Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 1 | """binhex - Macintosh binhex compression/decompression |
| 2 | easy interface: |
| 3 | binhex(inputfilename, outputfilename) |
| 4 | hexbin(inputfilename, outputfilename) |
| 5 | """ |
| 6 | |
| 7 | # |
| 8 | # Jack Jansen, CWI, August 1995. |
| 9 | # |
| 10 | # The module is supposed to be as compatible as possible. Especially the |
| 11 | # easy interface should work "as expected" on any platform. |
| 12 | # XXXX Note: currently, textfiles appear in mac-form on all platforms. |
| 13 | # We seem to lack a simple character-translate in python. |
| 14 | # (we should probably use ISO-Latin-1 on all but the mac platform). |
| 15 | # XXXX The simeple routines are too simple: they expect to hold the complete |
| 16 | # files in-core. Should be fixed. |
| 17 | # XXXX It would be nice to handle AppleDouble format on unix (for servers serving |
| 18 | # macs). |
| 19 | # XXXX I don't understand what happens when you get 0x90 times the same byte on |
| 20 | # input. The resulting code (xx 90 90) would appear to be interpreted as an |
| 21 | # escaped *value* of 0x90. All coders I've seen appear to ignore this nicety... |
| 22 | # |
| 23 | import sys |
| 24 | import os |
| 25 | import struct |
| 26 | import string |
| 27 | import binascii |
| 28 | |
Jack Jansen | 479c1b3 | 1995-08-14 12:41:20 +0000 | [diff] [blame] | 29 | DEBUG=0 |
| 30 | if DEBUG: |
| 31 | testf=open('@binhex.dbg.out', 'w') |
| 32 | |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 33 | Error = 'binhex.Error' |
| 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=48 # What we pass to hqx-coder at once |
| 41 | # *NOTE* Must be divisible by 3! |
| 42 | RUNCHAR=chr(0x90) # run-length introducer |
| 43 | |
| 44 | # |
| 45 | # The code is currently byte-order dependent |
| 46 | if struct.pack('i', 0177) != '\0\0\0\177': |
| 47 | raise ImportError, 'Module binhex is big-endian only' |
| 48 | |
| 49 | # |
| 50 | # Workarounds for non-mac machines. |
| 51 | if os.name == 'mac': |
| 52 | import macfs |
| 53 | |
| 54 | def FInfo(): |
| 55 | return macfs.FInfo() |
| 56 | |
| 57 | def getfileinfo(name): |
| 58 | finfo = macfs.FSSpec(name).GetFInfo() |
| 59 | dir, file = os.path.split(name) |
| 60 | # XXXX Get resource/data sizes |
| 61 | fp = open(name, 'rb') |
| 62 | fp.seek(0, 2) |
| 63 | dlen = fp.tell() |
| 64 | fp = open(name, '*rb') |
| 65 | fp.seek(0, 2) |
| 66 | rlen = fp.tell() |
| 67 | return file, finfo, dlen, rlen |
| 68 | |
| 69 | def openrsrc(name, *mode): |
| 70 | if mode: |
| 71 | mode = mode[0] |
| 72 | else: |
| 73 | mode = 'rb' |
| 74 | mode = '*' + mode |
| 75 | return open(name, mode) |
| 76 | |
| 77 | else: |
| 78 | # |
| 79 | # Glue code for non-macintosh useage |
| 80 | # |
| 81 | import regsub |
| 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 in string.whitespace and (c<' ' or ord(c) > 0177): |
| 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 = regsub.sub(':', '-', file) |
| 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 | |
| 126 | def write(self, data): |
| 127 | self.data = self.data + data |
| 128 | while len(self.data) > LINELEN: |
| 129 | hqxdata = binascii.b2a_hqx(self.data[:LINELEN]) |
| 130 | self.ofp.write(hqxdata+'\n') |
| 131 | self.data = self.data[LINELEN:] |
| 132 | |
| 133 | def close(self): |
| 134 | if self.data: |
| 135 | self.ofp.write(binascii.b2a_hqx(self.data)) |
| 136 | self.ofp.write(':\n') |
| 137 | self.ofp.close() |
| 138 | |
| 139 | class _Rlecoderengine: |
| 140 | """Write data to the RLE-coder in suitably large chunks""" |
| 141 | |
| 142 | def __init__(self, ofp): |
| 143 | self.ofp = ofp |
| 144 | self.data = '' |
| 145 | |
| 146 | def write(self, data): |
Jack Jansen | ac7c0df | 1995-08-17 14:17:39 +0000 | [diff] [blame] | 147 | if DEBUG: |
Jack Jansen | 479c1b3 | 1995-08-14 12:41:20 +0000 | [diff] [blame] | 148 | testf.write(data) # XXXX |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 149 | self.data = self.data + data |
| 150 | if len(self.data) < REASONABLY_LARGE: |
| 151 | return |
| 152 | rledata = binascii.rlecode_hqx(self.data) |
| 153 | self.ofp.write(rledata) |
| 154 | self.data = '' |
| 155 | |
| 156 | def close(self): |
| 157 | if self.data: |
| 158 | rledata = binascii.rlecode_hqx(self.data) |
| 159 | self.ofp.write(rledata) |
| 160 | self.ofp.close() |
| 161 | |
| 162 | class BinHex: |
| 163 | def __init__(self, (name, finfo, dlen, rlen), ofp): |
| 164 | if type(ofp) == type(''): |
| 165 | ofname = ofp |
| 166 | ofp = open(ofname, 'w') |
| 167 | if os.name == 'mac': |
| 168 | fss = macfs.FSSpec(ofname) |
| 169 | fss.SetCreatorType('BnHq', 'TEXT') |
| 170 | ofp.write('(This file may be decompressed with BinHex 4.0)\n\n:') |
| 171 | hqxer = _Hqxcoderengine(ofp) |
| 172 | self.ofp = _Rlecoderengine(hqxer) |
| 173 | self.crc = 0 |
| 174 | if finfo == None: |
| 175 | finfo = FInfo() |
| 176 | self.dlen = dlen |
| 177 | self.rlen = rlen |
| 178 | self._writeinfo(name, finfo) |
| 179 | self.state = _DID_HEADER |
| 180 | |
| 181 | def _writeinfo(self, name, finfo): |
| 182 | if DEBUG: |
| 183 | print 'binhex info:', name, finfo.Type, finfo.Creator, self.dlen, self.rlen |
| 184 | name = name |
| 185 | nl = len(name) |
| 186 | if nl > 63: |
| 187 | raise Error, 'Filename too long' |
| 188 | d = chr(nl) + name + '\0' |
| 189 | d2 = finfo.Type + finfo.Creator |
| 190 | d3 = struct.pack('h', finfo.Flags) |
| 191 | d4 = struct.pack('ii', self.dlen, self.rlen) |
| 192 | info = d + d2 + d3 + d4 |
| 193 | self._write(info) |
| 194 | self._writecrc() |
| 195 | |
| 196 | def _write(self, data): |
| 197 | self.crc = binascii.crc_hqx(data, self.crc) |
| 198 | self.ofp.write(data) |
| 199 | |
| 200 | def _writecrc(self): |
Jack Jansen | 685e16d | 1995-10-03 14:41:15 +0000 | [diff] [blame^] | 201 | ## self.crc = binascii.crc_hqx('\0\0', self.crc) # XXXX Should this be here?? |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 202 | self.ofp.write(struct.pack('h', self.crc)) |
| 203 | self.crc = 0 |
| 204 | |
| 205 | def write(self, data): |
| 206 | if self.state != _DID_HEADER: |
| 207 | raise Error, 'Writing data at the wrong time' |
| 208 | self.dlen = self.dlen - len(data) |
| 209 | self._write(data) |
| 210 | |
| 211 | def close_data(self): |
| 212 | if self.dlen <> 0: |
| 213 | raise Error, 'Incorrect data size, diff='+`self.rlen` |
| 214 | self._writecrc() |
| 215 | self.state = _DID_DATA |
| 216 | |
| 217 | def write_rsrc(self, data): |
| 218 | if self.state < _DID_DATA: |
| 219 | self.close_data() |
| 220 | if self.state != _DID_DATA: |
| 221 | raise Error, 'Writing resource data at the wrong time' |
| 222 | self.rlen = self.rlen - len(data) |
| 223 | self._write(data) |
| 224 | |
| 225 | def close(self): |
| 226 | if self.state < _DID_DATA: |
| 227 | self.close_data() |
| 228 | if self.state != _DID_DATA: |
| 229 | raise Error, 'Close at the wrong time' |
| 230 | if self.rlen <> 0: |
| 231 | raise Error, "Incorrect resource-datasize, diff="+`self.rlen` |
| 232 | self._writecrc() |
| 233 | self.ofp.close() |
| 234 | self.state = None |
| 235 | |
| 236 | def binhex(inp, out): |
| 237 | """(infilename, outfilename) - Create binhex-encoded copy of a file""" |
| 238 | finfo = getfileinfo(inp) |
| 239 | ofp = BinHex(finfo, out) |
| 240 | |
| 241 | ifp = open(inp, 'rb') |
| 242 | # XXXX Do textfile translation on non-mac systems |
| 243 | d = ifp.read() |
| 244 | ofp.write(d) |
| 245 | ofp.close_data() |
| 246 | ifp.close() |
| 247 | |
| 248 | ifp = openrsrc(inp, 'rb') |
| 249 | d = ifp.read() |
| 250 | ofp.write_rsrc(d) |
| 251 | ofp.close() |
| 252 | ifp.close() |
| 253 | |
| 254 | class _Hqxdecoderengine: |
| 255 | """Read data via the decoder in 4-byte chunks""" |
| 256 | |
| 257 | def __init__(self, ifp): |
| 258 | self.ifp = ifp |
| 259 | self.eof = 0 |
| 260 | |
Jack Jansen | 685e16d | 1995-10-03 14:41:15 +0000 | [diff] [blame^] | 261 | def read(self, totalwtd): |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 262 | """Read at least wtd bytes (or until EOF)""" |
| 263 | decdata = '' |
Jack Jansen | 685e16d | 1995-10-03 14:41:15 +0000 | [diff] [blame^] | 264 | wtd = totalwtd |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 265 | # |
| 266 | # The loop here is convoluted, since we don't really now how much |
| 267 | # to decode: there may be newlines in the incoming data. |
| 268 | while wtd > 0: |
| 269 | if self.eof: return decdata |
| 270 | wtd = ((wtd+2)/3)*4 |
| 271 | data = self.ifp.read(wtd) |
| 272 | # |
| 273 | # Next problem: there may not be a complete number of bytes in what we |
| 274 | # pass to a2b. Solve by yet another loop. |
| 275 | # |
| 276 | while 1: |
| 277 | try: |
| 278 | decdatacur, self.eof = binascii.a2b_hqx(data) |
| 279 | if self.eof: print 'EOF' |
| 280 | break |
| 281 | except binascii.Incomplete: |
| 282 | pass |
| 283 | newdata = self.ifp.read(1) |
| 284 | if not newdata: |
| 285 | raise Error, 'Premature EOF on binhex file' |
| 286 | data = data + newdata |
| 287 | decdata = decdata + decdatacur |
Jack Jansen | 685e16d | 1995-10-03 14:41:15 +0000 | [diff] [blame^] | 288 | wtd = totalwtd - len(decdata) |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 289 | if not decdata and not self.eof: |
| 290 | raise Error, 'Premature EOF on binhex file' |
| 291 | return decdata |
| 292 | |
| 293 | def close(self): |
| 294 | self.ifp.close() |
| 295 | |
| 296 | class _Rledecoderengine: |
| 297 | """Read data via the RLE-coder""" |
| 298 | |
| 299 | def __init__(self, ifp): |
| 300 | self.ifp = ifp |
| 301 | self.pre_buffer = '' |
| 302 | self.post_buffer = '' |
| 303 | self.eof = 0 |
| 304 | |
| 305 | def read(self, wtd): |
| 306 | if wtd > len(self.post_buffer): |
| 307 | self._fill(wtd-len(self.post_buffer)) |
| 308 | rv = self.post_buffer[:wtd] |
| 309 | self.post_buffer = self.post_buffer[wtd:] |
| 310 | print 'WTD', wtd, 'GOT', len(rv) |
| 311 | return rv |
| 312 | |
| 313 | def _fill(self, wtd): |
| 314 | # |
| 315 | # Obfuscated code ahead. We keep at least one byte in the pre_buffer, |
| 316 | # so we don't stumble over an orphaned RUNCHAR later on. If the |
| 317 | # last or second-last char is a RUNCHAR we keep more bytes. |
| 318 | # |
| 319 | self.pre_buffer = self.pre_buffer + self.ifp.read(wtd+2) |
| 320 | if self.ifp.eof: |
| 321 | self.post_buffer = self.post_buffer + \ |
| 322 | binascii.rledecode_hqx(self.pre_buffer) |
| 323 | self.pre_buffer = '' |
| 324 | return |
| 325 | |
| 326 | lastrle = string.rfind(self.pre_buffer, RUNCHAR) |
| 327 | if lastrle > 0 and lastrle == len(self.pre_buffer)-1: |
| 328 | # Last byte is an RLE, keep two bytes |
| 329 | mark = len(self.pre_buffer)-2 |
| 330 | elif lastrle > 0 and lastrle == len(self.pre_buffer)-2: |
| 331 | # second-last byte is an RLE. Decode all. |
| 332 | mark = len(self.pre_buffer) |
| 333 | else: |
| 334 | mark = len(self.pre_buffer)-1 |
| 335 | self.post_buffer = self.post_buffer + \ |
| 336 | binascii.rledecode_hqx(self.pre_buffer[:mark]) |
| 337 | self.pre_buffer = self.pre_buffer[mark:] |
| 338 | |
| 339 | def close(self): |
| 340 | self.ifp.close() |
| 341 | |
| 342 | class HexBin: |
| 343 | def __init__(self, ifp): |
| 344 | if type(ifp) == type(''): |
| 345 | ifp = open(ifp) |
| 346 | # |
| 347 | # Find initial colon. |
| 348 | # |
| 349 | while 1: |
| 350 | ch = ifp.read(1) |
| 351 | if not ch: |
| 352 | raise Error, "No binhex data found" |
| 353 | if ch == ':': |
| 354 | break |
| 355 | if ch != '\n': |
| 356 | dummy = ifp.readline() |
| 357 | if DEBUG: |
| 358 | print 'SKIP:', ch+dummy |
| 359 | |
| 360 | hqxifp = _Hqxdecoderengine(ifp) |
| 361 | self.ifp = _Rledecoderengine(hqxifp) |
| 362 | self.crc = 0 |
| 363 | self._readheader() |
| 364 | |
| 365 | def _read(self, len): |
| 366 | data = self.ifp.read(len) |
| 367 | self.crc = binascii.crc_hqx(data, self.crc) |
| 368 | return data |
| 369 | |
| 370 | def _checkcrc(self): |
| 371 | filecrc = struct.unpack('h', self.ifp.read(2))[0] & 0xffff |
Jack Jansen | 685e16d | 1995-10-03 14:41:15 +0000 | [diff] [blame^] | 372 | ## self.crc = binascii.crc_hqx('\0\0', self.crc) # XXXX Is this needed?? |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 373 | self.crc = self.crc & 0xffff |
| 374 | if DEBUG: |
| 375 | print 'DBG CRC %x %x'%(self.crc, filecrc) |
Jack Jansen | 685e16d | 1995-10-03 14:41:15 +0000 | [diff] [blame^] | 376 | if filecrc != self.crc: |
| 377 | raise Error, 'CRC error, computed %x, read %x'%(self.crc, filecrc) |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 378 | self.crc = 0 |
| 379 | |
| 380 | def _readheader(self): |
| 381 | len = self._read(1) |
| 382 | fname = self._read(ord(len)) |
| 383 | rest = self._read(1+4+4+2+4+4) |
| 384 | self._checkcrc() |
| 385 | |
| 386 | type = rest[1:5] |
| 387 | creator = rest[5:9] |
| 388 | flags = struct.unpack('h', rest[9:11])[0] |
| 389 | self.dlen = struct.unpack('l', rest[11:15])[0] |
| 390 | self.rlen = struct.unpack('l', rest[15:19])[0] |
| 391 | |
| 392 | if DEBUG: |
| 393 | print 'DATA, RLEN', self.dlen, self.rlen |
| 394 | |
| 395 | self.FName = fname |
| 396 | self.FInfo = FInfo() |
| 397 | self.FInfo.Creator = creator |
| 398 | self.FInfo.Type = type |
| 399 | self.FInfo.Flags = flags |
| 400 | |
| 401 | self.state = _DID_HEADER |
| 402 | |
| 403 | def read(self, *n): |
| 404 | if self.state != _DID_HEADER: |
| 405 | raise Error, 'Read data at wrong time' |
| 406 | if n: |
| 407 | n = n[0] |
| 408 | n = min(n, self.dlen) |
| 409 | else: |
| 410 | n = self.dlen |
| 411 | self.dlen = self.dlen - n |
| 412 | return self._read(n) |
| 413 | |
| 414 | def close_data(self): |
| 415 | if self.state != _DID_HEADER: |
| 416 | raise Error, 'close_data at wrong time' |
| 417 | if self.dlen: |
| 418 | dummy = self._read(self.dlen) |
| 419 | self._checkcrc() |
| 420 | self.state = _DID_DATA |
| 421 | |
| 422 | def read_rsrc(self, *n): |
| 423 | if self.state == _DID_HEADER: |
| 424 | self.close_data() |
| 425 | if self.state != _DID_DATA: |
| 426 | raise Error, 'Read resource data at wrong time' |
| 427 | if n: |
| 428 | n = n[0] |
| 429 | n = min(n, self.rlen) |
| 430 | else: |
| 431 | n = self.rlen |
| 432 | self.rlen = self.rlen - n |
| 433 | return self._read(n) |
| 434 | |
| 435 | def close(self): |
| 436 | if self.rlen: |
| 437 | dummy = self.read_rsrc(self.rlen) |
| 438 | self._checkcrc() |
| 439 | self.state = _DID_RSRC |
| 440 | self.ifp.close() |
| 441 | |
| 442 | def hexbin(inp, out): |
| 443 | """(infilename, outfilename) - Decode binhexed file""" |
| 444 | ifp = HexBin(inp) |
| 445 | finfo = ifp.FInfo |
| 446 | if not out: |
| 447 | out = ifp.FName |
| 448 | if os.name == 'mac': |
| 449 | ofss = macfs.FSSpec(out) |
| 450 | out = ofss.as_pathname() |
| 451 | |
| 452 | ofp = open(out, 'wb') |
| 453 | # XXXX Do translation on non-mac systems |
| 454 | d = ifp.read() |
| 455 | ofp.write(d) |
| 456 | ofp.close() |
| 457 | ifp.close_data() |
| 458 | |
| 459 | d = ifp.read_rsrc() |
| 460 | if d: |
| 461 | ofp = openrsrc(out, 'wb') |
| 462 | ofp.write(d) |
| 463 | ofp.close() |
| 464 | |
| 465 | if os.name == 'mac': |
| 466 | nfinfo = ofss.GetFInfo() |
| 467 | nfinfo.Creator = finfo.Creator |
| 468 | nfinfo.Type = finfo.Type |
| 469 | nfinfo.Flags = finfo.Flags |
| 470 | ofss.SetFInfo(nfinfo) |
| 471 | |
| 472 | ifp.close() |
| 473 | |
| 474 | def _test(): |
| 475 | if os.name == 'mac': |
Jack Jansen | 479c1b3 | 1995-08-14 12:41:20 +0000 | [diff] [blame] | 476 | fss, ok = macfs.PromptGetFile('File to convert:') |
Jack Jansen | fcdffea | 1995-08-07 14:36:51 +0000 | [diff] [blame] | 477 | if not ok: |
| 478 | sys.exit(0) |
| 479 | fname = fss.as_pathname() |
| 480 | else: |
| 481 | fname = sys.argv[1] |
| 482 | #binhex(fname, fname+'.hqx') |
| 483 | #hexbin(fname+'.hqx', fname+'.viahqx') |
| 484 | hexbin(fname, fname+'.unpacked') |
| 485 | sys.exit(1) |
| 486 | |
| 487 | if __name__ == '__main__': |
| 488 | _test() |
| 489 | |