binhex.py 14 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518
  1. """Macintosh binhex compression/decompression.
  2. easy interface:
  3. binhex(inputfilename, outputfilename)
  4. hexbin(inputfilename, outputfilename)
  5. """
  6. #
  7. # Jack Jansen, CWI, August 1995.
  8. #
  9. # The module is supposed to be as compatible as possible. Especially the
  10. # easy interface should work "as expected" on any platform.
  11. # XXXX Note: currently, textfiles appear in mac-form on all platforms.
  12. # We seem to lack a simple character-translate in python.
  13. # (we should probably use ISO-Latin-1 on all but the mac platform).
  14. # XXXX The simple routines are too simple: they expect to hold the complete
  15. # files in-core. Should be fixed.
  16. # XXXX It would be nice to handle AppleDouble format on unix
  17. # (for servers serving macs).
  18. # XXXX I don't understand what happens when you get 0x90 times the same byte on
  19. # input. The resulting code (xx 90 90) would appear to be interpreted as an
  20. # escaped *value* of 0x90. All coders I've seen appear to ignore this nicety...
  21. #
  22. import sys
  23. import os
  24. import struct
  25. import binascii
  26. __all__ = ["binhex","hexbin","Error"]
  27. class Error(Exception):
  28. pass
  29. # States (what have we written)
  30. _DID_HEADER = 0
  31. _DID_DATA = 1
  32. # Various constants
  33. REASONABLY_LARGE=32768 # Minimal amount we pass the rle-coder
  34. LINELEN=64
  35. RUNCHAR=chr(0x90) # run-length introducer
  36. #
  37. # This code is no longer byte-order dependent
  38. #
  39. # Workarounds for non-mac machines.
  40. try:
  41. from Carbon.File import FSSpec, FInfo
  42. from MacOS import openrf
  43. def getfileinfo(name):
  44. finfo = FSSpec(name).FSpGetFInfo()
  45. dir, file = os.path.split(name)
  46. # XXX Get resource/data sizes
  47. fp = open(name, 'rb')
  48. fp.seek(0, 2)
  49. dlen = fp.tell()
  50. fp = openrf(name, '*rb')
  51. fp.seek(0, 2)
  52. rlen = fp.tell()
  53. return file, finfo, dlen, rlen
  54. def openrsrc(name, *mode):
  55. if not mode:
  56. mode = '*rb'
  57. else:
  58. mode = '*' + mode[0]
  59. return openrf(name, mode)
  60. except ImportError:
  61. #
  62. # Glue code for non-macintosh usage
  63. #
  64. class FInfo:
  65. def __init__(self):
  66. self.Type = '????'
  67. self.Creator = '????'
  68. self.Flags = 0
  69. def getfileinfo(name):
  70. finfo = FInfo()
  71. # Quick check for textfile
  72. fp = open(name)
  73. data = open(name).read(256)
  74. for c in data:
  75. if not c.isspace() and (c<' ' or ord(c) > 0x7f):
  76. break
  77. else:
  78. finfo.Type = 'TEXT'
  79. fp.seek(0, 2)
  80. dsize = fp.tell()
  81. fp.close()
  82. dir, file = os.path.split(name)
  83. file = file.replace(':', '-', 1)
  84. return file, finfo, dsize, 0
  85. class openrsrc:
  86. def __init__(self, *args):
  87. pass
  88. def read(self, *args):
  89. return ''
  90. def write(self, *args):
  91. pass
  92. def close(self):
  93. pass
  94. class _Hqxcoderengine:
  95. """Write data to the coder in 3-byte chunks"""
  96. def __init__(self, ofp):
  97. self.ofp = ofp
  98. self.data = ''
  99. self.hqxdata = ''
  100. self.linelen = LINELEN-1
  101. def write(self, data):
  102. self.data = self.data + data
  103. datalen = len(self.data)
  104. todo = (datalen//3)*3
  105. data = self.data[:todo]
  106. self.data = self.data[todo:]
  107. if not data:
  108. return
  109. self.hqxdata = self.hqxdata + binascii.b2a_hqx(data)
  110. self._flush(0)
  111. def _flush(self, force):
  112. first = 0
  113. while first <= len(self.hqxdata)-self.linelen:
  114. last = first + self.linelen
  115. self.ofp.write(self.hqxdata[first:last]+'\n')
  116. self.linelen = LINELEN
  117. first = last
  118. self.hqxdata = self.hqxdata[first:]
  119. if force:
  120. self.ofp.write(self.hqxdata + ':\n')
  121. def close(self):
  122. if self.data:
  123. self.hqxdata = \
  124. self.hqxdata + binascii.b2a_hqx(self.data)
  125. self._flush(1)
  126. self.ofp.close()
  127. del self.ofp
  128. class _Rlecoderengine:
  129. """Write data to the RLE-coder in suitably large chunks"""
  130. def __init__(self, ofp):
  131. self.ofp = ofp
  132. self.data = ''
  133. def write(self, data):
  134. self.data = self.data + data
  135. if len(self.data) < REASONABLY_LARGE:
  136. return
  137. rledata = binascii.rlecode_hqx(self.data)
  138. self.ofp.write(rledata)
  139. self.data = ''
  140. def close(self):
  141. if self.data:
  142. rledata = binascii.rlecode_hqx(self.data)
  143. self.ofp.write(rledata)
  144. self.ofp.close()
  145. del self.ofp
  146. class BinHex:
  147. def __init__(self, name_finfo_dlen_rlen, ofp):
  148. name, finfo, dlen, rlen = name_finfo_dlen_rlen
  149. if type(ofp) == type(''):
  150. ofname = ofp
  151. ofp = open(ofname, 'w')
  152. ofp.write('(This file must be converted with BinHex 4.0)\n\n:')
  153. hqxer = _Hqxcoderengine(ofp)
  154. self.ofp = _Rlecoderengine(hqxer)
  155. self.crc = 0
  156. if finfo is None:
  157. finfo = FInfo()
  158. self.dlen = dlen
  159. self.rlen = rlen
  160. self._writeinfo(name, finfo)
  161. self.state = _DID_HEADER
  162. def _writeinfo(self, name, finfo):
  163. nl = len(name)
  164. if nl > 63:
  165. raise Error, 'Filename too long'
  166. d = chr(nl) + name + '\0'
  167. d2 = finfo.Type + finfo.Creator
  168. # Force all structs to be packed with big-endian
  169. d3 = struct.pack('>h', finfo.Flags)
  170. d4 = struct.pack('>ii', self.dlen, self.rlen)
  171. info = d + d2 + d3 + d4
  172. self._write(info)
  173. self._writecrc()
  174. def _write(self, data):
  175. self.crc = binascii.crc_hqx(data, self.crc)
  176. self.ofp.write(data)
  177. def _writecrc(self):
  178. # XXXX Should this be here??
  179. # self.crc = binascii.crc_hqx('\0\0', self.crc)
  180. if self.crc < 0:
  181. fmt = '>h'
  182. else:
  183. fmt = '>H'
  184. self.ofp.write(struct.pack(fmt, self.crc))
  185. self.crc = 0
  186. def write(self, data):
  187. if self.state != _DID_HEADER:
  188. raise Error, 'Writing data at the wrong time'
  189. self.dlen = self.dlen - len(data)
  190. self._write(data)
  191. def close_data(self):
  192. if self.dlen != 0:
  193. raise Error, 'Incorrect data size, diff=%r' % (self.rlen,)
  194. self._writecrc()
  195. self.state = _DID_DATA
  196. def write_rsrc(self, data):
  197. if self.state < _DID_DATA:
  198. self.close_data()
  199. if self.state != _DID_DATA:
  200. raise Error, 'Writing resource data at the wrong time'
  201. self.rlen = self.rlen - len(data)
  202. self._write(data)
  203. def close(self):
  204. if self.state is None:
  205. return
  206. try:
  207. if self.state < _DID_DATA:
  208. self.close_data()
  209. if self.state != _DID_DATA:
  210. raise Error, 'Close at the wrong time'
  211. if self.rlen != 0:
  212. raise Error, \
  213. "Incorrect resource-datasize, diff=%r" % (self.rlen,)
  214. self._writecrc()
  215. finally:
  216. self.state = None
  217. ofp = self.ofp
  218. del self.ofp
  219. ofp.close()
  220. def binhex(inp, out):
  221. """(infilename, outfilename) - Create binhex-encoded copy of a file"""
  222. finfo = getfileinfo(inp)
  223. ofp = BinHex(finfo, out)
  224. ifp = open(inp, 'rb')
  225. # XXXX Do textfile translation on non-mac systems
  226. while 1:
  227. d = ifp.read(128000)
  228. if not d: break
  229. ofp.write(d)
  230. ofp.close_data()
  231. ifp.close()
  232. ifp = openrsrc(inp, 'rb')
  233. while 1:
  234. d = ifp.read(128000)
  235. if not d: break
  236. ofp.write_rsrc(d)
  237. ofp.close()
  238. ifp.close()
  239. class _Hqxdecoderengine:
  240. """Read data via the decoder in 4-byte chunks"""
  241. def __init__(self, ifp):
  242. self.ifp = ifp
  243. self.eof = 0
  244. def read(self, totalwtd):
  245. """Read at least wtd bytes (or until EOF)"""
  246. decdata = ''
  247. wtd = totalwtd
  248. #
  249. # The loop here is convoluted, since we don't really now how
  250. # much to decode: there may be newlines in the incoming data.
  251. while wtd > 0:
  252. if self.eof: return decdata
  253. wtd = ((wtd+2)//3)*4
  254. data = self.ifp.read(wtd)
  255. #
  256. # Next problem: there may not be a complete number of
  257. # bytes in what we pass to a2b. Solve by yet another
  258. # loop.
  259. #
  260. while 1:
  261. try:
  262. decdatacur, self.eof = \
  263. binascii.a2b_hqx(data)
  264. break
  265. except binascii.Incomplete:
  266. pass
  267. newdata = self.ifp.read(1)
  268. if not newdata:
  269. raise Error, \
  270. 'Premature EOF on binhex file'
  271. data = data + newdata
  272. decdata = decdata + decdatacur
  273. wtd = totalwtd - len(decdata)
  274. if not decdata and not self.eof:
  275. raise Error, 'Premature EOF on binhex file'
  276. return decdata
  277. def close(self):
  278. self.ifp.close()
  279. class _Rledecoderengine:
  280. """Read data via the RLE-coder"""
  281. def __init__(self, ifp):
  282. self.ifp = ifp
  283. self.pre_buffer = ''
  284. self.post_buffer = ''
  285. self.eof = 0
  286. def read(self, wtd):
  287. if wtd > len(self.post_buffer):
  288. self._fill(wtd-len(self.post_buffer))
  289. rv = self.post_buffer[:wtd]
  290. self.post_buffer = self.post_buffer[wtd:]
  291. return rv
  292. def _fill(self, wtd):
  293. self.pre_buffer = self.pre_buffer + self.ifp.read(wtd+4)
  294. if self.ifp.eof:
  295. self.post_buffer = self.post_buffer + \
  296. binascii.rledecode_hqx(self.pre_buffer)
  297. self.pre_buffer = ''
  298. return
  299. #
  300. # Obfuscated code ahead. We have to take care that we don't
  301. # end up with an orphaned RUNCHAR later on. So, we keep a couple
  302. # of bytes in the buffer, depending on what the end of
  303. # the buffer looks like:
  304. # '\220\0\220' - Keep 3 bytes: repeated \220 (escaped as \220\0)
  305. # '?\220' - Keep 2 bytes: repeated something-else
  306. # '\220\0' - Escaped \220: Keep 2 bytes.
  307. # '?\220?' - Complete repeat sequence: decode all
  308. # otherwise: keep 1 byte.
  309. #
  310. mark = len(self.pre_buffer)
  311. if self.pre_buffer[-3:] == RUNCHAR + '\0' + RUNCHAR:
  312. mark = mark - 3
  313. elif self.pre_buffer[-1] == RUNCHAR:
  314. mark = mark - 2
  315. elif self.pre_buffer[-2:] == RUNCHAR + '\0':
  316. mark = mark - 2
  317. elif self.pre_buffer[-2] == RUNCHAR:
  318. pass # Decode all
  319. else:
  320. mark = mark - 1
  321. self.post_buffer = self.post_buffer + \
  322. binascii.rledecode_hqx(self.pre_buffer[:mark])
  323. self.pre_buffer = self.pre_buffer[mark:]
  324. def close(self):
  325. self.ifp.close()
  326. class HexBin:
  327. def __init__(self, ifp):
  328. if type(ifp) == type(''):
  329. ifp = open(ifp)
  330. #
  331. # Find initial colon.
  332. #
  333. while 1:
  334. ch = ifp.read(1)
  335. if not ch:
  336. raise Error, "No binhex data found"
  337. # Cater for \r\n terminated lines (which show up as \n\r, hence
  338. # all lines start with \r)
  339. if ch == '\r':
  340. continue
  341. if ch == ':':
  342. break
  343. if ch != '\n':
  344. dummy = ifp.readline()
  345. hqxifp = _Hqxdecoderengine(ifp)
  346. self.ifp = _Rledecoderengine(hqxifp)
  347. self.crc = 0
  348. self._readheader()
  349. def _read(self, len):
  350. data = self.ifp.read(len)
  351. self.crc = binascii.crc_hqx(data, self.crc)
  352. return data
  353. def _checkcrc(self):
  354. filecrc = struct.unpack('>h', self.ifp.read(2))[0] & 0xffff
  355. #self.crc = binascii.crc_hqx('\0\0', self.crc)
  356. # XXXX Is this needed??
  357. self.crc = self.crc & 0xffff
  358. if filecrc != self.crc:
  359. raise Error, 'CRC error, computed %x, read %x' \
  360. %(self.crc, filecrc)
  361. self.crc = 0
  362. def _readheader(self):
  363. len = self._read(1)
  364. fname = self._read(ord(len))
  365. rest = self._read(1+4+4+2+4+4)
  366. self._checkcrc()
  367. type = rest[1:5]
  368. creator = rest[5:9]
  369. flags = struct.unpack('>h', rest[9:11])[0]
  370. self.dlen = struct.unpack('>l', rest[11:15])[0]
  371. self.rlen = struct.unpack('>l', rest[15:19])[0]
  372. self.FName = fname
  373. self.FInfo = FInfo()
  374. self.FInfo.Creator = creator
  375. self.FInfo.Type = type
  376. self.FInfo.Flags = flags
  377. self.state = _DID_HEADER
  378. def read(self, *n):
  379. if self.state != _DID_HEADER:
  380. raise Error, 'Read data at wrong time'
  381. if n:
  382. n = n[0]
  383. n = min(n, self.dlen)
  384. else:
  385. n = self.dlen
  386. rv = ''
  387. while len(rv) < n:
  388. rv = rv + self._read(n-len(rv))
  389. self.dlen = self.dlen - n
  390. return rv
  391. def close_data(self):
  392. if self.state != _DID_HEADER:
  393. raise Error, 'close_data at wrong time'
  394. if self.dlen:
  395. dummy = self._read(self.dlen)
  396. self._checkcrc()
  397. self.state = _DID_DATA
  398. def read_rsrc(self, *n):
  399. if self.state == _DID_HEADER:
  400. self.close_data()
  401. if self.state != _DID_DATA:
  402. raise Error, 'Read resource data at wrong time'
  403. if n:
  404. n = n[0]
  405. n = min(n, self.rlen)
  406. else:
  407. n = self.rlen
  408. self.rlen = self.rlen - n
  409. return self._read(n)
  410. def close(self):
  411. if self.state is None:
  412. return
  413. try:
  414. if self.rlen:
  415. dummy = self.read_rsrc(self.rlen)
  416. self._checkcrc()
  417. finally:
  418. self.state = None
  419. self.ifp.close()
  420. def hexbin(inp, out):
  421. """(infilename, outfilename) - Decode binhexed file"""
  422. ifp = HexBin(inp)
  423. finfo = ifp.FInfo
  424. if not out:
  425. out = ifp.FName
  426. ofp = open(out, 'wb')
  427. # XXXX Do translation on non-mac systems
  428. while 1:
  429. d = ifp.read(128000)
  430. if not d: break
  431. ofp.write(d)
  432. ofp.close()
  433. ifp.close_data()
  434. d = ifp.read_rsrc(128000)
  435. if d:
  436. ofp = openrsrc(out, 'wb')
  437. ofp.write(d)
  438. while 1:
  439. d = ifp.read_rsrc(128000)
  440. if not d: break
  441. ofp.write(d)
  442. ofp.close()
  443. ifp.close()
  444. def _test():
  445. fname = sys.argv[1]
  446. binhex(fname, fname+'.hqx')
  447. hexbin(fname+'.hqx', fname+'.viahqx')
  448. #hexbin(fname, fname+'.unpacked')
  449. sys.exit(1)
  450. if __name__ == '__main__':
  451. _test()