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 | ofp.write('(This file must be converted with BinHex 4.0)\n\n:')
|
---|
179 | hqxer = _Hqxcoderengine(ofp)
|
---|
180 | self.ofp = _Rlecoderengine(hqxer)
|
---|
181 | self.crc = 0
|
---|
182 | if finfo is None:
|
---|
183 | finfo = FInfo()
|
---|
184 | self.dlen = dlen
|
---|
185 | self.rlen = rlen
|
---|
186 | self._writeinfo(name, finfo)
|
---|
187 | self.state = _DID_HEADER
|
---|
188 |
|
---|
189 | def _writeinfo(self, name, finfo):
|
---|
190 | nl = len(name)
|
---|
191 | if nl > 63:
|
---|
192 | raise Error, 'Filename too long'
|
---|
193 | d = chr(nl) + name + '\0'
|
---|
194 | d2 = finfo.Type + finfo.Creator
|
---|
195 |
|
---|
196 | # Force all structs to be packed with big-endian
|
---|
197 | d3 = struct.pack('>h', finfo.Flags)
|
---|
198 | d4 = struct.pack('>ii', self.dlen, self.rlen)
|
---|
199 | info = d + d2 + d3 + d4
|
---|
200 | self._write(info)
|
---|
201 | self._writecrc()
|
---|
202 |
|
---|
203 | def _write(self, data):
|
---|
204 | self.crc = binascii.crc_hqx(data, self.crc)
|
---|
205 | self.ofp.write(data)
|
---|
206 |
|
---|
207 | def _writecrc(self):
|
---|
208 | # XXXX Should this be here??
|
---|
209 | # self.crc = binascii.crc_hqx('\0\0', self.crc)
|
---|
210 | if self.crc < 0:
|
---|
211 | fmt = '>h'
|
---|
212 | else:
|
---|
213 | fmt = '>H'
|
---|
214 | self.ofp.write(struct.pack(fmt, self.crc))
|
---|
215 | self.crc = 0
|
---|
216 |
|
---|
217 | def write(self, data):
|
---|
218 | if self.state != _DID_HEADER:
|
---|
219 | raise Error, 'Writing data at the wrong time'
|
---|
220 | self.dlen = self.dlen - len(data)
|
---|
221 | self._write(data)
|
---|
222 |
|
---|
223 | def close_data(self):
|
---|
224 | if self.dlen != 0:
|
---|
225 | raise Error, 'Incorrect data size, diff=%r' % (self.rlen,)
|
---|
226 | self._writecrc()
|
---|
227 | self.state = _DID_DATA
|
---|
228 |
|
---|
229 | def write_rsrc(self, data):
|
---|
230 | if self.state < _DID_DATA:
|
---|
231 | self.close_data()
|
---|
232 | if self.state != _DID_DATA:
|
---|
233 | raise Error, 'Writing resource data at the wrong time'
|
---|
234 | self.rlen = self.rlen - len(data)
|
---|
235 | self._write(data)
|
---|
236 |
|
---|
237 | def close(self):
|
---|
238 | if self.state < _DID_DATA:
|
---|
239 | self.close_data()
|
---|
240 | if self.state != _DID_DATA:
|
---|
241 | raise Error, 'Close at the wrong time'
|
---|
242 | if self.rlen != 0:
|
---|
243 | raise Error, \
|
---|
244 | "Incorrect resource-datasize, diff=%r" % (self.rlen,)
|
---|
245 | self._writecrc()
|
---|
246 | self.ofp.close()
|
---|
247 | self.state = None
|
---|
248 | del self.ofp
|
---|
249 |
|
---|
250 | def binhex(inp, out):
|
---|
251 | """(infilename, outfilename) - Create binhex-encoded copy of a file"""
|
---|
252 | finfo = getfileinfo(inp)
|
---|
253 | ofp = BinHex(finfo, out)
|
---|
254 |
|
---|
255 | ifp = open(inp, 'rb')
|
---|
256 | # XXXX Do textfile translation on non-mac systems
|
---|
257 | while 1:
|
---|
258 | d = ifp.read(128000)
|
---|
259 | if not d: break
|
---|
260 | ofp.write(d)
|
---|
261 | ofp.close_data()
|
---|
262 | ifp.close()
|
---|
263 |
|
---|
264 | ifp = openrsrc(inp, 'rb')
|
---|
265 | while 1:
|
---|
266 | d = ifp.read(128000)
|
---|
267 | if not d: break
|
---|
268 | ofp.write_rsrc(d)
|
---|
269 | ofp.close()
|
---|
270 | ifp.close()
|
---|
271 |
|
---|
272 | class _Hqxdecoderengine:
|
---|
273 | """Read data via the decoder in 4-byte chunks"""
|
---|
274 |
|
---|
275 | def __init__(self, ifp):
|
---|
276 | self.ifp = ifp
|
---|
277 | self.eof = 0
|
---|
278 |
|
---|
279 | def read(self, totalwtd):
|
---|
280 | """Read at least wtd bytes (or until EOF)"""
|
---|
281 | decdata = ''
|
---|
282 | wtd = totalwtd
|
---|
283 | #
|
---|
284 | # The loop here is convoluted, since we don't really now how
|
---|
285 | # much to decode: there may be newlines in the incoming data.
|
---|
286 | while wtd > 0:
|
---|
287 | if self.eof: return decdata
|
---|
288 | wtd = ((wtd+2)//3)*4
|
---|
289 | data = self.ifp.read(wtd)
|
---|
290 | #
|
---|
291 | # Next problem: there may not be a complete number of
|
---|
292 | # bytes in what we pass to a2b. Solve by yet another
|
---|
293 | # loop.
|
---|
294 | #
|
---|
295 | while 1:
|
---|
296 | try:
|
---|
297 | decdatacur, self.eof = \
|
---|
298 | binascii.a2b_hqx(data)
|
---|
299 | break
|
---|
300 | except binascii.Incomplete:
|
---|
301 | pass
|
---|
302 | newdata = self.ifp.read(1)
|
---|
303 | if not newdata:
|
---|
304 | raise Error, \
|
---|
305 | 'Premature EOF on binhex file'
|
---|
306 | data = data + newdata
|
---|
307 | decdata = decdata + decdatacur
|
---|
308 | wtd = totalwtd - len(decdata)
|
---|
309 | if not decdata and not self.eof:
|
---|
310 | raise Error, 'Premature EOF on binhex file'
|
---|
311 | return decdata
|
---|
312 |
|
---|
313 | def close(self):
|
---|
314 | self.ifp.close()
|
---|
315 |
|
---|
316 | class _Rledecoderengine:
|
---|
317 | """Read data via the RLE-coder"""
|
---|
318 |
|
---|
319 | def __init__(self, ifp):
|
---|
320 | self.ifp = ifp
|
---|
321 | self.pre_buffer = ''
|
---|
322 | self.post_buffer = ''
|
---|
323 | self.eof = 0
|
---|
324 |
|
---|
325 | def read(self, wtd):
|
---|
326 | if wtd > len(self.post_buffer):
|
---|
327 | self._fill(wtd-len(self.post_buffer))
|
---|
328 | rv = self.post_buffer[:wtd]
|
---|
329 | self.post_buffer = self.post_buffer[wtd:]
|
---|
330 | return rv
|
---|
331 |
|
---|
332 | def _fill(self, wtd):
|
---|
333 | self.pre_buffer = self.pre_buffer + self.ifp.read(wtd+4)
|
---|
334 | if self.ifp.eof:
|
---|
335 | self.post_buffer = self.post_buffer + \
|
---|
336 | binascii.rledecode_hqx(self.pre_buffer)
|
---|
337 | self.pre_buffer = ''
|
---|
338 | return
|
---|
339 |
|
---|
340 | #
|
---|
341 | # Obfuscated code ahead. We have to take care that we don't
|
---|
342 | # end up with an orphaned RUNCHAR later on. So, we keep a couple
|
---|
343 | # of bytes in the buffer, depending on what the end of
|
---|
344 | # the buffer looks like:
|
---|
345 | # '\220\0\220' - Keep 3 bytes: repeated \220 (escaped as \220\0)
|
---|
346 | # '?\220' - Keep 2 bytes: repeated something-else
|
---|
347 | # '\220\0' - Escaped \220: Keep 2 bytes.
|
---|
348 | # '?\220?' - Complete repeat sequence: decode all
|
---|
349 | # otherwise: keep 1 byte.
|
---|
350 | #
|
---|
351 | mark = len(self.pre_buffer)
|
---|
352 | if self.pre_buffer[-3:] == RUNCHAR + '\0' + RUNCHAR:
|
---|
353 | mark = mark - 3
|
---|
354 | elif self.pre_buffer[-1] == RUNCHAR:
|
---|
355 | mark = mark - 2
|
---|
356 | elif self.pre_buffer[-2:] == RUNCHAR + '\0':
|
---|
357 | mark = mark - 2
|
---|
358 | elif self.pre_buffer[-2] == RUNCHAR:
|
---|
359 | pass # Decode all
|
---|
360 | else:
|
---|
361 | mark = mark - 1
|
---|
362 |
|
---|
363 | self.post_buffer = self.post_buffer + \
|
---|
364 | binascii.rledecode_hqx(self.pre_buffer[:mark])
|
---|
365 | self.pre_buffer = self.pre_buffer[mark:]
|
---|
366 |
|
---|
367 | def close(self):
|
---|
368 | self.ifp.close()
|
---|
369 |
|
---|
370 | class HexBin:
|
---|
371 | def __init__(self, ifp):
|
---|
372 | if type(ifp) == type(''):
|
---|
373 | ifp = open(ifp)
|
---|
374 | #
|
---|
375 | # Find initial colon.
|
---|
376 | #
|
---|
377 | while 1:
|
---|
378 | ch = ifp.read(1)
|
---|
379 | if not ch:
|
---|
380 | raise Error, "No binhex data found"
|
---|
381 | # Cater for \r\n terminated lines (which show up as \n\r, hence
|
---|
382 | # all lines start with \r)
|
---|
383 | if ch == '\r':
|
---|
384 | continue
|
---|
385 | if ch == ':':
|
---|
386 | break
|
---|
387 | if ch != '\n':
|
---|
388 | dummy = ifp.readline()
|
---|
389 |
|
---|
390 | hqxifp = _Hqxdecoderengine(ifp)
|
---|
391 | self.ifp = _Rledecoderengine(hqxifp)
|
---|
392 | self.crc = 0
|
---|
393 | self._readheader()
|
---|
394 |
|
---|
395 | def _read(self, len):
|
---|
396 | data = self.ifp.read(len)
|
---|
397 | self.crc = binascii.crc_hqx(data, self.crc)
|
---|
398 | return data
|
---|
399 |
|
---|
400 | def _checkcrc(self):
|
---|
401 | filecrc = struct.unpack('>h', self.ifp.read(2))[0] & 0xffff
|
---|
402 | #self.crc = binascii.crc_hqx('\0\0', self.crc)
|
---|
403 | # XXXX Is this needed??
|
---|
404 | self.crc = self.crc & 0xffff
|
---|
405 | if filecrc != self.crc:
|
---|
406 | raise Error, 'CRC error, computed %x, read %x' \
|
---|
407 | %(self.crc, filecrc)
|
---|
408 | self.crc = 0
|
---|
409 |
|
---|
410 | def _readheader(self):
|
---|
411 | len = self._read(1)
|
---|
412 | fname = self._read(ord(len))
|
---|
413 | rest = self._read(1+4+4+2+4+4)
|
---|
414 | self._checkcrc()
|
---|
415 |
|
---|
416 | type = rest[1:5]
|
---|
417 | creator = rest[5:9]
|
---|
418 | flags = struct.unpack('>h', rest[9:11])[0]
|
---|
419 | self.dlen = struct.unpack('>l', rest[11:15])[0]
|
---|
420 | self.rlen = struct.unpack('>l', rest[15:19])[0]
|
---|
421 |
|
---|
422 | self.FName = fname
|
---|
423 | self.FInfo = FInfo()
|
---|
424 | self.FInfo.Creator = creator
|
---|
425 | self.FInfo.Type = type
|
---|
426 | self.FInfo.Flags = flags
|
---|
427 |
|
---|
428 | self.state = _DID_HEADER
|
---|
429 |
|
---|
430 | def read(self, *n):
|
---|
431 | if self.state != _DID_HEADER:
|
---|
432 | raise Error, 'Read data at wrong time'
|
---|
433 | if n:
|
---|
434 | n = n[0]
|
---|
435 | n = min(n, self.dlen)
|
---|
436 | else:
|
---|
437 | n = self.dlen
|
---|
438 | rv = ''
|
---|
439 | while len(rv) < n:
|
---|
440 | rv = rv + self._read(n-len(rv))
|
---|
441 | self.dlen = self.dlen - n
|
---|
442 | return rv
|
---|
443 |
|
---|
444 | def close_data(self):
|
---|
445 | if self.state != _DID_HEADER:
|
---|
446 | raise Error, 'close_data at wrong time'
|
---|
447 | if self.dlen:
|
---|
448 | dummy = self._read(self.dlen)
|
---|
449 | self._checkcrc()
|
---|
450 | self.state = _DID_DATA
|
---|
451 |
|
---|
452 | def read_rsrc(self, *n):
|
---|
453 | if self.state == _DID_HEADER:
|
---|
454 | self.close_data()
|
---|
455 | if self.state != _DID_DATA:
|
---|
456 | raise Error, 'Read resource data at wrong time'
|
---|
457 | if n:
|
---|
458 | n = n[0]
|
---|
459 | n = min(n, self.rlen)
|
---|
460 | else:
|
---|
461 | n = self.rlen
|
---|
462 | self.rlen = self.rlen - n
|
---|
463 | return self._read(n)
|
---|
464 |
|
---|
465 | def close(self):
|
---|
466 | if self.rlen:
|
---|
467 | dummy = self.read_rsrc(self.rlen)
|
---|
468 | self._checkcrc()
|
---|
469 | self.state = _DID_RSRC
|
---|
470 | self.ifp.close()
|
---|
471 |
|
---|
472 | def hexbin(inp, out):
|
---|
473 | """(infilename, outfilename) - Decode binhexed file"""
|
---|
474 | ifp = HexBin(inp)
|
---|
475 | finfo = ifp.FInfo
|
---|
476 | if not out:
|
---|
477 | out = ifp.FName
|
---|
478 |
|
---|
479 | ofp = open(out, 'wb')
|
---|
480 | # XXXX Do translation on non-mac systems
|
---|
481 | while 1:
|
---|
482 | d = ifp.read(128000)
|
---|
483 | if not d: break
|
---|
484 | ofp.write(d)
|
---|
485 | ofp.close()
|
---|
486 | ifp.close_data()
|
---|
487 |
|
---|
488 | d = ifp.read_rsrc(128000)
|
---|
489 | if d:
|
---|
490 | ofp = openrsrc(out, 'wb')
|
---|
491 | ofp.write(d)
|
---|
492 | while 1:
|
---|
493 | d = ifp.read_rsrc(128000)
|
---|
494 | if not d: break
|
---|
495 | ofp.write(d)
|
---|
496 | ofp.close()
|
---|
497 |
|
---|
498 | ifp.close()
|
---|
499 |
|
---|
500 | def _test():
|
---|
501 | fname = sys.argv[1]
|
---|
502 | binhex(fname, fname+'.hqx')
|
---|
503 | hexbin(fname+'.hqx', fname+'.viahqx')
|
---|
504 | #hexbin(fname, fname+'.unpacked')
|
---|
505 | sys.exit(1)
|
---|
506 |
|
---|
507 | if __name__ == '__main__':
|
---|
508 | _test()
|
---|