1"""Macintosh binhex compression/decompression. 2 3easy interface: 4binhex(inputfilename, outputfilename) 5hexbin(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# 24import binascii 25import contextlib 26import io 27import os 28import struct 29import warnings 30 31warnings.warn('the binhex module is deprecated', DeprecationWarning, 32 stacklevel=2) 33 34 35__all__ = ["binhex","hexbin","Error"] 36 37class Error(Exception): 38 pass 39 40# States (what have we written) 41_DID_HEADER = 0 42_DID_DATA = 1 43 44# Various constants 45REASONABLY_LARGE = 32768 # Minimal amount we pass the rle-coder 46LINELEN = 64 47RUNCHAR = b"\x90" 48 49# 50# This code is no longer byte-order dependent 51 52 53class FInfo: 54 def __init__(self): 55 self.Type = '????' 56 self.Creator = '????' 57 self.Flags = 0 58 59def getfileinfo(name): 60 finfo = FInfo() 61 with io.open(name, 'rb') as fp: 62 # Quick check for textfile 63 data = fp.read(512) 64 if 0 not in data: 65 finfo.Type = 'TEXT' 66 fp.seek(0, 2) 67 dsize = fp.tell() 68 dir, file = os.path.split(name) 69 file = file.replace(':', '-', 1) 70 return file, finfo, dsize, 0 71 72class openrsrc: 73 def __init__(self, *args): 74 pass 75 76 def read(self, *args): 77 return b'' 78 79 def write(self, *args): 80 pass 81 82 def close(self): 83 pass 84 85 86# DeprecationWarning is already emitted on "import binhex". There is no need 87# to repeat the warning at each call to deprecated binascii functions. 88@contextlib.contextmanager 89def _ignore_deprecation_warning(): 90 with warnings.catch_warnings(): 91 warnings.filterwarnings('ignore', '', DeprecationWarning) 92 yield 93 94 95class _Hqxcoderengine: 96 """Write data to the coder in 3-byte chunks""" 97 98 def __init__(self, ofp): 99 self.ofp = ofp 100 self.data = b'' 101 self.hqxdata = b'' 102 self.linelen = LINELEN - 1 103 104 def write(self, data): 105 self.data = self.data + data 106 datalen = len(self.data) 107 todo = (datalen // 3) * 3 108 data = self.data[:todo] 109 self.data = self.data[todo:] 110 if not data: 111 return 112 with _ignore_deprecation_warning(): 113 self.hqxdata = self.hqxdata + binascii.b2a_hqx(data) 114 self._flush(0) 115 116 def _flush(self, force): 117 first = 0 118 while first <= len(self.hqxdata) - self.linelen: 119 last = first + self.linelen 120 self.ofp.write(self.hqxdata[first:last] + b'\r') 121 self.linelen = LINELEN 122 first = last 123 self.hqxdata = self.hqxdata[first:] 124 if force: 125 self.ofp.write(self.hqxdata + b':\r') 126 127 def close(self): 128 if self.data: 129 with _ignore_deprecation_warning(): 130 self.hqxdata = self.hqxdata + binascii.b2a_hqx(self.data) 131 self._flush(1) 132 self.ofp.close() 133 del self.ofp 134 135class _Rlecoderengine: 136 """Write data to the RLE-coder in suitably large chunks""" 137 138 def __init__(self, ofp): 139 self.ofp = ofp 140 self.data = b'' 141 142 def write(self, data): 143 self.data = self.data + data 144 if len(self.data) < REASONABLY_LARGE: 145 return 146 with _ignore_deprecation_warning(): 147 rledata = binascii.rlecode_hqx(self.data) 148 self.ofp.write(rledata) 149 self.data = b'' 150 151 def close(self): 152 if self.data: 153 with _ignore_deprecation_warning(): 154 rledata = binascii.rlecode_hqx(self.data) 155 self.ofp.write(rledata) 156 self.ofp.close() 157 del self.ofp 158 159class BinHex: 160 def __init__(self, name_finfo_dlen_rlen, ofp): 161 name, finfo, dlen, rlen = name_finfo_dlen_rlen 162 close_on_error = False 163 if isinstance(ofp, str): 164 ofname = ofp 165 ofp = io.open(ofname, 'wb') 166 close_on_error = True 167 try: 168 ofp.write(b'(This file must be converted with BinHex 4.0)\r\r:') 169 hqxer = _Hqxcoderengine(ofp) 170 self.ofp = _Rlecoderengine(hqxer) 171 self.crc = 0 172 if finfo is None: 173 finfo = FInfo() 174 self.dlen = dlen 175 self.rlen = rlen 176 self._writeinfo(name, finfo) 177 self.state = _DID_HEADER 178 except: 179 if close_on_error: 180 ofp.close() 181 raise 182 183 def _writeinfo(self, name, finfo): 184 nl = len(name) 185 if nl > 63: 186 raise Error('Filename too long') 187 d = bytes([nl]) + name.encode("latin-1") + b'\0' 188 tp, cr = finfo.Type, finfo.Creator 189 if isinstance(tp, str): 190 tp = tp.encode("latin-1") 191 if isinstance(cr, str): 192 cr = cr.encode("latin-1") 193 d2 = tp + cr 194 195 # Force all structs to be packed with big-endian 196 d3 = struct.pack('>h', finfo.Flags) 197 d4 = struct.pack('>ii', self.dlen, self.rlen) 198 info = d + d2 + d3 + d4 199 self._write(info) 200 self._writecrc() 201 202 def _write(self, data): 203 self.crc = binascii.crc_hqx(data, self.crc) 204 self.ofp.write(data) 205 206 def _writecrc(self): 207 # XXXX Should this be here?? 208 # self.crc = binascii.crc_hqx('\0\0', self.crc) 209 if self.crc < 0: 210 fmt = '>h' 211 else: 212 fmt = '>H' 213 self.ofp.write(struct.pack(fmt, self.crc)) 214 self.crc = 0 215 216 def write(self, data): 217 if self.state != _DID_HEADER: 218 raise Error('Writing data at the wrong time') 219 self.dlen = self.dlen - len(data) 220 self._write(data) 221 222 def close_data(self): 223 if self.dlen != 0: 224 raise Error('Incorrect data size, diff=%r' % (self.rlen,)) 225 self._writecrc() 226 self.state = _DID_DATA 227 228 def write_rsrc(self, data): 229 if self.state < _DID_DATA: 230 self.close_data() 231 if self.state != _DID_DATA: 232 raise Error('Writing resource data at the wrong time') 233 self.rlen = self.rlen - len(data) 234 self._write(data) 235 236 def close(self): 237 if self.state is None: 238 return 239 try: 240 if self.state < _DID_DATA: 241 self.close_data() 242 if self.state != _DID_DATA: 243 raise Error('Close at the wrong time') 244 if self.rlen != 0: 245 raise Error("Incorrect resource-datasize, diff=%r" % (self.rlen,)) 246 self._writecrc() 247 finally: 248 self.state = None 249 ofp = self.ofp 250 del self.ofp 251 ofp.close() 252 253def binhex(inp, out): 254 """binhex(infilename, outfilename): create binhex-encoded copy of a file""" 255 finfo = getfileinfo(inp) 256 ofp = BinHex(finfo, out) 257 258 with io.open(inp, 'rb') as ifp: 259 # XXXX Do textfile translation on non-mac systems 260 while True: 261 d = ifp.read(128000) 262 if not d: break 263 ofp.write(d) 264 ofp.close_data() 265 266 ifp = openrsrc(inp, 'rb') 267 while True: 268 d = ifp.read(128000) 269 if not d: break 270 ofp.write_rsrc(d) 271 ofp.close() 272 ifp.close() 273 274class _Hqxdecoderengine: 275 """Read data via the decoder in 4-byte chunks""" 276 277 def __init__(self, ifp): 278 self.ifp = ifp 279 self.eof = 0 280 281 def read(self, totalwtd): 282 """Read at least wtd bytes (or until EOF)""" 283 decdata = b'' 284 wtd = totalwtd 285 # 286 # The loop here is convoluted, since we don't really now how 287 # much to decode: there may be newlines in the incoming data. 288 while wtd > 0: 289 if self.eof: return decdata 290 wtd = ((wtd + 2) // 3) * 4 291 data = self.ifp.read(wtd) 292 # 293 # Next problem: there may not be a complete number of 294 # bytes in what we pass to a2b. Solve by yet another 295 # loop. 296 # 297 while True: 298 try: 299 with _ignore_deprecation_warning(): 300 decdatacur, self.eof = binascii.a2b_hqx(data) 301 break 302 except binascii.Incomplete: 303 pass 304 newdata = self.ifp.read(1) 305 if not newdata: 306 raise Error('Premature EOF on binhex file') 307 data = data + newdata 308 decdata = decdata + decdatacur 309 wtd = totalwtd - len(decdata) 310 if not decdata and not self.eof: 311 raise Error('Premature EOF on binhex file') 312 return decdata 313 314 def close(self): 315 self.ifp.close() 316 317class _Rledecoderengine: 318 """Read data via the RLE-coder""" 319 320 def __init__(self, ifp): 321 self.ifp = ifp 322 self.pre_buffer = b'' 323 self.post_buffer = b'' 324 self.eof = 0 325 326 def read(self, wtd): 327 if wtd > len(self.post_buffer): 328 self._fill(wtd - len(self.post_buffer)) 329 rv = self.post_buffer[:wtd] 330 self.post_buffer = self.post_buffer[wtd:] 331 return rv 332 333 def _fill(self, wtd): 334 self.pre_buffer = self.pre_buffer + self.ifp.read(wtd + 4) 335 if self.ifp.eof: 336 with _ignore_deprecation_warning(): 337 self.post_buffer = self.post_buffer + \ 338 binascii.rledecode_hqx(self.pre_buffer) 339 self.pre_buffer = b'' 340 return 341 342 # 343 # Obfuscated code ahead. We have to take care that we don't 344 # end up with an orphaned RUNCHAR later on. So, we keep a couple 345 # of bytes in the buffer, depending on what the end of 346 # the buffer looks like: 347 # '\220\0\220' - Keep 3 bytes: repeated \220 (escaped as \220\0) 348 # '?\220' - Keep 2 bytes: repeated something-else 349 # '\220\0' - Escaped \220: Keep 2 bytes. 350 # '?\220?' - Complete repeat sequence: decode all 351 # otherwise: keep 1 byte. 352 # 353 mark = len(self.pre_buffer) 354 if self.pre_buffer[-3:] == RUNCHAR + b'\0' + RUNCHAR: 355 mark = mark - 3 356 elif self.pre_buffer[-1:] == RUNCHAR: 357 mark = mark - 2 358 elif self.pre_buffer[-2:] == RUNCHAR + b'\0': 359 mark = mark - 2 360 elif self.pre_buffer[-2:-1] == RUNCHAR: 361 pass # Decode all 362 else: 363 mark = mark - 1 364 365 with _ignore_deprecation_warning(): 366 self.post_buffer = self.post_buffer + \ 367 binascii.rledecode_hqx(self.pre_buffer[:mark]) 368 self.pre_buffer = self.pre_buffer[mark:] 369 370 def close(self): 371 self.ifp.close() 372 373class HexBin: 374 def __init__(self, ifp): 375 if isinstance(ifp, str): 376 ifp = io.open(ifp, 'rb') 377 # 378 # Find initial colon. 379 # 380 while True: 381 ch = ifp.read(1) 382 if not ch: 383 raise Error("No binhex data found") 384 # Cater for \r\n terminated lines (which show up as \n\r, hence 385 # all lines start with \r) 386 if ch == b'\r': 387 continue 388 if ch == b':': 389 break 390 391 hqxifp = _Hqxdecoderengine(ifp) 392 self.ifp = _Rledecoderengine(hqxifp) 393 self.crc = 0 394 self._readheader() 395 396 def _read(self, len): 397 data = self.ifp.read(len) 398 self.crc = binascii.crc_hqx(data, self.crc) 399 return data 400 401 def _checkcrc(self): 402 filecrc = struct.unpack('>h', self.ifp.read(2))[0] & 0xffff 403 #self.crc = binascii.crc_hqx('\0\0', self.crc) 404 # XXXX Is this needed?? 405 self.crc = self.crc & 0xffff 406 if filecrc != self.crc: 407 raise Error('CRC error, computed %x, read %x' 408 % (self.crc, filecrc)) 409 self.crc = 0 410 411 def _readheader(self): 412 len = self._read(1) 413 fname = self._read(ord(len)) 414 rest = self._read(1 + 4 + 4 + 2 + 4 + 4) 415 self._checkcrc() 416 417 type = rest[1:5] 418 creator = rest[5:9] 419 flags = struct.unpack('>h', rest[9:11])[0] 420 self.dlen = struct.unpack('>l', rest[11:15])[0] 421 self.rlen = struct.unpack('>l', rest[15:19])[0] 422 423 self.FName = fname 424 self.FInfo = FInfo() 425 self.FInfo.Creator = creator 426 self.FInfo.Type = type 427 self.FInfo.Flags = flags 428 429 self.state = _DID_HEADER 430 431 def read(self, *n): 432 if self.state != _DID_HEADER: 433 raise Error('Read data at wrong time') 434 if n: 435 n = n[0] 436 n = min(n, self.dlen) 437 else: 438 n = self.dlen 439 rv = b'' 440 while len(rv) < n: 441 rv = rv + self._read(n-len(rv)) 442 self.dlen = self.dlen - n 443 return rv 444 445 def close_data(self): 446 if self.state != _DID_HEADER: 447 raise Error('close_data at wrong time') 448 if self.dlen: 449 dummy = self._read(self.dlen) 450 self._checkcrc() 451 self.state = _DID_DATA 452 453 def read_rsrc(self, *n): 454 if self.state == _DID_HEADER: 455 self.close_data() 456 if self.state != _DID_DATA: 457 raise Error('Read resource data at wrong time') 458 if n: 459 n = n[0] 460 n = min(n, self.rlen) 461 else: 462 n = self.rlen 463 self.rlen = self.rlen - n 464 return self._read(n) 465 466 def close(self): 467 if self.state is None: 468 return 469 try: 470 if self.rlen: 471 dummy = self.read_rsrc(self.rlen) 472 self._checkcrc() 473 finally: 474 self.state = None 475 self.ifp.close() 476 477def hexbin(inp, out): 478 """hexbin(infilename, outfilename) - Decode binhexed file""" 479 ifp = HexBin(inp) 480 finfo = ifp.FInfo 481 if not out: 482 out = ifp.FName 483 484 with io.open(out, 'wb') as ofp: 485 # XXXX Do translation on non-mac systems 486 while True: 487 d = ifp.read(128000) 488 if not d: break 489 ofp.write(d) 490 ifp.close_data() 491 492 d = ifp.read_rsrc(128000) 493 if d: 494 ofp = openrsrc(out, 'wb') 495 ofp.write(d) 496 while True: 497 d = ifp.read_rsrc(128000) 498 if not d: break 499 ofp.write(d) 500 ofp.close() 501 502 ifp.close() 503