Lib/binhex.py

   1 """Macintosh binhex compression/decompression.
   2
   3 easy interface:
   4 binhex(inputfilename, outputfilename)
   5 hexbin(inputfilename, outputfilename)
   6 """
   7
   8 #
   9 # Jack Jansen, CWI, August 1995.
  10 #
  11 # The module is supposed to be as compatible as possible. Especially the
  12 # easy interface should work "as expected" on any platform.
  13 # XXXX Note: currently, textfiles appear in mac-form on all platforms.
  14 # We seem to lack a simple character-translate in python.
  15 # (we should probably use ISO-Latin-1 on all but the mac platform).
  16 # XXXX The simple routines are too simple: they expect to hold the complete
  17 # files in-core. Should be fixed.
  18 # XXXX It would be nice to handle AppleDouble format on unix
  19 # (for servers serving macs).
  20 # XXXX I don't understand what happens when you get 0x90 times the same byte on
  21 # input. The resulting code (xx 90 90) would appear to be interpreted as an
  22 # escaped *value* of 0x90. All coders I've seen appear to ignore this nicety...
  23 #
  24 import sys
  25 import os
  26 import struct
  27 import binascii
  28
  29 __all__ = ["binhex","hexbin","Error"]
  30
  31 class Error(Exception):
  32     pass
  33
  34 # States (what have we written)
  35 [_DID_HEADER, _DID_DATA, _DID_RSRC] = range(3)
  36
  37 # Various constants
  38 REASONABLY_LARGE=32768  # Minimal amount we pass the rle-coder
  39 LINELEN=64
  40 RUNCHAR=chr(0x90)   # run-length introducer
  41
  42 #
  43 # This code is no longer byte-order dependent
  44
  45 #
  46 # Workarounds for non-mac machines.
  47 try:
  48     from Carbon.File import FSSpec, FInfo
  49     from MacOS import openrf
  50
  51     def getfileinfo(name):
  52         finfo = FSSpec(name).FSpGetFInfo()
  53         dir, file = os.path.split(name)
  54         # XXX Get resource/data sizes
  55         fp = open(name, 'rb')
  56         fp.seek(0, 2)
  57         dlen = fp.tell()
  58         fp = openrf(name, '*rb')
  59         fp.seek(0, 2)
  60         rlen = fp.tell()
  61         return file, finfo, dlen, rlen
  62
  63     def openrsrc(name, *mode):
  64         if not mode:
  65             mode = '*rb'
  66         else:
  67             mode = '*' + mode[0]
  68         return openrf(name, mode)
  69
  70 except ImportError:
  71     #
  72     # Glue code for non-macintosh usage
  73     #
  74
  75     class FInfo:
  76         def __init__(self):
  77             self.Type = '????'
  78             self.Creator = '????'
  79             self.Flags = 0
  80
  81     def getfileinfo(name):
  82         finfo = FInfo()
  83         # Quick check for textfile
  84         fp = open(name)
  85         data = open(name).read(256)
  86         for c in data:
  87             if not c.isspace() and (c<' ' or ord(c) > 0x7f):
  88                 break
  89         else:
  90             finfo.Type = 'TEXT'
  91         fp.seek(0, 2)
  92         dsize = fp.tell()
  93         fp.close()
  94         dir, file = os.path.split(name)
  95         file = file.replace(':', '-', 1)
  96         return file, finfo, dsize, 0
  97
  98     class openrsrc:
  99         def __init__(self, *args):
 100             pass
 101
 102         def read(self, *args):
 103             return ''
 104
 105         def write(self, *args):
 106             pass
 107
 108         def close(self):
 109             pass
 110
 111 class _Hqxcoderengine:
 112     """Write data to the coder in 3-byte chunks"""
 113
 114     def __init__(self, ofp):
 115         self.ofp = ofp
 116         self.data = ''
 117         self.hqxdata = ''
 118         self.linelen = LINELEN-1
 119
 120     def write(self, data):
 121         self.data = self.data + data
 122         datalen = len(self.data)
 123         todo = (datalen//3)*3
 124         data = self.data[:todo]
 125         self.data = self.data[todo:]
 126         if not data:
 127             return
 128         self.hqxdata = self.hqxdata + binascii.b2a_hqx(data)
 129         self._flush(0)
 130
 131     def _flush(self, force):
 132         first = 0
 133         while first <= len(self.hqxdata)-self.linelen:
 134             last = first + self.linelen
 135             self.ofp.write(self.hqxdata[first:last]+'\n')
 136             self.linelen = LINELEN
 137             first = last
 138         self.hqxdata = self.hqxdata[first:]
 139         if force:
 140             self.ofp.write(self.hqxdata + ':\n')
 141
 142     def close(self):
 143         if self.data:
 144             self.hqxdata = \
 145                  self.hqxdata + binascii.b2a_hqx(self.data)
 146         self._flush(1)
 147         self.ofp.close()
 148         del self.ofp
 149
 150 class _Rlecoderengine:
 151     """Write data to the RLE-coder in suitably large chunks"""
 152
 153     def __init__(self, ofp):
 154         self.ofp = ofp
 155         self.data = ''
 156
 157     def write(self, data):
 158         self.data = self.data + data
 159         if len(self.data) < REASONABLY_LARGE:
 160             return
 161         rledata = binascii.rlecode_hqx(self.data)
 162         self.ofp.write(rledata)
 163         self.data = ''
 164
 165     def close(self):
 166         if self.data:
 167             rledata = binascii.rlecode_hqx(self.data)
 168             self.ofp.write(rledata)
 169         self.ofp.close()
 170         del self.ofp
 171
 172 class BinHex:
 173     def __init__(self, name_finfo_dlen_rlen, ofp):
 174         name, finfo, dlen, rlen = name_finfo_dlen_rlen
 175         if type(ofp) == type(''):
 176             ofname = ofp
 177             ofp = open(ofname, 'w')
 178             if os.name == 'mac':
 179                 fss = FSSpec(ofname)
 180                 fss.SetCreatorType('BnHq', 'TEXT')
 181         ofp.write('(This file must be converted with BinHex 4.0)\n\n:')
 182         hqxer = _Hqxcoderengine(ofp)
 183         self.ofp = _Rlecoderengine(hqxer)
 184         self.crc = 0
 185         if finfo is None:
 186             finfo = FInfo()
 187         self.dlen = dlen
 188         self.rlen = rlen
 189         self._writeinfo(name, finfo)
 190         self.state = _DID_HEADER
 191
 192     def _writeinfo(self, name, finfo):
 193         nl = len(name)
 194         if nl > 63:
 195             raise Error, 'Filename too long'
 196         d = chr(nl) + name + '\0'
 197         d2 = finfo.Type + finfo.Creator
 198
 199         # Force all structs to be packed with big-endian
 200         d3 = struct.pack('>h', finfo.Flags)
 201         d4 = struct.pack('>ii', self.dlen, self.rlen)
 202         info = d + d2 + d3 + d4
 203         self._write(info)
 204         self._writecrc()
 205
 206     def _write(self, data):
 207         self.crc = binascii.crc_hqx(data, self.crc)
 208         self.ofp.write(data)
 209
 210     def _writecrc(self):
 211         # XXXX Should this be here??
 212         # self.crc = binascii.crc_hqx('\0\0', self.crc)
 213         if self.crc < 0:
 214             fmt = '>h'
 215         else:
 216             fmt = '>H'
 217         self.ofp.write(struct.pack(fmt, self.crc))
 218         self.crc = 0
 219
 220     def write(self, data):
 221         if self.state != _DID_HEADER:
 222             raise Error, 'Writing data at the wrong time'
 223         self.dlen = self.dlen - len(data)
 224         self._write(data)
 225
 226     def close_data(self):
 227         if self.dlen != 0:
 228             raise Error, 'Incorrect data size, diff=%r' % (self.rlen,)
 229         self._writecrc()
 230         self.state = _DID_DATA
 231
 232     def write_rsrc(self, data):
 233         if self.state < _DID_DATA:
 234             self.close_data()
 235         if self.state != _DID_DATA:
 236             raise Error, 'Writing resource data at the wrong time'
 237         self.rlen = self.rlen - len(data)
 238         self._write(data)
 239
 240     def close(self):
 241         if self.state < _DID_DATA:
 242             self.close_data()
 243         if self.state != _DID_DATA:
 244             raise Error, 'Close at the wrong time'
 245         if self.rlen != 0:
 246             raise Error, \
 247                   "Incorrect resource-datasize, diff=%r" % (self.rlen,)
 248         self._writecrc()
 249         self.ofp.close()
 250         self.state = None
 251         del self.ofp
 252
 253 def binhex(inp, out):
 254     """(infilename, outfilename) - Create binhex-encoded copy of a file"""
 255     finfo = getfileinfo(inp)
 256     ofp = BinHex(finfo, out)
 257
 258     ifp = open(inp, 'rb')
 259     # XXXX Do textfile translation on non-mac systems
 260     while 1:
 261         d = ifp.read(128000)
 262         if not d: break
 263         ofp.write(d)
 264     ofp.close_data()
 265     ifp.close()
 266
 267     ifp = openrsrc(inp, 'rb')
 268     while 1:
 269         d = ifp.read(128000)
 270         if not d: break
 271         ofp.write_rsrc(d)
 272     ofp.close()
 273     ifp.close()
 274
 275 class _Hqxdecoderengine:
 276     """Read data via the decoder in 4-byte chunks"""
 277
 278     def __init__(self, ifp):
 279         self.ifp = ifp
 280         self.eof = 0
 281
 282     def read(self, totalwtd):
 283         """Read at least wtd bytes (or until EOF)"""
 284         decdata = ''
 285         wtd = totalwtd
 286         #
 287         # The loop here is convoluted, since we don't really now how
 288         # much to decode: there may be newlines in the incoming data.
 289         while wtd > 0:
 290             if self.eof: return decdata
 291             wtd = ((wtd+2)//3)*4
 292             data = self.ifp.read(wtd)
 293             #
 294             # Next problem: there may not be a complete number of
 295             # bytes in what we pass to a2b. Solve by yet another
 296             # loop.
 297             #
 298             while 1:
 299                 try:
 300                     decdatacur, self.eof = \
 301                             binascii.a2b_hqx(data)
 302                     break
 303                 except binascii.Incomplete:
 304                     pass
 305                 newdata = self.ifp.read(1)
 306                 if not newdata:
 307                     raise Error, \
 308                           'Premature EOF on binhex file'
 309                 data = data + newdata
 310             decdata = decdata + decdatacur
 311             wtd = totalwtd - len(decdata)
 312             if not decdata and not self.eof:
 313                 raise Error, 'Premature EOF on binhex file'
 314         return decdata
 315
 316     def close(self):
 317         self.ifp.close()
 318
 319 class _Rledecoderengine:
 320     """Read data via the RLE-coder"""
 321
 322     def __init__(self, ifp):
 323         self.ifp = ifp
 324         self.pre_buffer = ''
 325         self.post_buffer = ''
 326         self.eof = 0
 327
 328     def read(self, wtd):
 329         if wtd > len(self.post_buffer):
 330             self._fill(wtd-len(self.post_buffer))
 331         rv = self.post_buffer[:wtd]
 332         self.post_buffer = self.post_buffer[wtd:]
 333         return rv
 334
 335     def _fill(self, wtd):
 336         self.pre_buffer = self.pre_buffer + self.ifp.read(wtd+4)
 337         if self.ifp.eof:
 338             self.post_buffer = self.post_buffer + \
 339                 binascii.rledecode_hqx(self.pre_buffer)
 340             self.pre_buffer = ''
 341             return
 342
 343         #
 344         # Obfuscated code ahead. We have to take care that we don't
 345         # end up with an orphaned RUNCHAR later on. So, we keep a couple
 346         # of bytes in the buffer, depending on what the end of
 347         # the buffer looks like:
 348         # '\220\0\220' - Keep 3 bytes: repeated \220 (escaped as \220\0)
 349         # '?\220' - Keep 2 bytes: repeated something-else
 350         # '\220\0' - Escaped \220: Keep 2 bytes.
 351         # '?\220?' - Complete repeat sequence: decode all
 352         # otherwise: keep 1 byte.
 353         #
 354         mark = len(self.pre_buffer)
 355         if self.pre_buffer[-3:] == RUNCHAR + '\0' + RUNCHAR:
 356             mark = mark - 3
 357         elif self.pre_buffer[-1] == RUNCHAR:
 358             mark = mark - 2
 359         elif self.pre_buffer[-2:] == RUNCHAR + '\0':
 360             mark = mark - 2
 361         elif self.pre_buffer[-2] == RUNCHAR:
 362             pass # Decode all
 363         else:
 364             mark = mark - 1
 365
 366         self.post_buffer = self.post_buffer + \
 367             binascii.rledecode_hqx(self.pre_buffer[:mark])
 368         self.pre_buffer = self.pre_buffer[mark:]
 369
 370     def close(self):
 371         self.ifp.close()
 372
 373 class HexBin:
 374     def __init__(self, ifp):
 375         if type(ifp) == type(''):
 376             ifp = open(ifp)
 377         #
 378         # Find initial colon.
 379         #
 380         while 1:
 381             ch = ifp.read(1)
 382             if not ch:
 383                 raise Error, "No binhex data found"
 384             # Cater for \r\n terminated lines (which show up as \n\r, hence
 385             # all lines start with \r)
 386             if ch == '\r':
 387                 continue
 388             if ch == ':':
 389                 break
 390             if ch != '\n':
 391                 dummy = ifp.readline()
 392
 393         hqxifp = _Hqxdecoderengine(ifp)
 394         self.ifp = _Rledecoderengine(hqxifp)
 395         self.crc = 0
 396         self._readheader()
 397
 398     def _read(self, len):
 399         data = self.ifp.read(len)
 400         self.crc = binascii.crc_hqx(data, self.crc)
 401         return data
 402
 403     def _checkcrc(self):
 404         filecrc = struct.unpack('>h', self.ifp.read(2))[0] & 0xffff
 405         #self.crc = binascii.crc_hqx('\0\0', self.crc)
 406         # XXXX Is this needed??
 407         self.crc = self.crc & 0xffff
 408         if filecrc != self.crc:
 409             raise Error, 'CRC error, computed %x, read %x' \
 410                   %(self.crc, filecrc)
 411         self.crc = 0
 412
 413     def _readheader(self):
 414         len = self._read(1)
 415         fname = self._read(ord(len))
 416         rest = self._read(1+4+4+2+4+4)
 417         self._checkcrc()
 418
 419         type = rest[1:5]
 420         creator = rest[5:9]
 421         flags = struct.unpack('>h', rest[9:11])[0]
 422         self.dlen = struct.unpack('>l', rest[11:15])[0]
 423         self.rlen = struct.unpack('>l', rest[15:19])[0]
 424
 425         self.FName = fname
 426         self.FInfo = FInfo()
 427         self.FInfo.Creator = creator
 428         self.FInfo.Type = type
 429         self.FInfo.Flags = flags
 430
 431         self.state = _DID_HEADER
 432
 433     def read(self, *n):
 434         if self.state != _DID_HEADER:
 435             raise Error, 'Read data at wrong time'
 436         if n:
 437             n = n[0]
 438             n = min(n, self.dlen)
 439         else:
 440             n = self.dlen
 441         rv = ''
 442         while len(rv) < n:
 443             rv = rv + self._read(n-len(rv))
 444         self.dlen = self.dlen - n
 445         return rv
 446
 447     def close_data(self):
 448         if self.state != _DID_HEADER:
 449             raise Error, 'close_data at wrong time'
 450         if self.dlen:
 451             dummy = self._read(self.dlen)
 452         self._checkcrc()
 453         self.state = _DID_DATA
 454
 455     def read_rsrc(self, *n):
 456         if self.state == _DID_HEADER:
 457             self.close_data()
 458         if self.state != _DID_DATA:
 459             raise Error, 'Read resource data at wrong time'
 460         if n:
 461             n = n[0]
 462             n = min(n, self.rlen)
 463         else:
 464             n = self.rlen
 465         self.rlen = self.rlen - n
 466         return self._read(n)
 467
 468     def close(self):
 469         if self.rlen:
 470             dummy = self.read_rsrc(self.rlen)
 471         self._checkcrc()
 472         self.state = _DID_RSRC
 473         self.ifp.close()
 474
 475 def hexbin(inp, out):
 476     """(infilename, outfilename) - Decode binhexed file"""
 477     ifp = HexBin(inp)
 478     finfo = ifp.FInfo
 479     if not out:
 480         out = ifp.FName
 481     if os.name == 'mac':
 482         ofss = FSSpec(out)
 483         out = ofss.as_pathname()
 484
 485     ofp = open(out, 'wb')
 486     # XXXX Do translation on non-mac systems
 487     while 1:
 488         d = ifp.read(128000)
 489         if not d: break
 490         ofp.write(d)
 491     ofp.close()
 492     ifp.close_data()
 493
 494     d = ifp.read_rsrc(128000)
 495     if d:
 496         ofp = openrsrc(out, 'wb')
 497         ofp.write(d)
 498         while 1:
 499             d = ifp.read_rsrc(128000)
 500             if not d: break
 501             ofp.write(d)
 502         ofp.close()
 503
 504     if os.name == 'mac':
 505         nfinfo = ofss.GetFInfo()
 506         nfinfo.Creator = finfo.Creator
 507         nfinfo.Type = finfo.Type
 508         nfinfo.Flags = finfo.Flags
 509         ofss.SetFInfo(nfinfo)
 510
 511     ifp.close()
 512
 513 def _test():
 514     fname = sys.argv[1]
 515     binhex(fname, fname+'.hqx')
 516     hexbin(fname+'.hqx', fname+'.viahqx')
 517     #hexbin(fname, fname+'.unpacked')
 518     sys.exit(1)
 519
 520 if __name__ == '__main__':
 521     _test()