appnote.txt from PKZIP 1.1

appnote.txt

Disclaimer
----------

Although PKWARE will attempt to supply current and accurate
information relating to its file formats, algorithms, and the
subject programs, the possibility of error can not be eliminated.
PKWARE therefore expressly disclaims any warranty that the
information contained in the associated materials relating to the
subject programs and/or the format of the files created or
accessed by the subject programs and/or the algorithms used by
the subject programs, or any other matter, is current, correct or
accurate as delivered.  Any risk of damage due to any possible
inaccurate information is assumed by the user of the information.
Furthermore, the information relating to the subject programs
and/or the file formats created or accessed by the subject
programs and/or the algorithms used by the subject programs is
subject to change without notice.


General Format of a ZIP file
----------------------------

  Files stored in arbitrary order.  Large zipfiles can span multiple
  diskette media.

  Overall zipfile format:

    [local file header+file data] . . .
    [central directory] end of central directory record


  A.  Local file header:

	local file header signature	4 bytes  (0x04034b50)
	version needed to extract	2 bytes
	general purpose bit flag	2 bytes
	compression method		2 bytes
	last mod file time 		2 bytes
	last mod file date		2 bytes
	crc-32   			4 bytes
	compressed size			4 bytes
	uncompressed size		4 bytes
	filename length			2 bytes
	extra field length		2 bytes

	filename (variable size)
	extra field (variable size)


  B.  Central directory structure:

      [file header] . . .  end of central dir record

      File header:

	central file header signature	4 bytes  (0x02014b50)
	version made by			2 bytes
	version needed to extract	2 bytes
	general purpose bit flag	2 bytes
	compression method		2 bytes
	last mod file time 		2 bytes
	last mod file date		2 bytes
	crc-32   			4 bytes
	compressed size			4 bytes
	uncompressed size		4 bytes
	filename length			2 bytes
	extra field length		2 bytes
	file comment length		2 bytes
	disk number start		2 bytes
	internal file attributes	2 bytes
	external file attributes	4 bytes
	relative offset of local header	4 bytes

	filename (variable size)
	extra field (variable size)
	file comment (variable size)

      End of central dir record:

	end of central dir signature	4 bytes  (0x06054b50)
	number of this disk		2 bytes
	number of the disk with the
	start of the central directory	2 bytes
	total number of entries in
	the central dir on this disk	2 bytes
	total number of entries in
	the central dir			2 bytes
	size of the central directory   4 bytes
	offset of start of central
	directory with respect to
	the starting disk number	4 bytes
	zipfile comment length		2 bytes
	zipfile comment (variable size)




  C.  Explanation of fields:

      version made by

	  The upper byte indicates the host system (OS) for the
	  file.  Software can use this information to determine
	  the line record format for text files etc.  The current
	  mappings are:

	  0 - MS-DOS and OS/2 (F.A.T. file systems)
	  1 - Amiga			2 - VMS
	  3 - *nix			4 - VM/CMS
	  5 - Atari ST                  6 - OS/2 H.P.F.S.
	  7 - Macintosh			8 - Z-System
	  9 - CP/M			10 thru 255 - unused

	  The lower byte indicates the version number of the
	  software used to encode the file.  The value/10
	  indicates the major version number, and the value
	  mod 10 is the minor version number.

      version needed to extract

	  The minimum software version needed to extract the
	  file, mapped as above.

      general purpose bit flag:

          bit 0: If set, indicates that the file is encrypted.
          bit 1: If the compression method used was type 6,
		 Imploding, then this bit, if set, indicates
		 an 8K sliding dictionary was used.  If clear,
		 then a 4K sliding dictionary was used.
          bit 2: If the compression method used was type 6,
		 Imploding, then this bit, if set, indicates
		 an 3 Shannon-Fano trees were used to encode the
		 sliding dictionary output.  If clear, then 2
		 Shannon-Fano trees were used.
	  Note:  Bits 1 and 2 are undefined if the compression
		 method is other than type 6 (Imploding).

          The upper three bits are reserved and used internally
	  by the software when processing the zipfile.  The
	  remaining bits are unused in version 1.0.

      compression method:

	  (see accompanying documentation for algorithm
	  descriptions)

	  0 - The file is stored (no compression)
	  1 - The file is Shrunk
	  2 - The file is Reduced with compression factor 1
	  3 - The file is Reduced with compression factor 2
	  4 - The file is Reduced with compression factor 3
	  5 - The file is Reduced with compression factor 4
          6 - The file is Imploded

      date and time fields:

	  The date and time are encoded in standard MS-DOS
	  format.

      CRC-32:

	  The CRC-32 algorithm was generously contributed by
	  David Schwaderer and can be found in his excellent
	  book "C Programmers Guide to NetBIOS" published by
	  Howard W. Sams & Co. Inc.  The 'magic number' for
	  the CRC is 0xdebb20e3.  The proper CRC pre and post
	  conditioning is used, meaning that the CRC register
	  is pre-conditioned with all ones (a starting value
	  of 0xffffffff) and the value is post-conditioned by
	  taking the one's complement of the CRC residual.
	
      compressed size:
      uncompressed size:

	  The size of the file compressed and uncompressed,
	  respectively.

      filename length:
      extra field length:
      file comment length:

	  The length of the filename, extra field, and comment
	  fields respectively.  The combined length of any
	  directory record and these three fields should not
	  generally exceed 65,535 bytes.

      disk number start:

	  The number of the disk on which this file begins.

      internal file attributes:

	  The lowest bit of this field indicates, if set, that
	  the file is apparently an ASCII or text file.  If not
	  set, that the file apparently contains binary data.
	  The remaining bits are unused in version 1.0.

      external file attributes:

	  The mapping of the external attributes is
	  host-system dependent (see 'version made by').  For
	  MS-DOS, the low order byte is the MS-DOS directory
	  attribute byte.

      relative offset of local header:

	  This is the offset from the start of the first disk on
	  which this file appears, to where the local header should
	  be found.

      filename:

	  The name of the file, with optional relative path.
	  The path stored should not contain a drive or
	  device letter, or a leading slash.  All slashes
	  should be forward slashes '/' as opposed to
	  backwards slashes '\' for compatibility with Amiga
	  and Unix file systems etc.

      extra field:

	  This is for future expansion.  If additional information
	  needs to be stored in the future, it should be stored
	  here.  Earlier versions of the software can then safely
	  skip this file, and find the next file or header.  This
	  field will be 0 length in version 1.0.

	  In order to allow different programs and different types 
	  of information to be stored in the 'extra' field in .ZIP 
	  files, the following structure should be used for all 
	  programs storing data in this field:

	  header1+data1 + header2+data2 . . .

	  Each header should consist of:

	    Header ID - 2 bytes
	    Data Size - 2 bytes

	  Note: all fields stored in Intel low-byte/high-byte order.

	  The Header ID field indicates the type of data that is in 
	  the following data block.
      
	  Header ID's of 0 thru 31 are reserved for use by PKWARE.  
	  The remaining ID's can be used by third party vendors for 
	  proprietary usage.

	  The Data Size field indicates the size of the following 
	  data block. Programs can use this value to skip to the 
	  next header block, passing over any data blocks that are 
	  not of interest.

	  Note: As stated above, the size of the entire .ZIP file
		header, including the filename, comment, and extra
		field should not exceed 64K in size.

	  In case two different programs should appropriate the same 
	  Header ID value, it is strongly recommended that each 
	  program place a unique signature of at least two bytes in 
	  size (and preferably 4 bytes or bigger) at the start of 
	  each data area.  Every program should verify that it's 
	  unique signature is present, in addition to the Header ID 
	  value being correct, before assuming that it is a block of 
	  known type.

      file comment:

	  The comment for this file.

      number of this disk:

	  The number of this disk, which contains central
	  directory end record.

      number of the disk with the start of the central directory:

	  The number of the disk on which the central
	  directory starts.

      total number of entries in the central dir on this disk:

	  The number of central directory entries on this disk.
	
      total number of entries in the central dir:

	  The total number of files in the zipfile.


      size of the central directory:

	  The size (in bytes) of the entire central directory.

      offset of start of central directory with respect to
      the starting disk number:

	  Offset of the start of the central direcory on the
	  disk on which the central directory starts.

      zipfile comment length:

	  The length of the comment for this zipfile.

      zipfile comment:

	  The comment for this zipfile.


  D.  General notes:

      1)  All fields unless otherwise noted are unsigned and stored
	  in Intel low-byte:high-byte, low-word:high-word order.

      2)  String fields are not null terminated, since the
	  length is given explicitly.

      3)  Local headers should not span disk boundries.  Also, even
	  though the central directory can span disk boundries, no
	  single record in the central directory should be split
	  across disks.

      4)  The entries in the central directory may not necessarily
	  be in the same order that files appear in the zipfile.

UnShrinking
-----------

Shrinking is a Dynamic Ziv-Lempel-Welch compression algorithm
with partial clearing.  The initial code size is 9 bits, and
the maximum code size is 13 bits.  Shrinking differs from
conventional Dynamic Ziv-lempel-Welch implementations in several
respects:

1)  The code size is controlled by the compressor, and is not
    automatically increased when codes larger than the current
    code size are created (but not necessarily used).  When
    the decompressor encounters the code sequence 256
    (decimal) followed by 1, it should increase the code size
    read from the input stream to the next bit size.  No
    blocking of the codes is performed, so the next code at
    the increased size should be read from the input stream
    immediately after where the previous code at the smaller
    bit size was read.  Again, the decompressor should not
    increase the code size used until the sequence 256,1 is
    encountered.

2)  When the table becomes full, total clearing is not
    performed.  Rather, when the compresser emits the code
    sequence 256,2 (decimal), the decompressor should clear
    all leaf nodes from the Ziv-Lempel tree, and continue to
    use the current code size.  The nodes that are cleared
    from the Ziv-Lempel tree are then re-used, with the lowest
    code value re-used first, and the highest code value
    re-used last.  The compressor can emit the sequence 256,2
    at any time.



Expanding
---------

The Reducing algorithm is actually a combination of two
distinct algorithms.  The first algorithm compresses repeated
byte sequences, and the second algorithm takes the compressed
stream from the first algorithm and applies a probabilistic
compression method.

The probabilistic compression stores an array of 'follower
sets' S(j), for j=0 to 255, corresponding to each possible
ASCII character.  Each set contains between 0 and 32
characters, to be denoted as S(j)[0],...,S(j)[m], where m<32.
The sets are stored at the beginning of the data area for a
Reduced file, in reverse order, with S(255) first, and S(0)
last.

The sets are encoded as { N(j), S(j)[0],...,S(j)[N(j)-1] },
where N(j) is the size of set S(j).  N(j) can be 0, in which
case the follower set for S(j) is empty.  Each N(j) value is
encoded in 6 bits, followed by N(j) eight bit character values
corresponding to S(j)[0] to S(j)[N(j)-1] respectively.  If
N(j) is 0, then no values for S(j) are stored, and the value
for N(j-1) immediately follows.

Immediately after the follower sets, is the compressed data
stream.  The compressed data stream can be interpreted for the
probabilistic decompression as follows:


let Last-Character <- 0.
loop until done
    if the follower set S(Last-Character) is empty then
	read 8 bits from the input stream, and copy this
	value to the output stream.
    otherwise if the follower set S(Last-Character) is non-empty then
	read 1 bit from the input stream.
	if this bit is not zero then
	    read 8 bits from the input stream, and copy this
	    value to the output stream.
	otherwise if this bit is zero then
	    read B(N(Last-Character)) bits from the input
	    stream, and assign this value to I.
	    Copy the value of S(Last-Character)[I] to the
	    output stream.
	
    assign the last value placed on the output stream to
    Last-Character.
end loop


B(N(j)) is defined as the minimal number of bits required to
encode the value N(j)-1.


The decompressed stream from above can then be expanded to
re-create the original file as follows:


let State <- 0.

loop until done
    read 8 bits from the input stream into C.
    case State of
	0:  if C is not equal to DLE (144 decimal) then
		copy C to the output stream.
	    otherwise if C is equal to DLE then
		let State <- 1.

	1:  if C is non-zero then
		let V <- C.
		let Len <- L(V)
		let State <- F(Len).
	    otherwise if C is zero then
		copy the value 144 (decimal) to the output stream.
		let State <- 0

	2:  let Len <- Len + C
	    let State <- 3.

	3:  move backwards D(V,C) bytes in the output stream
	    (if this position is before the start of the output
	    stream, then assume that all the data before the
	    start of the output stream is filled with zeros).
	    copy Len+3 bytes from this position to the output stream.
	    let State <- 0.
    end case
end loop


The functions F,L, and D are dependent on the 'compression
factor', 1 through 4, and are defined as follows:

For compression factor 1:
    L(X) equals the lower 7 bits of X.
    F(X) equals 2 if X equals 127 otherwise F(X) equals 3.
    D(X,Y) equals the (upper 1 bit of X) * 256 + Y + 1.
For compression factor 2:
    L(X) equals the lower 6 bits of X.
    F(X) equals 2 if X equals 63 otherwise F(X) equals 3.
    D(X,Y) equals the (upper 2 bits of X) * 256 + Y + 1.
For compression factor 3:
    L(X) equals the lower 5 bits of X.
    F(X) equals 2 if X equals 31 otherwise F(X) equals 3.
    D(X,Y) equals the (upper 3 bits of X) * 256 + Y + 1.
For compression factor 4:
    L(X) equals the lower 4 bits of X.
    F(X) equals 2 if X equals 15 otherwise F(X) equals 3.
    D(X,Y) equals the (upper 4 bits of X) * 256 + Y + 1.


Imploding
---------

The Imploding algorithm is actually a combination of two distinct
algorithms.  The first algorithm compresses repeated byte
sequences using a sliding dictionary.  The second algorithm is
used to compress the encoding of the sliding dictionary ouput,
using multiple Shannon-Fano trees.

The Imploding algorithm can use a 4K or 8K sliding dictionary
size. The dictionary size used can be determined by bit 1 in the
general purpose flag word, a 0 bit indicates a 4K dictionary
while a 1 bit indicates an 8K dictionary.

The Shannon-Fano trees are stored at the start of the compressed
file. The number of trees stored is defined by bit 2 in the
general purpose flag word, a 0 bit indicates two trees stored, a
1 bit indicates three trees are stored.  If 3 trees are stored,
the first Shannon-Fano tree represents the encoding of the
Literal characters, the second tree represents the encoding of
the Length information, the third represents the encoding of the
Distance information.  When 2 Shannon-Fano trees are stored, the
Length tree is stored first, followed by the Distance tree.

The Literal Shannon-Fano tree, if present is used to represent
the entire ASCII character set, and contains 256 values.  This
tree is used to compress any data not compressed by the sliding
dictionary algorithm.  When this tree is present, the Minimum
Match Length for the sliding dictionary is 3.  If this tree is
not present, the Minimum Match Length is 2.

The Length Shannon-Fano tree is used to compress the Length part
of the (length,distance) pairs from the sliding dictionary
output.  The Length tree contains 64 values, ranging from the
Minimum Match Length, to 63 plus the Minimum Match Length.

The Distance Shannon-Fano tree is used to compress the Distance
part of the (length,distance) pairs from the sliding dictionary
output. The Distance tree contains 64 values, ranging from 0 to
63, representing the upper 6 bits of the distance value.  The
distance values themselves will be between 0 and the sliding
dictionary size, either 4K or 8K.

The Shannon-Fano trees themselves are stored in a compressed
format. The first byte of the tree data represents the number of
bytes of data representing the (compressed) Shannon-Fano tree
minus 1.  The remaining bytes represent the Shannon-Fano tree
data encoded as:

    High 4 bits: Number of values at this bit length + 1. (1 - 16)
    Low  4 bits: Bit Length needed to represent value + 1. (1 - 16)

The Shannon-Fano codes can be constructed from the bit lengths
using the following algorithm:

1)  Sort the Bit Lengths in ascending order, while retaining the
    order of the original lengths stored in the file.

2)  Generate the Shannon-Fano trees:

    Code <- 0
    CodeIncrement <- 0
    LastBitLength <- 0
    i <- number of Shannon-Fano codes - 1   (either 255 or 63)

    loop while i >= 0
	Code = Code + CodeIncrement
	if BitLength(i) <> LastBitLength then
	    LastBitLength=BitLength(i)
	    CodeIncrement = 1 shifted left (16 - LastBitLength)
	ShannonCode(i) = Code
	i <- i - 1
    end loop


3)  Reverse the order of all the bits in the above ShannonCode()
    vector, so that the most significant bit becomes the least
    significant bit.  For example, the value 0x1234 (hex) would
    become 0x2C48 (hex).

4)  Restore the order of Shannon-Fano codes as originally stored
    within the file.

Example:

    This example will show the encoding of a Shannon-Fano tree
    of size 8.  Notice that the actual Shannon-Fano trees used
    for Imploding are either 64 or 256 entries in size.

Example:   0x02, 0x42, 0x01, 0x13

    The first byte indicates 3 values in this table.  Decoding the
    bytes:
	    0x42 = 5 codes of 3 bits long
	    0x01 = 1 code  of 2 bits long
	    0x13 = 2 codes of 4 bits long

    This would generate the original bit length array of:
    (3, 3, 3, 3, 3, 2, 4, 4)

    There are 8 codes in this table for the values 0 thru 7.  Using the
    algorithm to obtain the Shannon-Fano codes produces:

                                  Reversed     Order     Original
Val  Sorted   Constructed Code      Value     Restored    Length
---  ------   -----------------   --------    --------    ------
0:     2      1100000000000000        11       101          3
1:     3      1010000000000000       101       001          3
2:     3      1000000000000000       001       110          3
3:     3      0110000000000000       110       010          3
4:     3      0100000000000000       010       100          3
5:     3      0010000000000000       100        11          2
6:     4      0001000000000000      1000      1000          4
7:     4      0000000000000000      0000      0000          4


The values in the Val, Order Restored and Original Length columns
now represent the Shannon-Fano encoding tree that can be used for
decoding the Shannon-Fano encoded data.  How to parse the
variable length Shannon-Fano values from the data stream is beyond the
scope of this document.  (See the references listed at the end of
this document for more information.)  However, traditional decoding
schemes used for Huffman variable length decoding, such as the
Greenlaw algorithm, can be succesfully applied.

The compressed data stream begins immediately after the
compressed Shannon-Fano data.  The compressed data stream can be
interpreted as follows:

loop until done
    read 1 bit from input stream.

    if this bit is non-zero then       (encoded data is literal data)
	if Literal Shannon-Fano tree is present
	    read and decode character using Literal Shannon-Fano tree.
	otherwise
	    read 8 bits from input stream.
	copy character to the output stream.
    otherwise                   (encoded data is sliding dictionary match)
	if 8K dictionary size
	    read 7 bits for offset Distance (lower 7 bits of offset).
	otherwise
	    read 6 bits for offset Distance (lower 6 bits of offset).
	
	using the Distance Shannon-Fano tree, read and decode the
	  upper 6 bits of the Distance value.

	using the Length Shannon-Fano tree, read and decode
	  the Length value.
	
	Length <- Length + Minimum Match Length
	
	if Length = 63 + Minimum Match Length
	    read 8 bits from the input stream,
	    add this value to Length.

	move backwards Distance+1 bytes in the output stream, and
	copy Length characters from this position to the output
	stream.  (if this position is before the start of the output
	stream, then assume that all the data before the start of
	the output stream is filled with zeros).
end loop

Decryption
----------

The encryption used in PKZIP was generously supplied by Roger
Schlafly.  PKWARE is grateful to Mr. Schlafly for his expert
help and advice in the field of data encryption.

PKZIP encrypts the compressed data stream.  Encrypted files must
be decrypted before they can be extracted.

Each encrypted file has an extra 12 bytes stored at the start of
the data area defining the encryption header for that file.  The
encryption header is originally set to random values, and then
itself encrypted, using 3, 32-bit keys.  The key values are 
initialized using the supplied encryption password.  After each byte
is encrypted, the keys are then updated using psuedo-random number
generation techniques in combination with the same CRC-32 algorithm 
used in PKZIP and described elsewhere in this document.

The following is the basic steps required to decrypt a file:

1) Initialize the three 32-bit keys with the password.
2) Read and decrypt the 12-byte encryption header, further
   initializing the encryption keys.
3) Read and decrypt the compressed data stream using the
   encryption keys.


Step 1 - Initializing the encryption keys
-----------------------------------------

Key(0) <- 305419896
Key(1) <- 591751049
Key(2) <- 878082192

loop for i <- 0 to length(password)-1
    update_keys(password(i))
end loop


Where update_keys() is defined as:


update_keys(char):
  Key(0) <- crc32(key(0),char)
  Key(1) <- Key(1) + (Key(0) & 000000ffH)
  Key(1) <- Key(1) * 134775813 + 1
  Key(2) <- crc32(key(2),key(1) >> 24)
end update_keys


Where crc32(old_crc,char) is a routine that given a CRC value and a 
character, returns an updated CRC value after applying the CRC-32 
algorithm described elsewhere in this document.


Step 2 - Decrypting the encryption header
-----------------------------------------

The purpose of this step is to further initialize the encryption
keys, based on random data, to render a plaintext attack on the
data ineffective.


Read the 12-byte encryption header into Buffer, in locations
Buffer(0) thru Buffer(11).

loop for i <- 0 to 11
    C <- buffer(i) ^ decrypt_byte()
    update_keys(C)
    buffer(i) <- C
end loop


Where decrypt_byte() is defined as:


unsigned char decrypt_byte()
    local unsigned short temp
    temp <- Key(2) | 2
    decrypt_byte <- (temp * (temp ^ 1)) >> 8
end decrypt_byte


After the header is decrypted, the last two bytes in Buffer
should be the high-order word of the CRC for the file being
decrypted, stored in Intel low-byte/high-byte order.  This can
be used to test if the password supplied is correct or not.


Step 3 - Decrypting the compressed data stream
----------------------------------------------

The compressed data stream can be decrypted as follows:


loop until done
    read a charcter into C
    Temp <- C ^ decrypt_byte()
    update_keys(temp)
    output Temp
end loop


In addition to the above mentioned contributors to PKZIP and PKUNZIP, 
I would like to extend special thanks to Robert Mahoney for suggesting 
the extension .ZIP for this software.


References:

    Storer, James A. "Data Compression, Methods and Theory",
       Computer Science Press, 1988
    
    Held, Gilbert  "Data Compression, Techniques and Applications,
		    Hardware and Software Considerations"
       John Wiley & Sons, 1987