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 if os.name == 'mac':
  48     import macfs
  49     import MacOS
  50     try:
  51         openrf = MacOS.openrf
  52     except AttributeError:
  53         # Backward compatibility
  54         openrf = open
  55
  56     def FInfo():
  57         return macfs.FInfo()
  58
  59     def getfileinfo(name):
  60         finfo = macfs.FSSpec(name).GetFInfo()
  61         dir, file = os.path.split(name)
  62         # XXXX Get resource/data sizes
  63         fp = open(name, 'rb')
  64         fp.seek(0, 2)
  65         dlen = fp.tell()
  66         fp = openrf(name, '*rb')
  67         fp.seek(0, 2)
  68         rlen = fp.tell()
  69         return file, finfo, dlen, rlen
  70
  71     def openrsrc(name, *mode):
  72         if not mode:
  73             mode = '*rb'
  74         else:
  75             mode = '*' + mode[0]
  76         return openrf(name, mode)
  77
  78 else:
  79     #
  80     # Glue code for non-macintosh usage
  81     #
  82
  83     class FInfo:
  84         def __init__(self):
  85             self.Type = '????'
  86             self.Creator = '????'
  87             self.Flags = 0
  88
  89     def getfileinfo(name):
  90         finfo = FInfo()
  91         # Quick check for textfile
  92         fp = open(name)
  93         data = open(name).read(256)
  94         for c in data:
  95             if not c.isspace() and (c<' ' or ord(c) > 0x7f):
  96                 break
  97         else:
  98             finfo.Type = 'TEXT'
  99         fp.seek(0, 2)
 100         dsize = fp.tell()
 101         fp.close()
 102         dir, file = os.path.split(name)
 103         file = file.replace(':', '-', 1)
 104         return file, finfo, dsize, 0
 105
 106     class openrsrc:
 107         def __init__(self, *args):
 108             pass
 109
 110         def read(self, *args):
 111             return ''
 112
 113         def write(self, *args):
 114             pass
 115
 116         def close(self):
 117             pass
 118
 119 class _Hqxcoderengine:
 120     """Write data to the coder in 3-byte chunks"""
 121
 122     def __init__(self, ofp):
 123         self.ofp = ofp
 124         self.data = ''
 125         self.hqxdata = ''
 126         self.linelen = LINELEN-1
 127
 128     def write(self, data):
 129         self.data = self.data + data
 130         datalen = len(self.data)
 131         todo = (datalen//3)*3
 132         data = self.data[:todo]
 133         self.data = self.data[todo:]
 134         if not data:
 135             return
 136         self.hqxdata = self.hqxdata + binascii.b2a_hqx(data)
 137         self._flush(0)
 138
 139     def _flush(self, force):
 140         first = 0
 141         while first <= len(self.hqxdata)-self.linelen:
 142             last = first + self.linelen
 143             self.ofp.write(self.hqxdata[first:last]+'\n')
 144             self.linelen = LINELEN
 145             first = last
 146         self.hqxdata = self.hqxdata[first:]
 147         if force:
 148             self.ofp.write(self.hqxdata + ':\n')
 149
 150     def close(self):
 151         if self.data:
 152             self.hqxdata = \
 153                  self.hqxdata + binascii.b2a_hqx(self.data)
 154         self._flush(1)
 155         self.ofp.close()
 156         del self.ofp
 157
 158 class _Rlecoderengine:
 159     """Write data to the RLE-coder in suitably large chunks"""
 160
 161     def __init__(self, ofp):
 162         self.ofp = ofp
 163         self.data = ''
 164
 165     def write(self, data):
 166         self.data = self.data + data
 167         if len(self.data) < REASONABLY_LARGE:
 168             return
 169         rledata = binascii.rlecode_hqx(self.data)
 170         self.ofp.write(rledata)
 171         self.data = ''
 172
 173     def close(self):
 174         if self.data:
 175             rledata = binascii.rlecode_hqx(self.data)
 176             self.ofp.write(rledata)
 177         self.ofp.close()
 178         del self.ofp
 179
 180 class BinHex:
 181     def __init__(self, (name, finfo, dlen, rlen), ofp):
 182         if type(ofp) == type(''):
 183             ofname = ofp
 184             ofp = open(ofname, 'w')
 185             if os.name == 'mac':
 186                 fss = macfs.FSSpec(ofname)
 187                 fss.SetCreatorType('BnHq', 'TEXT')
 188         ofp.write('(This file must be converted with BinHex 4.0)\n\n:')
 189         hqxer = _Hqxcoderengine(ofp)
 190         self.ofp = _Rlecoderengine(hqxer)
 191         self.crc = 0
 192         if finfo is None:
 193             finfo = FInfo()
 194         self.dlen = dlen
 195         self.rlen = rlen
 196         self._writeinfo(name, finfo)
 197         self.state = _DID_HEADER
 198
 199     def _writeinfo(self, name, finfo):
 200         nl = len(name)
 201         if nl > 63:
 202             raise Error, 'Filename too long'
 203         d = chr(nl) + name + '\0'
 204         d2 = finfo.Type + finfo.Creator
 205
 206         # Force all structs to be packed with big-endian
 207         d3 = struct.pack('>h', finfo.Flags)
 208         d4 = struct.pack('>ii', self.dlen, self.rlen)
 209         info = d + d2 + d3 + d4
 210         self._write(info)
 211         self._writecrc()
 212
 213     def _write(self, data):
 214         self.crc = binascii.crc_hqx(data, self.crc)
 215         self.ofp.write(data)
 216
 217     def _writecrc(self):
 218         # XXXX Should this be here??
 219         # self.crc = binascii.crc_hqx('\0\0', self.crc)
 220         if self.crc < 0:
 221             fmt = '>h'
 222         else:
 223             fmt = '>H'
 224         self.ofp.write(struct.pack(fmt, self.crc))
 225         self.crc = 0
 226
 227     def write(self, data):
 228         if self.state != _DID_HEADER:
 229             raise Error, 'Writing data at the wrong time'
 230         self.dlen = self.dlen - len(data)
 231         self._write(data)
 232
 233     def close_data(self):
 234         if self.dlen != 0:
 235             raise Error, 'Incorrect data size, diff=%r' % (self.rlen,)
 236         self._writecrc()
 237         self.state = _DID_DATA
 238
 239     def write_rsrc(self, data):
 240         if self.state < _DID_DATA:
 241             self.close_data()
 242         if self.state != _DID_DATA:
 243             raise Error, 'Writing resource data at the wrong time'
 244         self.rlen = self.rlen - len(data)
 245         self._write(data)
 246
 247     def close(self):
 248         if self.state < _DID_DATA:
 249             self.close_data()
 250         if self.state != _DID_DATA:
 251             raise Error, 'Close at the wrong time'
 252         if self.rlen != 0:
 253             raise Error, \
 254                   "Incorrect resource-datasize, diff=%r" % (self.rlen,)
 255         self._writecrc()
 256         self.ofp.close()
 257         self.state = None
 258         del self.ofp
 259
 260 def binhex(inp, out):
 261     """(infilename, outfilename) - Create binhex-encoded copy of a file"""
 262     finfo = getfileinfo(inp)
 263     ofp = BinHex(finfo, out)
 264
 265     ifp = open(inp, 'rb')
 266     # XXXX Do textfile translation on non-mac systems
 267     while 1:
 268         d = ifp.read(128000)
 269         if not d: break
 270         ofp.write(d)
 271     ofp.close_data()
 272     ifp.close()
 273
 274     ifp = openrsrc(inp, 'rb')
 275     while 1:
 276         d = ifp.read(128000)
 277         if not d: break
 278         ofp.write_rsrc(d)
 279     ofp.close()
 280     ifp.close()
 281
 282 class _Hqxdecoderengine:
 283     """Read data via the decoder in 4-byte chunks"""
 284
 285     def __init__(self, ifp):
 286         self.ifp = ifp
 287         self.eof = 0
 288
 289     def read(self, totalwtd):
 290         """Read at least wtd bytes (or until EOF)"""
 291         decdata = ''
 292         wtd = totalwtd
 293         #
 294         # The loop here is convoluted, since we don't really now how
 295         # much to decode: there may be newlines in the incoming data.
 296         while wtd > 0:
 297             if self.eof: return decdata
 298             wtd = ((wtd+2)//3)*4
 299             data = self.ifp.read(wtd)
 300             #
 301             # Next problem: there may not be a complete number of
 302             # bytes in what we pass to a2b. Solve by yet another
 303             # loop.
 304             #
 305             while 1:
 306                 try:
 307                     decdatacur, self.eof = \
 308                             binascii.a2b_hqx(data)
 309                     break
 310                 except binascii.Incomplete:
 311                     pass
 312                 newdata = self.ifp.read(1)
 313                 if not newdata:
 314                     raise Error, \
 315                           'Premature EOF on binhex file'
 316                 data = data + newdata
 317             decdata = decdata + decdatacur
 318             wtd = totalwtd - len(decdata)
 319             if not decdata and not self.eof:
 320                 raise Error, 'Premature EOF on binhex file'
 321         return decdata
 322
 323     def close(self):
 324         self.ifp.close()
 325
 326 class _Rledecoderengine:
 327     """Read data via the RLE-coder"""
 328
 329     def __init__(self, ifp):
 330         self.ifp = ifp
 331         self.pre_buffer = ''
 332         self.post_buffer = ''
 333         self.eof = 0
 334
 335     def read(self, wtd):
 336         if wtd > len(self.post_buffer):
 337             self._fill(wtd-len(self.post_buffer))
 338         rv = self.post_buffer[:wtd]
 339         self.post_buffer = self.post_buffer[wtd:]
 340         return rv
 341
 342     def _fill(self, wtd):
 343         self.pre_buffer = self.pre_buffer + self.ifp.read(wtd+4)
 344         if self.ifp.eof:
 345             self.post_buffer = self.post_buffer + \
 346                 binascii.rledecode_hqx(self.pre_buffer)
 347             self.pre_buffer = ''
 348             return
 349
 350         #
 351         # Obfuscated code ahead. We have to take care that we don't
 352         # end up with an orphaned RUNCHAR later on. So, we keep a couple
 353         # of bytes in the buffer, depending on what the end of
 354         # the buffer looks like:
 355         # '\220\0\220' - Keep 3 bytes: repeated \220 (escaped as \220\0)
 356         # '?\220' - Keep 2 bytes: repeated something-else
 357         # '\220\0' - Escaped \220: Keep 2 bytes.
 358         # '?\220?' - Complete repeat sequence: decode all
 359         # otherwise: keep 1 byte.
 360         #
 361         mark = len(self.pre_buffer)
 362         if self.pre_buffer[-3:] == RUNCHAR + '\0' + RUNCHAR:
 363             mark = mark - 3
 364         elif self.pre_buffer[-1] == RUNCHAR:
 365             mark = mark - 2
 366         elif self.pre_buffer[-2:] == RUNCHAR + '\0':
 367             mark = mark - 2
 368         elif self.pre_buffer[-2] == RUNCHAR:
 369             pass # Decode all
 370         else:
 371             mark = mark - 1
 372
 373         self.post_buffer = self.post_buffer + \
 374             binascii.rledecode_hqx(self.pre_buffer[:mark])
 375         self.pre_buffer = self.pre_buffer[mark:]
 376
 377     def close(self):
 378         self.ifp.close()
 379
 380 class HexBin:
 381     def __init__(self, ifp):
 382         if type(ifp) == type(''):
 383             ifp = open(ifp)
 384         #
 385         # Find initial colon.
 386         #
 387         while 1:
 388             ch = ifp.read(1)
 389             if not ch:
 390                 raise Error, "No binhex data found"
 391             # Cater for \r\n terminated lines (which show up as \n\r, hence
 392             # all lines start with \r)
 393             if ch == '\r':
 394                 continue
 395             if ch == ':':
 396                 break
 397             if ch != '\n':
 398                 dummy = ifp.readline()
 399
 400         hqxifp = _Hqxdecoderengine(ifp)
 401         self.ifp = _Rledecoderengine(hqxifp)
 402         self.crc = 0
 403         self._readheader()
 404
 405     def _read(self, len):
 406         data = self.ifp.read(len)
 407         self.crc = binascii.crc_hqx(data, self.crc)
 408         return data
 409
 410     def _checkcrc(self):
 411         filecrc = struct.unpack('>h', self.ifp.read(2))[0] & 0xffff
 412         #self.crc = binascii.crc_hqx('\0\0', self.crc)
 413         # XXXX Is this needed??
 414         self.crc = self.crc & 0xffff
 415         if filecrc != self.crc:
 416             raise Error, 'CRC error, computed %x, read %x' \
 417                   %(self.crc, filecrc)
 418         self.crc = 0
 419
 420     def _readheader(self):
 421         len = self._read(1)
 422         fname = self._read(ord(len))
 423         rest = self._read(1+4+4+2+4+4)
 424         self._checkcrc()
 425
 426         type = rest[1:5]
 427         creator = rest[5:9]
 428         flags = struct.unpack('>h', rest[9:11])[0]
 429         self.dlen = struct.unpack('>l', rest[11:15])[0]
 430         self.rlen = struct.unpack('>l', rest[15:19])[0]
 431
 432         self.FName = fname
 433         self.FInfo = FInfo()
 434         self.FInfo.Creator = creator
 435         self.FInfo.Type = type
 436         self.FInfo.Flags = flags
 437
 438         self.state = _DID_HEADER
 439
 440     def read(self, *n):
 441         if self.state != _DID_HEADER:
 442             raise Error, 'Read data at wrong time'
 443         if n:
 444             n = n[0]
 445             n = min(n, self.dlen)
 446         else:
 447             n = self.dlen
 448         rv = ''
 449         while len(rv) < n:
 450             rv = rv + self._read(n-len(rv))
 451         self.dlen = self.dlen - n
 452         return rv
 453
 454     def close_data(self):
 455         if self.state != _DID_HEADER:
 456             raise Error, 'close_data at wrong time'
 457         if self.dlen:
 458             dummy = self._read(self.dlen)
 459         self._checkcrc()
 460         self.state = _DID_DATA
 461
 462     def read_rsrc(self, *n):
 463         if self.state == _DID_HEADER:
 464             self.close_data()
 465         if self.state != _DID_DATA:
 466             raise Error, 'Read resource data at wrong time'
 467         if n:
 468             n = n[0]
 469             n = min(n, self.rlen)
 470         else:
 471             n = self.rlen
 472         self.rlen = self.rlen - n
 473         return self._read(n)
 474
 475     def close(self):
 476         if self.rlen:
 477             dummy = self.read_rsrc(self.rlen)
 478         self._checkcrc()
 479         self.state = _DID_RSRC
 480         self.ifp.close()
 481
 482 def hexbin(inp, out):
 483     """(infilename, outfilename) - Decode binhexed file"""
 484     ifp = HexBin(inp)
 485     finfo = ifp.FInfo
 486     if not out:
 487         out = ifp.FName
 488     if os.name == 'mac':
 489         ofss = macfs.FSSpec(out)
 490         out = ofss.as_pathname()
 491
 492     ofp = open(out, 'wb')
 493     # XXXX Do translation on non-mac systems
 494     while 1:
 495         d = ifp.read(128000)
 496         if not d: break
 497         ofp.write(d)
 498     ofp.close()
 499     ifp.close_data()
 500
 501     d = ifp.read_rsrc(128000)
 502     if d:
 503         ofp = openrsrc(out, 'wb')
 504         ofp.write(d)
 505         while 1:
 506             d = ifp.read_rsrc(128000)
 507             if not d: break
 508             ofp.write(d)
 509         ofp.close()
 510
 511     if os.name == 'mac':
 512         nfinfo = ofss.GetFInfo()
 513         nfinfo.Creator = finfo.Creator
 514         nfinfo.Type = finfo.Type
 515         nfinfo.Flags = finfo.Flags
 516         ofss.SetFInfo(nfinfo)
 517
 518     ifp.close()
 519
 520 def _test():
 521     if os.name == 'mac':
 522         fss, ok = macfs.PromptGetFile('File to convert:')
 523         if not ok:
 524             sys.exit(0)
 525         fname = fss.as_pathname()
 526     else:
 527         fname = sys.argv[1]
 528     binhex(fname, fname+'.hqx')
 529     hexbin(fname+'.hqx', fname+'.viahqx')
 530     #hexbin(fname, fname+'.unpacked')
 531     sys.exit(1)
 532
 533 if __name__ == '__main__':
 534     _test()