compress 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394
  1. #------------------------------------------------------------------------------
  2. # $File: compress,v 1.77 2019/10/08 20:25:13 christos Exp $
  3. # compress: file(1) magic for pure-compression formats (no archives)
  4. #
  5. # compress, gzip, pack, compact, huf, squeeze, crunch, freeze, yabba, etc.
  6. #
  7. # Formats for various forms of compressed data
  8. # Formats for "compress" proper have been moved into "compress.c",
  9. # because it tries to uncompress it to figure out what's inside.
  10. # standard unix compress
  11. 0 string \037\235 compress'd data
  12. !:mime application/x-compress
  13. !:apple LZIVZIVU
  14. >2 byte&0x80 >0 block compressed
  15. >2 byte&0x1f x %d bits
  16. # gzip (GNU zip, not to be confused with Info-ZIP or PKWARE zip archiver)
  17. # URL: https://en.wikipedia.org/wiki/Gzip
  18. # Reference: https://tools.ietf.org/html/rfc1952
  19. # Update: Joerg Jenderek, Apr 2019
  20. # Edited by Chris Chittleborough <cchittleborough@yahoo.com.au>, March 2002
  21. # * Original filename is only at offset 10 if "extra field" absent
  22. # * Produce shorter output - notably, only report compression methods
  23. # other than 8 ("deflate", the only method defined in RFC 1952).
  24. # Note: find defs -iname '*.trid.xml' -exec grep -q '<Bytes>1F8B08' {} \; -ls
  25. # TODO:
  26. # FBR Blueberry FlashBack screen Record https://www.flashbackrecorder.com/
  27. # KPR KOffice/Calligra KPresenter application/x-kpresenter
  28. # KPT KOffice/Calligra KPresenter template? application/x-kpresenter
  29. # SAV Diggles Saved Game File http://www.innonics.com
  30. # SAV FarCry (demo) saved game http://www.farcry-thegame.com
  31. # DAT ZOAGZIP game data format http://en.wikipedia.org/wiki/SD_Gundam_Capsule_Fighter
  32. 0 string \037\213
  33. # to display gzip compressed (strength=100=2*50) before other (strength=50)?
  34. #!:strength * 2
  35. # no FNAME and FCOMMENT bit implies no file name/comment. That means only binary
  36. >3 byte&0x18 =0
  37. # For binary gzipped no ASCII text should occur
  38. # mcd-monu-cad.trid.xml
  39. >>10 string MCD Monu-Cad Drawing, Component or Font
  40. #>>36 string Created\ with\ MONU-CAD
  41. #!:mime application/octet-stream
  42. # http://fileformats.archiveteam.org/wiki/Monu-CAD
  43. # http://www.monucad.com/downloads/FullDemo-2005.EXE
  44. # /HANDS96.MCC Component
  45. # /DEMO_DD01.MCD Drawing
  46. # /MCALF020.FNT Font
  47. !:ext mcc/mcd/fnt
  48. # http://www.generalcadd.com
  49. >>10 string GXD General CADD, Drawing or Component
  50. #!:mime application/octet-stream
  51. # /gxc/BUILDINGEDGE.gxc Component
  52. # /gxd/HOCKETT-STPAUL-WRHSE.gxd Drawing
  53. # /gxd/POWERLAND-MILL-ADD-11.gxd Drawing v9.1.06
  54. !:ext gxc/gxd
  55. #>>>13 ubyte 0 \b, version 0
  56. >>>13 string 09 \b, version 9
  57. # other gzipped binary like gzipped tar, VirtualBox extension package,...
  58. >>10 default x gzip compressed data
  59. !:mime application/gzip
  60. >>>0 use gzip-info
  61. # size of the original (uncompressed) input data modulo 2^32
  62. >>>-4 ulelong x \b, original size modulo 2^32 %u
  63. # gzipped TAR or VirtualBox extension package
  64. #!:mime application/x-compressed-tar
  65. #!:mime application/x-virtualbox-vbox-extpack
  66. # https://www.w3.org/TR/SVG/mimereg.html
  67. #!:mime image/image/svg+xml-compressed
  68. # zlib.3.gz
  69. # microcode-20180312.tgz
  70. # tpz same as tgz
  71. # lua-md5_1.2-1_i386_i486.ipk https://en.wikipedia.org/wiki/Opkg
  72. # Oracle_VM_VirtualBox_Extension_Pack-5.0.12-104815.vbox-extpack
  73. !:ext gz/tgz/tpz/ipk/vbox-extpack/svgz
  74. # FNAME/FCOMMENT bit implies file name/comment as iso-8859-1 text
  75. >3 byte&0x18 >0 gzip compressed data
  76. !:mime application/gzip
  77. # gzipped tar, gzipped Abiword document
  78. #!:mime application/x-compressed-tar
  79. #!:mime application/x-abiword-compressed
  80. #!:mime image/image/svg+xml-compressed
  81. # kleopatra_splashscreen.svgz gzipped .svg
  82. !:ext gz/tgz/tpz/zabw/svgz
  83. >>0 use gzip-info
  84. # size of the original (uncompressed) input data modulo 2^32
  85. >>-4 ulelong x \b, original size modulo 2^32 %u
  86. # display information of gzip compressed files
  87. 0 name gzip-info
  88. #>2 byte x THIS iS GZIP
  89. >2 byte <8 \b, reserved method
  90. >2 byte >8 \b, unknown method
  91. >3 byte &0x01 \b, ASCII
  92. >3 byte &0x02 \b, has CRC
  93. >3 byte &0x04 \b, extra field
  94. >3 byte&0xC =0x08
  95. >>10 string x \b, was "%s"
  96. >3 byte &0x10 \b, has comment
  97. >3 byte &0x20 \b, encrypted
  98. >4 ledate >0 \b, last modified: %s
  99. >8 byte 2 \b, max compression
  100. >8 byte 4 \b, max speed
  101. >9 byte =0x00 \b, from FAT filesystem (MS-DOS, OS/2, NT)
  102. >9 byte =0x01 \b, from Amiga
  103. >9 byte =0x02 \b, from VMS
  104. >9 byte =0x03 \b, from Unix
  105. >9 byte =0x04 \b, from VM/CMS
  106. >9 byte =0x05 \b, from Atari
  107. >9 byte =0x06 \b, from HPFS filesystem (OS/2, NT)
  108. >9 byte =0x07 \b, from MacOS
  109. >9 byte =0x08 \b, from Z-System
  110. >9 byte =0x09 \b, from CP/M
  111. >9 byte =0x0A \b, from TOPS/20
  112. >9 byte =0x0B \b, from NTFS filesystem (NT)
  113. >9 byte =0x0C \b, from QDOS
  114. >9 byte =0x0D \b, from Acorn RISCOS
  115. # size of the original (uncompressed) input data modulo 2^32
  116. #>-4 ulelong x \b, original size modulo 2^32 %u
  117. #ERROR: line 114: non zero offset 1048572 at level 1
  118. # packed data, Huffman (minimum redundancy) codes on a byte-by-byte basis
  119. 0 string \037\036 packed data
  120. !:mime application/octet-stream
  121. >2 belong >1 \b, %d characters originally
  122. >2 belong =1 \b, %d character originally
  123. #
  124. # This magic number is byte-order-independent.
  125. 0 short 0x1f1f old packed data
  126. !:mime application/octet-stream
  127. # XXX - why *two* entries for "compacted data", one of which is
  128. # byte-order independent, and one of which is byte-order dependent?
  129. #
  130. 0 short 0x1fff compacted data
  131. !:mime application/octet-stream
  132. # This string is valid for SunOS (BE) and a matching "short" is listed
  133. # in the Ultrix (LE) magic file.
  134. 0 string \377\037 compacted data
  135. !:mime application/octet-stream
  136. 0 short 0145405 huf output
  137. !:mime application/octet-stream
  138. # bzip2
  139. 0 string BZh bzip2 compressed data
  140. !:mime application/x-bzip2
  141. >3 byte >47 \b, block size = %c00k
  142. # bzip a block-sorting file compressor
  143. # by Julian Seward <sewardj@cs.man.ac.uk> and others
  144. 0 string BZ0 bzip compressed data
  145. !:mime application/x-bzip
  146. >3 byte >47 \b, block size = %c00k
  147. # lzip
  148. 0 string LZIP lzip compressed data
  149. !:mime application/x-lzip
  150. >4 byte x \b, version: %d
  151. # squeeze and crunch
  152. # Michael Haardt <michael@cantor.informatik.rwth-aachen.de>
  153. 0 beshort 0x76FF squeezed data,
  154. >4 string x original name %s
  155. 0 beshort 0x76FE crunched data,
  156. >2 string x original name %s
  157. 0 beshort 0x76FD LZH compressed data,
  158. >2 string x original name %s
  159. # Freeze
  160. 0 string \037\237 frozen file 2.1
  161. 0 string \037\236 frozen file 1.0 (or gzip 0.5)
  162. # SCO compress -H (LZH)
  163. 0 string \037\240 SCO compress -H (LZH) data
  164. # European GSM 06.10 is a provisional standard for full-rate speech
  165. # transcoding, prI-ETS 300 036, which uses RPE/LTP (residual pulse
  166. # excitation/long term prediction) coding at 13 kbit/s.
  167. #
  168. # There's only a magic nibble (4 bits); that nibble repeats every 33
  169. # bytes. This isn't suited for use, but maybe we can use it someday.
  170. #
  171. # This will cause very short GSM files to be declared as data and
  172. # mismatches to be declared as data too!
  173. #0 byte&0xF0 0xd0 data
  174. #>33 byte&0xF0 0xd0
  175. #>66 byte&0xF0 0xd0
  176. #>99 byte&0xF0 0xd0
  177. #>132 byte&0xF0 0xd0 GSM 06.10 compressed audio
  178. # lzop from <markus.oberhumer@jk.uni-linz.ac.at>
  179. 0 string \x89\x4c\x5a\x4f\x00\x0d\x0a\x1a\x0a lzop compressed data
  180. >9 beshort <0x0940
  181. >>9 byte&0xf0 =0x00 - version 0.
  182. >>9 beshort&0x0fff x \b%03x,
  183. >>13 byte 1 LZO1X-1,
  184. >>13 byte 2 LZO1X-1(15),
  185. >>13 byte 3 LZO1X-999,
  186. ## >>22 bedate >0 last modified: %s,
  187. >>14 byte =0x00 os: MS-DOS
  188. >>14 byte =0x01 os: Amiga
  189. >>14 byte =0x02 os: VMS
  190. >>14 byte =0x03 os: Unix
  191. >>14 byte =0x05 os: Atari
  192. >>14 byte =0x06 os: OS/2
  193. >>14 byte =0x07 os: MacOS
  194. >>14 byte =0x0A os: Tops/20
  195. >>14 byte =0x0B os: WinNT
  196. >>14 byte =0x0E os: Win32
  197. >9 beshort >0x0939
  198. >>9 byte&0xf0 =0x00 - version 0.
  199. >>9 byte&0xf0 =0x10 - version 1.
  200. >>9 byte&0xf0 =0x20 - version 2.
  201. >>9 beshort&0x0fff x \b%03x,
  202. >>15 byte 1 LZO1X-1,
  203. >>15 byte 2 LZO1X-1(15),
  204. >>15 byte 3 LZO1X-999,
  205. ## >>25 bedate >0 last modified: %s,
  206. >>17 byte =0x00 os: MS-DOS
  207. >>17 byte =0x01 os: Amiga
  208. >>17 byte =0x02 os: VMS
  209. >>17 byte =0x03 os: Unix
  210. >>17 byte =0x05 os: Atari
  211. >>17 byte =0x06 os: OS/2
  212. >>17 byte =0x07 os: MacOS
  213. >>17 byte =0x0A os: Tops/20
  214. >>17 byte =0x0B os: WinNT
  215. >>17 byte =0x0E os: Win32
  216. # 4.3BSD-Quasijarus Strong Compression
  217. # https://minnie.tuhs.org/Quasijarus/compress.html
  218. 0 string \037\241 Quasijarus strong compressed data
  219. # From: Cory Dikkers <cdikkers@swbell.net>
  220. 0 string XPKF Amiga xpkf.library compressed data
  221. 0 string PP11 Power Packer 1.1 compressed data
  222. 0 string PP20 Power Packer 2.0 compressed data,
  223. >4 belong 0x09090909 fast compression
  224. >4 belong 0x090A0A0A mediocre compression
  225. >4 belong 0x090A0B0B good compression
  226. >4 belong 0x090A0C0C very good compression
  227. >4 belong 0x090A0C0D best compression
  228. # 7-zip archiver, from Thomas Klausner (wiz@danbala.tuwien.ac.at)
  229. # https://www.7-zip.org or DOC/7zFormat.txt
  230. #
  231. 0 string 7z\274\257\047\034 7-zip archive data,
  232. >6 byte x version %d
  233. >7 byte x \b.%d
  234. !:mime application/x-7z-compressed
  235. !:ext 7z/cb7
  236. # Type: LZMA
  237. 0 lelong&0xffffff =0x5d
  238. >12 leshort 0xff LZMA compressed data,
  239. !:mime application/x-lzma
  240. >>5 lequad =0xffffffffffffffff streamed
  241. >>5 lequad !0xffffffffffffffff non-streamed, size %lld
  242. >12 leshort 0 LZMA compressed data,
  243. >>5 lequad =0xffffffffffffffff streamed
  244. >>5 lequad !0xffffffffffffffff non-streamed, size %lld
  245. # http://tukaani.org/xz/xz-file-format.txt
  246. 0 ustring \xFD7zXZ\x00 XZ compressed data
  247. !:strength * 2
  248. !:mime application/x-xz
  249. # https://github.com/ckolivas/lrzip/blob/master/doc/magic.header.txt
  250. 0 string LRZI LRZIP compressed data
  251. >4 byte x - version %d
  252. >5 byte x \b.%d
  253. !:mime application/x-lrzip
  254. # https://fastcompression.blogspot.fi/2013/04/lz4-streaming-format-final.html
  255. 0 lelong 0x184d2204 LZ4 compressed data (v1.4+)
  256. !:mime application/x-lz4
  257. # Added by osm0sis@xda-developers.com
  258. 0 lelong 0x184c2103 LZ4 compressed data (v1.0-v1.3)
  259. !:mime application/x-lz4
  260. 0 lelong 0x184c2102 LZ4 compressed data (v0.1-v0.9)
  261. !:mime application/x-lz4
  262. # Zstandard/LZ4 skippable frames
  263. # https://github.com/facebook/zstd/blob/dev/zstd_compression_format.md
  264. 0 lelong&0xFFFFFFF0 0x184D2A50
  265. >(4.l+8) indirect x
  266. # Zstandard Dictionary ID subroutine
  267. 0 name zstd-dictionary-id
  268. # Single Segment = True
  269. >0 byte &0x20 \b, Dictionary ID:
  270. >>0 byte&0x03 0 None
  271. >>0 byte&0x03 1
  272. >>>1 byte x %u
  273. >>0 byte&0x03 2
  274. >>>1 leshort x %u
  275. >>0 byte&0x03 3
  276. >>>1 lelong x %u
  277. # Single Segment = False
  278. >0 byte ^0x20 \b, Dictionary ID:
  279. >>0 byte&0x03 0 None
  280. >>0 byte&0x03 1
  281. >>>2 byte x %u
  282. >>0 byte&0x03 2
  283. >>>2 leshort x %u
  284. >>0 byte&0x03 3
  285. >>>2 lelong x %u
  286. # Zstandard compressed data
  287. # https://github.com/facebook/zstd/blob/dev/zstd_compression_format.md
  288. 0 lelong 0xFD2FB522 Zstandard compressed data (v0.2)
  289. !:mime application/zstd
  290. 0 lelong 0xFD2FB523 Zstandard compressed data (v0.3)
  291. !:mime application/zstd
  292. 0 lelong 0xFD2FB524 Zstandard compressed data (v0.4)
  293. !:mime application/zstd
  294. 0 lelong 0xFD2FB525 Zstandard compressed data (v0.5)
  295. !:mime application/zstd
  296. 0 lelong 0xFD2FB526 Zstandard compressed data (v0.6)
  297. !:mime application/zstd
  298. 0 lelong 0xFD2FB527 Zstandard compressed data (v0.7)
  299. !:mime application/zstd
  300. >4 use zstd-dictionary-id
  301. 0 lelong 0xFD2FB528 Zstandard compressed data (v0.8+)
  302. !:mime application/zstd
  303. >4 use zstd-dictionary-id
  304. # https://github.com/facebook/zstd/blob/dev/zstd_compression_format.md
  305. 0 lelong 0xEC30A437 Zstandard dictionary
  306. !:mime application/x-std-dictionary
  307. >4 lelong x (ID %u)
  308. # AFX compressed files (Wolfram Kleff)
  309. 2 string -afx- AFX compressed file data
  310. # Supplementary magic data for the file(1) command to support
  311. # rzip(1). The format is described in magic(5).
  312. #
  313. # Copyright (C) 2003 by Andrew Tridgell. You may do whatever you want with
  314. # this file.
  315. #
  316. 0 string RZIP rzip compressed data
  317. >4 byte x - version %d
  318. >5 byte x \b.%d
  319. >6 belong x (%d bytes)
  320. 0 string ArC\x01 FreeArc archive <http://freearc.org>
  321. # Type: DACT compressed files
  322. 0 long 0x444354C3 DACT compressed data
  323. >4 byte >-1 (version %i.
  324. >5 byte >-1 %i.
  325. >6 byte >-1 %i)
  326. >7 long >0 , original size: %i bytes
  327. >15 long >30 , block size: %i bytes
  328. # Valve Pack (VPK) files
  329. 0 lelong 0x55aa1234 Valve Pak file
  330. >0x4 lelong x \b, version %u
  331. >0x8 lelong x \b, %u entries
  332. # Snappy framing format
  333. # https://code.google.com/p/snappy/source/browse/trunk/framing_format.txt
  334. 0 string \377\006\0\0sNaPpY snappy framed data
  335. !:mime application/x-snappy-framed
  336. # qpress, https://www.quicklz.com/
  337. 0 string qpress10 qpress compressed data
  338. !:mime application/x-qpress
  339. # Zlib https://www.ietf.org/rfc/rfc6713.txt
  340. 0 string/b x
  341. >0 beshort%31 =0
  342. >>0 byte&0xf =8
  343. >>>0 byte&0x80 =0 zlib compressed data
  344. !:mime application/zlib
  345. # BWC compression
  346. 0 string BWC
  347. >3 byte 0 BWC compressed data
  348. # UCL compression
  349. 0 bequad 0x00e955434cff011a UCL compressed data
  350. # Softlib archive
  351. 0 string SLIB Softlib archive
  352. >4 leshort x \b, version %d
  353. >6 leshort x (contains %d files)
  354. # URL: https://github.com/lzfse/lzfse/blob/master/src/lzfse_internal.h#L276
  355. # From: Eric Hall <eric.hall@darkart.com>
  356. 0 string bvx- lzfse encoded, no compression
  357. 0 string bvx1 lzfse compressed, uncompressed tables
  358. 0 string bvx2 lzfse compressed, compressed tables
  359. 0 string bvxn lzfse encoded, lzvn compressed