Gustavo Niemeyer | f8ca836 | 2002-11-05 16:50:05 +0000 | [diff] [blame^] | 1 | /* |
| 2 | |
| 3 | python-bz2 - python bz2 library interface |
| 4 | |
| 5 | Copyright (c) 2002 Gustavo Niemeyer <niemeyer@conectiva.com> |
| 6 | Copyright (c) 2002 Python Software Foundation; All Rights Reserved |
| 7 | |
| 8 | */ |
| 9 | |
| 10 | #include <stdio.h> |
| 11 | #include <bzlib.h> |
| 12 | #include "Python.h" |
| 13 | #include "structmember.h" |
| 14 | |
| 15 | #ifdef WITH_THREAD |
| 16 | #include "pythread.h" |
| 17 | #endif |
| 18 | |
| 19 | static char __author__[] = |
| 20 | "The bz2 python module was written by:\n\ |
| 21 | \n\ |
| 22 | Gustavo Niemeyer <niemeyer@conectiva.com>\n\ |
| 23 | "; |
| 24 | |
| 25 | #define BUF(v) PyString_AS_STRING((PyStringObject *)v) |
| 26 | |
| 27 | #define MODE_CLOSED 0 |
| 28 | #define MODE_READ 1 |
| 29 | #define MODE_READ_EOF 2 |
| 30 | #define MODE_WRITE 3 |
| 31 | |
| 32 | #define BZ2FileObject_Check(v) ((v)->ob_type == &BZ2File_Type) |
| 33 | |
| 34 | #if SIZEOF_LONG >= 8 |
| 35 | #define BZS_TOTAL_OUT(bzs) \ |
| 36 | (((long)bzs->total_out_hi32 << 32) + bzs->total_out_lo32) |
| 37 | #elif SIZEOF_LONG_LONG >= 8 |
| 38 | #define BZS_TOTAL_OUT(bzs) \ |
| 39 | (((long long)bzs->total_out_hi32 << 32) + bzs->total_out_lo32) |
| 40 | #else |
| 41 | #define BZS_TOTAL_OUT(bzs) \ |
| 42 | bzs->total_out_lo32; |
| 43 | #endif |
| 44 | |
| 45 | #ifdef WITH_THREAD |
| 46 | #define ACQUIRE_LOCK(obj) PyThread_acquire_lock(obj->lock, 1) |
| 47 | #define RELEASE_LOCK(obj) PyThread_release_lock(obj->lock) |
| 48 | #else |
| 49 | #define ACQUIRE_LOCK(obj) |
| 50 | #define RELEASE_LOCK(obj) |
| 51 | #endif |
| 52 | |
| 53 | #ifdef WITH_UNIVERSAL_NEWLINES |
| 54 | /* Bits in f_newlinetypes */ |
| 55 | #define NEWLINE_UNKNOWN 0 /* No newline seen, yet */ |
| 56 | #define NEWLINE_CR 1 /* \r newline seen */ |
| 57 | #define NEWLINE_LF 2 /* \n newline seen */ |
| 58 | #define NEWLINE_CRLF 4 /* \r\n newline seen */ |
| 59 | #endif |
| 60 | |
| 61 | /* ===================================================================== */ |
| 62 | /* Structure definitions. */ |
| 63 | |
| 64 | typedef struct { |
| 65 | PyFileObject file; |
| 66 | BZFILE *fp; |
| 67 | int mode; |
| 68 | long pos; |
| 69 | long size; |
| 70 | #ifdef WITH_THREAD |
| 71 | PyThread_type_lock lock; |
| 72 | #endif |
| 73 | } BZ2FileObject; |
| 74 | |
| 75 | typedef struct { |
| 76 | PyObject_HEAD |
| 77 | bz_stream bzs; |
| 78 | int running; |
| 79 | #ifdef WITH_THREAD |
| 80 | PyThread_type_lock lock; |
| 81 | #endif |
| 82 | } BZ2CompObject; |
| 83 | |
| 84 | typedef struct { |
| 85 | PyObject_HEAD |
| 86 | bz_stream bzs; |
| 87 | int running; |
| 88 | PyObject *unused_data; |
| 89 | #ifdef WITH_THREAD |
| 90 | PyThread_type_lock lock; |
| 91 | #endif |
| 92 | } BZ2DecompObject; |
| 93 | |
| 94 | /* ===================================================================== */ |
| 95 | /* Utility functions. */ |
| 96 | |
| 97 | static int |
| 98 | Util_CatchBZ2Error(int bzerror) |
| 99 | { |
| 100 | int ret = 0; |
| 101 | switch(bzerror) { |
| 102 | case BZ_OK: |
| 103 | case BZ_STREAM_END: |
| 104 | break; |
| 105 | |
| 106 | case BZ_CONFIG_ERROR: |
| 107 | PyErr_SetString(PyExc_SystemError, |
| 108 | "the bz2 library was not compiled " |
| 109 | "correctly"); |
| 110 | ret = 1; |
| 111 | break; |
| 112 | |
| 113 | case BZ_PARAM_ERROR: |
| 114 | PyErr_SetString(PyExc_ValueError, |
| 115 | "the bz2 library has received wrong " |
| 116 | "parameters"); |
| 117 | ret = 1; |
| 118 | break; |
| 119 | |
| 120 | case BZ_MEM_ERROR: |
| 121 | PyErr_NoMemory(); |
| 122 | ret = 1; |
| 123 | break; |
| 124 | |
| 125 | case BZ_DATA_ERROR: |
| 126 | case BZ_DATA_ERROR_MAGIC: |
| 127 | PyErr_SetString(PyExc_IOError, "invalid data stream"); |
| 128 | ret = 1; |
| 129 | break; |
| 130 | |
| 131 | case BZ_IO_ERROR: |
| 132 | PyErr_SetString(PyExc_IOError, "unknown IO error"); |
| 133 | ret = 1; |
| 134 | break; |
| 135 | |
| 136 | case BZ_UNEXPECTED_EOF: |
| 137 | PyErr_SetString(PyExc_EOFError, |
| 138 | "compressed file ended before the " |
| 139 | "logical end-of-stream was detected"); |
| 140 | ret = 1; |
| 141 | break; |
| 142 | |
| 143 | case BZ_SEQUENCE_ERROR: |
| 144 | PyErr_SetString(PyExc_RuntimeError, |
| 145 | "wrong sequence of bz2 library " |
| 146 | "commands used"); |
| 147 | ret = 1; |
| 148 | break; |
| 149 | } |
| 150 | return ret; |
| 151 | } |
| 152 | |
| 153 | #if BUFSIZ < 8192 |
| 154 | #define SMALLCHUNK 8192 |
| 155 | #else |
| 156 | #define SMALLCHUNK BUFSIZ |
| 157 | #endif |
| 158 | |
| 159 | #if SIZEOF_INT < 4 |
| 160 | #define BIGCHUNK (512 * 32) |
| 161 | #else |
| 162 | #define BIGCHUNK (512 * 1024) |
| 163 | #endif |
| 164 | |
| 165 | /* This is a hacked version of Python's fileobject.c:new_buffersize(). */ |
| 166 | static size_t |
| 167 | Util_NewBufferSize(size_t currentsize) |
| 168 | { |
| 169 | if (currentsize > SMALLCHUNK) { |
| 170 | /* Keep doubling until we reach BIGCHUNK; |
| 171 | then keep adding BIGCHUNK. */ |
| 172 | if (currentsize <= BIGCHUNK) |
| 173 | return currentsize + currentsize; |
| 174 | else |
| 175 | return currentsize + BIGCHUNK; |
| 176 | } |
| 177 | return currentsize + SMALLCHUNK; |
| 178 | } |
| 179 | |
| 180 | /* This is a hacked version of Python's fileobject.c:get_line(). */ |
| 181 | static PyObject * |
| 182 | Util_GetLine(BZ2FileObject *self, int n) |
| 183 | { |
| 184 | char c; |
| 185 | char *buf, *end; |
| 186 | size_t total_v_size; /* total # of slots in buffer */ |
| 187 | size_t used_v_size; /* # used slots in buffer */ |
| 188 | size_t increment; /* amount to increment the buffer */ |
| 189 | PyObject *v; |
| 190 | int bzerror; |
| 191 | #ifdef WITH_UNIVERSAL_NEWLINES |
| 192 | int newlinetypes = ((PyFileObject*)self)->f_newlinetypes; |
| 193 | int skipnextlf = ((PyFileObject*)self)->f_skipnextlf; |
| 194 | int univ_newline = ((PyFileObject*)self)->f_univ_newline; |
| 195 | #endif |
| 196 | |
| 197 | total_v_size = n > 0 ? n : 100; |
| 198 | v = PyString_FromStringAndSize((char *)NULL, total_v_size); |
| 199 | if (v == NULL) |
| 200 | return NULL; |
| 201 | |
| 202 | buf = BUF(v); |
| 203 | end = buf + total_v_size; |
| 204 | |
| 205 | for (;;) { |
| 206 | Py_BEGIN_ALLOW_THREADS |
| 207 | #ifdef WITH_UNIVERSAL_NEWLINES |
| 208 | if (univ_newline) { |
| 209 | while (1) { |
| 210 | BZ2_bzRead(&bzerror, self->fp, &c, 1); |
| 211 | self->pos++; |
| 212 | if (bzerror != BZ_OK || buf == end) |
| 213 | break; |
| 214 | if (skipnextlf ) { |
| 215 | skipnextlf = 0; |
| 216 | if (c == '\n') { |
| 217 | /* Seeing a \n here with |
| 218 | * skipnextlf true means we |
| 219 | * saw a \r before. |
| 220 | */ |
| 221 | newlinetypes |= NEWLINE_CRLF; |
| 222 | BZ2_bzRead(&bzerror, self->fp, |
| 223 | &c, 1); |
| 224 | if (bzerror != BZ_OK) |
| 225 | break; |
| 226 | } else { |
| 227 | newlinetypes |= NEWLINE_CR; |
| 228 | } |
| 229 | } |
| 230 | if (c == '\r') { |
| 231 | skipnextlf = 1; |
| 232 | c = '\n'; |
| 233 | } else if ( c == '\n') |
| 234 | newlinetypes |= NEWLINE_LF; |
| 235 | *buf++ = c; |
| 236 | if (c == '\n') break; |
| 237 | } |
| 238 | if (bzerror == BZ_STREAM_END && skipnextlf) |
| 239 | newlinetypes |= NEWLINE_CR; |
| 240 | } else /* If not universal newlines use the normal loop */ |
| 241 | #endif |
| 242 | do { |
| 243 | BZ2_bzRead(&bzerror, self->fp, &c, 1); |
| 244 | self->pos++; |
| 245 | *buf++ = c; |
| 246 | } while (bzerror == BZ_OK && c != '\n' && buf != end); |
| 247 | Py_END_ALLOW_THREADS |
| 248 | #ifdef WITH_UNIVERSAL_NEWLINES |
| 249 | ((PyFileObject*)self)->f_newlinetypes = newlinetypes; |
| 250 | ((PyFileObject*)self)->f_skipnextlf = skipnextlf; |
| 251 | #endif |
| 252 | if (bzerror == BZ_STREAM_END) { |
| 253 | self->size = self->pos; |
| 254 | self->mode = MODE_READ_EOF; |
| 255 | break; |
| 256 | } else if (bzerror != BZ_OK) { |
| 257 | Util_CatchBZ2Error(bzerror); |
| 258 | Py_DECREF(v); |
| 259 | return NULL; |
| 260 | } |
| 261 | if (c == '\n') |
| 262 | break; |
| 263 | /* Must be because buf == end */ |
| 264 | if (n > 0) |
| 265 | break; |
| 266 | used_v_size = total_v_size; |
| 267 | increment = total_v_size >> 2; /* mild exponential growth */ |
| 268 | total_v_size += increment; |
| 269 | if (total_v_size > INT_MAX) { |
| 270 | PyErr_SetString(PyExc_OverflowError, |
| 271 | "line is longer than a Python string can hold"); |
| 272 | Py_DECREF(v); |
| 273 | return NULL; |
| 274 | } |
| 275 | if (_PyString_Resize(&v, total_v_size) < 0) |
| 276 | return NULL; |
| 277 | buf = BUF(v) + used_v_size; |
| 278 | end = BUF(v) + total_v_size; |
| 279 | } |
| 280 | |
| 281 | used_v_size = buf - BUF(v); |
| 282 | if (used_v_size != total_v_size) |
| 283 | _PyString_Resize(&v, used_v_size); |
| 284 | return v; |
| 285 | } |
| 286 | |
| 287 | #ifndef WITH_UNIVERSAL_NEWLINES |
| 288 | #define Util_UnivNewlineRead(a,b,c,d,e) BZ2_bzRead(a,b,c,d) |
| 289 | #else |
| 290 | /* This is a hacked version of Python's |
| 291 | * fileobject.c:Py_UniversalNewlineFread(). */ |
| 292 | size_t |
| 293 | Util_UnivNewlineRead(int *bzerror, BZFILE *stream, |
| 294 | char* buf, size_t n, BZ2FileObject *fobj) |
| 295 | { |
| 296 | char *dst = buf; |
| 297 | PyFileObject *f = (PyFileObject *)fobj; |
| 298 | int newlinetypes, skipnextlf; |
| 299 | |
| 300 | assert(buf != NULL); |
| 301 | assert(stream != NULL); |
| 302 | |
| 303 | if (!f->f_univ_newline) |
| 304 | return BZ2_bzRead(bzerror, stream, buf, n); |
| 305 | |
| 306 | newlinetypes = f->f_newlinetypes; |
| 307 | skipnextlf = f->f_skipnextlf; |
| 308 | |
| 309 | /* Invariant: n is the number of bytes remaining to be filled |
| 310 | * in the buffer. |
| 311 | */ |
| 312 | while (n) { |
| 313 | size_t nread; |
| 314 | int shortread; |
| 315 | char *src = dst; |
| 316 | |
| 317 | nread = BZ2_bzRead(bzerror, stream, dst, n); |
| 318 | assert(nread <= n); |
| 319 | n -= nread; /* assuming 1 byte out for each in; will adjust */ |
| 320 | shortread = n != 0; /* true iff EOF or error */ |
| 321 | while (nread--) { |
| 322 | char c = *src++; |
| 323 | if (c == '\r') { |
| 324 | /* Save as LF and set flag to skip next LF. */ |
| 325 | *dst++ = '\n'; |
| 326 | skipnextlf = 1; |
| 327 | } |
| 328 | else if (skipnextlf && c == '\n') { |
| 329 | /* Skip LF, and remember we saw CR LF. */ |
| 330 | skipnextlf = 0; |
| 331 | newlinetypes |= NEWLINE_CRLF; |
| 332 | ++n; |
| 333 | } |
| 334 | else { |
| 335 | /* Normal char to be stored in buffer. Also |
| 336 | * update the newlinetypes flag if either this |
| 337 | * is an LF or the previous char was a CR. |
| 338 | */ |
| 339 | if (c == '\n') |
| 340 | newlinetypes |= NEWLINE_LF; |
| 341 | else if (skipnextlf) |
| 342 | newlinetypes |= NEWLINE_CR; |
| 343 | *dst++ = c; |
| 344 | skipnextlf = 0; |
| 345 | } |
| 346 | } |
| 347 | if (shortread) { |
| 348 | /* If this is EOF, update type flags. */ |
| 349 | if (skipnextlf && *bzerror == BZ_STREAM_END) |
| 350 | newlinetypes |= NEWLINE_CR; |
| 351 | break; |
| 352 | } |
| 353 | } |
| 354 | f->f_newlinetypes = newlinetypes; |
| 355 | f->f_skipnextlf = skipnextlf; |
| 356 | return dst - buf; |
| 357 | } |
| 358 | #endif |
| 359 | |
| 360 | /* This is a hacked version of Python's fileobject.c:drop_readahead(). */ |
| 361 | static void |
| 362 | Util_DropReadAhead(BZ2FileObject *self) |
| 363 | { |
| 364 | PyFileObject *f = (PyFileObject*)self; |
| 365 | if (f->f_buf != NULL) { |
| 366 | PyMem_Free(f->f_buf); |
| 367 | f->f_buf = NULL; |
| 368 | } |
| 369 | } |
| 370 | |
| 371 | /* This is a hacked version of Python's fileobject.c:readahead(). */ |
| 372 | static int |
| 373 | Util_ReadAhead(BZ2FileObject *self, int bufsize) |
| 374 | { |
| 375 | int chunksize; |
| 376 | int bzerror; |
| 377 | PyFileObject *f = (PyFileObject*)self; |
| 378 | |
| 379 | if (f->f_buf != NULL) { |
| 380 | if((f->f_bufend - f->f_bufptr) >= 1) |
| 381 | return 0; |
| 382 | else |
| 383 | Util_DropReadAhead(self); |
| 384 | } |
| 385 | if (self->mode == MODE_READ_EOF) { |
| 386 | return -1; |
| 387 | } |
| 388 | if ((f->f_buf = PyMem_Malloc(bufsize)) == NULL) { |
| 389 | return -1; |
| 390 | } |
| 391 | Py_BEGIN_ALLOW_THREADS |
| 392 | chunksize = Util_UnivNewlineRead(&bzerror, self->fp, f->f_buf, |
| 393 | bufsize, self); |
| 394 | Py_END_ALLOW_THREADS |
| 395 | self->pos += chunksize; |
| 396 | if (bzerror == BZ_STREAM_END) { |
| 397 | self->size = self->pos; |
| 398 | self->mode = MODE_READ_EOF; |
| 399 | } else if (bzerror != BZ_OK) { |
| 400 | Util_CatchBZ2Error(bzerror); |
| 401 | Util_DropReadAhead(self); |
| 402 | return -1; |
| 403 | } |
| 404 | f->f_bufptr = f->f_buf; |
| 405 | f->f_bufend = f->f_buf + chunksize; |
| 406 | return 0; |
| 407 | } |
| 408 | |
| 409 | /* This is a hacked version of Python's |
| 410 | * fileobject.c:readahead_get_line_skip(). */ |
| 411 | static PyStringObject * |
| 412 | Util_ReadAheadGetLineSkip(BZ2FileObject *bf, int skip, int bufsize) |
| 413 | { |
| 414 | PyFileObject *f = (PyFileObject*)bf; |
| 415 | PyStringObject* s; |
| 416 | char *bufptr; |
| 417 | char *buf; |
| 418 | int len; |
| 419 | |
| 420 | if (f->f_buf == NULL) |
| 421 | if (Util_ReadAhead(bf, bufsize) < 0) |
| 422 | return NULL; |
| 423 | |
| 424 | len = f->f_bufend - f->f_bufptr; |
| 425 | if (len == 0) |
| 426 | return (PyStringObject *) |
| 427 | PyString_FromStringAndSize(NULL, skip); |
| 428 | bufptr = memchr(f->f_bufptr, '\n', len); |
| 429 | if (bufptr != NULL) { |
| 430 | bufptr++; /* Count the '\n' */ |
| 431 | len = bufptr - f->f_bufptr; |
| 432 | s = (PyStringObject *) |
| 433 | PyString_FromStringAndSize(NULL, skip+len); |
| 434 | if (s == NULL) |
| 435 | return NULL; |
| 436 | memcpy(PyString_AS_STRING(s)+skip, f->f_bufptr, len); |
| 437 | f->f_bufptr = bufptr; |
| 438 | if (bufptr == f->f_bufend) |
| 439 | Util_DropReadAhead(bf); |
| 440 | } else { |
| 441 | bufptr = f->f_bufptr; |
| 442 | buf = f->f_buf; |
| 443 | f->f_buf = NULL; /* Force new readahead buffer */ |
| 444 | s = Util_ReadAheadGetLineSkip( |
| 445 | bf, skip+len, bufsize + (bufsize>>2) ); |
| 446 | if (s == NULL) { |
| 447 | PyMem_Free(buf); |
| 448 | return NULL; |
| 449 | } |
| 450 | memcpy(PyString_AS_STRING(s)+skip, bufptr, len); |
| 451 | PyMem_Free(buf); |
| 452 | } |
| 453 | return s; |
| 454 | } |
| 455 | |
| 456 | /* ===================================================================== */ |
| 457 | /* Methods of BZ2File. */ |
| 458 | |
| 459 | PyDoc_STRVAR(BZ2File_read__doc__, |
| 460 | "read([size]) -> string\n\ |
| 461 | \n\ |
| 462 | Read at most size uncompressed bytes, returned as a string. If the size\n\ |
| 463 | argument is negative or omitted, read until EOF is reached.\n\ |
| 464 | "); |
| 465 | |
| 466 | /* This is a hacked version of Python's fileobject.c:file_read(). */ |
| 467 | static PyObject * |
| 468 | BZ2File_read(BZ2FileObject *self, PyObject *args) |
| 469 | { |
| 470 | long bytesrequested = -1; |
| 471 | size_t bytesread, buffersize, chunksize; |
| 472 | int bzerror; |
| 473 | PyObject *ret = NULL; |
| 474 | |
| 475 | if (!PyArg_ParseTuple(args, "|l:read", &bytesrequested)) |
| 476 | return NULL; |
| 477 | |
| 478 | ACQUIRE_LOCK(self); |
| 479 | switch (self->mode) { |
| 480 | case MODE_READ: |
| 481 | break; |
| 482 | case MODE_READ_EOF: |
| 483 | ret = PyString_FromString(""); |
| 484 | goto cleanup; |
| 485 | case MODE_CLOSED: |
| 486 | PyErr_SetString(PyExc_ValueError, |
| 487 | "I/O operation on closed file"); |
| 488 | goto cleanup; |
| 489 | default: |
| 490 | PyErr_SetString(PyExc_IOError, |
| 491 | "file is not ready for reading"); |
| 492 | goto cleanup; |
| 493 | } |
| 494 | |
| 495 | if (bytesrequested < 0) |
| 496 | buffersize = Util_NewBufferSize((size_t)0); |
| 497 | else |
| 498 | buffersize = bytesrequested; |
| 499 | if (buffersize > INT_MAX) { |
| 500 | PyErr_SetString(PyExc_OverflowError, |
| 501 | "requested number of bytes is more than a Python string can hold"); |
| 502 | goto cleanup; |
| 503 | } |
| 504 | ret = PyString_FromStringAndSize((char *)NULL, buffersize); |
| 505 | if (ret == NULL) |
| 506 | goto cleanup; |
| 507 | bytesread = 0; |
| 508 | |
| 509 | for (;;) { |
| 510 | Py_BEGIN_ALLOW_THREADS |
| 511 | chunksize = Util_UnivNewlineRead(&bzerror, self->fp, |
| 512 | BUF(ret)+bytesread, |
| 513 | buffersize-bytesread, |
| 514 | self); |
| 515 | self->pos += chunksize; |
| 516 | Py_END_ALLOW_THREADS |
| 517 | bytesread += chunksize; |
| 518 | if (bzerror == BZ_STREAM_END) { |
| 519 | self->size = self->pos; |
| 520 | self->mode = MODE_READ_EOF; |
| 521 | break; |
| 522 | } else if (bzerror != BZ_OK) { |
| 523 | Util_CatchBZ2Error(bzerror); |
| 524 | Py_DECREF(ret); |
| 525 | ret = NULL; |
| 526 | goto cleanup; |
| 527 | } |
| 528 | if (bytesrequested < 0) { |
| 529 | buffersize = Util_NewBufferSize(buffersize); |
| 530 | if (_PyString_Resize(&ret, buffersize) < 0) |
| 531 | goto cleanup; |
| 532 | } else { |
| 533 | break; |
| 534 | } |
| 535 | } |
| 536 | if (bytesread != buffersize) |
| 537 | _PyString_Resize(&ret, bytesread); |
| 538 | |
| 539 | cleanup: |
| 540 | RELEASE_LOCK(self); |
| 541 | return ret; |
| 542 | } |
| 543 | |
| 544 | PyDoc_STRVAR(BZ2File_readline__doc__, |
| 545 | "readline([size]) -> string\n\ |
| 546 | \n\ |
| 547 | Return the next line from the file, as a string, retaining newline.\n\ |
| 548 | A non-negative size argument will limit the maximum number of bytes to\n\ |
| 549 | return (an incomplete line may be returned then). Return an empty\n\ |
| 550 | string at EOF.\n\ |
| 551 | "); |
| 552 | |
| 553 | static PyObject * |
| 554 | BZ2File_readline(BZ2FileObject *self, PyObject *args) |
| 555 | { |
| 556 | PyObject *ret = NULL; |
| 557 | int sizehint = -1; |
| 558 | |
| 559 | if (!PyArg_ParseTuple(args, "|i:readline", &sizehint)) |
| 560 | return NULL; |
| 561 | |
| 562 | ACQUIRE_LOCK(self); |
| 563 | switch (self->mode) { |
| 564 | case MODE_READ: |
| 565 | break; |
| 566 | case MODE_READ_EOF: |
| 567 | ret = PyString_FromString(""); |
| 568 | goto cleanup; |
| 569 | case MODE_CLOSED: |
| 570 | PyErr_SetString(PyExc_ValueError, |
| 571 | "I/O operation on closed file"); |
| 572 | goto cleanup; |
| 573 | default: |
| 574 | PyErr_SetString(PyExc_IOError, |
| 575 | "file is not ready for reading"); |
| 576 | goto cleanup; |
| 577 | } |
| 578 | |
| 579 | if (sizehint == 0) |
| 580 | ret = PyString_FromString(""); |
| 581 | else |
| 582 | ret = Util_GetLine(self, (sizehint < 0) ? 0 : sizehint); |
| 583 | |
| 584 | cleanup: |
| 585 | RELEASE_LOCK(self); |
| 586 | return ret; |
| 587 | } |
| 588 | |
| 589 | PyDoc_STRVAR(BZ2File_readlines__doc__, |
| 590 | "readlines([size]) -> list\n\ |
| 591 | \n\ |
| 592 | Call readline() repeatedly and return a list of lines read.\n\ |
| 593 | The optional size argument, if given, is an approximate bound on the\n\ |
| 594 | total number of bytes in the lines returned.\n\ |
| 595 | "); |
| 596 | |
| 597 | /* This is a hacked version of Python's fileobject.c:file_readlines(). */ |
| 598 | static PyObject * |
| 599 | BZ2File_readlines(BZ2FileObject *self, PyObject *args) |
| 600 | { |
| 601 | long sizehint = 0; |
| 602 | PyObject *list = NULL; |
| 603 | PyObject *line; |
| 604 | char small_buffer[SMALLCHUNK]; |
| 605 | char *buffer = small_buffer; |
| 606 | size_t buffersize = SMALLCHUNK; |
| 607 | PyObject *big_buffer = NULL; |
| 608 | size_t nfilled = 0; |
| 609 | size_t nread; |
| 610 | size_t totalread = 0; |
| 611 | char *p, *q, *end; |
| 612 | int err; |
| 613 | int shortread = 0; |
| 614 | int bzerror; |
| 615 | |
| 616 | if (!PyArg_ParseTuple(args, "|l:readlines", &sizehint)) |
| 617 | return NULL; |
| 618 | |
| 619 | ACQUIRE_LOCK(self); |
| 620 | switch (self->mode) { |
| 621 | case MODE_READ: |
| 622 | break; |
| 623 | case MODE_READ_EOF: |
| 624 | list = PyList_New(0); |
| 625 | goto cleanup; |
| 626 | case MODE_CLOSED: |
| 627 | PyErr_SetString(PyExc_ValueError, |
| 628 | "I/O operation on closed file"); |
| 629 | goto cleanup; |
| 630 | default: |
| 631 | PyErr_SetString(PyExc_IOError, |
| 632 | "file is not ready for reading"); |
| 633 | goto cleanup; |
| 634 | } |
| 635 | |
| 636 | if ((list = PyList_New(0)) == NULL) |
| 637 | goto cleanup; |
| 638 | |
| 639 | for (;;) { |
| 640 | Py_BEGIN_ALLOW_THREADS |
| 641 | nread = Util_UnivNewlineRead(&bzerror, self->fp, |
| 642 | buffer+nfilled, |
| 643 | buffersize-nfilled, self); |
| 644 | self->pos += nread; |
| 645 | Py_END_ALLOW_THREADS |
| 646 | if (bzerror == BZ_STREAM_END) { |
| 647 | self->size = self->pos; |
| 648 | self->mode = MODE_READ_EOF; |
| 649 | if (nread == 0) { |
| 650 | sizehint = 0; |
| 651 | break; |
| 652 | } |
| 653 | shortread = 1; |
| 654 | } else if (bzerror != BZ_OK) { |
| 655 | Util_CatchBZ2Error(bzerror); |
| 656 | error: |
| 657 | Py_DECREF(list); |
| 658 | list = NULL; |
| 659 | goto cleanup; |
| 660 | } |
| 661 | totalread += nread; |
| 662 | p = memchr(buffer+nfilled, '\n', nread); |
| 663 | if (p == NULL) { |
| 664 | /* Need a larger buffer to fit this line */ |
| 665 | nfilled += nread; |
| 666 | buffersize *= 2; |
| 667 | if (buffersize > INT_MAX) { |
| 668 | PyErr_SetString(PyExc_OverflowError, |
| 669 | "line is longer than a Python string can hold"); |
| 670 | goto error; |
| 671 | } |
| 672 | if (big_buffer == NULL) { |
| 673 | /* Create the big buffer */ |
| 674 | big_buffer = PyString_FromStringAndSize( |
| 675 | NULL, buffersize); |
| 676 | if (big_buffer == NULL) |
| 677 | goto error; |
| 678 | buffer = PyString_AS_STRING(big_buffer); |
| 679 | memcpy(buffer, small_buffer, nfilled); |
| 680 | } |
| 681 | else { |
| 682 | /* Grow the big buffer */ |
| 683 | _PyString_Resize(&big_buffer, buffersize); |
| 684 | buffer = PyString_AS_STRING(big_buffer); |
| 685 | } |
| 686 | continue; |
| 687 | } |
| 688 | end = buffer+nfilled+nread; |
| 689 | q = buffer; |
| 690 | do { |
| 691 | /* Process complete lines */ |
| 692 | p++; |
| 693 | line = PyString_FromStringAndSize(q, p-q); |
| 694 | if (line == NULL) |
| 695 | goto error; |
| 696 | err = PyList_Append(list, line); |
| 697 | Py_DECREF(line); |
| 698 | if (err != 0) |
| 699 | goto error; |
| 700 | q = p; |
| 701 | p = memchr(q, '\n', end-q); |
| 702 | } while (p != NULL); |
| 703 | /* Move the remaining incomplete line to the start */ |
| 704 | nfilled = end-q; |
| 705 | memmove(buffer, q, nfilled); |
| 706 | if (sizehint > 0) |
| 707 | if (totalread >= (size_t)sizehint) |
| 708 | break; |
| 709 | if (shortread) { |
| 710 | sizehint = 0; |
| 711 | break; |
| 712 | } |
| 713 | } |
| 714 | if (nfilled != 0) { |
| 715 | /* Partial last line */ |
| 716 | line = PyString_FromStringAndSize(buffer, nfilled); |
| 717 | if (line == NULL) |
| 718 | goto error; |
| 719 | if (sizehint > 0) { |
| 720 | /* Need to complete the last line */ |
| 721 | PyObject *rest = Util_GetLine(self, 0); |
| 722 | if (rest == NULL) { |
| 723 | Py_DECREF(line); |
| 724 | goto error; |
| 725 | } |
| 726 | PyString_Concat(&line, rest); |
| 727 | Py_DECREF(rest); |
| 728 | if (line == NULL) |
| 729 | goto error; |
| 730 | } |
| 731 | err = PyList_Append(list, line); |
| 732 | Py_DECREF(line); |
| 733 | if (err != 0) |
| 734 | goto error; |
| 735 | } |
| 736 | |
| 737 | cleanup: |
| 738 | RELEASE_LOCK(self); |
| 739 | if (big_buffer) { |
| 740 | Py_DECREF(big_buffer); |
| 741 | } |
| 742 | return list; |
| 743 | } |
| 744 | |
| 745 | PyDoc_STRVAR(BZ2File_write__doc__, |
| 746 | "write(data) -> None\n\ |
| 747 | \n\ |
| 748 | Write the 'data' string to file. Note that due to buffering, close() may\n\ |
| 749 | be needed before the file on disk reflects the data written.\n\ |
| 750 | "); |
| 751 | |
| 752 | /* This is a hacked version of Python's fileobject.c:file_write(). */ |
| 753 | static PyObject * |
| 754 | BZ2File_write(BZ2FileObject *self, PyObject *args) |
| 755 | { |
| 756 | PyObject *ret = NULL; |
| 757 | char *buf; |
| 758 | int len; |
| 759 | int bzerror; |
| 760 | |
| 761 | if (!PyArg_ParseTuple(args, "s#", &buf, &len)) |
| 762 | return NULL; |
| 763 | |
| 764 | ACQUIRE_LOCK(self); |
| 765 | switch (self->mode) { |
| 766 | case MODE_WRITE: |
| 767 | break; |
| 768 | |
| 769 | case MODE_CLOSED: |
| 770 | PyErr_SetString(PyExc_ValueError, |
| 771 | "I/O operation on closed file"); |
| 772 | goto cleanup;; |
| 773 | |
| 774 | default: |
| 775 | PyErr_SetString(PyExc_IOError, |
| 776 | "file is not ready for writing"); |
| 777 | goto cleanup;; |
| 778 | } |
| 779 | |
| 780 | PyFile_SoftSpace((PyObject*)self, 0); |
| 781 | |
| 782 | Py_BEGIN_ALLOW_THREADS |
| 783 | BZ2_bzWrite (&bzerror, self->fp, buf, len); |
| 784 | self->pos += len; |
| 785 | Py_END_ALLOW_THREADS |
| 786 | |
| 787 | if (bzerror != BZ_OK) { |
| 788 | Util_CatchBZ2Error(bzerror); |
| 789 | goto cleanup; |
| 790 | } |
| 791 | |
| 792 | Py_INCREF(Py_None); |
| 793 | ret = Py_None; |
| 794 | |
| 795 | cleanup: |
| 796 | RELEASE_LOCK(self); |
| 797 | return ret; |
| 798 | } |
| 799 | |
| 800 | PyDoc_STRVAR(BZ2File_writelines__doc__, |
| 801 | "writelines(sequence_of_strings) -> None\n\ |
| 802 | \n\ |
| 803 | Write the sequence of strings to the file. Note that newlines are not\n\ |
| 804 | added. The sequence can be any iterable object producing strings. This is\n\ |
| 805 | equivalent to calling write() for each string.\n\ |
| 806 | "); |
| 807 | |
| 808 | /* This is a hacked version of Python's fileobject.c:file_writelines(). */ |
| 809 | static PyObject * |
| 810 | BZ2File_writelines(BZ2FileObject *self, PyObject *seq) |
| 811 | { |
| 812 | #define CHUNKSIZE 1000 |
| 813 | PyObject *list = NULL; |
| 814 | PyObject *iter = NULL; |
| 815 | PyObject *ret = NULL; |
| 816 | PyObject *line; |
| 817 | int i, j, index, len, islist; |
| 818 | int bzerror; |
| 819 | |
| 820 | ACQUIRE_LOCK(self); |
| 821 | islist = PyList_Check(seq); |
| 822 | if (!islist) { |
| 823 | iter = PyObject_GetIter(seq); |
| 824 | if (iter == NULL) { |
| 825 | PyErr_SetString(PyExc_TypeError, |
| 826 | "writelines() requires an iterable argument"); |
| 827 | goto error; |
| 828 | } |
| 829 | list = PyList_New(CHUNKSIZE); |
| 830 | if (list == NULL) |
| 831 | goto error; |
| 832 | } |
| 833 | |
| 834 | /* Strategy: slurp CHUNKSIZE lines into a private list, |
| 835 | checking that they are all strings, then write that list |
| 836 | without holding the interpreter lock, then come back for more. */ |
| 837 | for (index = 0; ; index += CHUNKSIZE) { |
| 838 | if (islist) { |
| 839 | Py_XDECREF(list); |
| 840 | list = PyList_GetSlice(seq, index, index+CHUNKSIZE); |
| 841 | if (list == NULL) |
| 842 | goto error; |
| 843 | j = PyList_GET_SIZE(list); |
| 844 | } |
| 845 | else { |
| 846 | for (j = 0; j < CHUNKSIZE; j++) { |
| 847 | line = PyIter_Next(iter); |
| 848 | if (line == NULL) { |
| 849 | if (PyErr_Occurred()) |
| 850 | goto error; |
| 851 | break; |
| 852 | } |
| 853 | PyList_SetItem(list, j, line); |
| 854 | } |
| 855 | } |
| 856 | if (j == 0) |
| 857 | break; |
| 858 | |
| 859 | /* Check that all entries are indeed strings. If not, |
| 860 | apply the same rules as for file.write() and |
| 861 | convert the rets to strings. This is slow, but |
| 862 | seems to be the only way since all conversion APIs |
| 863 | could potentially execute Python code. */ |
| 864 | for (i = 0; i < j; i++) { |
| 865 | PyObject *v = PyList_GET_ITEM(list, i); |
| 866 | if (!PyString_Check(v)) { |
| 867 | const char *buffer; |
| 868 | int len; |
| 869 | if (PyObject_AsCharBuffer(v, &buffer, &len)) { |
| 870 | PyErr_SetString(PyExc_TypeError, |
| 871 | "writelines() " |
| 872 | "argument must be " |
| 873 | "a sequence of " |
| 874 | "strings"); |
| 875 | goto error; |
| 876 | } |
| 877 | line = PyString_FromStringAndSize(buffer, |
| 878 | len); |
| 879 | if (line == NULL) |
| 880 | goto error; |
| 881 | Py_DECREF(v); |
| 882 | PyList_SET_ITEM(list, i, line); |
| 883 | } |
| 884 | } |
| 885 | |
| 886 | PyFile_SoftSpace((PyObject*)self, 0); |
| 887 | |
| 888 | /* Since we are releasing the global lock, the |
| 889 | following code may *not* execute Python code. */ |
| 890 | Py_BEGIN_ALLOW_THREADS |
| 891 | for (i = 0; i < j; i++) { |
| 892 | line = PyList_GET_ITEM(list, i); |
| 893 | len = PyString_GET_SIZE(line); |
| 894 | BZ2_bzWrite (&bzerror, self->fp, |
| 895 | PyString_AS_STRING(line), len); |
| 896 | if (bzerror != BZ_OK) { |
| 897 | Py_BLOCK_THREADS |
| 898 | Util_CatchBZ2Error(bzerror); |
| 899 | goto error; |
| 900 | } |
| 901 | } |
| 902 | Py_END_ALLOW_THREADS |
| 903 | |
| 904 | if (j < CHUNKSIZE) |
| 905 | break; |
| 906 | } |
| 907 | |
| 908 | Py_INCREF(Py_None); |
| 909 | ret = Py_None; |
| 910 | |
| 911 | error: |
| 912 | RELEASE_LOCK(self); |
| 913 | Py_XDECREF(list); |
| 914 | Py_XDECREF(iter); |
| 915 | return ret; |
| 916 | #undef CHUNKSIZE |
| 917 | } |
| 918 | |
| 919 | PyDoc_STRVAR(BZ2File_seek__doc__, |
| 920 | "seek(offset [, whence]) -> None\n\ |
| 921 | \n\ |
| 922 | Move to new file position. Argument offset is a byte count. Optional\n\ |
| 923 | argument whence defaults to 0 (offset from start of file, offset\n\ |
| 924 | should be >= 0); other values are 1 (move relative to current position,\n\ |
| 925 | positive or negative), and 2 (move relative to end of file, usually\n\ |
| 926 | negative, although many platforms allow seeking beyond the end of a file).\n\ |
| 927 | \n\ |
| 928 | Note that seeking of bz2 files is emulated, and depending on the parameters\n\ |
| 929 | the operation may be extremely slow.\n\ |
| 930 | "); |
| 931 | |
| 932 | static PyObject * |
| 933 | BZ2File_seek(BZ2FileObject *self, PyObject *args) |
| 934 | { |
| 935 | int where = 0; |
| 936 | long offset; |
| 937 | char small_buffer[SMALLCHUNK]; |
| 938 | char *buffer = small_buffer; |
| 939 | size_t buffersize = SMALLCHUNK; |
| 940 | int bytesread = 0; |
| 941 | int readsize; |
| 942 | int chunksize; |
| 943 | int bzerror; |
| 944 | int rewind = 0; |
| 945 | PyObject *func; |
| 946 | PyObject *ret = NULL; |
| 947 | |
| 948 | if (!PyArg_ParseTuple(args, "l|i:seek", &offset, &where)) |
| 949 | return NULL; |
| 950 | |
| 951 | ACQUIRE_LOCK(self); |
| 952 | Util_DropReadAhead(self); |
| 953 | switch (self->mode) { |
| 954 | case MODE_READ: |
| 955 | case MODE_READ_EOF: |
| 956 | break; |
| 957 | |
| 958 | case MODE_CLOSED: |
| 959 | PyErr_SetString(PyExc_ValueError, |
| 960 | "I/O operation on closed file"); |
| 961 | goto cleanup;; |
| 962 | |
| 963 | default: |
| 964 | PyErr_SetString(PyExc_IOError, |
| 965 | "seek works only while reading"); |
| 966 | goto cleanup;; |
| 967 | } |
| 968 | |
| 969 | if (offset < 0) { |
| 970 | if (where == 1) { |
| 971 | offset = self->pos + offset; |
| 972 | rewind = 1; |
| 973 | } else if (where == 2) { |
| 974 | if (self->size == -1) { |
| 975 | assert(self->mode != MODE_READ_EOF); |
| 976 | for (;;) { |
| 977 | Py_BEGIN_ALLOW_THREADS |
| 978 | chunksize = Util_UnivNewlineRead( |
| 979 | &bzerror, self->fp, |
| 980 | buffer, buffersize, |
| 981 | self); |
| 982 | self->pos += chunksize; |
| 983 | Py_END_ALLOW_THREADS |
| 984 | |
| 985 | bytesread += chunksize; |
| 986 | if (bzerror == BZ_STREAM_END) { |
| 987 | break; |
| 988 | } else if (bzerror != BZ_OK) { |
| 989 | Util_CatchBZ2Error(bzerror); |
| 990 | goto cleanup; |
| 991 | } |
| 992 | } |
| 993 | self->mode = MODE_READ_EOF; |
| 994 | self->size = self->pos; |
| 995 | bytesread = 0; |
| 996 | } |
| 997 | offset = self->size + offset; |
| 998 | if (offset >= self->pos) |
| 999 | offset -= self->pos; |
| 1000 | else |
| 1001 | rewind = 1; |
| 1002 | } |
| 1003 | if (offset < 0) |
| 1004 | offset = 0; |
| 1005 | } else if (where == 0) { |
| 1006 | if (offset >= self->pos) |
| 1007 | offset -= self->pos; |
| 1008 | else |
| 1009 | rewind = 1; |
| 1010 | } |
| 1011 | |
| 1012 | if (rewind) { |
| 1013 | BZ2_bzReadClose(&bzerror, self->fp); |
| 1014 | func = Py_FindMethod(PyFile_Type.tp_methods, (PyObject*)self, |
| 1015 | "seek"); |
| 1016 | if (bzerror != BZ_OK) { |
| 1017 | Util_CatchBZ2Error(bzerror); |
| 1018 | goto cleanup; |
| 1019 | } |
| 1020 | if (!func) { |
| 1021 | PyErr_SetString(PyExc_RuntimeError, |
| 1022 | "can't find file.seek method"); |
| 1023 | goto cleanup; |
| 1024 | } |
| 1025 | ret = PyObject_CallFunction(func, "(i)", 0); |
| 1026 | if (!ret) |
| 1027 | goto cleanup; |
| 1028 | Py_DECREF(ret); |
| 1029 | ret = NULL; |
| 1030 | self->pos = 0; |
| 1031 | self->fp = BZ2_bzReadOpen(&bzerror, |
| 1032 | PyFile_AsFile((PyObject*)self), |
| 1033 | 0, 0, NULL, 0); |
| 1034 | if (bzerror != BZ_OK) { |
| 1035 | Util_CatchBZ2Error(bzerror); |
| 1036 | goto cleanup; |
| 1037 | } |
| 1038 | self->mode = MODE_READ; |
| 1039 | } else if (self->mode == MODE_READ_EOF) { |
| 1040 | goto exit; |
| 1041 | } |
| 1042 | |
| 1043 | if (offset == 0) |
| 1044 | goto exit; |
| 1045 | |
| 1046 | /* Before getting here, offset must be set to the number of bytes |
| 1047 | * to walk forward. */ |
| 1048 | for (;;) { |
| 1049 | if (offset-bytesread > buffersize) |
| 1050 | readsize = buffersize; |
| 1051 | else |
| 1052 | readsize = offset-bytesread; |
| 1053 | Py_BEGIN_ALLOW_THREADS |
| 1054 | chunksize = Util_UnivNewlineRead(&bzerror, self->fp, |
| 1055 | buffer, readsize, self); |
| 1056 | self->pos += chunksize; |
| 1057 | Py_END_ALLOW_THREADS |
| 1058 | bytesread += chunksize; |
| 1059 | if (bzerror == BZ_STREAM_END) { |
| 1060 | self->size = self->pos; |
| 1061 | self->mode = MODE_READ_EOF; |
| 1062 | break; |
| 1063 | } else if (bzerror != BZ_OK) { |
| 1064 | Util_CatchBZ2Error(bzerror); |
| 1065 | goto cleanup; |
| 1066 | } |
| 1067 | if (bytesread == offset) |
| 1068 | break; |
| 1069 | } |
| 1070 | |
| 1071 | exit: |
| 1072 | Py_INCREF(Py_None); |
| 1073 | ret = Py_None; |
| 1074 | |
| 1075 | cleanup: |
| 1076 | RELEASE_LOCK(self); |
| 1077 | return ret; |
| 1078 | } |
| 1079 | |
| 1080 | PyDoc_STRVAR(BZ2File_tell__doc__, |
| 1081 | "tell() -> int\n\ |
| 1082 | \n\ |
| 1083 | Return the current file position, an integer (may be a long integer).\n\ |
| 1084 | "); |
| 1085 | |
| 1086 | static PyObject * |
| 1087 | BZ2File_tell(BZ2FileObject *self, PyObject *args) |
| 1088 | { |
| 1089 | PyObject *ret = NULL; |
| 1090 | |
| 1091 | if (self->mode == MODE_CLOSED) { |
| 1092 | PyErr_SetString(PyExc_ValueError, |
| 1093 | "I/O operation on closed file"); |
| 1094 | goto cleanup; |
| 1095 | } |
| 1096 | |
| 1097 | ret = PyInt_FromLong(self->pos); |
| 1098 | |
| 1099 | cleanup: |
| 1100 | return ret; |
| 1101 | } |
| 1102 | |
| 1103 | PyDoc_STRVAR(BZ2File_notsup__doc__, |
| 1104 | "Operation not supported.\n\ |
| 1105 | "); |
| 1106 | |
| 1107 | static PyObject * |
| 1108 | BZ2File_notsup(BZ2FileObject *self, PyObject *args) |
| 1109 | { |
| 1110 | PyErr_SetString(PyExc_IOError, "operation not supported"); |
| 1111 | return NULL; |
| 1112 | } |
| 1113 | |
| 1114 | PyDoc_STRVAR(BZ2File_close__doc__, |
| 1115 | "close() -> None or (perhaps) an integer\n\ |
| 1116 | \n\ |
| 1117 | Close the file. Sets data attribute .closed to true. A closed file\n\ |
| 1118 | cannot be used for further I/O operations. close() may be called more\n\ |
| 1119 | than once without error.\n\ |
| 1120 | "); |
| 1121 | |
| 1122 | static PyObject * |
| 1123 | BZ2File_close(BZ2FileObject *self) |
| 1124 | { |
| 1125 | PyObject *file_close; |
| 1126 | PyObject *ret = NULL; |
| 1127 | int bzerror = BZ_OK; |
| 1128 | |
| 1129 | ACQUIRE_LOCK(self); |
| 1130 | switch (self->mode) { |
| 1131 | case MODE_READ: |
| 1132 | case MODE_READ_EOF: |
| 1133 | BZ2_bzReadClose(&bzerror, self->fp); |
| 1134 | break; |
| 1135 | case MODE_WRITE: |
| 1136 | BZ2_bzWriteClose(&bzerror, self->fp, |
| 1137 | 0, NULL, NULL); |
| 1138 | break; |
| 1139 | } |
| 1140 | self->mode = MODE_CLOSED; |
| 1141 | file_close = Py_FindMethod(PyFile_Type.tp_methods, (PyObject*)self, |
| 1142 | "close"); |
| 1143 | if (!file_close) { |
| 1144 | PyErr_SetString(PyExc_RuntimeError, |
| 1145 | "can't find file.close method"); |
| 1146 | goto cleanup; |
| 1147 | } |
| 1148 | ret = PyObject_CallObject(file_close, NULL); |
| 1149 | if (bzerror != BZ_OK) { |
| 1150 | Util_CatchBZ2Error(bzerror); |
| 1151 | Py_XDECREF(ret); |
| 1152 | ret = NULL; |
| 1153 | goto cleanup; |
| 1154 | } |
| 1155 | |
| 1156 | cleanup: |
| 1157 | RELEASE_LOCK(self); |
| 1158 | return ret; |
| 1159 | } |
| 1160 | |
| 1161 | static PyMethodDef BZ2File_methods[] = { |
| 1162 | {"read", (PyCFunction)BZ2File_read, METH_VARARGS, BZ2File_read__doc__}, |
| 1163 | {"readline", (PyCFunction)BZ2File_readline, METH_VARARGS, BZ2File_readline__doc__}, |
| 1164 | {"readlines", (PyCFunction)BZ2File_readlines, METH_VARARGS, BZ2File_readlines__doc__}, |
| 1165 | {"write", (PyCFunction)BZ2File_write, METH_VARARGS, BZ2File_write__doc__}, |
| 1166 | {"writelines", (PyCFunction)BZ2File_writelines, METH_O, BZ2File_writelines__doc__}, |
| 1167 | {"seek", (PyCFunction)BZ2File_seek, METH_VARARGS, BZ2File_seek__doc__}, |
| 1168 | {"tell", (PyCFunction)BZ2File_tell, METH_NOARGS, BZ2File_tell__doc__}, |
| 1169 | {"truncate", (PyCFunction)BZ2File_notsup, METH_VARARGS, BZ2File_notsup__doc__}, |
| 1170 | {"readinto", (PyCFunction)BZ2File_notsup, METH_VARARGS, BZ2File_notsup__doc__}, |
| 1171 | {"close", (PyCFunction)BZ2File_close, METH_NOARGS, BZ2File_close__doc__}, |
| 1172 | {NULL, NULL} /* sentinel */ |
| 1173 | }; |
| 1174 | |
| 1175 | |
| 1176 | /* ===================================================================== */ |
| 1177 | /* Slot definitions for BZ2File_Type. */ |
| 1178 | |
| 1179 | static int |
| 1180 | BZ2File_init(BZ2FileObject *self, PyObject *args, PyObject *kwargs) |
| 1181 | { |
| 1182 | PyObject *file_args = NULL; |
| 1183 | static char *kwlist[] = {"filename", "mode", "buffering", |
| 1184 | "compresslevel", 0}; |
| 1185 | char *name = NULL; |
| 1186 | char *mode = "r"; |
| 1187 | int buffering = -1; |
| 1188 | int compresslevel = 9; |
| 1189 | int bzerror; |
| 1190 | int mode_char = 0; |
| 1191 | int univ_newline = 0; |
| 1192 | |
| 1193 | self->size = -1; |
| 1194 | |
| 1195 | if (!PyArg_ParseTupleAndKeywords(args, kwargs, "et|sii:BZ2File", |
| 1196 | kwlist, Py_FileSystemDefaultEncoding, |
| 1197 | &name, &mode, &buffering, |
| 1198 | &compresslevel)) |
| 1199 | return -1; |
| 1200 | |
| 1201 | if (compresslevel < 1 || compresslevel > 9) { |
| 1202 | PyErr_SetString(PyExc_ValueError, |
| 1203 | "compresslevel must be between 1 and 9"); |
| 1204 | return -1; |
| 1205 | } |
| 1206 | |
| 1207 | for (;;) { |
| 1208 | int error = 0; |
| 1209 | switch (*mode) { |
| 1210 | case 'r': |
| 1211 | case 'w': |
| 1212 | if (mode_char) |
| 1213 | error = 1; |
| 1214 | mode_char = *mode; |
| 1215 | break; |
| 1216 | |
| 1217 | case 'b': |
| 1218 | break; |
| 1219 | |
| 1220 | case 'U': |
| 1221 | univ_newline = 1; |
| 1222 | break; |
| 1223 | |
| 1224 | default: |
| 1225 | error = 1; |
| 1226 | } |
| 1227 | if (error) { |
| 1228 | PyErr_SetString(PyExc_ValueError, "invalid mode"); |
| 1229 | return -1; |
| 1230 | } |
| 1231 | mode++; |
| 1232 | if (*mode == 0) |
| 1233 | break; |
| 1234 | } |
| 1235 | |
| 1236 | if (mode_char == 'r') |
| 1237 | mode = univ_newline ? "rbU" : "rb"; |
| 1238 | else |
| 1239 | mode = univ_newline ? "wbU" : "wb"; |
| 1240 | |
| 1241 | file_args = Py_BuildValue("(ssi)", name, mode, buffering); |
| 1242 | if (!file_args) |
| 1243 | goto error; |
| 1244 | |
| 1245 | if (PyFile_Type.tp_init((PyObject *)self, file_args, NULL) < 0) |
| 1246 | goto error; |
| 1247 | |
| 1248 | #ifdef WITH_THREAD |
| 1249 | self->lock = PyThread_allocate_lock(); |
| 1250 | if (!self->lock) |
| 1251 | goto error; |
| 1252 | #endif |
| 1253 | |
| 1254 | if (mode_char == 'r') |
| 1255 | self->fp = BZ2_bzReadOpen(&bzerror, |
| 1256 | PyFile_AsFile((PyObject*)self), |
| 1257 | 0, 0, NULL, 0); |
| 1258 | else |
| 1259 | self->fp = BZ2_bzWriteOpen(&bzerror, |
| 1260 | PyFile_AsFile((PyObject*)self), |
| 1261 | compresslevel, 0, 0); |
| 1262 | |
| 1263 | if (bzerror != BZ_OK) { |
| 1264 | Util_CatchBZ2Error(bzerror); |
| 1265 | goto error; |
| 1266 | } |
| 1267 | |
| 1268 | self->mode = (mode_char == 'r') ? MODE_READ : MODE_WRITE; |
| 1269 | |
| 1270 | Py_XDECREF(file_args); |
| 1271 | PyMem_Free(name); |
| 1272 | return 0; |
| 1273 | |
| 1274 | error: |
| 1275 | #ifdef WITH_THREAD |
| 1276 | if (self->lock) |
| 1277 | PyThread_free_lock(self->lock); |
| 1278 | #endif |
| 1279 | Py_XDECREF(file_args); |
| 1280 | PyMem_Free(name); |
| 1281 | return -1; |
| 1282 | } |
| 1283 | |
| 1284 | static void |
| 1285 | BZ2File_dealloc(BZ2FileObject *self) |
| 1286 | { |
| 1287 | int bzerror; |
| 1288 | #ifdef WITH_THREAD |
| 1289 | if (self->lock) |
| 1290 | PyThread_free_lock(self->lock); |
| 1291 | #endif |
| 1292 | switch (self->mode) { |
| 1293 | case MODE_READ: |
| 1294 | case MODE_READ_EOF: |
| 1295 | BZ2_bzReadClose(&bzerror, self->fp); |
| 1296 | break; |
| 1297 | case MODE_WRITE: |
| 1298 | BZ2_bzWriteClose(&bzerror, self->fp, |
| 1299 | 0, NULL, NULL); |
| 1300 | break; |
| 1301 | } |
| 1302 | ((PyObject*)self)->ob_type->tp_free((PyObject *)self); |
| 1303 | } |
| 1304 | |
| 1305 | /* This is a hacked version of Python's fileobject.c:file_getiter(). */ |
| 1306 | static PyObject * |
| 1307 | BZ2File_getiter(BZ2FileObject *self) |
| 1308 | { |
| 1309 | if (self->mode == MODE_CLOSED) { |
| 1310 | PyErr_SetString(PyExc_ValueError, |
| 1311 | "I/O operation on closed file"); |
| 1312 | return NULL; |
| 1313 | } |
| 1314 | Py_INCREF((PyObject*)self); |
| 1315 | return (PyObject *)self; |
| 1316 | } |
| 1317 | |
| 1318 | /* This is a hacked version of Python's fileobject.c:file_iternext(). */ |
| 1319 | #define READAHEAD_BUFSIZE 8192 |
| 1320 | static PyObject * |
| 1321 | BZ2File_iternext(BZ2FileObject *self) |
| 1322 | { |
| 1323 | PyStringObject* ret; |
| 1324 | ACQUIRE_LOCK(self); |
| 1325 | if (self->mode == MODE_CLOSED) { |
| 1326 | PyErr_SetString(PyExc_ValueError, |
| 1327 | "I/O operation on closed file"); |
| 1328 | return NULL; |
| 1329 | } |
| 1330 | ret = Util_ReadAheadGetLineSkip(self, 0, READAHEAD_BUFSIZE); |
| 1331 | RELEASE_LOCK(self); |
| 1332 | if (ret == NULL || PyString_GET_SIZE(ret) == 0) { |
| 1333 | Py_XDECREF(ret); |
| 1334 | return NULL; |
| 1335 | } |
| 1336 | return (PyObject *)ret; |
| 1337 | } |
| 1338 | |
| 1339 | /* ===================================================================== */ |
| 1340 | /* BZ2File_Type definition. */ |
| 1341 | |
| 1342 | PyDoc_VAR(BZ2File__doc__) = |
| 1343 | PyDoc_STR( |
| 1344 | "BZ2File(name [, mode='r', buffering=0, compresslevel=9]) -> file object\n\ |
| 1345 | \n\ |
| 1346 | Open a bz2 file. The mode can be 'r' or 'w', for reading (default) or\n\ |
| 1347 | writing. When opened for writing, the file will be created if it doesn't\n\ |
| 1348 | exist, and truncated otherwise. If the buffering argument is given, 0 means\n\ |
| 1349 | unbuffered, and larger numbers specify the buffer size. If compresslevel\n\ |
| 1350 | is given, must be a number between 1 and 9.\n\ |
| 1351 | ") |
| 1352 | #ifdef WITH_UNIVERSAL_NEWLINES |
| 1353 | PyDoc_STR( |
| 1354 | "\n\ |
| 1355 | Add a 'U' to mode to open the file for input with universal newline\n\ |
| 1356 | support. Any line ending in the input file will be seen as a '\\n' in\n\ |
| 1357 | Python. Also, a file so opened gains the attribute 'newlines'; the value\n\ |
| 1358 | for this attribute is one of None (no newline read yet), '\\r', '\\n',\n\ |
| 1359 | '\\r\\n' or a tuple containing all the newline types seen. Universal\n\ |
| 1360 | newlines are available only when reading.\n\ |
| 1361 | ") |
| 1362 | #endif |
| 1363 | ; |
| 1364 | |
| 1365 | statichere PyTypeObject BZ2File_Type = { |
| 1366 | PyObject_HEAD_INIT(NULL) |
| 1367 | 0, /*ob_size*/ |
| 1368 | "bz2.BZ2File", /*tp_name*/ |
| 1369 | sizeof(BZ2FileObject), /*tp_basicsize*/ |
| 1370 | 0, /*tp_itemsize*/ |
| 1371 | (destructor)BZ2File_dealloc, /*tp_dealloc*/ |
| 1372 | 0, /*tp_print*/ |
| 1373 | 0, /*tp_getattr*/ |
| 1374 | 0, /*tp_setattr*/ |
| 1375 | 0, /*tp_compare*/ |
| 1376 | 0, /*tp_repr*/ |
| 1377 | 0, /*tp_as_number*/ |
| 1378 | 0, /*tp_as_sequence*/ |
| 1379 | 0, /*tp_as_mapping*/ |
| 1380 | 0, /*tp_hash*/ |
| 1381 | 0, /*tp_call*/ |
| 1382 | 0, /*tp_str*/ |
| 1383 | PyObject_GenericGetAttr,/*tp_getattro*/ |
| 1384 | PyObject_GenericSetAttr,/*tp_setattro*/ |
| 1385 | 0, /*tp_as_buffer*/ |
| 1386 | Py_TPFLAGS_DEFAULT|Py_TPFLAGS_BASETYPE, /*tp_flags*/ |
| 1387 | BZ2File__doc__, /*tp_doc*/ |
| 1388 | 0, /*tp_traverse*/ |
| 1389 | 0, /*tp_clear*/ |
| 1390 | 0, /*tp_richcompare*/ |
| 1391 | 0, /*tp_weaklistoffset*/ |
| 1392 | (getiterfunc)BZ2File_getiter, /*tp_iter*/ |
| 1393 | (iternextfunc)BZ2File_iternext, /*tp_iternext*/ |
| 1394 | BZ2File_methods, /*tp_methods*/ |
| 1395 | 0, /*tp_members*/ |
| 1396 | 0, /*tp_getset*/ |
| 1397 | 0, /*tp_base*/ |
| 1398 | 0, /*tp_dict*/ |
| 1399 | 0, /*tp_descr_get*/ |
| 1400 | 0, /*tp_descr_set*/ |
| 1401 | 0, /*tp_dictoffset*/ |
| 1402 | (initproc)BZ2File_init, /*tp_init*/ |
| 1403 | PyType_GenericAlloc, /*tp_alloc*/ |
| 1404 | 0, /*tp_new*/ |
| 1405 | _PyObject_Del, /*tp_free*/ |
| 1406 | 0, /*tp_is_gc*/ |
| 1407 | }; |
| 1408 | |
| 1409 | |
| 1410 | /* ===================================================================== */ |
| 1411 | /* Methods of BZ2Comp. */ |
| 1412 | |
| 1413 | PyDoc_STRVAR(BZ2Comp_compress__doc__, |
| 1414 | "compress(data) -> string\n\ |
| 1415 | \n\ |
| 1416 | Provide more data to the compressor object. It will return chunks of\n\ |
| 1417 | compressed data whenever possible. When you've finished providing data\n\ |
| 1418 | to compress, call the flush() method to finish the compression process,\n\ |
| 1419 | and return what is left in the internal buffers.\n\ |
| 1420 | "); |
| 1421 | |
| 1422 | static PyObject * |
| 1423 | BZ2Comp_compress(BZ2CompObject *self, PyObject *args) |
| 1424 | { |
| 1425 | char *data; |
| 1426 | int datasize; |
| 1427 | int bufsize = SMALLCHUNK; |
| 1428 | long totalout; |
| 1429 | PyObject *ret; |
| 1430 | bz_stream *bzs = &self->bzs; |
| 1431 | int bzerror; |
| 1432 | |
| 1433 | if (!PyArg_ParseTuple(args, "s#", &data, &datasize)) |
| 1434 | return NULL; |
| 1435 | |
| 1436 | ACQUIRE_LOCK(self); |
| 1437 | if (!self->running) { |
| 1438 | PyErr_SetString(PyExc_ValueError, "this object was already " |
| 1439 | "flushed"); |
| 1440 | goto error; |
| 1441 | } |
| 1442 | |
| 1443 | ret = PyString_FromStringAndSize(NULL, bufsize); |
| 1444 | if (!ret) |
| 1445 | goto error; |
| 1446 | |
| 1447 | bzs->next_in = data; |
| 1448 | bzs->avail_in = datasize; |
| 1449 | bzs->next_out = BUF(ret); |
| 1450 | bzs->avail_out = bufsize; |
| 1451 | |
| 1452 | totalout = BZS_TOTAL_OUT(bzs); |
| 1453 | |
| 1454 | for (;;) { |
| 1455 | Py_BEGIN_ALLOW_THREADS |
| 1456 | bzerror = BZ2_bzCompress(bzs, BZ_RUN); |
| 1457 | Py_END_ALLOW_THREADS |
| 1458 | if (bzerror != BZ_RUN_OK) { |
| 1459 | Util_CatchBZ2Error(bzerror); |
| 1460 | goto error; |
| 1461 | } |
| 1462 | if (bzs->avail_out == 0) { |
| 1463 | bufsize = Util_NewBufferSize(bufsize); |
| 1464 | if (_PyString_Resize(&ret, bufsize) < 0) { |
| 1465 | BZ2_bzCompressEnd(bzs); |
| 1466 | goto error; |
| 1467 | } |
| 1468 | bzs->next_out = BUF(ret) + (BZS_TOTAL_OUT(bzs) |
| 1469 | - totalout); |
| 1470 | bzs->avail_out = bufsize - (bzs->next_out - BUF(ret)); |
| 1471 | } else if (bzs->avail_in == 0) { |
| 1472 | break; |
| 1473 | } |
| 1474 | } |
| 1475 | |
| 1476 | _PyString_Resize(&ret, BZS_TOTAL_OUT(bzs) - totalout); |
| 1477 | |
| 1478 | RELEASE_LOCK(self); |
| 1479 | return ret; |
| 1480 | |
| 1481 | error: |
| 1482 | RELEASE_LOCK(self); |
| 1483 | Py_XDECREF(ret); |
| 1484 | return NULL; |
| 1485 | } |
| 1486 | |
| 1487 | PyDoc_STRVAR(BZ2Comp_flush__doc__, |
| 1488 | "flush() -> string\n\ |
| 1489 | \n\ |
| 1490 | Finish the compression process and return what is left in internal buffers.\n\ |
| 1491 | You must not use the compressor object after calling this method.\n\ |
| 1492 | "); |
| 1493 | |
| 1494 | static PyObject * |
| 1495 | BZ2Comp_flush(BZ2CompObject *self) |
| 1496 | { |
| 1497 | int bufsize = SMALLCHUNK; |
| 1498 | PyObject *ret; |
| 1499 | bz_stream *bzs = &self->bzs; |
| 1500 | int totalout; |
| 1501 | int bzerror; |
| 1502 | |
| 1503 | ACQUIRE_LOCK(self); |
| 1504 | if (!self->running) { |
| 1505 | PyErr_SetString(PyExc_ValueError, "object was already " |
| 1506 | "flushed"); |
| 1507 | goto error; |
| 1508 | } |
| 1509 | self->running = 0; |
| 1510 | |
| 1511 | ret = PyString_FromStringAndSize(NULL, bufsize); |
| 1512 | if (!ret) |
| 1513 | goto error; |
| 1514 | |
| 1515 | bzs->next_out = BUF(ret); |
| 1516 | bzs->avail_out = bufsize; |
| 1517 | |
| 1518 | totalout = BZS_TOTAL_OUT(bzs); |
| 1519 | |
| 1520 | for (;;) { |
| 1521 | Py_BEGIN_ALLOW_THREADS |
| 1522 | bzerror = BZ2_bzCompress(bzs, BZ_FINISH); |
| 1523 | Py_END_ALLOW_THREADS |
| 1524 | if (bzerror == BZ_STREAM_END) { |
| 1525 | break; |
| 1526 | } else if (bzerror != BZ_FINISH_OK) { |
| 1527 | Util_CatchBZ2Error(bzerror); |
| 1528 | goto error; |
| 1529 | } |
| 1530 | if (bzs->avail_out == 0) { |
| 1531 | bufsize = Util_NewBufferSize(bufsize); |
| 1532 | if (_PyString_Resize(&ret, bufsize) < 0) |
| 1533 | goto error; |
| 1534 | bzs->next_out = BUF(ret); |
| 1535 | bzs->next_out = BUF(ret) + (BZS_TOTAL_OUT(bzs) |
| 1536 | - totalout); |
| 1537 | bzs->avail_out = bufsize - (bzs->next_out - BUF(ret)); |
| 1538 | } |
| 1539 | } |
| 1540 | |
| 1541 | if (bzs->avail_out != 0) |
| 1542 | _PyString_Resize(&ret, BZS_TOTAL_OUT(bzs) - totalout); |
| 1543 | |
| 1544 | RELEASE_LOCK(self); |
| 1545 | return ret; |
| 1546 | |
| 1547 | error: |
| 1548 | RELEASE_LOCK(self); |
| 1549 | Py_XDECREF(ret); |
| 1550 | return NULL; |
| 1551 | } |
| 1552 | |
| 1553 | static PyMethodDef BZ2Comp_methods[] = { |
| 1554 | {"compress", (PyCFunction)BZ2Comp_compress, METH_VARARGS, BZ2Comp_compress__doc__}, |
| 1555 | {"flush", (PyCFunction)BZ2Comp_flush, METH_NOARGS, BZ2Comp_flush__doc__}, |
| 1556 | {NULL, NULL} /* sentinel */ |
| 1557 | }; |
| 1558 | |
| 1559 | |
| 1560 | /* ===================================================================== */ |
| 1561 | /* Slot definitions for BZ2Comp_Type. */ |
| 1562 | |
| 1563 | static int |
| 1564 | BZ2Comp_init(BZ2CompObject *self, PyObject *args, PyObject *kwargs) |
| 1565 | { |
| 1566 | int compresslevel = 9; |
| 1567 | int bzerror; |
| 1568 | static char *kwlist[] = {"compresslevel", 0}; |
| 1569 | |
| 1570 | if (!PyArg_ParseTupleAndKeywords(args, kwargs, "|i:BZ2Compressor", |
| 1571 | kwlist, &compresslevel)) |
| 1572 | return -1; |
| 1573 | |
| 1574 | if (compresslevel < 1 || compresslevel > 9) { |
| 1575 | PyErr_SetString(PyExc_ValueError, |
| 1576 | "compresslevel must be between 1 and 9"); |
| 1577 | goto error; |
| 1578 | } |
| 1579 | |
| 1580 | #ifdef WITH_THREAD |
| 1581 | self->lock = PyThread_allocate_lock(); |
| 1582 | if (!self->lock) |
| 1583 | goto error; |
| 1584 | #endif |
| 1585 | |
| 1586 | memset(&self->bzs, 0, sizeof(bz_stream)); |
| 1587 | bzerror = BZ2_bzCompressInit(&self->bzs, compresslevel, 0, 0); |
| 1588 | if (bzerror != BZ_OK) { |
| 1589 | Util_CatchBZ2Error(bzerror); |
| 1590 | goto error; |
| 1591 | } |
| 1592 | |
| 1593 | self->running = 1; |
| 1594 | |
| 1595 | return 0; |
| 1596 | error: |
| 1597 | #ifdef WITH_THREAD |
| 1598 | if (self->lock) |
| 1599 | PyThread_free_lock(self->lock); |
| 1600 | #endif |
| 1601 | return -1; |
| 1602 | } |
| 1603 | |
| 1604 | static void |
| 1605 | BZ2Comp_dealloc(BZ2CompObject *self) |
| 1606 | { |
| 1607 | #ifdef WITH_THREAD |
| 1608 | if (self->lock) |
| 1609 | PyThread_free_lock(self->lock); |
| 1610 | #endif |
| 1611 | BZ2_bzCompressEnd(&self->bzs); |
| 1612 | ((PyObject*)self)->ob_type->tp_free((PyObject *)self); |
| 1613 | } |
| 1614 | |
| 1615 | |
| 1616 | /* ===================================================================== */ |
| 1617 | /* BZ2Comp_Type definition. */ |
| 1618 | |
| 1619 | PyDoc_STRVAR(BZ2Comp__doc__, |
| 1620 | "BZ2Compressor([compresslevel=9]) -> compressor object\n\ |
| 1621 | \n\ |
| 1622 | Create a new compressor object. This object may be used to compress\n\ |
| 1623 | data sequentially. If you want to compress data in one shot, use the\n\ |
| 1624 | compress() function instead. The compresslevel parameter, if given,\n\ |
| 1625 | must be a number between 1 and 9.\n\ |
| 1626 | "); |
| 1627 | |
| 1628 | statichere PyTypeObject BZ2Comp_Type = { |
| 1629 | PyObject_HEAD_INIT(NULL) |
| 1630 | 0, /*ob_size*/ |
| 1631 | "bz2.BZ2Compressor", /*tp_name*/ |
| 1632 | sizeof(BZ2CompObject), /*tp_basicsize*/ |
| 1633 | 0, /*tp_itemsize*/ |
| 1634 | (destructor)BZ2Comp_dealloc, /*tp_dealloc*/ |
| 1635 | 0, /*tp_print*/ |
| 1636 | 0, /*tp_getattr*/ |
| 1637 | 0, /*tp_setattr*/ |
| 1638 | 0, /*tp_compare*/ |
| 1639 | 0, /*tp_repr*/ |
| 1640 | 0, /*tp_as_number*/ |
| 1641 | 0, /*tp_as_sequence*/ |
| 1642 | 0, /*tp_as_mapping*/ |
| 1643 | 0, /*tp_hash*/ |
| 1644 | 0, /*tp_call*/ |
| 1645 | 0, /*tp_str*/ |
| 1646 | PyObject_GenericGetAttr,/*tp_getattro*/ |
| 1647 | PyObject_GenericSetAttr,/*tp_setattro*/ |
| 1648 | 0, /*tp_as_buffer*/ |
| 1649 | Py_TPFLAGS_DEFAULT|Py_TPFLAGS_BASETYPE, /*tp_flags*/ |
| 1650 | BZ2Comp__doc__, /*tp_doc*/ |
| 1651 | 0, /*tp_traverse*/ |
| 1652 | 0, /*tp_clear*/ |
| 1653 | 0, /*tp_richcompare*/ |
| 1654 | 0, /*tp_weaklistoffset*/ |
| 1655 | 0, /*tp_iter*/ |
| 1656 | 0, /*tp_iternext*/ |
| 1657 | BZ2Comp_methods, /*tp_methods*/ |
| 1658 | 0, /*tp_members*/ |
| 1659 | 0, /*tp_getset*/ |
| 1660 | 0, /*tp_base*/ |
| 1661 | 0, /*tp_dict*/ |
| 1662 | 0, /*tp_descr_get*/ |
| 1663 | 0, /*tp_descr_set*/ |
| 1664 | 0, /*tp_dictoffset*/ |
| 1665 | (initproc)BZ2Comp_init, /*tp_init*/ |
| 1666 | PyType_GenericAlloc, /*tp_alloc*/ |
| 1667 | PyType_GenericNew, /*tp_new*/ |
| 1668 | _PyObject_Del, /*tp_free*/ |
| 1669 | 0, /*tp_is_gc*/ |
| 1670 | }; |
| 1671 | |
| 1672 | |
| 1673 | /* ===================================================================== */ |
| 1674 | /* Members of BZ2Decomp. */ |
| 1675 | |
| 1676 | #define OFF(x) offsetof(BZ2DecompObject, x) |
| 1677 | |
| 1678 | static PyMemberDef BZ2Decomp_members[] = { |
| 1679 | {"unused_data", T_OBJECT, OFF(unused_data), RO}, |
| 1680 | {NULL} /* Sentinel */ |
| 1681 | }; |
| 1682 | |
| 1683 | |
| 1684 | /* ===================================================================== */ |
| 1685 | /* Methods of BZ2Decomp. */ |
| 1686 | |
| 1687 | PyDoc_STRVAR(BZ2Decomp_decompress__doc__, |
| 1688 | "decompress(data) -> string\n\ |
| 1689 | \n\ |
| 1690 | Provide more data to the decompressor object. It will return chunks\n\ |
| 1691 | of decompressed data whenever possible. If you try to decompress data\n\ |
| 1692 | after the end of stream is found, EOFError will be raised. If any data\n\ |
| 1693 | was found after the end of stream, it'll be ignored and saved in\n\ |
| 1694 | unused_data attribute.\n\ |
| 1695 | "); |
| 1696 | |
| 1697 | static PyObject * |
| 1698 | BZ2Decomp_decompress(BZ2DecompObject *self, PyObject *args) |
| 1699 | { |
| 1700 | char *data; |
| 1701 | int datasize; |
| 1702 | int bufsize = SMALLCHUNK; |
| 1703 | long totalout; |
| 1704 | PyObject *ret; |
| 1705 | bz_stream *bzs = &self->bzs; |
| 1706 | int bzerror; |
| 1707 | |
| 1708 | if (!PyArg_ParseTuple(args, "s#", &data, &datasize)) |
| 1709 | return NULL; |
| 1710 | |
| 1711 | ACQUIRE_LOCK(self); |
| 1712 | if (!self->running) { |
| 1713 | PyErr_SetString(PyExc_EOFError, "end of stream was " |
| 1714 | "already found"); |
| 1715 | goto error; |
| 1716 | } |
| 1717 | |
| 1718 | ret = PyString_FromStringAndSize(NULL, bufsize); |
| 1719 | if (!ret) |
| 1720 | goto error; |
| 1721 | |
| 1722 | bzs->next_in = data; |
| 1723 | bzs->avail_in = datasize; |
| 1724 | bzs->next_out = BUF(ret); |
| 1725 | bzs->avail_out = bufsize; |
| 1726 | |
| 1727 | totalout = BZS_TOTAL_OUT(bzs); |
| 1728 | |
| 1729 | for (;;) { |
| 1730 | Py_BEGIN_ALLOW_THREADS |
| 1731 | bzerror = BZ2_bzDecompress(bzs); |
| 1732 | Py_END_ALLOW_THREADS |
| 1733 | if (bzerror == BZ_STREAM_END) { |
| 1734 | if (bzs->avail_in != 0) { |
| 1735 | Py_DECREF(self->unused_data); |
| 1736 | self->unused_data = |
| 1737 | PyString_FromStringAndSize(bzs->next_in, |
| 1738 | bzs->avail_in); |
| 1739 | } |
| 1740 | self->running = 0; |
| 1741 | break; |
| 1742 | } |
| 1743 | if (bzerror != BZ_OK) { |
| 1744 | Util_CatchBZ2Error(bzerror); |
| 1745 | goto error; |
| 1746 | } |
| 1747 | if (bzs->avail_out == 0) { |
| 1748 | bufsize = Util_NewBufferSize(bufsize); |
| 1749 | if (_PyString_Resize(&ret, bufsize) < 0) { |
| 1750 | BZ2_bzDecompressEnd(bzs); |
| 1751 | goto error; |
| 1752 | } |
| 1753 | bzs->next_out = BUF(ret); |
| 1754 | bzs->next_out = BUF(ret) + (BZS_TOTAL_OUT(bzs) |
| 1755 | - totalout); |
| 1756 | bzs->avail_out = bufsize - (bzs->next_out - BUF(ret)); |
| 1757 | } else if (bzs->avail_in == 0) { |
| 1758 | break; |
| 1759 | } |
| 1760 | } |
| 1761 | |
| 1762 | if (bzs->avail_out != 0) |
| 1763 | _PyString_Resize(&ret, BZS_TOTAL_OUT(bzs) - totalout); |
| 1764 | |
| 1765 | RELEASE_LOCK(self); |
| 1766 | return ret; |
| 1767 | |
| 1768 | error: |
| 1769 | RELEASE_LOCK(self); |
| 1770 | Py_XDECREF(ret); |
| 1771 | return NULL; |
| 1772 | } |
| 1773 | |
| 1774 | static PyMethodDef BZ2Decomp_methods[] = { |
| 1775 | {"decompress", (PyCFunction)BZ2Decomp_decompress, METH_VARARGS, BZ2Decomp_decompress__doc__}, |
| 1776 | {NULL, NULL} /* sentinel */ |
| 1777 | }; |
| 1778 | |
| 1779 | |
| 1780 | /* ===================================================================== */ |
| 1781 | /* Slot definitions for BZ2Decomp_Type. */ |
| 1782 | |
| 1783 | static int |
| 1784 | BZ2Decomp_init(BZ2DecompObject *self, PyObject *args, PyObject *kwargs) |
| 1785 | { |
| 1786 | int bzerror; |
| 1787 | |
| 1788 | if (!PyArg_ParseTuple(args, ":BZ2Decompressor")) |
| 1789 | return -1; |
| 1790 | |
| 1791 | #ifdef WITH_THREAD |
| 1792 | self->lock = PyThread_allocate_lock(); |
| 1793 | if (!self->lock) |
| 1794 | goto error; |
| 1795 | #endif |
| 1796 | |
| 1797 | self->unused_data = PyString_FromString(""); |
| 1798 | if (!self->unused_data) |
| 1799 | goto error; |
| 1800 | |
| 1801 | memset(&self->bzs, 0, sizeof(bz_stream)); |
| 1802 | bzerror = BZ2_bzDecompressInit(&self->bzs, 0, 0); |
| 1803 | if (bzerror != BZ_OK) { |
| 1804 | Util_CatchBZ2Error(bzerror); |
| 1805 | goto error; |
| 1806 | } |
| 1807 | |
| 1808 | self->running = 1; |
| 1809 | |
| 1810 | return 0; |
| 1811 | |
| 1812 | error: |
| 1813 | #ifdef WITH_THREAD |
| 1814 | if (self->lock) |
| 1815 | PyThread_free_lock(self->lock); |
| 1816 | #endif |
| 1817 | Py_XDECREF(self->unused_data); |
| 1818 | return -1; |
| 1819 | } |
| 1820 | |
| 1821 | static void |
| 1822 | BZ2Decomp_dealloc(BZ2DecompObject *self) |
| 1823 | { |
| 1824 | #ifdef WITH_THREAD |
| 1825 | if (self->lock) |
| 1826 | PyThread_free_lock(self->lock); |
| 1827 | #endif |
| 1828 | Py_XDECREF(self->unused_data); |
| 1829 | BZ2_bzDecompressEnd(&self->bzs); |
| 1830 | ((PyObject*)self)->ob_type->tp_free((PyObject *)self); |
| 1831 | } |
| 1832 | |
| 1833 | |
| 1834 | /* ===================================================================== */ |
| 1835 | /* BZ2Decomp_Type definition. */ |
| 1836 | |
| 1837 | PyDoc_STRVAR(BZ2Decomp__doc__, |
| 1838 | "BZ2Decompressor() -> decompressor object\n\ |
| 1839 | \n\ |
| 1840 | Create a new decompressor object. This object may be used to decompress\n\ |
| 1841 | data sequentially. If you want to decompress data in one shot, use the\n\ |
| 1842 | decompress() function instead.\n\ |
| 1843 | "); |
| 1844 | |
| 1845 | statichere PyTypeObject BZ2Decomp_Type = { |
| 1846 | PyObject_HEAD_INIT(NULL) |
| 1847 | 0, /*ob_size*/ |
| 1848 | "bz2.BZ2Decompressor", /*tp_name*/ |
| 1849 | sizeof(BZ2DecompObject), /*tp_basicsize*/ |
| 1850 | 0, /*tp_itemsize*/ |
| 1851 | (destructor)BZ2Decomp_dealloc, /*tp_dealloc*/ |
| 1852 | 0, /*tp_print*/ |
| 1853 | 0, /*tp_getattr*/ |
| 1854 | 0, /*tp_setattr*/ |
| 1855 | 0, /*tp_compare*/ |
| 1856 | 0, /*tp_repr*/ |
| 1857 | 0, /*tp_as_number*/ |
| 1858 | 0, /*tp_as_sequence*/ |
| 1859 | 0, /*tp_as_mapping*/ |
| 1860 | 0, /*tp_hash*/ |
| 1861 | 0, /*tp_call*/ |
| 1862 | 0, /*tp_str*/ |
| 1863 | PyObject_GenericGetAttr,/*tp_getattro*/ |
| 1864 | PyObject_GenericSetAttr,/*tp_setattro*/ |
| 1865 | 0, /*tp_as_buffer*/ |
| 1866 | Py_TPFLAGS_DEFAULT|Py_TPFLAGS_BASETYPE, /*tp_flags*/ |
| 1867 | BZ2Decomp__doc__, /*tp_doc*/ |
| 1868 | 0, /*tp_traverse*/ |
| 1869 | 0, /*tp_clear*/ |
| 1870 | 0, /*tp_richcompare*/ |
| 1871 | 0, /*tp_weaklistoffset*/ |
| 1872 | 0, /*tp_iter*/ |
| 1873 | 0, /*tp_iternext*/ |
| 1874 | BZ2Decomp_methods, /*tp_methods*/ |
| 1875 | BZ2Decomp_members, /*tp_members*/ |
| 1876 | 0, /*tp_getset*/ |
| 1877 | 0, /*tp_base*/ |
| 1878 | 0, /*tp_dict*/ |
| 1879 | 0, /*tp_descr_get*/ |
| 1880 | 0, /*tp_descr_set*/ |
| 1881 | 0, /*tp_dictoffset*/ |
| 1882 | (initproc)BZ2Decomp_init, /*tp_init*/ |
| 1883 | PyType_GenericAlloc, /*tp_alloc*/ |
| 1884 | PyType_GenericNew, /*tp_new*/ |
| 1885 | _PyObject_Del, /*tp_free*/ |
| 1886 | 0, /*tp_is_gc*/ |
| 1887 | }; |
| 1888 | |
| 1889 | |
| 1890 | /* ===================================================================== */ |
| 1891 | /* Module functions. */ |
| 1892 | |
| 1893 | PyDoc_STRVAR(bz2_compress__doc__, |
| 1894 | "compress(data [, compresslevel=9]) -> string\n\ |
| 1895 | \n\ |
| 1896 | Compress data in one shot. If you want to compress data sequentially,\n\ |
| 1897 | use an instance of BZ2Compressor instead. The compresslevel parameter, if\n\ |
| 1898 | given, must be a number between 1 and 9.\n\ |
| 1899 | "); |
| 1900 | |
| 1901 | static PyObject * |
| 1902 | bz2_compress(PyObject *self, PyObject *args, PyObject *kwargs) |
| 1903 | { |
| 1904 | int compresslevel=9; |
| 1905 | char *data; |
| 1906 | int datasize; |
| 1907 | int bufsize; |
| 1908 | PyObject *ret; |
| 1909 | bz_stream _bzs; |
| 1910 | bz_stream *bzs = &_bzs; |
| 1911 | int bzerror; |
| 1912 | static char *kwlist[] = {"data", "compresslevel", 0}; |
| 1913 | |
| 1914 | if (!PyArg_ParseTupleAndKeywords(args, kwargs, "s#|i", |
| 1915 | kwlist, &data, &datasize, |
| 1916 | &compresslevel)) |
| 1917 | return NULL; |
| 1918 | |
| 1919 | if (compresslevel < 1 || compresslevel > 9) { |
| 1920 | PyErr_SetString(PyExc_ValueError, |
| 1921 | "compresslevel must be between 1 and 9"); |
| 1922 | return NULL; |
| 1923 | } |
| 1924 | |
| 1925 | /* Conforming to bz2 manual, this is large enough to fit compressed |
| 1926 | * data in one shot. We will check it later anyway. */ |
| 1927 | bufsize = datasize + (datasize/100+1) + 600; |
| 1928 | |
| 1929 | ret = PyString_FromStringAndSize(NULL, bufsize); |
| 1930 | if (!ret) |
| 1931 | return NULL; |
| 1932 | |
| 1933 | memset(bzs, 0, sizeof(bz_stream)); |
| 1934 | |
| 1935 | bzs->next_in = data; |
| 1936 | bzs->avail_in = datasize; |
| 1937 | bzs->next_out = BUF(ret); |
| 1938 | bzs->avail_out = bufsize; |
| 1939 | |
| 1940 | bzerror = BZ2_bzCompressInit(bzs, compresslevel, 0, 0); |
| 1941 | if (bzerror != BZ_OK) { |
| 1942 | Util_CatchBZ2Error(bzerror); |
| 1943 | Py_DECREF(ret); |
| 1944 | return NULL; |
| 1945 | } |
| 1946 | |
| 1947 | for (;;) { |
| 1948 | Py_BEGIN_ALLOW_THREADS |
| 1949 | bzerror = BZ2_bzCompress(bzs, BZ_FINISH); |
| 1950 | Py_END_ALLOW_THREADS |
| 1951 | if (bzerror == BZ_STREAM_END) { |
| 1952 | break; |
| 1953 | } else if (bzerror != BZ_FINISH_OK) { |
| 1954 | BZ2_bzCompressEnd(bzs); |
| 1955 | Util_CatchBZ2Error(bzerror); |
| 1956 | Py_DECREF(ret); |
| 1957 | return NULL; |
| 1958 | } |
| 1959 | if (bzs->avail_out == 0) { |
| 1960 | bufsize = Util_NewBufferSize(bufsize); |
| 1961 | if (_PyString_Resize(&ret, bufsize) < 0) { |
| 1962 | BZ2_bzCompressEnd(bzs); |
| 1963 | Py_DECREF(ret); |
| 1964 | return NULL; |
| 1965 | } |
| 1966 | bzs->next_out = BUF(ret) + BZS_TOTAL_OUT(bzs); |
| 1967 | bzs->avail_out = bufsize - (bzs->next_out - BUF(ret)); |
| 1968 | } |
| 1969 | } |
| 1970 | |
| 1971 | if (bzs->avail_out != 0) |
| 1972 | _PyString_Resize(&ret, BZS_TOTAL_OUT(bzs)); |
| 1973 | BZ2_bzCompressEnd(bzs); |
| 1974 | |
| 1975 | return ret; |
| 1976 | } |
| 1977 | |
| 1978 | PyDoc_STRVAR(bz2_decompress__doc__, |
| 1979 | "decompress(data) -> decompressed data\n\ |
| 1980 | \n\ |
| 1981 | Decompress data in one shot. If you want to decompress data sequentially,\n\ |
| 1982 | use an instance of BZ2Decompressor instead.\n\ |
| 1983 | "); |
| 1984 | |
| 1985 | static PyObject * |
| 1986 | bz2_decompress(PyObject *self, PyObject *args) |
| 1987 | { |
| 1988 | char *data; |
| 1989 | int datasize; |
| 1990 | int bufsize = SMALLCHUNK; |
| 1991 | PyObject *ret; |
| 1992 | bz_stream _bzs; |
| 1993 | bz_stream *bzs = &_bzs; |
| 1994 | int bzerror; |
| 1995 | |
| 1996 | if (!PyArg_ParseTuple(args, "s#", &data, &datasize)) |
| 1997 | return NULL; |
| 1998 | |
| 1999 | if (datasize == 0) |
| 2000 | return PyString_FromString(""); |
| 2001 | |
| 2002 | ret = PyString_FromStringAndSize(NULL, bufsize); |
| 2003 | if (!ret) |
| 2004 | return NULL; |
| 2005 | |
| 2006 | memset(bzs, 0, sizeof(bz_stream)); |
| 2007 | |
| 2008 | bzs->next_in = data; |
| 2009 | bzs->avail_in = datasize; |
| 2010 | bzs->next_out = BUF(ret); |
| 2011 | bzs->avail_out = bufsize; |
| 2012 | |
| 2013 | bzerror = BZ2_bzDecompressInit(bzs, 0, 0); |
| 2014 | if (bzerror != BZ_OK) { |
| 2015 | Util_CatchBZ2Error(bzerror); |
| 2016 | Py_DECREF(ret); |
| 2017 | return NULL; |
| 2018 | } |
| 2019 | |
| 2020 | for (;;) { |
| 2021 | Py_BEGIN_ALLOW_THREADS |
| 2022 | bzerror = BZ2_bzDecompress(bzs); |
| 2023 | Py_END_ALLOW_THREADS |
| 2024 | if (bzerror == BZ_STREAM_END) { |
| 2025 | break; |
| 2026 | } else if (bzerror != BZ_OK) { |
| 2027 | BZ2_bzDecompressEnd(bzs); |
| 2028 | Util_CatchBZ2Error(bzerror); |
| 2029 | Py_DECREF(ret); |
| 2030 | return NULL; |
| 2031 | } |
| 2032 | if (bzs->avail_out == 0) { |
| 2033 | bufsize = Util_NewBufferSize(bufsize); |
| 2034 | if (_PyString_Resize(&ret, bufsize) < 0) { |
| 2035 | BZ2_bzDecompressEnd(bzs); |
| 2036 | Py_DECREF(ret); |
| 2037 | return NULL; |
| 2038 | } |
| 2039 | bzs->next_out = BUF(ret) + BZS_TOTAL_OUT(bzs); |
| 2040 | bzs->avail_out = bufsize - (bzs->next_out - BUF(ret)); |
| 2041 | } else if (bzs->avail_in == 0) { |
| 2042 | BZ2_bzDecompressEnd(bzs); |
| 2043 | PyErr_SetString(PyExc_ValueError, |
| 2044 | "couldn't find end of stream"); |
| 2045 | Py_DECREF(ret); |
| 2046 | return NULL; |
| 2047 | } |
| 2048 | } |
| 2049 | |
| 2050 | if (bzs->avail_out != 0) |
| 2051 | _PyString_Resize(&ret, BZS_TOTAL_OUT(bzs)); |
| 2052 | BZ2_bzDecompressEnd(bzs); |
| 2053 | |
| 2054 | return ret; |
| 2055 | } |
| 2056 | |
| 2057 | static PyMethodDef bz2_methods[] = { |
| 2058 | {"compress", (PyCFunction) bz2_compress, METH_VARARGS|METH_KEYWORDS, |
| 2059 | bz2_compress__doc__}, |
| 2060 | {"decompress", (PyCFunction) bz2_decompress, METH_VARARGS, |
| 2061 | bz2_decompress__doc__}, |
| 2062 | {NULL, NULL} /* sentinel */ |
| 2063 | }; |
| 2064 | |
| 2065 | /* ===================================================================== */ |
| 2066 | /* Initialization function. */ |
| 2067 | |
| 2068 | PyDoc_STRVAR(bz2__doc__, |
| 2069 | "The python bz2 module provides a comprehensive interface for\n\ |
| 2070 | the bz2 compression library. It implements a complete file\n\ |
| 2071 | interface, one shot (de)compression functions, and types for\n\ |
| 2072 | sequential (de)compression.\n\ |
| 2073 | "); |
| 2074 | |
| 2075 | DL_EXPORT(void) |
| 2076 | initbz2(void) |
| 2077 | { |
| 2078 | PyObject *m; |
| 2079 | |
| 2080 | BZ2File_Type.ob_type = &PyType_Type; |
| 2081 | BZ2File_Type.tp_base = &PyFile_Type; |
| 2082 | BZ2File_Type.tp_new = PyFile_Type.tp_new; |
| 2083 | |
| 2084 | BZ2Comp_Type.ob_type = &PyType_Type; |
| 2085 | BZ2Decomp_Type.ob_type = &PyType_Type; |
| 2086 | |
| 2087 | m = Py_InitModule3("bz2", bz2_methods, bz2__doc__); |
| 2088 | |
| 2089 | PyModule_AddObject(m, "__author__", PyString_FromString(__author__)); |
| 2090 | |
| 2091 | Py_INCREF(&BZ2File_Type); |
| 2092 | PyModule_AddObject(m, "BZ2File", (PyObject *)&BZ2File_Type); |
| 2093 | |
| 2094 | Py_INCREF(&BZ2Comp_Type); |
| 2095 | PyModule_AddObject(m, "BZ2Compressor", (PyObject *)&BZ2Comp_Type); |
| 2096 | |
| 2097 | Py_INCREF(&BZ2Decomp_Type); |
| 2098 | PyModule_AddObject(m, "BZ2Decompressor", (PyObject *)&BZ2Decomp_Type); |
| 2099 | } |