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         name = name
 201         nl = len(name)
 202         if nl > 63:
 203             raise Error, 'Filename too long'
 204         d = chr(nl) + name + '\0'
 205         d2 = finfo.Type + finfo.Creator
 206
 207         # Force all structs to be packed with big-endian
 208         d3 = struct.pack('>h', finfo.Flags)
 209         d4 = struct.pack('>ii', self.dlen, self.rlen)
 210         info = d + d2 + d3 + d4
 211         self._write(info)
 212         self._writecrc()
 213
 214     def _write(self, data):
 215         self.crc = binascii.crc_hqx(data, self.crc)
 216         self.ofp.write(data)
 217
 218     def _writecrc(self):
 219         # XXXX Should this be here??
 220         # self.crc = binascii.crc_hqx('\0\0', self.crc)
 221         self.ofp.write(struct.pack('>h', self.crc))
 222         self.crc = 0
 223
 224     def write(self, data):
 225         if self.state != _DID_HEADER:
 226             raise Error, 'Writing data at the wrong time'
 227         self.dlen = self.dlen - len(data)
 228         self._write(data)
 229
 230     def close_data(self):
 231         if self.dlen != 0:
 232             raise Error, 'Incorrect data size, diff='+`self.rlen`
 233         self._writecrc()
 234         self.state = _DID_DATA
 235
 236     def write_rsrc(self, data):
 237         if self.state < _DID_DATA:
 238             self.close_data()
 239         if self.state != _DID_DATA:
 240             raise Error, 'Writing resource data at the wrong time'
 241         self.rlen = self.rlen - len(data)
 242         self._write(data)
 243
 244     def close(self):
 245         if self.state < _DID_DATA:
 246             self.close_data()
 247         if self.state != _DID_DATA:
 248             raise Error, 'Close at the wrong time'
 249         if self.rlen != 0:
 250             raise Error, \
 251                   "Incorrect resource-datasize, diff="+`self.rlen`
 252         self._writecrc()
 253         self.ofp.close()
 254         self.state = None
 255         del self.ofp
 256
 257 def binhex(inp, out):
 258     """(infilename, outfilename) - Create binhex-encoded copy of a file"""
 259     finfo = getfileinfo(inp)
 260     ofp = BinHex(finfo, out)
 261
 262     ifp = open(inp, 'rb')
 263     # XXXX Do textfile translation on non-mac systems
 264     while 1:
 265         d = ifp.read(128000)
 266         if not d: break
 267         ofp.write(d)
 268     ofp.close_data()
 269     ifp.close()
 270
 271     ifp = openrsrc(inp, 'rb')
 272     while 1:
 273         d = ifp.read(128000)
 274         if not d: break
 275         ofp.write_rsrc(d)
 276     ofp.close()
 277     ifp.close()
 278
 279 class _Hqxdecoderengine:
 280     """Read data via the decoder in 4-byte chunks"""
 281
 282     def __init__(self, ifp):
 283         self.ifp = ifp
 284         self.eof = 0
 285
 286     def read(self, totalwtd):
 287         """Read at least wtd bytes (or until EOF)"""
 288         decdata = ''
 289         wtd = totalwtd
 290         #
 291         # The loop here is convoluted, since we don't really now how
 292         # much to decode: there may be newlines in the incoming data.
 293         while wtd > 0:
 294             if self.eof: return decdata
 295             wtd = ((wtd+2)//3)*4
 296             data = self.ifp.read(wtd)
 297             #
 298             # Next problem: there may not be a complete number of
 299             # bytes in what we pass to a2b. Solve by yet another
 300             # loop.
 301             #
 302             while 1:
 303                 try:
 304                     decdatacur, self.eof = \
 305                             binascii.a2b_hqx(data)
 306                     break
 307                 except binascii.Incomplete:
 308                     pass
 309                 newdata = self.ifp.read(1)
 310                 if not newdata:
 311                     raise Error, \
 312                           'Premature EOF on binhex file'
 313                 data = data + newdata
 314             decdata = decdata + decdatacur
 315             wtd = totalwtd - len(decdata)
 316             if not decdata and not self.eof:
 317                 raise Error, 'Premature EOF on binhex file'
 318         return decdata
 319
 320     def close(self):
 321         self.ifp.close()
 322
 323 class _Rledecoderengine:
 324     """Read data via the RLE-coder"""
 325
 326     def __init__(self, ifp):
 327         self.ifp = ifp
 328         self.pre_buffer = ''
 329         self.post_buffer = ''
 330         self.eof = 0
 331
 332     def read(self, wtd):
 333         if wtd > len(self.post_buffer):
 334             self._fill(wtd-len(self.post_buffer))
 335         rv = self.post_buffer[:wtd]
 336         self.post_buffer = self.post_buffer[wtd:]
 337         return rv
 338
 339     def _fill(self, wtd):
 340         self.pre_buffer = self.pre_buffer + self.ifp.read(wtd+4)
 341         if self.ifp.eof:
 342             self.post_buffer = self.post_buffer + \
 343                 binascii.rledecode_hqx(self.pre_buffer)
 344             self.pre_buffer = ''
 345             return
 346
 347         #
 348         # Obfuscated code ahead. We have to take care that we don't
 349         # end up with an orphaned RUNCHAR later on. So, we keep a couple
 350         # of bytes in the buffer, depending on what the end of
 351         # the buffer looks like:
 352         # '\220\0\220' - Keep 3 bytes: repeated \220 (escaped as \220\0)
 353         # '?\220' - Keep 2 bytes: repeated something-else
 354         # '\220\0' - Escaped \220: Keep 2 bytes.
 355         # '?\220?' - Complete repeat sequence: decode all
 356         # otherwise: keep 1 byte.
 357         #
 358         mark = len(self.pre_buffer)
 359         if self.pre_buffer[-3:] == RUNCHAR + '\0' + RUNCHAR:
 360             mark = mark - 3
 361         elif self.pre_buffer[-1] == RUNCHAR:
 362             mark = mark - 2
 363         elif self.pre_buffer[-2:] == RUNCHAR + '\0':
 364             mark = mark - 2
 365         elif self.pre_buffer[-2] == RUNCHAR:
 366             pass # Decode all
 367         else:
 368             mark = mark - 1
 369
 370         self.post_buffer = self.post_buffer + \
 371             binascii.rledecode_hqx(self.pre_buffer[:mark])
 372         self.pre_buffer = self.pre_buffer[mark:]
 373
 374     def close(self):
 375         self.ifp.close()
 376
 377 class HexBin:
 378     def __init__(self, ifp):
 379         if type(ifp) == type(''):
 380             ifp = open(ifp)
 381         #
 382         # Find initial colon.
 383         #
 384         while 1:
 385             ch = ifp.read(1)
 386             if not ch:
 387                 raise Error, "No binhex data found"
 388             # Cater for \r\n terminated lines (which show up as \n\r, hence
 389             # all lines start with \r)
 390             if ch == '\r':
 391                 continue
 392             if ch == ':':
 393                 break
 394             if ch != '\n':
 395                 dummy = ifp.readline()
 396
 397         hqxifp = _Hqxdecoderengine(ifp)
 398         self.ifp = _Rledecoderengine(hqxifp)
 399         self.crc = 0
 400         self._readheader()
 401
 402     def _read(self, len):
 403         data = self.ifp.read(len)
 404         self.crc = binascii.crc_hqx(data, self.crc)
 405         return data
 406
 407     def _checkcrc(self):
 408         filecrc = struct.unpack('>h', self.ifp.read(2))[0] & 0xffff
 409         #self.crc = binascii.crc_hqx('\0\0', self.crc)
 410         # XXXX Is this needed??
 411         self.crc = self.crc & 0xffff
 412         if filecrc != self.crc:
 413             raise Error, 'CRC error, computed %x, read %x' \
 414                   %(self.crc, filecrc)
 415         self.crc = 0
 416
 417     def _readheader(self):
 418         len = self._read(1)
 419         fname = self._read(ord(len))
 420         rest = self._read(1+4+4+2+4+4)
 421         self._checkcrc()
 422
 423         type = rest[1:5]
 424         creator = rest[5:9]
 425         flags = struct.unpack('>h', rest[9:11])[0]
 426         self.dlen = struct.unpack('>l', rest[11:15])[0]
 427         self.rlen = struct.unpack('>l', rest[15:19])[0]
 428
 429         self.FName = fname
 430         self.FInfo = FInfo()
 431         self.FInfo.Creator = creator
 432         self.FInfo.Type = type
 433         self.FInfo.Flags = flags
 434
 435         self.state = _DID_HEADER
 436
 437     def read(self, *n):
 438         if self.state != _DID_HEADER:
 439             raise Error, 'Read data at wrong time'
 440         if n:
 441             n = n[0]
 442             n = min(n, self.dlen)
 443         else:
 444             n = self.dlen
 445         rv = ''
 446         while len(rv) < n:
 447             rv = rv + self._read(n-len(rv))
 448         self.dlen = self.dlen - n
 449         return rv
 450
 451     def close_data(self):
 452         if self.state != _DID_HEADER:
 453             raise Error, 'close_data at wrong time'
 454         if self.dlen:
 455             dummy = self._read(self.dlen)
 456         self._checkcrc()
 457         self.state = _DID_DATA
 458
 459     def read_rsrc(self, *n):
 460         if self.state == _DID_HEADER:
 461             self.close_data()
 462         if self.state != _DID_DATA:
 463             raise Error, 'Read resource data at wrong time'
 464         if n:
 465             n = n[0]
 466             n = min(n, self.rlen)
 467         else:
 468             n = self.rlen
 469         self.rlen = self.rlen - n
 470         return self._read(n)
 471
 472     def close(self):
 473         if self.rlen:
 474             dummy = self.read_rsrc(self.rlen)
 475         self._checkcrc()
 476         self.state = _DID_RSRC
 477         self.ifp.close()
 478
 479 def hexbin(inp, out):
 480     """(infilename, outfilename) - Decode binhexed file"""
 481     ifp = HexBin(inp)
 482     finfo = ifp.FInfo
 483     if not out:
 484         out = ifp.FName
 485     if os.name == 'mac':
 486         ofss = macfs.FSSpec(out)
 487         out = ofss.as_pathname()
 488
 489     ofp = open(out, 'wb')
 490     # XXXX Do translation on non-mac systems
 491     while 1:
 492         d = ifp.read(128000)
 493         if not d: break
 494         ofp.write(d)
 495     ofp.close()
 496     ifp.close_data()
 497
 498     d = ifp.read_rsrc(128000)
 499     if d:
 500         ofp = openrsrc(out, 'wb')
 501         ofp.write(d)
 502         while 1:
 503             d = ifp.read_rsrc(128000)
 504             if not d: break
 505             ofp.write(d)
 506         ofp.close()
 507
 508     if os.name == 'mac':
 509         nfinfo = ofss.GetFInfo()
 510         nfinfo.Creator = finfo.Creator
 511         nfinfo.Type = finfo.Type
 512         nfinfo.Flags = finfo.Flags
 513         ofss.SetFInfo(nfinfo)
 514
 515     ifp.close()
 516
 517 def _test():
 518     if os.name == 'mac':
 519         fss, ok = macfs.PromptGetFile('File to convert:')
 520         if not ok:
 521             sys.exit(0)
 522         fname = fss.as_pathname()
 523     else:
 524         fname = sys.argv[1]
 525     binhex(fname, fname+'.hqx')
 526     hexbin(fname+'.hqx', fname+'.viahqx')
 527     #hexbin(fname, fname+'.unpacked')
 528     sys.exit(1)
 529
 530 if __name__ == '__main__':
 531     _test()