metc file formats
=================
2009-10-28 .. 2010-02-06, solyga@absinth.net

1 Binaries
----------
Infos taken from
http://cpansearch.perl.org/src/KLIMKIN/P2P-pDonkey-0.01/doc/eDonkey-protocol,
the Linux aMule code and some reverse engineering.

1.1 known.met
-------------
Syntax (with some semantics):
(terminals start with 0 (zero), nonterminals start with a letter)

Format         = ( 0x0e | 0x0f )  FileStatusList .	(* met magic *)
FileStatusList = DWORD { FileStatus } .			(* number of items *)
FileStatus     = DATE HASH PartHashList TagList .	(* ed2k hash *)
PartHashList   = WORD { HASH } .			(* number of items *)
TagList        = DWORD { Tag } .			(* number of items *)
Tag            = 0x02 TagName String |			(* tag (value) type *)
                 0x03 TagName DWORD |
                 0x04 TagName FLOAT .
TagName        = ( 0x0001 SpecialTag ) |
                 ( 0x0002 TwoByteTag ) |
                 ( 0x0003 ThreeByteTag ) |
                 String .				(* strlen # 1..3 reqd *)
SpecialTag     = 0x01 |				(* name *)
                 0x02 | 			(* size *)
                 0x03 |				(* type *)
                 0x04 |				(* format *)
                 0x08 |				(* copied *)
TwoByteTag     = DWORD .
                 

String         = WORD { CHAR } .		(* number of chars *)

FLOAT          = DWORD .		(* 4 bytes, intel coding *)
DATE           = DWORD .		(* seconds since 1.1.1970)
CHAR           = BYTE .			(* some vals may be disallowed *)
WORD           = BYTE BYTE .		(* 2 bytes, little endian *)
DWORD          = WORD WORD .		(* 4 bytes, little endian *)
QuadWord       = DWORD DWORD .		(* 8 bytes, little endian *)
HASH           = QuadWord QuadWord .	(* 16 bytes, little endian (?) *)
BYTE           = 0x01 .. 0xff .

1.2 part.met
------------
Format         = 0xe0 FileStatus .


2 Sources
---------
2.1 known.met
-------------
Syntax (with some semantics):

Syntax    = Magic FileList .
Magic     = 'known' | 'knownLarge' .
FileList  = 'file' FileCount '(' File ')' .
File      = Date HASH HashList TagList .
Date      = Year '-' Month '-' Day Hour ':' Minute ':' Second DST .
HashList  = 'parts' HashCount '(' HASH ')' .
TagList   = 'tags' TagCount '(' Tag ')' .
Tag       = TagName '=' TagValue .
TagName   = STRING | NUMBER .
TagValue  = STRING | NUMBER .

FileCount = NUMBER .
HashCount = NUMBER .
TagCount  = NUMBER .
Year      = NUMBER .
Month     = NUMBER .
Day       = NUMBER .
Hour      = NUMBER .
Minute    = NUMBER .
Second    = NUMBER .
DST       = NUMBER .   (* DST # 0  <==> daylight saving time active *)

scanner output symbols:
  keywords: 'known', 'knowLarge', 'files', 'parts', 'tags'
  simple  : '(', ')', '-', ':', '=', '+'
  complex : NUMBER, STRING, HASH

NUMBER     = '$' HexNum | DecNum .
HexNum     = HexDigit { HexDigit } .
DecNum     = DecDigit { DecDigit } .
HexDigit   = DecDigit | 'a' .. 'f' | 'A' .. 'F' .
DecDigit   = '0' .. '9' .

STRING     = StringPart { '+' StringPart } .
StringPart = Single | Double .
Single     = '"' { Char - '"' } '"' .
Double     = "'" { Char - "'" } "'" .

HASH       = '#' HexDigit (* insert 62 digits here *) HexDigit .


Examples:

The shortest of all valid source texts is
----- snip -----
known files 0()
----- snip -----
End-of-line is not required.

A little bit longer was
----- snip -----
known
files 1 (
  2009-09-17 22:10:04 1 (* 4AB2979C *)
  #966282E09C3C715A6E6B310E6EF1B63D
  parts 0 (
  )
  tags 9 (
    $01 (* name *) = 'ÄÖÜäöüß'
    $01 (* name *) = 'ÄÖÜäöüß'
    $02 (* size *) = $0000000B (* 11 *)
    $50 (* xfdl *) = $00000000 (* 0 *)
    $54 (* xfdh *) = $00000000 (* 0 *)
    $51 (* rqed *) = $00000000 (* 0 *)
    $52 (* accd *) = $00000000 (* 0 *)
    $19 (* priu *) = $00000006 (* 6 *)
    $27 (* aich *) = 'VUW5JYNY3YMB2MSFWJ7S4FL4NQGIJMAB'
  )
)
----- snip -----

2.2 part.met
------------
Syntax    = 'part' FileList .