transcend-io · jimmywarting · Jul 10, 2019 · Jul 10, 2019 · Jul 11, 2019 · Jul 11, 2019
diff --git a/example/index.html b/example/index.html
@@ -5,9 +5,12 @@
  <title></title>
  </head>
  <body>
+ <!-- for the new blob reading methods -->
+ <script src="https://cdn.jsdelivr.net/gh/eligrey/Blob.js/Blob.js"></script>
  <script src="https://cdn.jsdelivr.net/npm/[email protected]/dist/polyfill.min.js"></script>
  <script type="module">
  import Zip from '../write.js'
+ import Reader from '../read.js'
 
  const { readable, writable } = new Zip()
  const writer = writable.getWriter()
@@ -24,10 +27,9 @@
  let o = 0; chunks.forEach(c => {uint8.set(c, o); o += c.length})
 
  const hash = await crypto.subtle.digest('sha-1', uint8)
- const expected = '18282460561088092989-11685288625477285951386915632'
+ const expected = '831739241-296827476-199608789416608489701811849431'
  const blob = new Blob([uint8], { type: 'application/zip' })
  const a = document.createElement('a')
-
  a.href = URL.createObjectURL(blob)
  a.innerText = a.download = 'Archive.zip'
 
@@ -44,6 +46,14 @@
  pump()
 
  writer.write({
+ comment: 'dog',
+ name: '/dog.txt',
+ lastModified: new Date(0),
+ stream: () => new Response('woff').body
+ })
+
+ writer.write({
+ comment: 'cat',
  name: '/cat.txt',
  lastModified: new Date(0),
  stream: () => new Response('mjau').body

diff --git a/example/reading.html b/example/reading.html
@@ -0,0 +1,79 @@
+<!DOCTYPE html>
+<html lang="en" dir="ltr">
+ <head>
+ <meta charset="utf-8">
+ <title></title>
+ </head>
+ <body>
+ <!-- for the new blob reading methods -->
+ <script src="https://cdn.jsdelivr.net/gh/eligrey/Blob.js/Blob.js"></script>
+ <script type="module">
+import reader from '../read.js'
+
+const jszip = 'https://cdn.jsdelivr.net/gh/Stuk/jszip/test/ref/'
+const urls = [
+ '/test/fixture/dog_cat.zip', // ok
+ jszip + 'all-stream.zip', // problem reading
+ jszip + 'all.7zip.zip', // problem reading
+ jszip + 'all.windows.zip', // problem reading a compressed image
+ jszip + 'all.zip', // problem reading
+ jszip + 'all_appended_bytes.zip', // problem reading
+ jszip + 'all_missing_bytes.zip', // fail as it should
+ jszip + 'all_prepended_bytes.zip', // problem reading
+ jszip + 'archive_comment.zip', // Ok
+ jszip + 'backslash.zip', // ok
+ jszip + 'data_descriptor.zip', // problem reading
+ jszip + 'deflate-stream.zip',
+ jszip + 'deflate.zip',
+ jszip + 'empty.zip', // ok
+ jszip + 'encrypted.zip',
+ jszip + 'extra_attributes.zip', // ???
+ jszip + 'folder.zip', // ok
+ jszip + 'image.zip', // ok
+ jszip + 'local_encoding_in_name.zip', // unsolvable?
+ jszip + 'nested.zip', // ok
+ jszip + 'nested_data_descriptor.zip',
+ jszip + 'nested_zip64.zip',
+ jszip + 'pile_of_poo.zip',
+ jszip + 'slashes_and_izarc.zip',
+ jszip + 'store-stream.zip',
+ jszip + 'store.zip',
+ jszip + 'subfolder.zip',
+ jszip + 'text.zip',
+ jszip + 'utf8.zip',
+ jszip + 'utf8_in_name.zip',
+ jszip + 'winrar_utf8_in_name.zip',
+ jszip + 'zip64.zip',
+ jszip + 'zip64_appended_bytes.zip',
+ jszip + 'zip64_missing_bytes.zip'
+]
+
+let url = urls[22]
+console.log(url)
+fetch(url).then(async res => {
+ const zip = await res.blob()
+ const entries = []
+
+ // console.log(urls[4])
+ for await (const entry of reader(zip)) {
+ const { comment, name, directory, compressionMethod, size, compressedSize } = entry
+ console.log(entry)
+ entries.push({
+ comment,
+ name,
+ directory,
+ compressionMethod,
+ size,
+ compressedSize,
+ ab: await entry.arrayBuffer(),
+ body: await entry.text(),
+ url: URL.createObjectURL(new Blob([await entry.arrayBuffer()], {type: 'd/d'}))
+ })
+
+ }
+
+ console.table(entries)
+})
+ </script>
+ </body>
+</html>
diff --git a/read.js b/read.js
@@ -1 +1,174 @@
 // TODO: later
+const ERR_BAD_FORMAT = 'File format is not recognized.'
+const ZIP_COMMENT_MAX = 65536
+const EOCDR_MIN = 22
+const EOCDR_MAX = EOCDR_MIN + ZIP_COMMENT_MAX
+
+const decoder = new TextDecoder()
+
+class Entry {
+ constructor (dataView) {
+ if (dataView.getUint32(0) !== 0x504b0102) {
+ throw new Error('ERR_BAD_FORMAT')
+ }
+ this.dataView = dataView
+ }
+ get version () {
+ return this.dataView.getUint16(6, true)
+ }
+ get bitFlag () {
+ return this.dataView.getUint16(8, true)
+ }
+ get encrypted () {
+ return (this.bitFlag & 1) === 1
+ }
+ get compressionMethod () {
+ return this.dataView.getUint16(10, true)
+ }
+ get lastModDateRaw () {
+ return this.dataView.getUint32(12, true)
+ }
+ get crc32 () {
+ return this.dataView.getUint32(16, true)
+ }
+ get filenameLength () {
+ return this.dataView.getUint16(28, true)
+ }
+ get extraFieldLength () {
+ return this.dataView.getUint16(30, true)
+ }
+ get commentLength () {
+ return this.dataView.getUint16(32, true)
+ }
+ get directory () {
+ return (this.dataView.getUint8(38) & 16) === 16
+ }
+ get offset () {
+ return this.dataView.getUint16(42, true)
+ }
+ get zip64 () {
+ return this.size === 0xFFFFFFFF ||
+ this.uncompressedSize === 0xFFFFFFFF
+ }
+
+ get comment () {
+ const dv = this.dataView
+ const uint8 = new Uint8Array(dv.buffer, dv.byteOffset + 46 + this.filenameLength + this.extraFieldLength, this.commentLength)
+ return decoder.decode(uint8)
+ }
+
+ // File like IDL methods
+
+ get lastModifiedDate () {
+ // TODO: conversion
+ return new Date(this.lastModDate)
+ }
+
+ get lastModified () {
+
+ }
+
+ get name () {
+ const dv = this.dataView
+ const uint8 = new Uint8Array(dv.buffer, dv.byteOffset + 46, this.filenameLength)
+ return decoder.decode(uint8)
+ }
+
+ get size () {
+ return this.dataView.getUint32(24, true)
+ }
+
+ stream () {
+ // TODO: Investigate
+ // From my understanding jszip tells me that extraFieldLength **might**
+ // vary from local and central dir?
+ // - wtf?!
+ // one guess is that if extraFieldLength is defined it will then also
+ // have a 4 byte added signature? reason:
+ //
+ // 4.3.11 Archive extra data record:
+ //
+ // archive extra data signature 4 bytes (0x08064b50)
+ // extra field length 4 bytes
+ // extra field data (variable size)
+ //
+ // But i can also be wrong. An example file i tried to read was
+ // https://cdn.jsdelivr.net/gh/Stuk/jszip/test/ref/extra_attributes.zip
+ // in the central dir the length was 24
+ // in local header it was 28
+ const extra = this.extraFieldLength
+
+ const start = this.offset + this.filenameLength + 30 + (extra ? extra + 4 : 0)
+ const end = start + this.compressedSize
+
+ return this
+ ._fileLike
+ .slice(start, end)
+ .stream()
+ // .pipeThrought(inflate) // TODO: optional inflate
+ // .pipeThrought(crc) // TODO: crc32 validate
+ }
+
+ arrayBuffer () {
+ return new Response(this.stream()).arrayBuffer()
+ }
+
+ text () {
+ return new Response(this.stream()).text()
+ }
+}
+
+async function * seekEOCDR (fileLike) {
+ // "End of central directory record" is the last part of a zip archive, and is at least 22 bytes long.
+ // Zip file comment is the last part of EOCDR and has max length of 64KB,
+ // so we only have to search the last 64K + 22 bytes of a archive for EOCDR signature (0x06054b50).
+ if (fileLike.size < EOCDR_MIN) throw new Error(ERR_BAD_FORMAT)
+
+ // In most cases, the EOCDR is EOCDR_MIN bytes long
+ const dv = await doSeek(EOCDR_MIN) ||
+ await doSeek(Math.min(EOCDR_MAX, fileLike.size))
+
+ if (!dv) throw new Error(ERR_BAD_FORMAT)
+
+ const datalength = dv.getUint32(16, true)
+ const fileslength = dv.getUint16(8, true)
+
+ if (datalength < 0 || datalength >= fileLike.size) {
+ throw new Error(ERR_BAD_FORMAT)
+ }
+
+ // const bytes = await fileLike.slice(fileLike.size - datalength).arrayBuffer()
+ const bytes = new Uint8Array(await fileLike.slice(datalength).arrayBuffer())
+
+ const uint16e = (b, n) => b[n] | (b[n + 1] << 8)
+
+ for (let i = 0, index = 0; i < fileslength; i++) {
+ const size =
+ uint16e(bytes, 28) + // filenameLength
+ uint16e(bytes, 30) + // extraFieldLength
+ uint16e(bytes, 32) + // commentLength
+ 46
+
+ yield new Entry(
+ new DataView(bytes.buffer, index, size),
+ fileLike
+ )
+
+ index += size
+ }
+
+ // seek last length bytes of file for EOCDR
+ async function doSeek (length) {
+ const ab = await fileLike.slice(fileLike.size - length).arrayBuffer()
+ const bytes = new Uint8Array(ab)
+ for (let i = bytes.length - EOCDR_MIN; i >= 0; i--) {
+ if (bytes[i] === 0x50 && bytes[i + 1] === 0x4b && bytes[i + 2] === 0x05 && bytes[i + 3] === 0x06) {
+ return new DataView(bytes.buffer, i, EOCDR_MIN)
+ }
+ }
+
+ return null
+ }
+}
+
+export default seekEOCDR
diff --git a/test/fixture/dog_cat.zip b/test/fixture/dog_cat.zip
diff --git a/write.js b/write.js
@@ -49,10 +49,6 @@ class ZipTransformer {
  const { header } = zipObject;
  const data = getDataHelper(30 + nameBuf.length);
 
- if (entry.level !== 0 && !entry.directory) {
- header.view.setUint16(4, 0x0800);
- }
-
  header.view.setUint32(0, 0x14000808);
  header.view.setUint16(6, (((date.getHours() << 6) | date.getMinutes()) << 5) | date.getSeconds() / 2, true);
  header.view.setUint16(8, ((((date.getFullYear() - 1980) << 4) | (date.getMonth() + 1)) << 5) | date.getDate(), true);