调整目录结构; 精简字符串处理;调整结果输出

README.md

@@ -4,6 +4,7 @@
 ### 版本同步状态: 
 
 pako: ![pako](https://img.shields.io/npm/v/pako.svg)
+
 @bytedo/gzip: ![@bytedo/gzip](https://img.shields.io/npm/v/@bytedo/gzip.svg)
 
 
@@ -22,10 +23,14 @@ import { gzip , ungzip } from '@bytedo/gzip' // use importmap
 // or
 import { gzip , ungzip } from '//jscdn.ink/@bytedo/gzip/latest/index.js'
 
+// 也可以单独引入
+import { gzip } from '//jscdn.ink/@bytedo/gzip/latest/gzip.js'
+import { ungzip } from '//jscdn.ink/@bytedo/gzip/latest/ungzip.js'
 
-let res = gzip('hello world')  // return Uint8Array object
 
-let txt = ungzip(res) // return Uint8Array object
+let base64Str = gzip('hello world')  // return base64 string
+
+let txt = ungzip(base64Str) // return hello world
 
 ```
 

index.js

@@ -1,2 +0,0 @@
-export { gzip } from './lib/deflate.js'
-export { ungzip } from './lib/inflate.js'

lib/utils/strings.js

@@ -1,197 +0,0 @@
-// Quick check if we can use fast array to bin string conversion
-//
-// - apply(Array) can fail on Android 2.2
-// - apply(Uint8Array) can fail on iOS 5.1 Safari
-//
-let STR_APPLY_UIA_OK = true
-
-try {
-  String.fromCharCode.apply(null, new Uint8Array(1))
-} catch (__) {
-  STR_APPLY_UIA_OK = false
-}
-
-// Table with utf8 lengths (calculated by first byte of sequence)
-// Note, that 5 & 6-byte values and some 4-byte values can not be represented in JS,
-// because max possible codepoint is 0x10ffff
-const _utf8len = new Uint8Array(256)
-for (let q = 0; q < 256; q++) {
-  _utf8len[q] =
-    q >= 252 ? 6 : q >= 248 ? 5 : q >= 240 ? 4 : q >= 224 ? 3 : q >= 192 ? 2 : 1
-}
-_utf8len[254] = _utf8len[254] = 1 // Invalid sequence start
-
-// convert string to array (typed, when possible)
-export const string2buf = str => {
-  if (typeof TextEncoder === 'function' && TextEncoder.prototype.encode) {
-    return new TextEncoder().encode(str)
-  }
-
-  let buf,
-    c,
-    c2,
-    m_pos,
-    i,
-    str_len = str.length,
-    buf_len = 0
-
-  // count binary size
-  for (m_pos = 0; m_pos < str_len; m_pos++) {
-    c = str.charCodeAt(m_pos)
-    if ((c & 0xfc00) === 0xd800 && m_pos + 1 < str_len) {
-      c2 = str.charCodeAt(m_pos + 1)
-      if ((c2 & 0xfc00) === 0xdc00) {
-        c = 0x10000 + ((c - 0xd800) << 10) + (c2 - 0xdc00)
-        m_pos++
-      }
-    }
-    buf_len += c < 0x80 ? 1 : c < 0x800 ? 2 : c < 0x10000 ? 3 : 4
-  }
-
-  // allocate buffer
-  buf = new Uint8Array(buf_len)
-
-  // convert
-  for (i = 0, m_pos = 0; i < buf_len; m_pos++) {
-    c = str.charCodeAt(m_pos)
-    if ((c & 0xfc00) === 0xd800 && m_pos + 1 < str_len) {
-      c2 = str.charCodeAt(m_pos + 1)
-      if ((c2 & 0xfc00) === 0xdc00) {
-        c = 0x10000 + ((c - 0xd800) << 10) + (c2 - 0xdc00)
-        m_pos++
-      }
-    }
-    if (c < 0x80) {
-      /* one byte */
-      buf[i++] = c
-    } else if (c < 0x800) {
-      /* two bytes */
-      buf[i++] = 0xc0 | (c >>> 6)
-      buf[i++] = 0x80 | (c & 0x3f)
-    } else if (c < 0x10000) {
-      /* three bytes */
-      buf[i++] = 0xe0 | (c >>> 12)
-      buf[i++] = 0x80 | ((c >>> 6) & 0x3f)
-      buf[i++] = 0x80 | (c & 0x3f)
-    } else {
-      /* four bytes */
-      buf[i++] = 0xf0 | (c >>> 18)
-      buf[i++] = 0x80 | ((c >>> 12) & 0x3f)
-      buf[i++] = 0x80 | ((c >>> 6) & 0x3f)
-      buf[i++] = 0x80 | (c & 0x3f)
-    }
-  }
-
-  return buf
-}
-
-// Helper
-const buf2binstring = (buf, len) => {
-  // On Chrome, the arguments in a function call that are allowed is `65534`.
-  // If the length of the buffer is smaller than that, we can use this optimization,
-  // otherwise we will take a slower path.
-  if (len < 65534) {
-    if (buf.subarray && STR_APPLY_UIA_OK) {
-      return String.fromCharCode.apply(
-        null,
-        buf.length === len ? buf : buf.subarray(0, len)
-      )
-    }
-  }
-
-  let result = ''
-  for (let i = 0; i < len; i++) {
-    result += String.fromCharCode(buf[i])
-  }
-  return result
-}
-
-// convert array to string
-export const buf2string = (buf, max) => {
-  const len = max || buf.length
-
-  if (typeof TextDecoder === 'function' && TextDecoder.prototype.decode) {
-    return new TextDecoder().decode(buf.subarray(0, max))
-  }
-
-  let i, out
-
-  // Reserve max possible length (2 words per char)
-  // NB: by unknown reasons, Array is significantly faster for
-  //     String.fromCharCode.apply than Uint16Array.
-  const utf16buf = new Array(len * 2)
-
-  for (out = 0, i = 0; i < len; ) {
-    let c = buf[i++]
-    // quick process ascii
-    if (c < 0x80) {
-      utf16buf[out++] = c
-      continue
-    }
-
-    let c_len = _utf8len[c]
-    // skip 5 & 6 byte codes
-    if (c_len > 4) {
-      utf16buf[out++] = 0xfffd
-      i += c_len - 1
-      continue
-    }
-
-    // apply mask on first byte
-    c &= c_len === 2 ? 0x1f : c_len === 3 ? 0x0f : 0x07
-    // join the rest
-    while (c_len > 1 && i < len) {
-      c = (c << 6) | (buf[i++] & 0x3f)
-      c_len--
-    }
-
-    // terminated by end of string?
-    if (c_len > 1) {
-      utf16buf[out++] = 0xfffd
-      continue
-    }
-
-    if (c < 0x10000) {
-      utf16buf[out++] = c
-    } else {
-      c -= 0x10000
-      utf16buf[out++] = 0xd800 | ((c >> 10) & 0x3ff)
-      utf16buf[out++] = 0xdc00 | (c & 0x3ff)
-    }
-  }
-
-  return buf2binstring(utf16buf, out)
-}
-
-// Calculate max possible position in utf8 buffer,
-// that will not break sequence. If that's not possible
-// - (very small limits) return max size as is.
-//
-// buf[] - utf8 bytes array
-// max   - length limit (mandatory);
-export const utf8border = (buf, max) => {
-  max = max || buf.length
-  if (max > buf.length) {
-    max = buf.length
-  }
-
-  // go back from last position, until start of sequence found
-  let pos = max - 1
-  while (pos >= 0 && (buf[pos] & 0xc0) === 0x80) {
-    pos--
-  }
-
-  // Very small and broken sequence,
-  // return max, because we should return something anyway.
-  if (pos < 0) {
-    return max
-  }
-
-  // If we came to start of buffer - that means buffer is too small,
-  // return max too.
-  if (pos === 0) {
-    return max
-  }
-
-  return pos + _utf8len[buf[pos]] > max ? pos : max
-}

package.json

@@ -1,23 +1,27 @@
 {
   "name": "@bytedo/gzip",
   "type": "module",
-  "description": "zlib port to javascript - fast, modularized, with browser support",
+  "description": "浏览器端的gzip库, fork于pako, 仅调整语法为esm, 并只保留gzip的导出。",
   "version": "2.1.0",
   "keywords": [
-    "zlib",
-    "deflate",
-    "inflate",
-    "gzip"
+    "gzip",
+    "ungzip"
   ],
   "files": [
     "dist/*"
   ],
-  "license": "(MIT AND Zlib)",
-  "repository": "bytedo/gzip",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/bytedo/gzip.git"
+  },
   "scripts": {
-    "build": "esbuild index.js --minify --bundle --format=esm --target=esnext --outfile=dist/index.js"
+    "build:gzip": "esbuild src/gzip.js --minify --bundle --format=esm --target=esnext --outfile=dist/gzip.js",
+    "build:ungzip": "esbuild src/ungzip.js --minify --bundle --format=esm --target=esnext --outfile=dist/ungzip.js",
+    "build:all": "esbuild src/index.js --format=esm --outfile=dist/index.js",
+    "build": "npm run build:gzip && npm run build:ungzip && npm run build:all"
   },
   "devDependencies": {
     "esbuild": "^0.17.16"
-  }
+  },
+  "license": "(MIT AND Zlib)"
 }

src/gzip.js

@@ -1,6 +1,6 @@
 import zlib_deflate from './zlib/deflate.js'
 import { flattenChunks } from './utils/common.js'
-import { string2buf } from './utils/strings.js'
+import { string2buf, buf2base64 } from './utils/strings.js'
 import msg from './zlib/messages.js'
 import ZStream from './zlib/zstream.js'
 
@@ -357,7 +357,7 @@ function deflate(input, options) {
 function gzip(input, options) {
   options = options || {}
   options.gzip = true
-  return deflate(input, options)
+  return buf2base64(deflate(input, options))
 }
 
 export { gzip }

src/index.js

@@ -0,0 +1,2 @@
+export { gzip } from './gzip.js'
+export { ungzip } from './ungzip.js'

src/ungzip.js

@@ -1,6 +1,11 @@
 import zlib_inflate from './zlib/inflate.js'
 import { flattenChunks } from './utils/common.js'
-import { string2buf, utf8border, buf2string } from './utils/strings.js'
+import {
+  utf8border,
+  string2buf,
+  buf2string,
+  base642buf
+} from './utils/strings.js'
 import msg from './zlib/messages.js'
 import ZStream from './zlib/zstream.js'
 import GZheader from './zlib/gzheader.js'
@@ -388,12 +393,12 @@ Inflate.prototype.onEnd = function (status) {
 function ungzip(input, options) {
   const inflator = new Inflate(options)
 
-  inflator.push(input)
+  inflator.push(base642buf(input))
 
   // That will never happens, if you don't cheat with options :)
   if (inflator.err) throw inflator.msg || msg[inflator.err]
 
-  return inflator.result
+  return buf2string(inflator.result)
 }
 
 /**

src/utils/strings.js

@@ -0,0 +1,75 @@
+const encoder = new TextEncoder()
+const decoder = new TextDecoder()
+
+// Table with utf8 lengths (calculated by first byte of sequence)
+// Note, that 5 & 6-byte values and some 4-byte values can not be represented in JS,
+// because max possible codepoint is 0x10ffff
+const _utf8len = new Uint8Array(256)
+for (let q = 0; q < 256; q++) {
+  _utf8len[q] =
+    q >= 252 ? 6 : q >= 248 ? 5 : q >= 240 ? 4 : q >= 224 ? 3 : q >= 192 ? 2 : 1
+}
+_utf8len[254] = _utf8len[254] = 1 // Invalid sequence start
+
+export function buf2base64(uint8) {
+  var bin = ''
+  for (var i = 0; i < uint8.length; i++) {
+    bin += String.fromCharCode(uint8[i])
+  }
+  return btoa(bin)
+}
+
+export function base642buf(base64) {
+  let bin = atob(base64)
+  let u8 = new Uint8Array(bin.length)
+
+  for (let i = 0; i < bin.length; i++) {
+    u8[i] = bin[i].codePointAt(0)
+  }
+
+  return u8
+}
+
+// convert string to array (typed, when possible)
+export const string2buf = str => {
+  return encoder.encode(str)
+}
+
+// convert array to string
+export const buf2string = (buf, max) => {
+  let len = max || buf.length
+  return decoder.decode(buf.subarray(0, max))
+}
+
+// Calculate max possible position in utf8 buffer,
+// that will not break sequence. If that's not possible
+// - (very small limits) return max size as is.
+//
+// buf[] - utf8 bytes array
+// max   - length limit (mandatory);
+export const utf8border = (buf, max) => {
+  max = max || buf.length
+  if (max > buf.length) {
+    max = buf.length
+  }
+
+  // go back from last position, until start of sequence found
+  let pos = max - 1
+  while (pos >= 0 && (buf[pos] & 0xc0) === 0x80) {
+    pos--
+  }
+
+  // Very small and broken sequence,
+  // return max, because we should return something anyway.
+  if (pos < 0) {
+    return max
+  }
+
+  // If we came to start of buffer - that means buffer is too small,
+  // return max too.
+  if (pos === 0) {
+    return max
+  }
+
+  return pos + _utf8len[buf[pos]] > max ? pos : max
+}

src/zlib/adler32.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // Note: adler32 takes 12% for level 0 and 2% for level 6.
 // It isn't worth it to make additional optimizations as in original.
 // Small size is preferable.

src/zlib/crc32.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // Note: we can't get significant speed boost here.
 // So write code to minimize size - no pregenerated tables
 // and array tools dependencies.

src/zlib/deflate.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // (C) 1995-2013 Jean-loup Gailly and Mark Adler
 // (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
 //

src/zlib/gzheader.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // (C) 1995-2013 Jean-loup Gailly and Mark Adler
 // (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
 //

src/zlib/inffast.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // (C) 1995-2013 Jean-loup Gailly and Mark Adler
 // (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
 //

src/zlib/inflate.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // (C) 1995-2013 Jean-loup Gailly and Mark Adler
 // (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
 //

src/zlib/inftrees.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // (C) 1995-2013 Jean-loup Gailly and Mark Adler
 // (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
 //

src/zlib/messages.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // (C) 1995-2013 Jean-loup Gailly and Mark Adler
 // (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
 //

src/zlib/trees.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // (C) 1995-2013 Jean-loup Gailly and Mark Adler
 // (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
 //

src/zlib/zstream.js

@@ -1,5 +1,3 @@
-'use strict'
-
 // (C) 1995-2013 Jean-loup Gailly and Mark Adler
 // (C) 2014-2017 Vitaly Puzrin and Andrey Tupitsin
 //