cborg 4.5.4 → 4.5.6

package/CHANGELOG.md

@@ -1,3 +1,11 @@
+## [4.5.6](https://github.com/rvagg/cborg/compare/v4.5.5...v4.5.6) (2026-01-21)
+
+## [4.5.5](https://github.com/rvagg/cborg/compare/v4.5.4...v4.5.5) (2026-01-20)
+
+### Trivial Changes
+
+* **bench:** output json to file ([ca32690](https://github.com/rvagg/cborg/commit/ca326908643ce9bc2ac56dd06b6c64502b8a4f03))
+
 ## [4.5.4](https://github.com/rvagg/cborg/compare/v4.5.3...v4.5.4) (2026-01-20)
 
 ## [4.5.3](https://github.com/rvagg/cborg/compare/v4.5.2...v4.5.3) (2026-01-20)

package/bench/bench.js

@@ -5,12 +5,12 @@
  * Works in both Node.js and browser environments.
  *
  * Usage:
- *   node bench/bench-new.js                    # run all benchmarks (dag-cbor mode)
- *   node bench/bench-new.js --mode=raw         # run with raw cborg (no tags)
- *   node bench/bench-new.js --suite=bsky       # run only bluesky suite
- *   node bench/bench-new.js --json             # output JSON for comparison
- *   node bench/bench-new.js --compare=baseline.json  # compare to baseline
- *   node bench/bench-new.js --encode-into      # use encodeInto instead of encode
+ *   node bench/bench.js                           # run all benchmarks (dag-cbor mode)
+ *   node bench/bench.js --mode=raw                # run with raw cborg (no tags)
+ *   node bench/bench.js --suite=bsky              # run only bluesky suite
+ *   node bench/bench.js --json=output.json        # write JSON results to file
+ *   node bench/bench.js --compare=baseline.json   # compare to baseline
+ *   node bench/bench.js --encode-into             # use encodeInto instead of encode
  */
 
 import { encode, decode, encodeInto, Token, Type } from '../cborg.js'
@@ -129,7 +129,7 @@ const FIXTURE_SEED = 12345
 // Parse CLI args (Node.js only, ignored in browser)
 const args = typeof process !== 'undefined' ? process.argv.slice(2) : []
 const opts = {
-  json: args.includes('--json'),
+  json: args.find(a => a.startsWith('--json='))?.split('=')[1] || null,
   suite: args.find(a => a.startsWith('--suite='))?.split('=')[1] || null,
   compare: args.find(a => a.startsWith('--compare='))?.split('=')[1] || null,
   duration: parseInt(args.find(a => a.startsWith('--duration='))?.split('=')[1] || DEFAULT_DURATION_MS),
@@ -184,11 +184,11 @@ function getOptions (suiteType = 'default') {
   }
 }
 
-// Output helpers
-const log = opts.json ? () => {} : console.log.bind(console)
+// Output helpers - always show progress to console
+const log = console.log.bind(console)
 const write = typeof process !== 'undefined' && process.stdout
   ? (s) => process.stdout.write(s)
-  : (s) => log(s)
+  : (s) => console.log(s)
 
 /**
  * Run a benchmark function for a duration, return ops/sec
@@ -364,7 +364,7 @@ async function main () {
   const avgDecode = Math.round(allDecodeRates.reduce((a, b) => a + b, 0) / allDecodeRates.length * 10) / 10
   log(`Average throughput: encode ${avgEncode} MB/s, decode ${avgDecode} MB/s`)
 
-  // JSON output
+  // JSON output to file
   if (opts.json) {
     const output = {
       timestamp: new Date().toISOString(),
@@ -375,7 +375,11 @@ async function main () {
       suites: allResults,
       summary: { avgEncodeMBps: avgEncode, avgDecodeMBps: avgDecode }
     }
-    console.log(JSON.stringify(output, null, 2))
+    if (typeof process !== 'undefined') {
+      const fs = await import('fs')
+      fs.writeFileSync(opts.json, JSON.stringify(output, null, 2) + '\n')
+      log(`\nResults written to ${opts.json}`)
+    }
   }
 
   // Compare to baseline

package/lib/2bytes.js

@@ -1,7 +1,7 @@
 import { Token, Type } from './token.js'
 import { assertEnoughData, decodeErrPrefix } from './common.js'
 import * as uint from './0uint.js'
-import { compare, fromString, slice } from './byte-utils.js'
+import { compare, fromString } from './byte-utils.js'
 
 /**
  * @typedef {import('../interface').ByteWriter} ByteWriter
@@ -17,7 +17,7 @@ import { compare, fromString, slice } from './byte-utils.js'
  */
 function toToken (data, pos, prefix, length) {
   assertEnoughData(data, pos, prefix + length)
-  const buf = slice(data, pos + prefix, pos + prefix + length)
+  const buf = data.slice(pos + prefix, pos + prefix + length)
   return new Token(Type.bytes, buf, prefix + length)
 }
 

package/lib/3string.js

@@ -2,13 +2,42 @@ import { Token, Type } from './token.js'
 import { assertEnoughData, decodeErrPrefix } from './common.js'
 import * as uint from './0uint.js'
 import { encodeBytes } from './2bytes.js'
-import { toString, slice } from './byte-utils.js'
+
+const textDecoder = new TextDecoder()
+
+// Threshold for ASCII fast-path vs TextDecoder. Short ASCII strings (common for
+// map keys) are faster to decode with a simple loop than TextDecoder overhead.
+const ASCII_THRESHOLD = 32
 
 /**
  * @typedef {import('../interface').ByteWriter} ByteWriter
  * @typedef {import('../interface').DecodeOptions} DecodeOptions
  */
 
+/**
+ * Decode UTF-8 bytes to string. For short ASCII strings (common case for map keys),
+ * a simple loop is faster than TextDecoder.
+ * @param {Uint8Array} bytes
+ * @param {number} start
+ * @param {number} end
+ * @returns {string}
+ */
+function toStr (bytes, start, end) {
+  const len = end - start
+  if (len < ASCII_THRESHOLD) {
+    let str = ''
+    for (let i = start; i < end; i++) {
+      const c = bytes[i]
+      if (c & 0x80) { // non-ASCII, fall back to TextDecoder
+        return textDecoder.decode(bytes.subarray(start, end))
+      }
+      str += String.fromCharCode(c)
+    }
+    return str
+  }
+  return textDecoder.decode(bytes.subarray(start, end))
+}
+
 /**
  * @param {Uint8Array} data
  * @param {number} pos
@@ -20,9 +49,9 @@ import { toString, slice } from './byte-utils.js'
 function toToken (data, pos, prefix, length, options) {
   const totLength = prefix + length
   assertEnoughData(data, pos, totLength)
-  const tok = new Token(Type.string, toString(data, pos + prefix, pos + totLength), totLength)
+  const tok = new Token(Type.string, toStr(data, pos + prefix, pos + totLength), totLength)
   if (options.retainStringBytes === true) {
-    tok.byteValue = slice(data, pos + prefix, pos + totLength)
+    tok.byteValue = data.slice(pos + prefix, pos + totLength)
   }
   return tok
 }

package/lib/byte-utils.js

@@ -10,7 +10,6 @@ export const useBuffer = globalThis.process &&
   // @ts-ignore
   typeof globalThis.Buffer.isBuffer === 'function'
 
-const textDecoder = new TextDecoder()
 const textEncoder = new TextEncoder()
 
 /**
@@ -34,37 +33,6 @@ export function asU8A (buf) {
   return isBuffer(buf) ? new Uint8Array(buf.buffer, buf.byteOffset, buf.byteLength) : buf
 }
 
-// Threshold for switching between manual utf8Slice and native TextDecoder/Buffer
-// Manual decoding has overhead from array building; native is fast for strings > 32 bytes
-const UTF8_THRESHOLD = 32
-
-export const toString = useBuffer
-  ? // eslint-disable-line operator-linebreak
-    /**
-     * @param {Uint8Array} bytes
-     * @param {number} start
-     * @param {number} end
-     */
-    (bytes, start, end) => {
-      return end - start > UTF8_THRESHOLD
-        ? // eslint-disable-line operator-linebreak
-      // @ts-ignore
-        globalThis.Buffer.from(bytes.subarray(start, end)).toString('utf8')
-        : utf8Slice(bytes, start, end)
-    }
-  /* c8 ignore next 11 */
-  : // eslint-disable-line operator-linebreak
-    /**
-     * @param {Uint8Array} bytes
-     * @param {number} start
-     * @param {number} end
-     */
-    (bytes, start, end) => {
-      return end - start > UTF8_THRESHOLD
-        ? textDecoder.decode(bytes.subarray(start, end))
-        : utf8Slice(bytes, start, end)
-    }
-
 export const fromString = useBuffer
   ? // eslint-disable-line operator-linebreak
     /**
@@ -102,6 +70,7 @@ export const slice = useBuffer
      * @param {number} start
      * @param {number} end
      */
+    // Buffer.slice() returns a view, not a copy, so we need special handling
     (bytes, start, end) => {
       if (isBuffer(bytes)) {
         return new Uint8Array(bytes.subarray(start, end))
@@ -317,85 +286,6 @@ function utf8ToBytes (str) {
   return out
 }
 
-// The below code is mostly taken from https://github.com/feross/buffer
-// Licensed MIT. Copyright (c) Feross Aboukhadijeh
-
-/**
- * @param {Uint8Array} buf
- * @param {number} offset
- * @param {number} end
- * @returns {string}
- */
-function utf8Slice (buf, offset, end) {
-  const res = []
-
-  while (offset < end) {
-    const firstByte = buf[offset]
-    let codePoint = null
-    let bytesPerSequence = (firstByte > 0xef) ? 4 : (firstByte > 0xdf) ? 3 : (firstByte > 0xbf) ? 2 : 1
-
-    if (offset + bytesPerSequence <= end) {
-      let secondByte, thirdByte, fourthByte, tempCodePoint
-
-      switch (bytesPerSequence) {
-        case 1:
-          if (firstByte < 0x80) {
-            codePoint = firstByte
-          }
-          break
-        case 2:
-          secondByte = buf[offset + 1]
-          if ((secondByte & 0xc0) === 0x80) {
-            tempCodePoint = (firstByte & 0x1f) << 0x6 | (secondByte & 0x3f)
-            if (tempCodePoint > 0x7f) {
-              codePoint = tempCodePoint
-            }
-          }
-          break
-        case 3:
-          secondByte = buf[offset + 1]
-          thirdByte = buf[offset + 2]
-          if ((secondByte & 0xc0) === 0x80 && (thirdByte & 0xc0) === 0x80) {
-            tempCodePoint = (firstByte & 0xf) << 0xc | (secondByte & 0x3f) << 0x6 | (thirdByte & 0x3f)
-            /* c8 ignore next 3 */
-            if (tempCodePoint > 0x7ff && (tempCodePoint < 0xd800 || tempCodePoint > 0xdfff)) {
-              codePoint = tempCodePoint
-            }
-          }
-          break
-        case 4:
-          secondByte = buf[offset + 1]
-          thirdByte = buf[offset + 2]
-          fourthByte = buf[offset + 3]
-          if ((secondByte & 0xc0) === 0x80 && (thirdByte & 0xc0) === 0x80 && (fourthByte & 0xc0) === 0x80) {
-            tempCodePoint = (firstByte & 0xf) << 0x12 | (secondByte & 0x3f) << 0xc | (thirdByte & 0x3f) << 0x6 | (fourthByte & 0x3f)
-            if (tempCodePoint > 0xffff && tempCodePoint < 0x110000) {
-              codePoint = tempCodePoint
-            }
-          }
-      }
-    }
-
-    /* c8 ignore next 5 */
-    if (codePoint === null) {
-      // we did not generate a valid codePoint so insert a
-      // replacement char (U+FFFD) and advance only 1 byte
-      codePoint = 0xfffd
-      bytesPerSequence = 1
-    } else if (codePoint > 0xffff) {
-      // encode to utf16 (surrogate pair dance)
-      codePoint -= 0x10000
-      res.push(codePoint >>> 10 & 0x3ff | 0xd800)
-      codePoint = 0xdc00 | codePoint & 0x3ff
-    }
-
-    res.push(codePoint)
-    offset += bytesPerSequence
-  }
-
-  return decodeCodePointsArray(res)
-}
-
 // Based on http://stackoverflow.com/a/22747272/680742, the browser with
 // the lowest limit is Chrome, with 0x10000 args.
 // We go 1 magnitude less, for safety

package/lib/decode.js

@@ -1,6 +1,7 @@
 import { decodeErrPrefix } from './common.js'
 import { Type } from './token.js'
 import { jump, quick } from './jump.js'
+import { asU8A } from './byte-utils.js'
 
 /**
  * @typedef {import('./token.js').Token} Token
@@ -183,7 +184,9 @@ function decodeFirst (data, options) {
     throw new Error(`${decodeErrPrefix} data to decode must be a Uint8Array`)
   }
   options = Object.assign({}, defaultDecodeOptions, options)
-  const tokeniser = options.tokenizer || new Tokeniser(data, options)
+  // Convert Buffer to plain Uint8Array for faster slicing in decode path
+  const u8aData = asU8A(data)
+  const tokeniser = options.tokenizer || new Tokeniser(u8aData, options)
   const decoded = tokensToObject(tokeniser, options)
   if (decoded === DONE) {
     throw new Error(`${decodeErrPrefix} did not find any content to decode`)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "cborg",
-  "version": "4.5.4",
+  "version": "4.5.6",
   "description": "Fast CBOR with a focus on strictness",
   "main": "cborg.js",
   "type": "module",

package/types/lib/3string.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"3string.d.ts","sourceRoot":"","sources":["../../lib/3string.js"],"names":[],"mappings":"AA6BA;;;;;;GAMG;AACH,0CANW,UAAU,OACV,MAAM,SACN,MAAM,WACN,aAAa,GACX,KAAK,CAIjB;AAED;;;;;;GAMG;AACH,oCANW,UAAU,OACV,MAAM,UACN,MAAM,WACN,aAAa,GACX,KAAK,CAIjB;AAED;;;;;;GAMG;AACH,qCANW,UAAU,OACV,MAAM,UACN,MAAM,WACN,aAAa,GACX,KAAK,CAIjB;AAED;;;;;;GAMG;AACH,qCANW,UAAU,OACV,MAAM,UACN,MAAM,WACN,aAAa,GACX,KAAK,CAIjB;AAGD;;;;;;GAMG;AACH,qCANW,UAAU,OACV,MAAM,UACN,MAAM,WACN,aAAa,GACX,KAAK,CAQjB;AAED,8CAAuC;yBAlF1B,OAAO,cAAc,EAAE,UAAU;4BACjC,OAAO,cAAc,EAAE,aAAa;sBARrB,YAAY;4BAGZ,aAAa"}
+{"version":3,"file":"3string.d.ts","sourceRoot":"","sources":["../../lib/3string.js"],"names":[],"mappings":"AA0DA;;;;;;GAMG;AACH,0CANW,UAAU,OACV,MAAM,SACN,MAAM,WACN,aAAa,GACX,KAAK,CAIjB;AAED;;;;;;GAMG;AACH,oCANW,UAAU,OACV,MAAM,UACN,MAAM,WACN,aAAa,GACX,KAAK,CAIjB;AAED;;;;;;GAMG;AACH,qCANW,UAAU,OACV,MAAM,UACN,MAAM,WACN,aAAa,GACX,KAAK,CAIjB;AAED;;;;;;GAMG;AACH,qCANW,UAAU,OACV,MAAM,UACN,MAAM,WACN,aAAa,GACX,KAAK,CAIjB;AAGD;;;;;;GAMG;AACH,qCANW,UAAU,OACV,MAAM,UACN,MAAM,WACN,aAAa,GACX,KAAK,CAQjB;AAED,8CAAuC;yBA1G1B,OAAO,cAAc,EAAE,UAAU;4BACjC,OAAO,cAAc,EAAE,aAAa;sBAbrB,YAAY;4BAGZ,aAAa"}

package/types/lib/byte-utils.d.ts

@@ -15,12 +15,6 @@ export function compare(b1: Uint8Array, b2: Uint8Array): number;
  */
 export function decodeCodePointsArray(codePoints: number[]): string;
 export const useBuffer: boolean;
-/**
- * @param {Uint8Array} bytes
- * @param {number} start
- * @param {number} end
- */
-export function toString(bytes: Uint8Array, start: number, end: number): string;
 export const fromString: ((string: string) => number[] | Buffer<ArrayBuffer>) | ((string: string) => number[] | Uint8Array<ArrayBuffer>);
 export function fromArray(arr: number[]): Uint8Array;
 /**

package/types/lib/byte-utils.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"byte-utils.d.ts","sourceRoot":"","sources":["../../lib/byte-utils.js"],"names":[],"mappings":"AAwBA;;;GAGG;AACH,2BAHW,UAAU,GAAC,MAAM,EAAE,GACjB,UAAU,CAQtB;AAkOD;;;;GAIG;AACH,4BAJW,UAAU,MACV,UAAU,GACR,MAAM,CAgBlB;AA4HD;;;GAGG;AACH,kDAHW,MAAM,EAAE,GACN,MAAM,CAkBlB;AAnaD,gCAMkD;AAgC9C;;;;GAIG;AACH,gCAJW,UAAU,SACV,MAAM,OACN,MAAM,UAQhB;AAcL,mCAGe,MAAM,iDAYN,MAAM,yCAIhB;AAOE,+BAHI,MAAM,EAAE,GACN,UAAU,CAItB;AAIG;;;;GAIG;AACH,6BAJW,UAAU,SACV,MAAM,OACN,MAAM,2BAOhB;AAcD;;;;GAIG;AACH,+BAJW,UAAU,EAAE,UACZ,MAAM,GACJ,UAAU,CActB;AAwBD;;;GAGG;AACH,4BAHW,MAAM,GACJ,UAAU,CAMtB;AAaD;;;GAGG;AACH,yBAHW,UAAU,GACR,MAAM,CAQlB;AAiBH;;;GAGG;AACD,6BAHS,MAAM,GAAC,UAAU,GACf,UAAU,CAQpB"}
+{"version":3,"file":"byte-utils.d.ts","sourceRoot":"","sources":["../../lib/byte-utils.js"],"names":[],"mappings":"AAuBA;;;GAGG;AACH,2BAHW,UAAU,GAAC,MAAM,EAAE,GACjB,UAAU,CAQtB;AAoMD;;;;GAIG;AACH,4BAJW,UAAU,MACV,UAAU,GACR,MAAM,CAgBlB;AA6CD;;;GAGG;AACH,kDAHW,MAAM,EAAE,GACN,MAAM,CAkBlB;AArTD,gCAMkD;AAyBlD,mCAGe,MAAM,iDAYN,MAAM,yCAIhB;AAOE,+BAHI,MAAM,EAAE,GACN,UAAU,CAItB;AAIG;;;;GAIG;AAEH,6BALW,UAAU,SACV,MAAM,OACN,MAAM,2BAQhB;AAcD;;;;GAIG;AACH,+BAJW,UAAU,EAAE,UACZ,MAAM,GACJ,UAAU,CActB;AAwBD;;;GAGG;AACH,4BAHW,MAAM,GACJ,UAAU,CAMtB;AAaD;;;GAGG;AACH,yBAHW,UAAU,GACR,MAAM,CAQlB;AAiBH;;;GAGG;AACD,6BAHS,MAAM,GAAC,UAAU,GACf,UAAU,CAQpB"}

package/types/lib/decode.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"decode.d.ts","sourceRoot":"","sources":["../../lib/decode.js"],"names":[],"mappings":"oBAKa,OAAO,YAAY,EAAE,KAAK;4BAC1B,OAAO,cAAc,EAAE,aAAa;8BACpC,OAAO,cAAc,EAAE,eAAe;AAUnD;;GAEG;AACH,kCAFgB,eAAe;IAG7B;;;OAGG;IACH,kBAHW,UAAU,YACV,aAAa,EAMvB;IAHC,aAAa;IACb,kCAAgB;IAChB,8CAAsB;IAGxB,cAEC;IAED,gBAEC;IAED,mCAgBC;CACF;AA8ED;;;;GAIG;AACH,0CAJW,eAAe,WACf,aAAa,GACX,GAAG,6BAAW,CAoC1B;AAuBD;;;;GAIG;AACH,6BAJW,UAAU,YACV,aAAa,GACX,GAAG,CAQf;AAhCD;;;;GAIG;AACH,kCAJW,UAAU,YACV,aAAa,GACX,CAAC,GAAG,EAAE,UAAU,CAAC,CAgB7B;AAvID,mCAAiC;AADjC,kCAA+B"}
+{"version":3,"file":"decode.d.ts","sourceRoot":"","sources":["../../lib/decode.js"],"names":[],"mappings":"oBAMa,OAAO,YAAY,EAAE,KAAK;4BAC1B,OAAO,cAAc,EAAE,aAAa;8BACpC,OAAO,cAAc,EAAE,eAAe;AAUnD;;GAEG;AACH,kCAFgB,eAAe;IAG7B;;;OAGG;IACH,kBAHW,UAAU,YACV,aAAa,EAMvB;IAHC,aAAa;IACb,kCAAgB;IAChB,8CAAsB;IAGxB,cAEC;IAED,gBAEC;IAED,mCAgBC;CACF;AA8ED;;;;GAIG;AACH,0CAJW,eAAe,WACf,aAAa,GACX,GAAG,6BAAW,CAoC1B;AAyBD;;;;GAIG;AACH,6BAJW,UAAU,YACV,aAAa,GACX,GAAG,CAQf;AAlCD;;;;GAIG;AACH,kCAJW,UAAU,YACV,aAAa,GACX,CAAC,GAAG,EAAE,UAAU,CAAC,CAkB7B;AAzID,mCAAiC;AADjC,kCAA+B"}