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 io
  25 import os
  26 import sys
  27 import struct
  28 import binascii
  29
  30 __all__ = ["binhex","hexbin","Error"]
  31
  32 class Error(Exception):
  33     pass
  34
  35 # States (what have we written)
  36 [_DID_HEADER, _DID_DATA, _DID_RSRC] = range(3)
  37
  38 # Various constants
  39 REASONABLY_LARGE = 32768  # Minimal amount we pass the rle-coder
  40 LINELEN = 64
  41 RUNCHAR = b"\x90"
  42
  43 #
  44 # This code is no longer byte-order dependent
  45
  46
  47 class FInfo:
  48     def __init__(self):
  49         self.Type = '????'
  50         self.Creator = '????'
  51         self.Flags = 0
  52
  53 def getfileinfo(name):
  54     finfo = FInfo()
  55     fp = io.open(name, 'rb')
  56     # Quick check for textfile
  57     data = fp.read(512)
  58     if 0 not in data:
  59         finfo.Type = 'TEXT'
  60     fp.seek(0, 2)
  61     dsize = fp.tell()
  62     fp.close()
  63     dir, file = os.path.split(name)
  64     file = file.replace(':', '-', 1)
  65     return file, finfo, dsize, 0
  66
  67 class openrsrc:
  68     def __init__(self, *args):
  69         pass
  70
  71     def read(self, *args):
  72         return b''
  73
  74     def write(self, *args):
  75         pass
  76
  77     def close(self):
  78         pass
  79
  80 class _Hqxcoderengine:
  81     """Write data to the coder in 3-byte chunks"""
  82
  83     def __init__(self, ofp):
  84         self.ofp = ofp
  85         self.data = b''
  86         self.hqxdata = b''
  87         self.linelen = LINELEN - 1
  88
  89     def write(self, data):
  90         self.data = self.data + data
  91         datalen = len(self.data)
  92         todo = (datalen // 3) * 3
  93         data = self.data[:todo]
  94         self.data = self.data[todo:]
  95         if not data:
  96             return
  97         self.hqxdata = self.hqxdata + binascii.b2a_hqx(data)
  98         self._flush(0)
  99
 100     def _flush(self, force):
 101         first = 0
 102         while first <= len(self.hqxdata) - self.linelen:
 103             last = first + self.linelen
 104             self.ofp.write(self.hqxdata[first:last] + b'\n')
 105             self.linelen = LINELEN
 106             first = last
 107         self.hqxdata = self.hqxdata[first:]
 108         if force:
 109             self.ofp.write(self.hqxdata + b':\n')
 110
 111     def close(self):
 112         if self.data:
 113             self.hqxdata = self.hqxdata + binascii.b2a_hqx(self.data)
 114         self._flush(1)
 115         self.ofp.close()
 116         del self.ofp
 117
 118 class _Rlecoderengine:
 119     """Write data to the RLE-coder in suitably large chunks"""
 120
 121     def __init__(self, ofp):
 122         self.ofp = ofp
 123         self.data = b''
 124
 125     def write(self, data):
 126         self.data = self.data + data
 127         if len(self.data) < REASONABLY_LARGE:
 128             return
 129         rledata = binascii.rlecode_hqx(self.data)
 130         self.ofp.write(rledata)
 131         self.data = b''
 132
 133     def close(self):
 134         if self.data:
 135             rledata = binascii.rlecode_hqx(self.data)
 136             self.ofp.write(rledata)
 137         self.ofp.close()
 138         del self.ofp
 139
 140 class BinHex:
 141     def __init__(self, name_finfo_dlen_rlen, ofp):
 142         name, finfo, dlen, rlen = name_finfo_dlen_rlen
 143         if isinstance(ofp, str):
 144             ofname = ofp
 145             ofp = io.open(ofname, 'wb')
 146         ofp.write(b'(This file must be converted with BinHex 4.0)\r\r:')
 147         hqxer = _Hqxcoderengine(ofp)
 148         self.ofp = _Rlecoderengine(hqxer)
 149         self.crc = 0
 150         if finfo is None:
 151             finfo = FInfo()
 152         self.dlen = dlen
 153         self.rlen = rlen
 154         self._writeinfo(name, finfo)
 155         self.state = _DID_HEADER
 156
 157     def _writeinfo(self, name, finfo):
 158         nl = len(name)
 159         if nl > 63:
 160             raise Error('Filename too long')
 161         d = bytes([nl]) + name.encode("latin-1") + b'\0'
 162         tp, cr = finfo.Type, finfo.Creator
 163         if isinstance(tp, str):
 164             tp = tp.encode("latin-1")
 165         if isinstance(cr, str):
 166             cr = cr.encode("latin-1")
 167         d2 = tp + cr
 168
 169         # Force all structs to be packed with big-endian
 170         d3 = struct.pack('>h', finfo.Flags)
 171         d4 = struct.pack('>ii', self.dlen, self.rlen)
 172         info = d + d2 + d3 + d4
 173         self._write(info)
 174         self._writecrc()
 175
 176     def _write(self, data):
 177         self.crc = binascii.crc_hqx(data, self.crc)
 178         self.ofp.write(data)
 179
 180     def _writecrc(self):
 181         # XXXX Should this be here??
 182         # self.crc = binascii.crc_hqx('\0\0', self.crc)
 183         if self.crc < 0:
 184             fmt = '>h'
 185         else:
 186             fmt = '>H'
 187         self.ofp.write(struct.pack(fmt, self.crc))
 188         self.crc = 0
 189
 190     def write(self, data):
 191         if self.state != _DID_HEADER:
 192             raise Error('Writing data at the wrong time')
 193         self.dlen = self.dlen - len(data)
 194         self._write(data)
 195
 196     def close_data(self):
 197         if self.dlen != 0:
 198             raise Error('Incorrect data size, diff=%r' % (self.rlen,))
 199         self._writecrc()
 200         self.state = _DID_DATA
 201
 202     def write_rsrc(self, data):
 203         if self.state < _DID_DATA:
 204             self.close_data()
 205         if self.state != _DID_DATA:
 206             raise Error('Writing resource data at the wrong time')
 207         self.rlen = self.rlen - len(data)
 208         self._write(data)
 209
 210     def close(self):
 211         if self.state < _DID_DATA:
 212             self.close_data()
 213         if self.state != _DID_DATA:
 214             raise Error('Close at the wrong time')
 215         if self.rlen != 0:
 216             raise Error("Incorrect resource-datasize, diff=%r" % (self.rlen,))
 217         self._writecrc()
 218         self.ofp.close()
 219         self.state = None
 220         del self.ofp
 221
 222 def binhex(inp, out):
 223     """binhex(infilename, outfilename): create binhex-encoded copy of a file"""
 224     finfo = getfileinfo(inp)
 225     ofp = BinHex(finfo, out)
 226
 227     ifp = io.open(inp, 'rb')
 228     # XXXX Do textfile translation on non-mac systems
 229     while True:
 230         d = ifp.read(128000)
 231         if not d: break
 232         ofp.write(d)
 233     ofp.close_data()
 234     ifp.close()
 235
 236     ifp = openrsrc(inp, 'rb')
 237     while True:
 238         d = ifp.read(128000)
 239         if not d: break
 240         ofp.write_rsrc(d)
 241     ofp.close()
 242     ifp.close()
 243
 244 class _Hqxdecoderengine:
 245     """Read data via the decoder in 4-byte chunks"""
 246
 247     def __init__(self, ifp):
 248         self.ifp = ifp
 249         self.eof = 0
 250
 251     def read(self, totalwtd):
 252         """Read at least wtd bytes (or until EOF)"""
 253         decdata = b''
 254         wtd = totalwtd
 255         #
 256         # The loop here is convoluted, since we don't really now how
 257         # much to decode: there may be newlines in the incoming data.
 258         while wtd > 0:
 259             if self.eof: return decdata
 260             wtd = ((wtd + 2) // 3) * 4
 261             data = self.ifp.read(wtd)
 262             #
 263             # Next problem: there may not be a complete number of
 264             # bytes in what we pass to a2b. Solve by yet another
 265             # loop.
 266             #
 267             while True:
 268                 try:
 269                     decdatacur, self.eof = binascii.a2b_hqx(data)
 270                     break
 271                 except binascii.Incomplete:
 272                     pass
 273                 newdata = self.ifp.read(1)
 274                 if not newdata:
 275                     raise Error('Premature EOF on binhex file')
 276                 data = data + newdata
 277             decdata = decdata + decdatacur
 278             wtd = totalwtd - len(decdata)
 279             if not decdata and not self.eof:
 280                 raise Error('Premature EOF on binhex file')
 281         return decdata
 282
 283     def close(self):
 284         self.ifp.close()
 285
 286 class _Rledecoderengine:
 287     """Read data via the RLE-coder"""
 288
 289     def __init__(self, ifp):
 290         self.ifp = ifp
 291         self.pre_buffer = b''
 292         self.post_buffer = b''
 293         self.eof = 0
 294
 295     def read(self, wtd):
 296         if wtd > len(self.post_buffer):
 297             self._fill(wtd - len(self.post_buffer))
 298         rv = self.post_buffer[:wtd]
 299         self.post_buffer = self.post_buffer[wtd:]
 300         return rv
 301
 302     def _fill(self, wtd):
 303         self.pre_buffer = self.pre_buffer + self.ifp.read(wtd + 4)
 304         if self.ifp.eof:
 305             self.post_buffer = self.post_buffer + \
 306                 binascii.rledecode_hqx(self.pre_buffer)
 307             self.pre_buffer = b''
 308             return
 309
 310         #
 311         # Obfuscated code ahead. We have to take care that we don't
 312         # end up with an orphaned RUNCHAR later on. So, we keep a couple
 313         # of bytes in the buffer, depending on what the end of
 314         # the buffer looks like:
 315         # '\220\0\220' - Keep 3 bytes: repeated \220 (escaped as \220\0)
 316         # '?\220' - Keep 2 bytes: repeated something-else
 317         # '\220\0' - Escaped \220: Keep 2 bytes.
 318         # '?\220?' - Complete repeat sequence: decode all
 319         # otherwise: keep 1 byte.
 320         #
 321         mark = len(self.pre_buffer)
 322         if self.pre_buffer[-3:] == RUNCHAR + b'\0' + RUNCHAR:
 323             mark = mark - 3
 324         elif self.pre_buffer[-1:] == RUNCHAR:
 325             mark = mark - 2
 326         elif self.pre_buffer[-2:] == RUNCHAR + b'\0':
 327             mark = mark - 2
 328         elif self.pre_buffer[-2:-1] == RUNCHAR:
 329             pass # Decode all
 330         else:
 331             mark = mark - 1
 332
 333         self.post_buffer = self.post_buffer + \
 334             binascii.rledecode_hqx(self.pre_buffer[:mark])
 335         self.pre_buffer = self.pre_buffer[mark:]
 336
 337     def close(self):
 338         self.ifp.close()
 339
 340 class HexBin:
 341     def __init__(self, ifp):
 342         if isinstance(ifp, str):
 343             ifp = io.open(ifp, 'rb')
 344         #
 345         # Find initial colon.
 346         #
 347         while True:
 348             ch = ifp.read(1)
 349             if not ch:
 350                 raise Error("No binhex data found")
 351             # Cater for \r\n terminated lines (which show up as \n\r, hence
 352             # all lines start with \r)
 353             if ch == b'\r':
 354                 continue
 355             if ch == b':':
 356                 break
 357
 358         hqxifp = _Hqxdecoderengine(ifp)
 359         self.ifp = _Rledecoderengine(hqxifp)
 360         self.crc = 0
 361         self._readheader()
 362
 363     def _read(self, len):
 364         data = self.ifp.read(len)
 365         self.crc = binascii.crc_hqx(data, self.crc)
 366         return data
 367
 368     def _checkcrc(self):
 369         filecrc = struct.unpack('>h', self.ifp.read(2))[0] & 0xffff
 370         #self.crc = binascii.crc_hqx('\0\0', self.crc)
 371         # XXXX Is this needed??
 372         self.crc = self.crc & 0xffff
 373         if filecrc != self.crc:
 374             raise Error('CRC error, computed %x, read %x'
 375                         % (self.crc, filecrc))
 376         self.crc = 0
 377
 378     def _readheader(self):
 379         len = self._read(1)
 380         fname = self._read(ord(len))
 381         rest = self._read(1 + 4 + 4 + 2 + 4 + 4)
 382         self._checkcrc()
 383
 384         type = rest[1:5]
 385         creator = rest[5:9]
 386         flags = struct.unpack('>h', rest[9:11])[0]
 387         self.dlen = struct.unpack('>l', rest[11:15])[0]
 388         self.rlen = struct.unpack('>l', rest[15:19])[0]
 389
 390         self.FName = fname
 391         self.FInfo = FInfo()
 392         self.FInfo.Creator = creator
 393         self.FInfo.Type = type
 394         self.FInfo.Flags = flags
 395
 396         self.state = _DID_HEADER
 397
 398     def read(self, *n):
 399         if self.state != _DID_HEADER:
 400             raise Error('Read data at wrong time')
 401         if n:
 402             n = n[0]
 403             n = min(n, self.dlen)
 404         else:
 405             n = self.dlen
 406         rv = b''
 407         while len(rv) < n:
 408             rv = rv + self._read(n-len(rv))
 409         self.dlen = self.dlen - n
 410         return rv
 411
 412     def close_data(self):
 413         if self.state != _DID_HEADER:
 414             raise Error('close_data at wrong time')
 415         if self.dlen:
 416             dummy = self._read(self.dlen)
 417         self._checkcrc()
 418         self.state = _DID_DATA
 419
 420     def read_rsrc(self, *n):
 421         if self.state == _DID_HEADER:
 422             self.close_data()
 423         if self.state != _DID_DATA:
 424             raise Error('Read resource data at wrong time')
 425         if n:
 426             n = n[0]
 427             n = min(n, self.rlen)
 428         else:
 429             n = self.rlen
 430         self.rlen = self.rlen - n
 431         return self._read(n)
 432
 433     def close(self):
 434         if self.rlen:
 435             dummy = self.read_rsrc(self.rlen)
 436         self._checkcrc()
 437         self.state = _DID_RSRC
 438         self.ifp.close()
 439
 440 def hexbin(inp, out):
 441     """hexbin(infilename, outfilename) - Decode binhexed file"""
 442     ifp = HexBin(inp)
 443     finfo = ifp.FInfo
 444     if not out:
 445         out = ifp.FName
 446
 447     ofp = io.open(out, 'wb')
 448     # XXXX Do translation on non-mac systems
 449     while True:
 450         d = ifp.read(128000)
 451         if not d: break
 452         ofp.write(d)
 453     ofp.close()
 454     ifp.close_data()
 455
 456     d = ifp.read_rsrc(128000)
 457     if d:
 458         ofp = openrsrc(out, 'wb')
 459         ofp.write(d)
 460         while True:
 461             d = ifp.read_rsrc(128000)
 462             if not d: break
 463             ofp.write(d)
 464         ofp.close()
 465
 466     ifp.close()