (0, _stringUtils.isDelimiter)(text.charAt(i));
-      if (!stopAtDelimiter && (hasEndQuote && !isAtEnd && !nextIsDelimiter || !hasEndQuote && isAtEnd)) {
-        i = iBefore;
-        output = output.substring(0, oBefore);
-        return parseString(true);
-      }
-      parseConcatenatedString();
-      return true;
     }
@@ -507,3 +533,3 @@ return false;
     const start = i;
-    while (i < text.length && !(0, _stringUtils.isDelimiter)(text[i])) {
+    while (i < text.length && !(0, _stringUtils.isDelimiter)(text[i]) && !(0, _stringUtils.isQuote)(text.charCodeAt(i))) {
       i++;
@@ -544,2 +570,9 @@ }
   }
+  function prevNonWhitespaceIndex(start) {
+    let prev = start;
+    while (prev > 0 && (0, _stringUtils.isWhitespace)(text.charCodeAt(prev))) {
+      prev--;
+    }
+    return prev;
+  }
   function expectDigit(start) {
@@ -546,0 +579,0 @@ if (!(0, _stringUtils.isDigit)(text.charCodeAt(i))) {

lib/cjs/streaming/buffer/OutputBuffer.js

@@ -54,2 +54,8 @@ "use strict";
   }
+  function insertAt(index, text) {
+    if (index < offset) {
+      throw new Error("Cannot insert: ".concat(flushedMessage));
+    }
+    buffer = buffer.substring(0, index - offset) + text + buffer.substring(index - offset);
+  }
   function length() {
@@ -103,2 +109,3 @@ return offset + buffer.length;
     remove,
+    insertAt,
     length,
@@ -105,0 +112,0 @@ flush,

lib/cjs/streaming/core.js

@@ -402,5 +402,2 @@ "use strict";
     let stopAtDelimiter = arguments.length > 0 && arguments[0] !== undefined ? arguments[0] : false;
-    // we may need to revert
-    const iBefore = i;
-    const oBefore = output.length();
     let skipEscapeChars = input.charCodeAt(i) === _stringUtils.codeBackslash;
@@ -417,12 +414,58 @@ if (skipEscapeChars) {
       // or any single-quote-like start with a single-quote-like end
-      const isEndQuote = (0, _stringUtils.isDoubleQuote)(input.charCodeAt(i)) ? _stringUtils.isDoubleQuote : (0, _stringUtils.isSingleQuote)(input.charCodeAt(i)) ? _stringUtils.isSingleQuote // eslint-disable-line indent
-      : (0, _stringUtils.isSingleQuoteLike)(input.charCodeAt(i)) // eslint-disable-line indent
-      ? _stringUtils.isSingleQuoteLike // eslint-disable-line indent
-      : _stringUtils.isDoubleQuoteLike; // eslint-disable-line indent
-
+      const isEndQuote = (0, _stringUtils.isDoubleQuote)(input.charCodeAt(i)) ? _stringUtils.isDoubleQuote : (0, _stringUtils.isSingleQuote)(input.charCodeAt(i)) ? _stringUtils.isSingleQuote : (0, _stringUtils.isSingleQuoteLike)(input.charCodeAt(i)) ? _stringUtils.isSingleQuoteLike : _stringUtils.isDoubleQuoteLike;
+      const iBefore = i;
+      const oBefore = output.length();
       output.push('"');
       i++;
-      const isEndOfString = stopAtDelimiter ? i => (0, _stringUtils.isDelimiter)(input.charAt(i)) : i => isEndQuote(input.charCodeAt(i));
-      while (!input.isEnd(i) && !isEndOfString(i)) {
-        if (input.charCodeAt(i) === _stringUtils.codeBackslash) {
+      while (true) {
+        if (input.isEnd(i)) {
+          // end of text, we have a missing quote somewhere
+          if (!stopAtDelimiter) {
+            i = iBefore;
+            output.remove(oBefore);
+            return parseString(true);
+          }
+
+          // repair missing quote
+          output.insertBeforeLastWhitespace('"');
+          return stack.update(_stack.Caret.afterValue);
+        } else if (isEndQuote(input.charCodeAt(i))) {
+          // end quote
+          // let us check what is before and after the quote to verify whether this is a legit end quote
+          const iQuote = i;
+          const oQuote = output.length();
+          output.push('"');
+          i++;
+          parseWhitespaceAndSkipComments();
+          if (stopAtDelimiter || input.isEnd(i) || (0, _stringUtils.isDelimiter)(input.charAt(i)) || (0, _stringUtils.isQuote)(input.charCodeAt(i))) {
+            // The quote is followed by a delimiter or the end of the text,
+            // so the quote is indeed the end of the string
+            parseConcatenatedString();
+            return stack.update(_stack.Caret.afterValue);
+          }
+          if ((0, _stringUtils.isDelimiter)(input.charAt(prevNonWhitespaceIndex(iQuote - 1)))) {
+            // This is not the right end quote: it is preceded by a delimiter,
+            // and NOT followed by a delimiter. So, there is an end quote missing
+            // parse the string again and then stop at the first next delimiter
+            i = iBefore;
+            output.remove(oBefore);
+            return parseString(true);
+          }
+
+          // revert to right after the quote but before any whitespace, and continue parsing the string
+          output.remove(oQuote + 1);
+          i = iQuote + 1;
+
+          // repair unescaped quote
+          output.insertAt(oQuote, '\\');
+        } else if (stopAtDelimiter && (0, _stringUtils.isDelimiter)(input.charAt(i))) {
+          // we're in the mode to stop the string at the first delimiter
+          // because there is an end quote missing
+
+          // repair missing quote
+          output.insertBeforeLastWhitespace('"');
+          parseConcatenatedString();
+          return stack.update(_stack.Caret.afterValue);
+        } else if (input.charCodeAt(i) === _stringUtils.codeBackslash) {
+          // handle escaped content like \n or \u2605
           const char = input.charAt(i + 1);
@@ -454,2 +497,3 @@ const escapeChar = escapeCharacters[char];
         } else {
+          // handle regular characters
           const char = input.charAt(i);
@@ -478,26 +522,2 @@ const code = char.charCodeAt(0);
       }
-      const hasEndQuote = (0, _stringUtils.isQuote)(input.charCodeAt(i));
-      if (hasEndQuote) {
-        output.push('"');
-        i++;
-      } else {
-        // repair missing quote
-        output.insertBeforeLastWhitespace('"');
-      }
-      parseWhitespaceAndSkipComments();
-
-      // See whether we have:
-      // (a) An end quote which is not followed by a valid delimiter
-      // (b) No end quote and reached the end of the input
-      // If so, revert parsing this string and try again, running in a more
-      // conservative mode, stopping at the first next delimiter
-      const isAtEnd = input.isEnd(i);
-      const nextIsDelimiter = (0, _stringUtils.isDelimiter)(input.charAt(i));
-      if (!stopAtDelimiter && (hasEndQuote && !isAtEnd && !nextIsDelimiter || !hasEndQuote && isAtEnd)) {
-        i = iBefore;
-        output.remove(oBefore);
-        return parseString(true);
-      }
-      parseConcatenatedString();
-      return stack.update(_stack.Caret.afterValue);
     }
@@ -621,3 +641,3 @@ return false;
     let j = i;
-    while (!input.isEnd(j) && !(0, _stringUtils.isDelimiter)(input.charAt(j))) {
+    while (!input.isEnd(j) && !(0, _stringUtils.isDelimiter)(input.charAt(j)) && !(0, _stringUtils.isQuote)(input.charCodeAt(j))) {
       j++;
@@ -627,2 +647,9 @@ }
   }
+  function prevNonWhitespaceIndex(start) {
+    let prev = start;
+    while (prev > 0 && (0, _stringUtils.isWhitespace)(input.charCodeAt(prev))) {
+      prev--;
+    }
+    return prev;
+  }
   function expectDigit(start) {
@@ -629,0 +656,0 @@ if (!(0, _stringUtils.isDigit)(input.charCodeAt(i))) {

lib/cjs/utils/stringUtils.js

@@ -78,5 +78,5 @@ "use strict";
 function isDelimiter(char) {
-  return regexDelimiter.test(char) || isQuote(char.charCodeAt(0));
+  return regexDelimiter.test(char);
 }
-const regexDelimiter = /^[,:[\]{}()\n+]$/;
+const regexDelimiter = /^[,:[\]/{}()\n+]$/;
 function isStartOfValue(char) {
@@ -83,0 +83,0 @@ return regexStartOfValue.test(char) || char && isQuote(char.charCodeAt(0));

lib/esm/regular/jsonrepair.js

@@ -313,5 +313,57 @@ import { JSONRepairError } from '../utils/JSONRepairError.js';
       i++;
-      const isEndOfString = stopAtDelimiter ? i => isDelimiter(text[i]) : i => isEndQuote(text.charCodeAt(i));
-      while (i < text.length && !isEndOfString(i)) {
-        if (text.charCodeAt(i) === codeBackslash) {
+      while (true) {
+        if (i >= text.length) {
+          // end of text, we are missing an end quote
+          if (!stopAtDelimiter) {
+            // retry parsing the string, stopping at the first next delimiter
+            i = iBefore;
+            output = output.substring(0, oBefore);
+            return parseString(true);
+          }
+
+          // repair missing quote
+          str = insertBeforeLastWhitespace(str, '"');
+          output += str;
+          return true;
+        } else if (isEndQuote(text.charCodeAt(i))) {
+          // end quote
+          // let us check what is before and after the quote to verify whether this is a legit end quote
+          const iQuote = i;
+          const oQuote = str.length;
+          str += '"';
+          i++;
+          output += str;
+          parseWhitespaceAndSkipComments();
+          if (stopAtDelimiter || i >= text.length || isDelimiter(text.charAt(i)) || isQuote(text.charCodeAt(i))) {
+            // The quote is followed by a delimiter or the end of the text,
+            // so the quote is indeed the end of the string
+            parseConcatenatedString();
+            return true;
+          }
+          if (isDelimiter(text.charAt(prevNonWhitespaceIndex(iQuote - 1)))) {
+            // This is not the right end quote: it is preceded by a delimiter,
+            // and NOT followed by a delimiter. So, there is an end quote missing
+            // parse the string again and then stop at the first next delimiter
+            i = iBefore;
+            output = output.substring(0, oBefore);
+            return parseString(true);
+          }
+
+          // revert to right after the quote but before any whitespace, and continue parsing the string
+          output = output.substring(0, oBefore);
+          i = iQuote + 1;
+
+          // repair unescaped quote
+          str = str.substring(0, oQuote) + '\\' + str.substring(oQuote);
+        } else if (stopAtDelimiter && isDelimiter(text[i])) {
+          // we're in the mode to stop the string at the first delimiter
+          // because there is an end quote missing
+
+          // repair missing quote
+          str = insertBeforeLastWhitespace(str, '"');
+          output += str;
+          parseConcatenatedString();
+          return true;
+        } else if (text.charCodeAt(i) === codeBackslash) {
+          // handle escaped content like \n or \u2605
           const char = text.charAt(i + 1);
@@ -343,2 +395,3 @@ const escapeChar = escapeCharacters[char];
         } else {
+          // handle regular characters
           const char = text.charAt(i);
@@ -363,33 +416,6 @@ const code = text.charCodeAt(i);
         if (skipEscapeChars) {
-          const processed = skipEscapeCharacter();
-          if (processed) {
-            // repair: skipped escape character (nothing to do)
-          }
+          // repair: skipped escape character (nothing to do)
+          skipEscapeCharacter();
         }
       }
-      const hasEndQuote = isQuote(text.charCodeAt(i));
-      if (hasEndQuote) {
-        str += '"';
-        i++;
-      } else {
-        // repair missing quote
-        str = insertBeforeLastWhitespace(str, '"');
-      }
-      output += str;
-      parseWhitespaceAndSkipComments();
-
-      // See whether we have:
-      // (a) An end quote which is not followed by a valid delimiter
-      // (b) No end quote and reached the end of the input
-      // If so, revert parsing this string and try again, running in a more
-      // conservative mode, stopping at the first next delimiter
-      const isAtEnd = i >= text.length;
-      const nextIsDelimiter = isDelimiter(text.charAt(i));
-      if (!stopAtDelimiter && (hasEndQuote && !isAtEnd && !nextIsDelimiter || !hasEndQuote && isAtEnd)) {
-        i = iBefore;
-        output = output.substring(0, oBefore);
-        return parseString(true);
-      }
-      parseConcatenatedString();
-      return true;
     }
@@ -501,3 +527,3 @@ return false;
     const start = i;
-    while (i < text.length && !isDelimiter(text[i])) {
+    while (i < text.length && !isDelimiter(text[i]) && !isQuote(text.charCodeAt(i))) {
       i++;
@@ -538,2 +564,9 @@ }
   }
+  function prevNonWhitespaceIndex(start) {
+    let prev = start;
+    while (prev > 0 && isWhitespace(text.charCodeAt(prev))) {
+      prev--;
+    }
+    return prev;
+  }
   function expectDigit(start) {
@@ -540,0 +573,0 @@ if (!isDigit(text.charCodeAt(i))) {

lib/esm/streaming/buffer/OutputBuffer.js

@@ -48,2 +48,8 @@ import { isWhitespace } from '../../utils/stringUtils.js';
   }
+  function insertAt(index, text) {
+    if (index < offset) {
+      throw new Error("Cannot insert: ".concat(flushedMessage));
+    }
+    buffer = buffer.substring(0, index - offset) + text + buffer.substring(index - offset);
+  }
   function length() {
@@ -97,2 +103,3 @@ return offset + buffer.length;
     remove,
+    insertAt,
     length,
@@ -99,0 +106,0 @@ flush,

lib/esm/streaming/core.js

@@ -396,5 +396,2 @@ import { createInputBuffer } from './buffer/InputBuffer.js';
     let stopAtDelimiter = arguments.length > 0 && arguments[0] !== undefined ? arguments[0] : false;
-    // we may need to revert
-    const iBefore = i;
-    const oBefore = output.length();
     let skipEscapeChars = input.charCodeAt(i) === codeBackslash;
@@ -411,12 +408,58 @@ if (skipEscapeChars) {
       // or any single-quote-like start with a single-quote-like end
-      const isEndQuote = isDoubleQuote(input.charCodeAt(i)) ? isDoubleQuote : isSingleQuote(input.charCodeAt(i)) ? isSingleQuote // eslint-disable-line indent
-      : isSingleQuoteLike(input.charCodeAt(i)) // eslint-disable-line indent
-      ? isSingleQuoteLike // eslint-disable-line indent
-      : isDoubleQuoteLike; // eslint-disable-line indent
-
+      const isEndQuote = isDoubleQuote(input.charCodeAt(i)) ? isDoubleQuote : isSingleQuote(input.charCodeAt(i)) ? isSingleQuote : isSingleQuoteLike(input.charCodeAt(i)) ? isSingleQuoteLike : isDoubleQuoteLike;
+      const iBefore = i;
+      const oBefore = output.length();
       output.push('"');
       i++;
-      const isEndOfString = stopAtDelimiter ? i => isDelimiter(input.charAt(i)) : i => isEndQuote(input.charCodeAt(i));
-      while (!input.isEnd(i) && !isEndOfString(i)) {
-        if (input.charCodeAt(i) === codeBackslash) {
+      while (true) {
+        if (input.isEnd(i)) {
+          // end of text, we have a missing quote somewhere
+          if (!stopAtDelimiter) {
+            i = iBefore;
+            output.remove(oBefore);
+            return parseString(true);
+          }
+
+          // repair missing quote
+          output.insertBeforeLastWhitespace('"');
+          return stack.update(Caret.afterValue);
+        } else if (isEndQuote(input.charCodeAt(i))) {
+          // end quote
+          // let us check what is before and after the quote to verify whether this is a legit end quote
+          const iQuote = i;
+          const oQuote = output.length();
+          output.push('"');
+          i++;
+          parseWhitespaceAndSkipComments();
+          if (stopAtDelimiter || input.isEnd(i) || isDelimiter(input.charAt(i)) || isQuote(input.charCodeAt(i))) {
+            // The quote is followed by a delimiter or the end of the text,
+            // so the quote is indeed the end of the string
+            parseConcatenatedString();
+            return stack.update(Caret.afterValue);
+          }
+          if (isDelimiter(input.charAt(prevNonWhitespaceIndex(iQuote - 1)))) {
+            // This is not the right end quote: it is preceded by a delimiter,
+            // and NOT followed by a delimiter. So, there is an end quote missing
+            // parse the string again and then stop at the first next delimiter
+            i = iBefore;
+            output.remove(oBefore);
+            return parseString(true);
+          }
+
+          // revert to right after the quote but before any whitespace, and continue parsing the string
+          output.remove(oQuote + 1);
+          i = iQuote + 1;
+
+          // repair unescaped quote
+          output.insertAt(oQuote, '\\');
+        } else if (stopAtDelimiter && isDelimiter(input.charAt(i))) {
+          // we're in the mode to stop the string at the first delimiter
+          // because there is an end quote missing
+
+          // repair missing quote
+          output.insertBeforeLastWhitespace('"');
+          parseConcatenatedString();
+          return stack.update(Caret.afterValue);
+        } else if (input.charCodeAt(i) === codeBackslash) {
+          // handle escaped content like \n or \u2605
           const char = input.charAt(i + 1);
@@ -448,2 +491,3 @@ const escapeChar = escapeCharacters[char];
         } else {
+          // handle regular characters
           const char = input.charAt(i);
@@ -472,26 +516,2 @@ const code = char.charCodeAt(0);
       }
-      const hasEndQuote = isQuote(input.charCodeAt(i));
-      if (hasEndQuote) {
-        output.push('"');
-        i++;
-      } else {
-        // repair missing quote
-        output.insertBeforeLastWhitespace('"');
-      }
-      parseWhitespaceAndSkipComments();
-
-      // See whether we have:
-      // (a) An end quote which is not followed by a valid delimiter
-      // (b) No end quote and reached the end of the input
-      // If so, revert parsing this string and try again, running in a more
-      // conservative mode, stopping at the first next delimiter
-      const isAtEnd = input.isEnd(i);
-      const nextIsDelimiter = isDelimiter(input.charAt(i));
-      if (!stopAtDelimiter && (hasEndQuote && !isAtEnd && !nextIsDelimiter || !hasEndQuote && isAtEnd)) {
-        i = iBefore;
-        output.remove(oBefore);
-        return parseString(true);
-      }
-      parseConcatenatedString();
-      return stack.update(Caret.afterValue);
     }
@@ -615,3 +635,3 @@ return false;
     let j = i;
-    while (!input.isEnd(j) && !isDelimiter(input.charAt(j))) {
+    while (!input.isEnd(j) && !isDelimiter(input.charAt(j)) && !isQuote(input.charCodeAt(j))) {
       j++;
@@ -621,2 +641,9 @@ }
   }
+  function prevNonWhitespaceIndex(start) {
+    let prev = start;
+    while (prev > 0 && isWhitespace(input.charCodeAt(prev))) {
+      prev--;
+    }
+    return prev;
+  }
   function expectDigit(start) {
@@ -623,0 +650,0 @@ if (!isDigit(input.charCodeAt(i))) {

lib/esm/utils/stringUtils.js

@@ -55,5 +55,5 @@ export const codeBackslash = 0x5c; // "\"
 export function isDelimiter(char) {
-  return regexDelimiter.test(char) || isQuote(char.charCodeAt(0));
+  return regexDelimiter.test(char);
 }
-const regexDelimiter = /^[,:[\]{}()\n+]$/;
+const regexDelimiter = /^[,:[\]/{}()\n+]$/;
 export function isStartOfValue(char) {
@@ -60,0 +60,0 @@ return regexStartOfValue.test(char) || char && isQuote(char.charCodeAt(0));

lib/types/streaming/buffer/OutputBuffer.d.ts

@@ -5,2 +5,3 @@ export interface OutputBuffer {
     remove: (start: number, end?: number) => void;
+    insertAt: (index: number, text: string) => void;
     length: () => number;
@@ -7,0 +8,0 @@ flush: () => void;

lib/umd/jsonrepair.js

@@ -68,5 +68,5 @@ (function (global, factory) {
   function isDelimiter(char) {
-    return regexDelimiter.test(char) || isQuote(char.charCodeAt(0));
+    return regexDelimiter.test(char);
   }
-  const regexDelimiter = /^[,:[\]{}()\n+]$/;
+  const regexDelimiter = /^[,:[\]/{}()\n+]$/;
   function isStartOfValue(char) {
@@ -480,5 +480,57 @@ return regexStartOfValue.test(char) || char && isQuote(char.charCodeAt(0));
         i++;
-        const isEndOfString = stopAtDelimiter ? i => isDelimiter(text[i]) : i => isEndQuote(text.charCodeAt(i));
-        while (i < text.length && !isEndOfString(i)) {
-          if (text.charCodeAt(i) === codeBackslash) {
+        while (true) {
+          if (i >= text.length) {
+            // end of text, we are missing an end quote
+            if (!stopAtDelimiter) {
+              // retry parsing the string, stopping at the first next delimiter
+              i = iBefore;
+              output = output.substring(0, oBefore);
+              return parseString(true);
+            }
+
+            // repair missing quote
+            str = insertBeforeLastWhitespace(str, '"');
+            output += str;
+            return true;
+          } else if (isEndQuote(text.charCodeAt(i))) {
+            // end quote
+            // let us check what is before and after the quote to verify whether this is a legit end quote
+            const iQuote = i;
+            const oQuote = str.length;
+            str += '"';
+            i++;
+            output += str;
+            parseWhitespaceAndSkipComments();
+            if (stopAtDelimiter || i >= text.length || isDelimiter(text.charAt(i)) || isQuote(text.charCodeAt(i))) {
+              // The quote is followed by a delimiter or the end of the text,
+              // so the quote is indeed the end of the string
+              parseConcatenatedString();
+              return true;
+            }
+            if (isDelimiter(text.charAt(prevNonWhitespaceIndex(iQuote - 1)))) {
+              // This is not the right end quote: it is preceded by a delimiter,
+              // and NOT followed by a delimiter. So, there is an end quote missing
+              // parse the string again and then stop at the first next delimiter
+              i = iBefore;
+              output = output.substring(0, oBefore);
+              return parseString(true);
+            }
+
+            // revert to right after the quote but before any whitespace, and continue parsing the string
+            output = output.substring(0, oBefore);
+            i = iQuote + 1;
+
+            // repair unescaped quote
+            str = str.substring(0, oQuote) + '\\' + str.substring(oQuote);
+          } else if (stopAtDelimiter && isDelimiter(text[i])) {
+            // we're in the mode to stop the string at the first delimiter
+            // because there is an end quote missing
+
+            // repair missing quote
+            str = insertBeforeLastWhitespace(str, '"');
+            output += str;
+            parseConcatenatedString();
+            return true;
+          } else if (text.charCodeAt(i) === codeBackslash) {
+            // handle escaped content like \n or \u2605
             const char = text.charAt(i + 1);
@@ -510,2 +562,3 @@ const escapeChar = escapeCharacters[char];
           } else {
+            // handle regular characters
             const char = text.charAt(i);
@@ -530,30 +583,6 @@ const code = text.charCodeAt(i);
           if (skipEscapeChars) {
+            // repair: skipped escape character (nothing to do)
             skipEscapeCharacter();
           }
         }
-        const hasEndQuote = isQuote(text.charCodeAt(i));
-        if (hasEndQuote) {
-          str += '"';
-          i++;
-        } else {
-          // repair missing quote
-          str = insertBeforeLastWhitespace(str, '"');
-        }
-        output += str;
-        parseWhitespaceAndSkipComments();
-
-        // See whether we have:
-        // (a) An end quote which is not followed by a valid delimiter
-        // (b) No end quote and reached the end of the input
-        // If so, revert parsing this string and try again, running in a more
-        // conservative mode, stopping at the first next delimiter
-        const isAtEnd = i >= text.length;
-        const nextIsDelimiter = isDelimiter(text.charAt(i));
-        if (!stopAtDelimiter && (hasEndQuote && !isAtEnd && !nextIsDelimiter || !hasEndQuote && isAtEnd)) {
-          i = iBefore;
-          output = output.substring(0, oBefore);
-          return parseString(true);
-        }
-        parseConcatenatedString();
-        return true;
       }
@@ -665,3 +694,3 @@ return false;
       const start = i;
-      while (i < text.length && !isDelimiter(text[i])) {
+      while (i < text.length && !isDelimiter(text[i]) && !isQuote(text.charCodeAt(i))) {
         i++;
@@ -702,2 +731,9 @@ }
     }
+    function prevNonWhitespaceIndex(start) {
+      let prev = start;
+      while (prev > 0 && isWhitespace(text.charCodeAt(prev))) {
+        prev--;
+      }
+      return prev;
+    }
     function expectDigit(start) {
@@ -704,0 +740,0 @@ if (!isDigit(text.charCodeAt(i))) {

lib/umd/jsonrepair.min.js

@@ -1,1 +0,1 @@
-!function(t,e){"object"==typeof exports&&"undefined"!=typeof module?e(exports):"function"==typeof define&&define.amd?define(["exports"],e):e((t="undefined"!=typeof globalThis?globalThis:t||self).JSONRepair={})}(this,function(t){"use strict";class y extends Error{constructor(t,e){super(t+" at position "+e),this.position=e}}const u=125,e=32,O=10,N=9,J=13,S=8,j=12,I=34,r=39,k=48,m=57,T=65,$=97,E=70,R=102,h=160,d=8192,l=8202,s=8239,A=8287,C=12288,n=8220,o=8221,c=8216,i=8217,f=96,a=180;function U(t){return t>=k&&t<=m}function F(t){return g.test(t)||B(t.charCodeAt(0))}const g=/^[,:[\]{}()\n+]$/;function q(t){return v.test(t)||t&&B(t.charCodeAt(0))}const v=/^[[{\w-]$/;function z(t){return t===e||t===O||t===N||t===J}function B(t){return D(t)||H(t)}function D(t){return t===I||t===n||t===o}function G(t){return t===I}function H(t){return t===r||t===c||t===i||t===f||t===a}function K(t){return t===r}function L(t,e,r){r=2<arguments.length&&void 0!==r&&r,e=t.lastIndexOf(e);return-1!==e?t.substring(0,e)+(r?"":t.substring(e+1)):t}function M(t,e){let r=t.length;if(!z(t.charCodeAt(r-1)))return t+e;for(;z(t.charCodeAt(r-1));)r--;return t.substring(0,r)+e+t.substring(r)}const P={"\b":"\\b","\f":"\\f","\n":"\\n","\r":"\\r","\t":"\\t"},Q={'"':'"',"\\":"\\","/":"/",b:"\b",f:"\f",n:"\n",r:"\r",t:"\t"};t.JSONRepairError=y,t.jsonrepair=function(g){let v=0,b="";if(!c())throw new y("Unexpected end of json string",g.length);var t=i(44);if(t&&p(),q(g[v])&&/[,\n][ \t\r]*$/.test(b)){t||(b=M(b,","));{let t=!0,e=!0;for(;e;)t?t=!1:i(44)||(b=M(b,",")),e=c();e||(b=L(b,","));b="[\n".concat(b,"\n]")}}else t&&(b=L(b,","));for(;g.charCodeAt(v)===u||93===g.charCodeAt(v);)v++,p();if(v>=g.length)return b;throw new y("Unexpected character "+JSON.stringify(g[v]),v);function c(){p();var t=function(){if(123!==g.charCodeAt(v))return!1;{b+="{",v++,p();let e=!0;for(;v<g.length&&g.charCodeAt(v)!==u;){let t;if(e?(t=!0,e=!1):((t=i(44))||(b=M(b,",")),p()),!(x()||f())){g.charCodeAt(v)===u||123===g.charCodeAt(v)||93===g.charCodeAt(v)||91===g.charCodeAt(v)||void 0===g[v]?b=L(b,","):function(){throw new y("Object key expected",v)}();break}p();var r=i(58),n=v>=g.length,o=(r||(q(g[v])||n?b=M(b,":"):a()),c());o||(r||n?b+="null":a())}return g.charCodeAt(v)===u?(b+="}",v++):b=M(b,"}"),!0}}()||function(){if(91!==g.charCodeAt(v))return!1;{b+="[",v++,p();let t=!0;for(;v<g.length&&93!==g.charCodeAt(v);){t?t=!1:i(44)||(b=M(b,","));var e=c();if(!e){b=L(b,",");break}}return 93===g.charCodeAt(v)?(b+="]",v++):b=M(b,"]"),!0}}()||x()||function(){var t=v;if(45===g.charCodeAt(v)&&(v++,o(t)))return!0;for(;U(g.charCodeAt(v));)v++;if(46===g.charCodeAt(v)){if(v++,o(t))return!0;for(;U(g.charCodeAt(v));)v++}if(101===g.charCodeAt(v)||69===g.charCodeAt(v)){if(v++,45!==g.charCodeAt(v)&&43!==g.charCodeAt(v)||v++,o(t))return!0;for(;U(g.charCodeAt(v));)v++}{var e;if(v>t)return t=g.slice(t,v),e=/^0\d/.test(t),b+=e?'"'.concat(t,'"'):t,!0}return!1}()||r("true","true")||r("false","false")||r("null","null")||r("True","true")||r("False","false")||r("None","null")||f();return p(),t}function p(){v;let t=e();for(;t=(t=function(){if(47===g.charCodeAt(v)&&42===g.charCodeAt(v+1)){for(;v<g.length&&!function(t,e){return"*"===t[e]&&"/"===t[e+1]}(g,v);)v++;v+=2}else{if(47!==g.charCodeAt(v)||47!==g.charCodeAt(v+1))return!1;for(;v<g.length&&g.charCodeAt(v)!==O;)v++}return!0}())&&e(););v}function e(){let t="";for(var e,r;(e=z(g.charCodeAt(v)))||(r=g.charCodeAt(v))===h||r>=d&&r<=l||r===s||r===A||r===C;)t+=e?g[v]:" ",v++;return 0<t.length&&(b+=t,!0)}function i(t){return g.charCodeAt(v)===t&&(b+=g[v],v++,!0)}function w(){92===g.charCodeAt(v)&&v++}function x(t){var r,t=0<arguments.length&&void 0!==t&&t;let n=92===g.charCodeAt(v);if(n&&(v++,n=!0),B(g.charCodeAt(v))){const C=G(g.charCodeAt(v))?G:K(g.charCodeAt(v))?K:H(g.charCodeAt(v))?H:D;var o=v,c=b.length;let e='"';v++;for(var i=t?t=>F(g[t]):t=>C(g.charCodeAt(t));v<g.length&&!i(v);){if(92===g.charCodeAt(v)){var f=g.charAt(v+1);if(void 0!==Q[f])e+=g.slice(v,v+2),v+=2;else if("u"===f){let t=2;for(;t<6&&((r=g.charCodeAt(v+t))>=k&&r<=m||r>=T&&r<=E||r>=$&&r<=R);)t++;if(6===t)e+=g.slice(v,v+6),v+=6;else{if(!(v+t>=g.length))throw u=void 0,u=g.slice(v,v+6),new y('Invalid unicode character "'.concat(u,'"'),v);v=g.length}}else e+=f,v+=2}else{var a,u=g.charAt(v),f=g.charCodeAt(v);if(f===I&&92!==g.charCodeAt(v-1))e+="\\"+u;else if((a=f)===O||a===J||a===N||a===S||a===j)e+=P[u];else{if(!(32<=(a=f)&&a<=1114111))throw a=void 0,a=u,new y("Invalid character "+JSON.stringify(a),v);e+=u}v++}n&&w()}var h=B(g.charCodeAt(v)),d=(h?(e+='"',v++):e=M(e,'"'),b+=e,p(),v>=g.length),l=F(g.charAt(v));if(!t&&(h&&!d&&!l||!h&&d))return v=o,b=b.substring(0,c),x(!0);{let t=!1;p();for(;43===g.charCodeAt(v);){t=!0,v++,p();var s=(b=L(b,'"',!0)).length,A=x();b=A?function(t,e,r){return t.substring(0,e)+t.substring(e+r)}(b,s,1):M(b,'"')}t}return!0}return!1}function r(t,e){return g.slice(v,v+t.length)===t&&(b+=e,v+=t.length,!0)}function f(){for(var t=v;v<g.length&&!F(g[v]);)v++;if(v>t){if(40===g.charCodeAt(v))v++,c(),41===g.charCodeAt(v)&&(v++,59===g.charCodeAt(v))&&v++;else{for(;z(g.charCodeAt(v-1))&&0<v;)v--;t=g.slice(t,v);b+="undefined"===t?"null":JSON.stringify(t),g.charCodeAt(v)===I&&v++}return!0}}function n(t){if(!U(g.charCodeAt(v)))throw t=g.slice(t,v),new y("Invalid number '".concat(t,"', expecting a digit ").concat(g[v]?"but got '".concat(g[v],"'"):"but reached end of input"),v)}function o(t){if(v>=g.length)return b+=g.slice(t,v)+"0",1;n(t)}function a(){throw new y("Colon expected",v)}}});
+!function(t,e){"object"==typeof exports&&"undefined"!=typeof module?e(exports):"function"==typeof define&&define.amd?define(["exports"],e):e((t="undefined"!=typeof globalThis?globalThis:t||self).JSONRepair={})}(this,function(t){"use strict";class w extends Error{constructor(t,e){super(t+" at position "+e),this.position=e}}const a=125,e=32,x=10,y=9,O=13,N=8,J=12,S=34,r=39,j=48,I=57,k=65,m=97,T=70,$=102,h=160,d=8192,l=8202,E=8239,R=8287,U=12288,n=8220,o=8221,c=8216,i=8217,f=96,u=180;function F(t){return t>=j&&t<=I}function q(t){return s.test(t)}const s=/^[,:[\]/{}()\n+]$/;function z(t){return A.test(t)||t&&D(t.charCodeAt(0))}const A=/^[[{\w-]$/;function B(t){return t===e||t===x||t===y||t===O}function D(t){return G(t)||K(t)}function G(t){return t===S||t===n||t===o}function H(t){return t===S}function K(t){return t===r||t===c||t===i||t===f||t===u}function L(t){return t===r}function M(t,e,r){r=2<arguments.length&&void 0!==r&&r,e=t.lastIndexOf(e);return-1!==e?t.substring(0,e)+(r?"":t.substring(e+1)):t}function P(t,e){let r=t.length;if(!B(t.charCodeAt(r-1)))return t+e;for(;B(t.charCodeAt(r-1));)r--;return t.substring(0,r)+e+t.substring(r)}const Q={"\b":"\\b","\f":"\\f","\n":"\\n","\r":"\\r","\t":"\\t"},V={'"':'"',"\\":"\\","/":"/",b:"\b",f:"\f",n:"\n",r:"\r",t:"\t"};t.JSONRepairError=w,t.jsonrepair=function(s){let A=0,C="";if(!c())throw new w("Unexpected end of json string",s.length);var t=i(44);if(t&&g(),z(s[A])&&/[,\n][ \t\r]*$/.test(C)){t||(C=P(C,","));{let t=!0,e=!0;for(;e;)t?t=!1:i(44)||(C=P(C,",")),e=c();e||(C=M(C,","));C="[\n".concat(C,"\n]")}}else t&&(C=M(C,","));for(;s.charCodeAt(A)===a||93===s.charCodeAt(A);)A++,g();if(A>=s.length)return C;throw new w("Unexpected character "+JSON.stringify(s[A]),A);function c(){g();var t=function(){if(123!==s.charCodeAt(A))return!1;{C+="{",A++,g();let e=!0;for(;A<s.length&&s.charCodeAt(A)!==a;){let t;if(e?(t=!0,e=!1):((t=i(44))||(C=P(C,",")),g()),!(v()||f())){s.charCodeAt(A)===a||123===s.charCodeAt(A)||93===s.charCodeAt(A)||91===s.charCodeAt(A)||void 0===s[A]?C=M(C,","):function(){throw new w("Object key expected",A)}();break}g();var r=i(58),n=A>=s.length,o=(r||(z(s[A])||n?C=P(C,":"):u()),c());o||(r||n?C+="null":u())}return s.charCodeAt(A)===a?(C+="}",A++):C=P(C,"}"),!0}}()||function(){if(91!==s.charCodeAt(A))return!1;{C+="[",A++,g();let t=!0;for(;A<s.length&&93!==s.charCodeAt(A);){t?t=!1:i(44)||(C=P(C,","));var e=c();if(!e){C=M(C,",");break}}return 93===s.charCodeAt(A)?(C+="]",A++):C=P(C,"]"),!0}}()||v()||function(){var t=A;if(45===s.charCodeAt(A)&&(A++,o(t)))return!0;for(;F(s.charCodeAt(A));)A++;if(46===s.charCodeAt(A)){if(A++,o(t))return!0;for(;F(s.charCodeAt(A));)A++}if(101===s.charCodeAt(A)||69===s.charCodeAt(A)){if(A++,45!==s.charCodeAt(A)&&43!==s.charCodeAt(A)||A++,o(t))return!0;for(;F(s.charCodeAt(A));)A++}{var e;if(A>t)return t=s.slice(t,A),e=/^0\d/.test(t),C+=e?'"'.concat(t,'"'):t,!0}return!1}()||r("true","true")||r("false","false")||r("null","null")||r("True","true")||r("False","false")||r("None","null")||f();return g(),t}function g(){A;let t=e();for(;t=(t=function(){if(47===s.charCodeAt(A)&&42===s.charCodeAt(A+1)){for(;A<s.length&&!function(t,e){return"*"===t[e]&&"/"===t[e+1]}(s,A);)A++;A+=2}else{if(47!==s.charCodeAt(A)||47!==s.charCodeAt(A+1))return!1;for(;A<s.length&&s.charCodeAt(A)!==x;)A++}return!0}())&&e(););A}function e(){let t="";for(var e,r;(e=B(s.charCodeAt(A)))||(r=s.charCodeAt(A))===h||r>=d&&r<=l||r===E||r===R||r===U;)t+=e?s[A]:" ",A++;return 0<t.length&&(C+=t,!0)}function i(t){return s.charCodeAt(A)===t&&(C+=s[A],A++,!0)}function b(){92===s.charCodeAt(A)&&A++}function v(t){var r,n,o=0<arguments.length&&void 0!==t&&t;let c=92===s.charCodeAt(A);if(c&&(A++,c=!0),D(s.charCodeAt(A))){var i=H(s.charCodeAt(A))?H:L(s.charCodeAt(A))?L:K(s.charCodeAt(A))?K:G,f=A,u=C.length;let e='"';for(A++;;){if(A>=s.length)return o?(e=P(e,'"'),C+=e,!0):(A=f,C=C.substring(0,u),v(!0));if(i(s.charCodeAt(A))){var a=A,h=e.length;if(e+='"',A++,C+=e,g(),o||A>=s.length||q(s.charAt(A))||D(s.charCodeAt(A)))return p(),!0;if(q(s.charAt(function(t){let e=t;for(;0<e&&B(s.charCodeAt(e));)e--;return e}(a-1))))return A=f,C=C.substring(0,u),v(!0);C=C.substring(0,u),A=a+1,e=e.substring(0,h)+"\\"+e.substring(h)}else{if(o&&q(s[A]))return e=P(e,'"'),C+=e,p(),!0;if(92===s.charCodeAt(A)){a=s.charAt(A+1);if(void 0!==V[a])e+=s.slice(A,A+2),A+=2;else if("u"===a){let t=2;for(;t<6&&((n=s.charCodeAt(A+t))>=j&&n<=I||n>=k&&n<=T||n>=m&&n<=$);)t++;if(6===t)e+=s.slice(A,A+6),A+=6;else{if(!(A+t>=s.length))throw h=void 0,h=s.slice(A,A+6),new w('Invalid unicode character "'.concat(h,'"'),A);A=s.length}}else e+=a,A+=2}else{var d=s.charAt(A),l=s.charCodeAt(A);if(l===S&&92!==s.charCodeAt(A-1))e+="\\"+d;else if((r=l)===x||r===O||r===y||r===N||r===J)e+=Q[d];else{if(!(32<=(r=l)&&r<=1114111))throw l=void 0,l=d,new w("Invalid character "+JSON.stringify(l),A);e+=d}A++}}c&&b()}}return!1}function p(){let t=!1;for(g();43===s.charCodeAt(A);){t=!0,A++,g();var e=(C=M(C,'"',!0)).length,r=v();C=r?(r=C,e=e,n=1,r.substring(0,e)+r.substring(e+n)):P(C,'"')}var n;t}function r(t,e){return s.slice(A,A+t.length)===t&&(C+=e,A+=t.length,!0)}function f(){for(var t=A;A<s.length&&!q(s[A])&&!D(s.charCodeAt(A));)A++;if(A>t){if(40===s.charCodeAt(A))A++,c(),41===s.charCodeAt(A)&&(A++,59===s.charCodeAt(A))&&A++;else{for(;B(s.charCodeAt(A-1))&&0<A;)A--;t=s.slice(t,A);C+="undefined"===t?"null":JSON.stringify(t),s.charCodeAt(A)===S&&A++}return!0}}function n(t){if(!F(s.charCodeAt(A)))throw t=s.slice(t,A),new w("Invalid number '".concat(t,"', expecting a digit ").concat(s[A]?"but got '".concat(s[A],"'"):"but reached end of input"),A)}function o(t){if(A>=s.length)return C+=s.slice(t,A)+"0",1;n(t)}function u(){throw new w("Colon expected",A)}}});

package.json

 {
   "name": "jsonrepair",
-  "version": "3.5.1",
+  "version": "3.6.0",
   "description": "Repair broken JSON documents",
@@ -60,3 +60,3 @@ "repository": {
     "release-dry-run": "npm run build-and-test && standard-version --dry-run",
-    "prepare": "husky install"
+    "prepare": "husky"
   },
@@ -71,12 +71,12 @@ "files": [
   "devDependencies": {
-    "@babel/cli": "7.23.4",
-    "@babel/core": "7.23.7",
+    "@babel/cli": "7.23.9",
+    "@babel/core": "7.23.9",
     "@babel/plugin-transform-typescript": "7.23.6",
-    "@babel/preset-env": "7.23.8",
+    "@babel/preset-env": "7.23.9",
     "@babel/preset-typescript": "7.23.3",
-    "@commitlint/cli": "18.4.4",
-    "@commitlint/config-conventional": "18.4.4",
-    "@types/node": "20.10.8",
-    "@typescript-eslint/eslint-plugin": "6.18.1",
-    "@typescript-eslint/parser": "6.18.1",
+    "@commitlint/cli": "18.6.0",
+    "@commitlint/config-conventional": "18.6.0",
+    "@types/node": "20.11.17",
+    "@typescript-eslint/eslint-plugin": "7.0.1",
+    "@typescript-eslint/parser": "7.0.1",
     "benchmark": "2.1.4",
@@ -91,6 +91,5 @@ "cpy-cli": "5.0.0",
     "eslint-plugin-promise": "6.1.1",
-    "husky": "8.0.3",
+    "husky": "9.0.10",
     "npm-run-all": "4.1.5",
-    "prettier": "3.1.1",
-    "rollup": "4.9.4",
+    "rollup": "4.10.0",
     "standard-version": "9.5.0",
@@ -100,4 +99,4 @@ "ts-node": "10.9.2",
     "uglify-js": "3.17.4",
-    "vitest": "1.1.3"
+    "vitest": "1.2.2"
   }
 }

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

469432

increased by4.29%

Dev dependency count

27

decreased by-3.57%

Lines of code

4491

increased by3.67%

No dependency changes