bson 6.5.0 → 6.7.0

package/src/utils/byte_utils.ts

@@ -1,10 +1,16 @@
 import { nodeJsByteUtils } from './node_byte_utils';
 import { webByteUtils } from './web_byte_utils';
 
-/** @internal */
+/**
+ * @public
+ * @experimental
+ *
+ * A collection of functions that help work with data in a Uint8Array.
+ * ByteUtils is configured at load time to use Node.js or Web based APIs for the internal implementations.
+ */
 export type ByteUtils = {
   /** Transforms the input to an instance of Buffer if running on node, otherwise Uint8Array */
-  toLocalBufferType(buffer: Uint8Array | ArrayBufferView | ArrayBuffer): Uint8Array;
+  toLocalBufferType: (buffer: Uint8Array | ArrayBufferView | ArrayBuffer) => Uint8Array;
   /** Create empty space of size */
   allocate: (size: number) => Uint8Array;
   /** Create empty space of size, use pooled memory when available */
@@ -30,9 +36,9 @@ export type ByteUtils = {
   /** Get the utf8 code unit count from a string if it were to be transformed to utf8 */
   utf8ByteLength: (input: string) => number;
   /** Encode UTF8 bytes generated from `source` string into `destination` at byteOffset. Returns the number of bytes encoded. */
-  encodeUTF8Into(destination: Uint8Array, source: string, byteOffset: number): number;
+  encodeUTF8Into: (destination: Uint8Array, source: string, byteOffset: number) => number;
   /** Generate a Uint8Array filled with random bytes with byteLength */
-  randomBytes(byteLength: number): Uint8Array;
+  randomBytes: (byteLength: number) => Uint8Array;
 };
 
 declare const Buffer: { new (): unknown; prototype?: { _isBuffer?: boolean } } | undefined;

package/src/utils/node_byte_utils.ts

@@ -1,5 +1,5 @@
 import { BSONError } from '../error';
-import { validateUtf8 } from '../validate_utf8';
+import { parseUtf8 } from '../parse_utf8';
 import { tryReadBasicLatin, tryWriteBasicLatin } from './latin';
 
 type NodeJsEncoding = 'base64' | 'hex' | 'utf8' | 'binary';
@@ -136,12 +136,9 @@ export const nodeJsByteUtils = {
 
     const string = nodeJsByteUtils.toLocalBufferType(buffer).toString('utf8', start, end);
     if (fatal) {
-      // TODO(NODE-4930): Insufficiently strict BSON UTF8 validation
       for (let i = 0; i < string.length; i++) {
         if (string.charCodeAt(i) === 0xfffd) {
-          if (!validateUtf8(buffer, start, end)) {
-            throw new BSONError('Invalid UTF-8 string in BSON document');
-          }
+          parseUtf8(buffer, start, end, true);
           break;
         }
       }

package/src/utils/number_utils.ts

@@ -6,12 +6,47 @@ FLOAT[0] = -1;
 // Big endian    [191, 240, 0, 0, 0, 0, 0, 0]
 const isBigEndian = FLOAT_BYTES[7] === 0;
 
+/**
+ * @experimental
+ * @public
+ *
+ * A collection of functions that get or set various numeric types and bit widths from a Uint8Array.
+ */
+export type NumberUtils = {
+  /**
+   * Parses a signed int32 at offset. Throws a `RangeError` if value is negative.
+   */
+  getNonnegativeInt32LE: (source: Uint8Array, offset: number) => number;
+  getInt32LE: (source: Uint8Array, offset: number) => number;
+  getUint32LE: (source: Uint8Array, offset: number) => number;
+  getUint32BE: (source: Uint8Array, offset: number) => number;
+  getBigInt64LE: (source: Uint8Array, offset: number) => bigint;
+  getFloat64LE: (source: Uint8Array, offset: number) => number;
+  setInt32BE: (destination: Uint8Array, offset: number, value: number) => 4;
+  setInt32LE: (destination: Uint8Array, offset: number, value: number) => 4;
+  setBigInt64LE: (destination: Uint8Array, offset: number, value: bigint) => 8;
+  setFloat64LE: (destination: Uint8Array, offset: number, value: number) => 8;
+};
+
 /**
  * Number parsing and serializing utilities.
  *
- * @internal
+ * @experimental
+ * @public
  */
-export const NumberUtils = {
+export const NumberUtils: NumberUtils = {
+  getNonnegativeInt32LE(source: Uint8Array, offset: number): number {
+    if (source[offset + 3] > 127) {
+      throw new RangeError(`Size cannot be negative at offset: ${offset}`);
+    }
+    return (
+      source[offset] |
+      (source[offset + 1] << 8) |
+      (source[offset + 2] << 16) |
+      (source[offset + 3] << 24)
+    );
+  },
+
   /** Reads a little-endian 32-bit integer from source */
   getInt32LE(source: Uint8Array, offset: number): number {
     return (

package/src/utils/string_utils.ts

@@ -0,0 +1,44 @@
+/**
+ * @internal
+ * Removes leading zeros and explicit plus from textual representation of a number.
+ */
+export function removeLeadingZerosAndExplicitPlus(str: string): string {
+  if (str === '') {
+    return str;
+  }
+
+  let startIndex = 0;
+
+  const isNegative = str[startIndex] === '-';
+  const isExplicitlyPositive = str[startIndex] === '+';
+
+  if (isExplicitlyPositive || isNegative) {
+    startIndex += 1;
+  }
+
+  let foundInsignificantZero = false;
+
+  for (; startIndex < str.length && str[startIndex] === '0'; ++startIndex) {
+    foundInsignificantZero = true;
+  }
+
+  if (!foundInsignificantZero) {
+    return isExplicitlyPositive ? str.slice(1) : str;
+  }
+
+  return `${isNegative ? '-' : ''}${str.length === startIndex ? '0' : str.slice(startIndex)}`;
+}
+
+/**
+ * @internal
+ * Returns false for an string that contains invalid characters for its radix, else returns the original string.
+ * @param str - The textual representation of the Long
+ * @param radix - The radix in which the text is written (2-36), defaults to 10
+ */
+export function validateStringCharacters(str: string, radix?: number): false | string {
+  radix = radix ?? 10;
+  const validCharacters = '0123456789abcdefghijklmnopqrstuvwxyz'.slice(0, radix);
+  // regex is case insensitive and checks that each character within the string is one of the validCharacters
+  const regex = new RegExp(`[^-+${validCharacters}]`, 'i');
+  return regex.test(str) ? false : str;
+}

package/src/utils/web_byte_utils.ts

@@ -1,5 +1,6 @@
 import { BSONError } from '../error';
 import { tryReadBasicLatin } from './latin';
+import { parseUtf8 } from '../parse_utf8';
 
 type TextDecoder = {
   readonly encoding: string;
@@ -179,14 +180,7 @@ export const webByteUtils = {
       return basicLatin;
     }
 
-    if (fatal) {
-      try {
-        return new TextDecoder('utf8', { fatal }).decode(uint8array.slice(start, end));
-      } catch (cause) {
-        throw new BSONError('Invalid UTF-8 string in BSON document', { cause });
-      }
-    }
-    return new TextDecoder('utf8', { fatal }).decode(uint8array.slice(start, end));
+    return parseUtf8(uint8array, start, end, fatal);
   },
 
   utf8ByteLength(input: string): number {

package/src/validate_utf8.ts

@@ -1,47 +0,0 @@
-const FIRST_BIT = 0x80;
-const FIRST_TWO_BITS = 0xc0;
-const FIRST_THREE_BITS = 0xe0;
-const FIRST_FOUR_BITS = 0xf0;
-const FIRST_FIVE_BITS = 0xf8;
-
-const TWO_BIT_CHAR = 0xc0;
-const THREE_BIT_CHAR = 0xe0;
-const FOUR_BIT_CHAR = 0xf0;
-const CONTINUING_CHAR = 0x80;
-
-/**
- * Determines if the passed in bytes are valid utf8
- * @param bytes - An array of 8-bit bytes. Must be indexable and have length property
- * @param start - The index to start validating
- * @param end - The index to end validating
- */
-export function validateUtf8(
-  bytes: { [index: number]: number },
-  start: number,
-  end: number
-): boolean {
-  let continuation = 0;
-
-  for (let i = start; i < end; i += 1) {
-    const byte = bytes[i];
-
-    if (continuation) {
-      if ((byte & FIRST_TWO_BITS) !== CONTINUING_CHAR) {
-        return false;
-      }
-      continuation -= 1;
-    } else if (byte & FIRST_BIT) {
-      if ((byte & FIRST_THREE_BITS) === TWO_BIT_CHAR) {
-        continuation = 1;
-      } else if ((byte & FIRST_FOUR_BITS) === THREE_BIT_CHAR) {
-        continuation = 2;
-      } else if ((byte & FIRST_FIVE_BITS) === FOUR_BIT_CHAR) {
-        continuation = 3;
-      } else {
-        return false;
-      }
-    }
-  }
-
-  return !continuation;
-}