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