Merged revisions 80402 via svnmerge from
[python/dscho.git] / Modules / bz2module.c
blob3e8eb8ba4010d33e213a756ac133a776207545b3
1 /*
3 python-bz2 - python bz2 library interface
5 Copyright (c) 2002 Gustavo Niemeyer <niemeyer@conectiva.com>
6 Copyright (c) 2002 Python Software Foundation; All Rights Reserved
8 */
10 #include "Python.h"
11 #include <stdio.h>
12 #include <bzlib.h>
13 #include "structmember.h"
15 #ifdef WITH_THREAD
16 #include "pythread.h"
17 #endif
19 static char __author__[] =
20 "The bz2 python module was written by:\n\
21 \n\
22 Gustavo Niemeyer <niemeyer@conectiva.com>\n\
25 /* Our very own off_t-like type, 64-bit if possible */
26 /* copied from Objects/fileobject.c */
27 #if !defined(HAVE_LARGEFILE_SUPPORT)
28 typedef off_t Py_off_t;
29 #elif SIZEOF_OFF_T >= 8
30 typedef off_t Py_off_t;
31 #elif SIZEOF_FPOS_T >= 8
32 typedef fpos_t Py_off_t;
33 #else
34 #error "Large file support, but neither off_t nor fpos_t is large enough."
35 #endif
37 #define BUF(v) PyBytes_AS_STRING(v)
39 #define MODE_CLOSED 0
40 #define MODE_READ 1
41 #define MODE_READ_EOF 2
42 #define MODE_WRITE 3
44 #define BZ2FileObject_Check(v) (Py_TYPE(v) == &BZ2File_Type)
47 #ifdef BZ_CONFIG_ERROR
49 #if SIZEOF_LONG >= 8
50 #define BZS_TOTAL_OUT(bzs) \
51 (((long)bzs->total_out_hi32 << 32) + bzs->total_out_lo32)
52 #elif SIZEOF_LONG_LONG >= 8
53 #define BZS_TOTAL_OUT(bzs) \
54 (((PY_LONG_LONG)bzs->total_out_hi32 << 32) + bzs->total_out_lo32)
55 #else
56 #define BZS_TOTAL_OUT(bzs) \
57 bzs->total_out_lo32
58 #endif
60 #else /* ! BZ_CONFIG_ERROR */
62 #define BZ2_bzRead bzRead
63 #define BZ2_bzReadOpen bzReadOpen
64 #define BZ2_bzReadClose bzReadClose
65 #define BZ2_bzWrite bzWrite
66 #define BZ2_bzWriteOpen bzWriteOpen
67 #define BZ2_bzWriteClose bzWriteClose
68 #define BZ2_bzCompress bzCompress
69 #define BZ2_bzCompressInit bzCompressInit
70 #define BZ2_bzCompressEnd bzCompressEnd
71 #define BZ2_bzDecompress bzDecompress
72 #define BZ2_bzDecompressInit bzDecompressInit
73 #define BZ2_bzDecompressEnd bzDecompressEnd
75 #define BZS_TOTAL_OUT(bzs) bzs->total_out
77 #endif /* ! BZ_CONFIG_ERROR */
80 #ifdef WITH_THREAD
81 #define ACQUIRE_LOCK(obj) do { \
82 if (!PyThread_acquire_lock(obj->lock, 0)) { \
83 Py_BEGIN_ALLOW_THREADS \
84 PyThread_acquire_lock(obj->lock, 1); \
85 Py_END_ALLOW_THREADS \
86 } } while(0)
87 #define RELEASE_LOCK(obj) PyThread_release_lock(obj->lock)
88 #else
89 #define ACQUIRE_LOCK(obj)
90 #define RELEASE_LOCK(obj)
91 #endif
93 /* Bits in f_newlinetypes */
94 #define NEWLINE_UNKNOWN 0 /* No newline seen, yet */
95 #define NEWLINE_CR 1 /* \r newline seen */
96 #define NEWLINE_LF 2 /* \n newline seen */
97 #define NEWLINE_CRLF 4 /* \r\n newline seen */
99 /* ===================================================================== */
100 /* Structure definitions. */
102 typedef struct {
103 PyObject_HEAD
104 FILE *rawfp;
106 char* f_buf; /* Allocated readahead buffer */
107 char* f_bufend; /* Points after last occupied position */
108 char* f_bufptr; /* Current buffer position */
110 BZFILE *fp;
111 int mode;
112 Py_off_t pos;
113 Py_off_t size;
114 #ifdef WITH_THREAD
115 PyThread_type_lock lock;
116 #endif
117 } BZ2FileObject;
119 typedef struct {
120 PyObject_HEAD
121 bz_stream bzs;
122 int running;
123 #ifdef WITH_THREAD
124 PyThread_type_lock lock;
125 #endif
126 } BZ2CompObject;
128 typedef struct {
129 PyObject_HEAD
130 bz_stream bzs;
131 int running;
132 PyObject *unused_data;
133 #ifdef WITH_THREAD
134 PyThread_type_lock lock;
135 #endif
136 } BZ2DecompObject;
138 /* ===================================================================== */
139 /* Utility functions. */
141 static int
142 Util_CatchBZ2Error(int bzerror)
144 int ret = 0;
145 switch(bzerror) {
146 case BZ_OK:
147 case BZ_STREAM_END:
148 break;
150 #ifdef BZ_CONFIG_ERROR
151 case BZ_CONFIG_ERROR:
152 PyErr_SetString(PyExc_SystemError,
153 "the bz2 library was not compiled "
154 "correctly");
155 ret = 1;
156 break;
157 #endif
159 case BZ_PARAM_ERROR:
160 PyErr_SetString(PyExc_ValueError,
161 "the bz2 library has received wrong "
162 "parameters");
163 ret = 1;
164 break;
166 case BZ_MEM_ERROR:
167 PyErr_NoMemory();
168 ret = 1;
169 break;
171 case BZ_DATA_ERROR:
172 case BZ_DATA_ERROR_MAGIC:
173 PyErr_SetString(PyExc_IOError, "invalid data stream");
174 ret = 1;
175 break;
177 case BZ_IO_ERROR:
178 PyErr_SetString(PyExc_IOError, "unknown IO error");
179 ret = 1;
180 break;
182 case BZ_UNEXPECTED_EOF:
183 PyErr_SetString(PyExc_EOFError,
184 "compressed file ended before the "
185 "logical end-of-stream was detected");
186 ret = 1;
187 break;
189 case BZ_SEQUENCE_ERROR:
190 PyErr_SetString(PyExc_RuntimeError,
191 "wrong sequence of bz2 library "
192 "commands used");
193 ret = 1;
194 break;
196 return ret;
199 #if BUFSIZ < 8192
200 #define SMALLCHUNK 8192
201 #else
202 #define SMALLCHUNK BUFSIZ
203 #endif
205 #if SIZEOF_INT < 4
206 #define BIGCHUNK (512 * 32)
207 #else
208 #define BIGCHUNK (512 * 1024)
209 #endif
211 /* This is a hacked version of Python's fileobject.c:new_buffersize(). */
212 static size_t
213 Util_NewBufferSize(size_t currentsize)
215 if (currentsize > SMALLCHUNK) {
216 /* Keep doubling until we reach BIGCHUNK;
217 then keep adding BIGCHUNK. */
218 if (currentsize <= BIGCHUNK)
219 return currentsize + currentsize;
220 else
221 return currentsize + BIGCHUNK;
223 return currentsize + SMALLCHUNK;
226 /* This is a hacked version of Python's fileobject.c:get_line(). */
227 static PyObject *
228 Util_GetLine(BZ2FileObject *f, int n)
230 char c;
231 char *buf, *end;
232 size_t total_v_size; /* total # of slots in buffer */
233 size_t used_v_size; /* # used slots in buffer */
234 size_t increment; /* amount to increment the buffer */
235 PyObject *v;
236 int bzerror;
237 int bytes_read;
239 total_v_size = n > 0 ? n : 100;
240 v = PyBytes_FromStringAndSize((char *)NULL, total_v_size);
241 if (v == NULL)
242 return NULL;
244 buf = BUF(v);
245 end = buf + total_v_size;
247 for (;;) {
248 Py_BEGIN_ALLOW_THREADS
249 do {
250 bytes_read = BZ2_bzRead(&bzerror, f->fp, &c, 1);
251 f->pos++;
252 if (bytes_read == 0)
253 break;
254 *buf++ = c;
255 } while (bzerror == BZ_OK && c != '\n' && buf != end);
256 Py_END_ALLOW_THREADS
257 if (bzerror == BZ_STREAM_END) {
258 f->size = f->pos;
259 f->mode = MODE_READ_EOF;
260 break;
261 } else if (bzerror != BZ_OK) {
262 Util_CatchBZ2Error(bzerror);
263 Py_DECREF(v);
264 return NULL;
266 if (c == '\n')
267 break;
268 /* Must be because buf == end */
269 if (n > 0)
270 break;
271 used_v_size = total_v_size;
272 increment = total_v_size >> 2; /* mild exponential growth */
273 total_v_size += increment;
274 if (total_v_size > INT_MAX) {
275 PyErr_SetString(PyExc_OverflowError,
276 "line is longer than a Python string can hold");
277 Py_DECREF(v);
278 return NULL;
280 if (_PyBytes_Resize(&v, total_v_size) < 0) {
281 return NULL;
283 buf = BUF(v) + used_v_size;
284 end = BUF(v) + total_v_size;
287 used_v_size = buf - BUF(v);
288 if (used_v_size != total_v_size) {
289 if (_PyBytes_Resize(&v, used_v_size) < 0) {
290 v = NULL;
293 return v;
296 /* This is a hacked version of Python's fileobject.c:drop_readahead(). */
297 static void
298 Util_DropReadAhead(BZ2FileObject *f)
300 if (f->f_buf != NULL) {
301 PyMem_Free(f->f_buf);
302 f->f_buf = NULL;
306 /* This is a hacked version of Python's fileobject.c:readahead(). */
307 static int
308 Util_ReadAhead(BZ2FileObject *f, int bufsize)
310 int chunksize;
311 int bzerror;
313 if (f->f_buf != NULL) {
314 if((f->f_bufend - f->f_bufptr) >= 1)
315 return 0;
316 else
317 Util_DropReadAhead(f);
319 if (f->mode == MODE_READ_EOF) {
320 f->f_bufptr = f->f_buf;
321 f->f_bufend = f->f_buf;
322 return 0;
324 if ((f->f_buf = PyMem_Malloc(bufsize)) == NULL) {
325 PyErr_NoMemory();
326 return -1;
328 Py_BEGIN_ALLOW_THREADS
329 chunksize = BZ2_bzRead(&bzerror, f->fp, f->f_buf, bufsize);
330 Py_END_ALLOW_THREADS
331 f->pos += chunksize;
332 if (bzerror == BZ_STREAM_END) {
333 f->size = f->pos;
334 f->mode = MODE_READ_EOF;
335 } else if (bzerror != BZ_OK) {
336 Util_CatchBZ2Error(bzerror);
337 Util_DropReadAhead(f);
338 return -1;
340 f->f_bufptr = f->f_buf;
341 f->f_bufend = f->f_buf + chunksize;
342 return 0;
345 /* This is a hacked version of Python's
346 * fileobject.c:readahead_get_line_skip(). */
347 static PyBytesObject *
348 Util_ReadAheadGetLineSkip(BZ2FileObject *f, int skip, int bufsize)
350 PyBytesObject* s;
351 char *bufptr;
352 char *buf;
353 int len;
355 if (f->f_buf == NULL)
356 if (Util_ReadAhead(f, bufsize) < 0)
357 return NULL;
359 len = f->f_bufend - f->f_bufptr;
360 if (len == 0)
361 return (PyBytesObject *)
362 PyBytes_FromStringAndSize(NULL, skip);
363 bufptr = memchr(f->f_bufptr, '\n', len);
364 if (bufptr != NULL) {
365 bufptr++; /* Count the '\n' */
366 len = bufptr - f->f_bufptr;
367 s = (PyBytesObject *)
368 PyBytes_FromStringAndSize(NULL, skip+len);
369 if (s == NULL)
370 return NULL;
371 memcpy(PyBytes_AS_STRING(s)+skip, f->f_bufptr, len);
372 f->f_bufptr = bufptr;
373 if (bufptr == f->f_bufend)
374 Util_DropReadAhead(f);
375 } else {
376 bufptr = f->f_bufptr;
377 buf = f->f_buf;
378 f->f_buf = NULL; /* Force new readahead buffer */
379 s = Util_ReadAheadGetLineSkip(f, skip+len,
380 bufsize + (bufsize>>2));
381 if (s == NULL) {
382 PyMem_Free(buf);
383 return NULL;
385 memcpy(PyBytes_AS_STRING(s)+skip, bufptr, len);
386 PyMem_Free(buf);
388 return s;
391 /* ===================================================================== */
392 /* Methods of BZ2File. */
394 PyDoc_STRVAR(BZ2File_read__doc__,
395 "read([size]) -> string\n\
397 Read at most size uncompressed bytes, returned as a string. If the size\n\
398 argument is negative or omitted, read until EOF is reached.\n\
401 /* This is a hacked version of Python's fileobject.c:file_read(). */
402 static PyObject *
403 BZ2File_read(BZ2FileObject *self, PyObject *args)
405 long bytesrequested = -1;
406 size_t bytesread, buffersize, chunksize;
407 int bzerror;
408 PyObject *ret = NULL;
410 if (!PyArg_ParseTuple(args, "|l:read", &bytesrequested))
411 return NULL;
413 ACQUIRE_LOCK(self);
414 switch (self->mode) {
415 case MODE_READ:
416 break;
417 case MODE_READ_EOF:
418 ret = PyBytes_FromStringAndSize("", 0);
419 goto cleanup;
420 case MODE_CLOSED:
421 PyErr_SetString(PyExc_ValueError,
422 "I/O operation on closed file");
423 goto cleanup;
424 default:
425 PyErr_SetString(PyExc_IOError,
426 "file is not ready for reading");
427 goto cleanup;
430 if (bytesrequested < 0)
431 buffersize = Util_NewBufferSize((size_t)0);
432 else
433 buffersize = bytesrequested;
434 if (buffersize > INT_MAX) {
435 PyErr_SetString(PyExc_OverflowError,
436 "requested number of bytes is "
437 "more than a Python string can hold");
438 goto cleanup;
440 ret = PyBytes_FromStringAndSize((char *)NULL, buffersize);
441 if (ret == NULL || buffersize == 0)
442 goto cleanup;
443 bytesread = 0;
445 for (;;) {
446 Py_BEGIN_ALLOW_THREADS
447 chunksize = BZ2_bzRead(&bzerror, self->fp,
448 BUF(ret)+bytesread,
449 buffersize-bytesread);
450 self->pos += chunksize;
451 Py_END_ALLOW_THREADS
452 bytesread += chunksize;
453 if (bzerror == BZ_STREAM_END) {
454 self->size = self->pos;
455 self->mode = MODE_READ_EOF;
456 break;
457 } else if (bzerror != BZ_OK) {
458 Util_CatchBZ2Error(bzerror);
459 Py_DECREF(ret);
460 ret = NULL;
461 goto cleanup;
463 if (bytesrequested < 0) {
464 buffersize = Util_NewBufferSize(buffersize);
465 if (_PyBytes_Resize(&ret, buffersize) < 0) {
466 ret = NULL;
467 goto cleanup;
469 } else {
470 break;
473 if (bytesread != buffersize) {
474 if (_PyBytes_Resize(&ret, bytesread) < 0) {
475 ret = NULL;
479 cleanup:
480 RELEASE_LOCK(self);
481 return ret;
484 PyDoc_STRVAR(BZ2File_readline__doc__,
485 "readline([size]) -> string\n\
487 Return the next line from the file, as a string, retaining newline.\n\
488 A non-negative size argument will limit the maximum number of bytes to\n\
489 return (an incomplete line may be returned then). Return an empty\n\
490 string at EOF.\n\
493 static PyObject *
494 BZ2File_readline(BZ2FileObject *self, PyObject *args)
496 PyObject *ret = NULL;
497 int sizehint = -1;
499 if (!PyArg_ParseTuple(args, "|i:readline", &sizehint))
500 return NULL;
502 ACQUIRE_LOCK(self);
503 switch (self->mode) {
504 case MODE_READ:
505 break;
506 case MODE_READ_EOF:
507 ret = PyBytes_FromStringAndSize("", 0);
508 goto cleanup;
509 case MODE_CLOSED:
510 PyErr_SetString(PyExc_ValueError,
511 "I/O operation on closed file");
512 goto cleanup;
513 default:
514 PyErr_SetString(PyExc_IOError,
515 "file is not ready for reading");
516 goto cleanup;
519 if (sizehint == 0)
520 ret = PyBytes_FromStringAndSize("", 0);
521 else
522 ret = Util_GetLine(self, (sizehint < 0) ? 0 : sizehint);
524 cleanup:
525 RELEASE_LOCK(self);
526 return ret;
529 PyDoc_STRVAR(BZ2File_readlines__doc__,
530 "readlines([size]) -> list\n\
532 Call readline() repeatedly and return a list of lines read.\n\
533 The optional size argument, if given, is an approximate bound on the\n\
534 total number of bytes in the lines returned.\n\
537 /* This is a hacked version of Python's fileobject.c:file_readlines(). */
538 static PyObject *
539 BZ2File_readlines(BZ2FileObject *self, PyObject *args)
541 long sizehint = 0;
542 PyObject *list = NULL;
543 PyObject *line;
544 char small_buffer[SMALLCHUNK];
545 char *buffer = small_buffer;
546 size_t buffersize = SMALLCHUNK;
547 PyObject *big_buffer = NULL;
548 size_t nfilled = 0;
549 size_t nread;
550 size_t totalread = 0;
551 char *p, *q, *end;
552 int err;
553 int shortread = 0;
554 int bzerror;
556 if (!PyArg_ParseTuple(args, "|l:readlines", &sizehint))
557 return NULL;
559 ACQUIRE_LOCK(self);
560 switch (self->mode) {
561 case MODE_READ:
562 break;
563 case MODE_READ_EOF:
564 list = PyList_New(0);
565 goto cleanup;
566 case MODE_CLOSED:
567 PyErr_SetString(PyExc_ValueError,
568 "I/O operation on closed file");
569 goto cleanup;
570 default:
571 PyErr_SetString(PyExc_IOError,
572 "file is not ready for reading");
573 goto cleanup;
576 if ((list = PyList_New(0)) == NULL)
577 goto cleanup;
579 for (;;) {
580 Py_BEGIN_ALLOW_THREADS
581 nread = BZ2_bzRead(&bzerror, self->fp,
582 buffer+nfilled, buffersize-nfilled);
583 self->pos += nread;
584 Py_END_ALLOW_THREADS
585 if (bzerror == BZ_STREAM_END) {
586 self->size = self->pos;
587 self->mode = MODE_READ_EOF;
588 if (nread == 0) {
589 sizehint = 0;
590 break;
592 shortread = 1;
593 } else if (bzerror != BZ_OK) {
594 Util_CatchBZ2Error(bzerror);
595 error:
596 Py_DECREF(list);
597 list = NULL;
598 goto cleanup;
600 totalread += nread;
601 p = memchr(buffer+nfilled, '\n', nread);
602 if (!shortread && p == NULL) {
603 /* Need a larger buffer to fit this line */
604 nfilled += nread;
605 buffersize *= 2;
606 if (buffersize > INT_MAX) {
607 PyErr_SetString(PyExc_OverflowError,
608 "line is longer than a Python string can hold");
609 goto error;
611 if (big_buffer == NULL) {
612 /* Create the big buffer */
613 big_buffer = PyBytes_FromStringAndSize(
614 NULL, buffersize);
615 if (big_buffer == NULL)
616 goto error;
617 buffer = PyBytes_AS_STRING(big_buffer);
618 memcpy(buffer, small_buffer, nfilled);
620 else {
621 /* Grow the big buffer */
622 if (_PyBytes_Resize(&big_buffer, buffersize) < 0){
623 big_buffer = NULL;
624 goto error;
626 buffer = PyBytes_AS_STRING(big_buffer);
628 continue;
630 end = buffer+nfilled+nread;
631 q = buffer;
632 while (p != NULL) {
633 /* Process complete lines */
634 p++;
635 line = PyBytes_FromStringAndSize(q, p-q);
636 if (line == NULL)
637 goto error;
638 err = PyList_Append(list, line);
639 Py_DECREF(line);
640 if (err != 0)
641 goto error;
642 q = p;
643 p = memchr(q, '\n', end-q);
645 /* Move the remaining incomplete line to the start */
646 nfilled = end-q;
647 memmove(buffer, q, nfilled);
648 if (sizehint > 0)
649 if (totalread >= (size_t)sizehint)
650 break;
651 if (shortread) {
652 sizehint = 0;
653 break;
656 if (nfilled != 0) {
657 /* Partial last line */
658 line = PyBytes_FromStringAndSize(buffer, nfilled);
659 if (line == NULL)
660 goto error;
661 if (sizehint > 0) {
662 /* Need to complete the last line */
663 PyObject *rest = Util_GetLine(self, 0);
664 if (rest == NULL) {
665 Py_DECREF(line);
666 goto error;
668 PyBytes_Concat(&line, rest);
669 Py_DECREF(rest);
670 if (line == NULL)
671 goto error;
673 err = PyList_Append(list, line);
674 Py_DECREF(line);
675 if (err != 0)
676 goto error;
679 cleanup:
680 RELEASE_LOCK(self);
681 if (big_buffer) {
682 Py_DECREF(big_buffer);
684 return list;
687 PyDoc_STRVAR(BZ2File_write__doc__,
688 "write(data) -> None\n\
690 Write the 'data' string to file. Note that due to buffering, close() may\n\
691 be needed before the file on disk reflects the data written.\n\
694 /* This is a hacked version of Python's fileobject.c:file_write(). */
695 static PyObject *
696 BZ2File_write(BZ2FileObject *self, PyObject *args)
698 PyObject *ret = NULL;
699 Py_buffer pbuf;
700 char *buf;
701 int len;
702 int bzerror;
704 if (!PyArg_ParseTuple(args, "y*:write", &pbuf))
705 return NULL;
706 buf = pbuf.buf;
707 len = pbuf.len;
709 ACQUIRE_LOCK(self);
710 switch (self->mode) {
711 case MODE_WRITE:
712 break;
714 case MODE_CLOSED:
715 PyErr_SetString(PyExc_ValueError,
716 "I/O operation on closed file");
717 goto cleanup;
719 default:
720 PyErr_SetString(PyExc_IOError,
721 "file is not ready for writing");
722 goto cleanup;
725 Py_BEGIN_ALLOW_THREADS
726 BZ2_bzWrite (&bzerror, self->fp, buf, len);
727 self->pos += len;
728 Py_END_ALLOW_THREADS
730 if (bzerror != BZ_OK) {
731 Util_CatchBZ2Error(bzerror);
732 goto cleanup;
735 Py_INCREF(Py_None);
736 ret = Py_None;
738 cleanup:
739 PyBuffer_Release(&pbuf);
740 RELEASE_LOCK(self);
741 return ret;
744 PyDoc_STRVAR(BZ2File_writelines__doc__,
745 "writelines(sequence_of_strings) -> None\n\
747 Write the sequence of strings to the file. Note that newlines are not\n\
748 added. The sequence can be any iterable object producing strings. This is\n\
749 equivalent to calling write() for each string.\n\
752 /* This is a hacked version of Python's fileobject.c:file_writelines(). */
753 static PyObject *
754 BZ2File_writelines(BZ2FileObject *self, PyObject *seq)
756 #define CHUNKSIZE 1000
757 PyObject *list = NULL;
758 PyObject *iter = NULL;
759 PyObject *ret = NULL;
760 PyObject *line;
761 int i, j, index, len, islist;
762 int bzerror;
764 ACQUIRE_LOCK(self);
765 switch (self->mode) {
766 case MODE_WRITE:
767 break;
769 case MODE_CLOSED:
770 PyErr_SetString(PyExc_ValueError,
771 "I/O operation on closed file");
772 goto error;
774 default:
775 PyErr_SetString(PyExc_IOError,
776 "file is not ready for writing");
777 goto error;
780 islist = PyList_Check(seq);
781 if (!islist) {
782 iter = PyObject_GetIter(seq);
783 if (iter == NULL) {
784 PyErr_SetString(PyExc_TypeError,
785 "writelines() requires an iterable argument");
786 goto error;
788 list = PyList_New(CHUNKSIZE);
789 if (list == NULL)
790 goto error;
793 /* Strategy: slurp CHUNKSIZE lines into a private list,
794 checking that they are all strings, then write that list
795 without holding the interpreter lock, then come back for more. */
796 for (index = 0; ; index += CHUNKSIZE) {
797 if (islist) {
798 Py_XDECREF(list);
799 list = PyList_GetSlice(seq, index, index+CHUNKSIZE);
800 if (list == NULL)
801 goto error;
802 j = PyList_GET_SIZE(list);
804 else {
805 for (j = 0; j < CHUNKSIZE; j++) {
806 line = PyIter_Next(iter);
807 if (line == NULL) {
808 if (PyErr_Occurred())
809 goto error;
810 break;
812 PyList_SetItem(list, j, line);
815 if (j == 0)
816 break;
818 /* Check that all entries are indeed byte strings. If not,
819 apply the same rules as for file.write() and
820 convert the rets to strings. This is slow, but
821 seems to be the only way since all conversion APIs
822 could potentially execute Python code. */
823 for (i = 0; i < j; i++) {
824 PyObject *v = PyList_GET_ITEM(list, i);
825 if (!PyBytes_Check(v)) {
826 const char *buffer;
827 Py_ssize_t len;
828 if (PyObject_AsCharBuffer(v, &buffer, &len)) {
829 PyErr_SetString(PyExc_TypeError,
830 "writelines() "
831 "argument must be "
832 "a sequence of "
833 "bytes objects");
834 goto error;
836 line = PyBytes_FromStringAndSize(buffer,
837 len);
838 if (line == NULL)
839 goto error;
840 Py_DECREF(v);
841 PyList_SET_ITEM(list, i, line);
845 /* Since we are releasing the global lock, the
846 following code may *not* execute Python code. */
847 Py_BEGIN_ALLOW_THREADS
848 for (i = 0; i < j; i++) {
849 line = PyList_GET_ITEM(list, i);
850 len = PyBytes_GET_SIZE(line);
851 BZ2_bzWrite (&bzerror, self->fp,
852 PyBytes_AS_STRING(line), len);
853 if (bzerror != BZ_OK) {
854 Py_BLOCK_THREADS
855 Util_CatchBZ2Error(bzerror);
856 goto error;
859 Py_END_ALLOW_THREADS
861 if (j < CHUNKSIZE)
862 break;
865 Py_INCREF(Py_None);
866 ret = Py_None;
868 error:
869 RELEASE_LOCK(self);
870 Py_XDECREF(list);
871 Py_XDECREF(iter);
872 return ret;
873 #undef CHUNKSIZE
876 PyDoc_STRVAR(BZ2File_seek__doc__,
877 "seek(offset [, whence]) -> None\n\
879 Move to new file position. Argument offset is a byte count. Optional\n\
880 argument whence defaults to 0 (offset from start of file, offset\n\
881 should be >= 0); other values are 1 (move relative to current position,\n\
882 positive or negative), and 2 (move relative to end of file, usually\n\
883 negative, although many platforms allow seeking beyond the end of a file).\n\
885 Note that seeking of bz2 files is emulated, and depending on the parameters\n\
886 the operation may be extremely slow.\n\
889 static PyObject *
890 BZ2File_seek(BZ2FileObject *self, PyObject *args)
892 int where = 0;
893 PyObject *offobj;
894 Py_off_t offset;
895 char small_buffer[SMALLCHUNK];
896 char *buffer = small_buffer;
897 size_t buffersize = SMALLCHUNK;
898 Py_off_t bytesread = 0;
899 size_t readsize;
900 int chunksize;
901 int bzerror;
902 PyObject *ret = NULL;
904 if (!PyArg_ParseTuple(args, "O|i:seek", &offobj, &where))
905 return NULL;
906 #if !defined(HAVE_LARGEFILE_SUPPORT)
907 offset = PyLong_AsLong(offobj);
908 #else
909 offset = PyLong_Check(offobj) ?
910 PyLong_AsLongLong(offobj) : PyLong_AsLong(offobj);
911 #endif
912 if (PyErr_Occurred())
913 return NULL;
915 ACQUIRE_LOCK(self);
916 Util_DropReadAhead(self);
917 switch (self->mode) {
918 case MODE_READ:
919 case MODE_READ_EOF:
920 break;
922 case MODE_CLOSED:
923 PyErr_SetString(PyExc_ValueError,
924 "I/O operation on closed file");
925 goto cleanup;
927 default:
928 PyErr_SetString(PyExc_IOError,
929 "seek works only while reading");
930 goto cleanup;
933 if (where == 2) {
934 if (self->size == -1) {
935 assert(self->mode != MODE_READ_EOF);
936 for (;;) {
937 Py_BEGIN_ALLOW_THREADS
938 chunksize = BZ2_bzRead(&bzerror, self->fp,
939 buffer, buffersize);
940 self->pos += chunksize;
941 Py_END_ALLOW_THREADS
943 bytesread += chunksize;
944 if (bzerror == BZ_STREAM_END) {
945 break;
946 } else if (bzerror != BZ_OK) {
947 Util_CatchBZ2Error(bzerror);
948 goto cleanup;
951 self->mode = MODE_READ_EOF;
952 self->size = self->pos;
953 bytesread = 0;
955 offset = self->size + offset;
956 } else if (where == 1) {
957 offset = self->pos + offset;
960 /* Before getting here, offset must be the absolute position the file
961 * pointer should be set to. */
963 if (offset >= self->pos) {
964 /* we can move forward */
965 offset -= self->pos;
966 } else {
967 /* we cannot move back, so rewind the stream */
968 BZ2_bzReadClose(&bzerror, self->fp);
969 if (bzerror != BZ_OK) {
970 Util_CatchBZ2Error(bzerror);
971 goto cleanup;
973 rewind(self->rawfp);
974 self->pos = 0;
975 self->fp = BZ2_bzReadOpen(&bzerror, self->rawfp,
976 0, 0, NULL, 0);
977 if (bzerror != BZ_OK) {
978 Util_CatchBZ2Error(bzerror);
979 goto cleanup;
981 self->mode = MODE_READ;
984 if (offset <= 0 || self->mode == MODE_READ_EOF)
985 goto exit;
987 /* Before getting here, offset must be set to the number of bytes
988 * to walk forward. */
989 for (;;) {
990 if (offset-bytesread > buffersize)
991 readsize = buffersize;
992 else
993 /* offset might be wider that readsize, but the result
994 * of the subtraction is bound by buffersize (see the
995 * condition above). buffersize is 8192. */
996 readsize = (size_t)(offset-bytesread);
997 Py_BEGIN_ALLOW_THREADS
998 chunksize = BZ2_bzRead(&bzerror, self->fp, buffer, readsize);
999 self->pos += chunksize;
1000 Py_END_ALLOW_THREADS
1001 bytesread += chunksize;
1002 if (bzerror == BZ_STREAM_END) {
1003 self->size = self->pos;
1004 self->mode = MODE_READ_EOF;
1005 break;
1006 } else if (bzerror != BZ_OK) {
1007 Util_CatchBZ2Error(bzerror);
1008 goto cleanup;
1010 if (bytesread == offset)
1011 break;
1014 exit:
1015 Py_INCREF(Py_None);
1016 ret = Py_None;
1018 cleanup:
1019 RELEASE_LOCK(self);
1020 return ret;
1023 PyDoc_STRVAR(BZ2File_tell__doc__,
1024 "tell() -> int\n\
1026 Return the current file position, an integer (may be a long integer).\n\
1029 static PyObject *
1030 BZ2File_tell(BZ2FileObject *self, PyObject *args)
1032 PyObject *ret = NULL;
1034 if (self->mode == MODE_CLOSED) {
1035 PyErr_SetString(PyExc_ValueError,
1036 "I/O operation on closed file");
1037 goto cleanup;
1040 #if !defined(HAVE_LARGEFILE_SUPPORT)
1041 ret = PyLong_FromLong(self->pos);
1042 #else
1043 ret = PyLong_FromLongLong(self->pos);
1044 #endif
1046 cleanup:
1047 return ret;
1050 PyDoc_STRVAR(BZ2File_close__doc__,
1051 "close() -> None or (perhaps) an integer\n\
1053 Close the file. Sets data attribute .closed to true. A closed file\n\
1054 cannot be used for further I/O operations. close() may be called more\n\
1055 than once without error.\n\
1058 static PyObject *
1059 BZ2File_close(BZ2FileObject *self)
1061 PyObject *ret = NULL;
1062 int bzerror = BZ_OK;
1064 if (self->mode == MODE_CLOSED) {
1065 Py_RETURN_NONE;
1068 ACQUIRE_LOCK(self);
1069 switch (self->mode) {
1070 case MODE_READ:
1071 case MODE_READ_EOF:
1072 BZ2_bzReadClose(&bzerror, self->fp);
1073 break;
1074 case MODE_WRITE:
1075 BZ2_bzWriteClose(&bzerror, self->fp,
1076 0, NULL, NULL);
1077 break;
1079 self->mode = MODE_CLOSED;
1080 fclose(self->rawfp);
1081 self->rawfp = NULL;
1082 if (bzerror == BZ_OK) {
1083 Py_INCREF(Py_None);
1084 ret = Py_None;
1086 else {
1087 Util_CatchBZ2Error(bzerror);
1090 RELEASE_LOCK(self);
1091 return ret;
1094 PyDoc_STRVAR(BZ2File_enter_doc,
1095 "__enter__() -> self.");
1097 static PyObject *
1098 BZ2File_enter(BZ2FileObject *self)
1100 if (self->mode == MODE_CLOSED) {
1101 PyErr_SetString(PyExc_ValueError,
1102 "I/O operation on closed file");
1103 return NULL;
1105 Py_INCREF(self);
1106 return (PyObject *) self;
1109 PyDoc_STRVAR(BZ2File_exit_doc,
1110 "__exit__(*excinfo) -> None. Closes the file.");
1112 static PyObject *
1113 BZ2File_exit(BZ2FileObject *self, PyObject *args)
1115 PyObject *ret = PyObject_CallMethod((PyObject *) self, "close", NULL);
1116 if (!ret)
1117 /* If error occurred, pass through */
1118 return NULL;
1119 Py_DECREF(ret);
1120 Py_RETURN_NONE;
1124 static PyObject *BZ2File_getiter(BZ2FileObject *self);
1126 static PyMethodDef BZ2File_methods[] = {
1127 {"read", (PyCFunction)BZ2File_read, METH_VARARGS, BZ2File_read__doc__},
1128 {"readline", (PyCFunction)BZ2File_readline, METH_VARARGS, BZ2File_readline__doc__},
1129 {"readlines", (PyCFunction)BZ2File_readlines, METH_VARARGS, BZ2File_readlines__doc__},
1130 {"write", (PyCFunction)BZ2File_write, METH_VARARGS, BZ2File_write__doc__},
1131 {"writelines", (PyCFunction)BZ2File_writelines, METH_O, BZ2File_writelines__doc__},
1132 {"seek", (PyCFunction)BZ2File_seek, METH_VARARGS, BZ2File_seek__doc__},
1133 {"tell", (PyCFunction)BZ2File_tell, METH_NOARGS, BZ2File_tell__doc__},
1134 {"close", (PyCFunction)BZ2File_close, METH_NOARGS, BZ2File_close__doc__},
1135 {"__enter__", (PyCFunction)BZ2File_enter, METH_NOARGS, BZ2File_enter_doc},
1136 {"__exit__", (PyCFunction)BZ2File_exit, METH_VARARGS, BZ2File_exit_doc},
1137 {NULL, NULL} /* sentinel */
1141 /* ===================================================================== */
1142 /* Getters and setters of BZ2File. */
1144 static PyObject *
1145 BZ2File_get_closed(BZ2FileObject *self, void *closure)
1147 return PyLong_FromLong(self->mode == MODE_CLOSED);
1150 static PyGetSetDef BZ2File_getset[] = {
1151 {"closed", (getter)BZ2File_get_closed, NULL,
1152 "True if the file is closed"},
1153 {NULL} /* Sentinel */
1157 /* ===================================================================== */
1158 /* Slot definitions for BZ2File_Type. */
1160 static int
1161 BZ2File_init(BZ2FileObject *self, PyObject *args, PyObject *kwargs)
1163 static char *kwlist[] = {"filename", "mode", "buffering",
1164 "compresslevel", 0};
1165 PyObject *name_obj = NULL;
1166 char *name;
1167 char *mode = "r";
1168 int buffering = -1;
1169 int compresslevel = 9;
1170 int bzerror;
1171 int mode_char = 0;
1173 self->size = -1;
1175 if (!PyArg_ParseTupleAndKeywords(args, kwargs, "O&|sii:BZ2File",
1176 kwlist, PyUnicode_FSConverter, &name_obj,
1177 &mode, &buffering,
1178 &compresslevel))
1179 return -1;
1181 if (PyBytes_Check(name_obj))
1182 name = PyBytes_AsString(name_obj);
1183 else
1184 name = PyByteArray_AsString(name_obj);
1185 if (compresslevel < 1 || compresslevel > 9) {
1186 PyErr_SetString(PyExc_ValueError,
1187 "compresslevel must be between 1 and 9");
1188 Py_DECREF(name_obj);
1189 return -1;
1192 for (;;) {
1193 int error = 0;
1194 switch (*mode) {
1195 case 'r':
1196 case 'w':
1197 if (mode_char)
1198 error = 1;
1199 mode_char = *mode;
1200 break;
1202 case 'b':
1203 break;
1205 default:
1206 error = 1;
1207 break;
1209 if (error) {
1210 PyErr_Format(PyExc_ValueError,
1211 "invalid mode char %c", *mode);
1212 Py_DECREF(name_obj);
1213 return -1;
1215 mode++;
1216 if (*mode == '\0')
1217 break;
1220 if (mode_char == 0) {
1221 mode_char = 'r';
1224 mode = (mode_char == 'r') ? "rb" : "wb";
1226 self->rawfp = fopen(name, mode);
1227 Py_DECREF(name_obj);
1228 if (self->rawfp == NULL) {
1229 PyErr_SetFromErrno(PyExc_IOError);
1230 return -1;
1232 /* XXX Ignore buffering */
1234 /* From now on, we have stuff to dealloc, so jump to error label
1235 * instead of returning */
1237 #ifdef WITH_THREAD
1238 self->lock = PyThread_allocate_lock();
1239 if (!self->lock) {
1240 PyErr_SetString(PyExc_MemoryError, "unable to allocate lock");
1241 goto error;
1243 #endif
1245 if (mode_char == 'r')
1246 self->fp = BZ2_bzReadOpen(&bzerror, self->rawfp,
1247 0, 0, NULL, 0);
1248 else
1249 self->fp = BZ2_bzWriteOpen(&bzerror, self->rawfp,
1250 compresslevel, 0, 0);
1252 if (bzerror != BZ_OK) {
1253 Util_CatchBZ2Error(bzerror);
1254 goto error;
1257 self->mode = (mode_char == 'r') ? MODE_READ : MODE_WRITE;
1259 return 0;
1261 error:
1262 fclose(self->rawfp);
1263 self->rawfp = NULL;
1264 #ifdef WITH_THREAD
1265 if (self->lock) {
1266 PyThread_free_lock(self->lock);
1267 self->lock = NULL;
1269 #endif
1270 return -1;
1273 static void
1274 BZ2File_dealloc(BZ2FileObject *self)
1276 int bzerror;
1277 #ifdef WITH_THREAD
1278 if (self->lock)
1279 PyThread_free_lock(self->lock);
1280 #endif
1281 switch (self->mode) {
1282 case MODE_READ:
1283 case MODE_READ_EOF:
1284 BZ2_bzReadClose(&bzerror, self->fp);
1285 break;
1286 case MODE_WRITE:
1287 BZ2_bzWriteClose(&bzerror, self->fp,
1288 0, NULL, NULL);
1289 break;
1291 Util_DropReadAhead(self);
1292 if (self->rawfp != NULL)
1293 fclose(self->rawfp);
1294 Py_TYPE(self)->tp_free((PyObject *)self);
1297 /* This is a hacked version of Python's fileobject.c:file_getiter(). */
1298 static PyObject *
1299 BZ2File_getiter(BZ2FileObject *self)
1301 if (self->mode == MODE_CLOSED) {
1302 PyErr_SetString(PyExc_ValueError,
1303 "I/O operation on closed file");
1304 return NULL;
1306 Py_INCREF((PyObject*)self);
1307 return (PyObject *)self;
1310 /* This is a hacked version of Python's fileobject.c:file_iternext(). */
1311 #define READAHEAD_BUFSIZE 8192
1312 static PyObject *
1313 BZ2File_iternext(BZ2FileObject *self)
1315 PyBytesObject* ret;
1316 ACQUIRE_LOCK(self);
1317 if (self->mode == MODE_CLOSED) {
1318 RELEASE_LOCK(self);
1319 PyErr_SetString(PyExc_ValueError,
1320 "I/O operation on closed file");
1321 return NULL;
1323 ret = Util_ReadAheadGetLineSkip(self, 0, READAHEAD_BUFSIZE);
1324 RELEASE_LOCK(self);
1325 if (ret == NULL || PyBytes_GET_SIZE(ret) == 0) {
1326 Py_XDECREF(ret);
1327 return NULL;
1329 return (PyObject *)ret;
1332 /* ===================================================================== */
1333 /* BZ2File_Type definition. */
1335 PyDoc_VAR(BZ2File__doc__) =
1336 PyDoc_STR(
1337 "BZ2File(name [, mode='r', buffering=0, compresslevel=9]) -> file object\n\
1339 Open a bz2 file. The mode can be 'r' or 'w', for reading (default) or\n\
1340 writing. When opened for writing, the file will be created if it doesn't\n\
1341 exist, and truncated otherwise. If the buffering argument is given, 0 means\n\
1342 unbuffered, and larger numbers specify the buffer size. If compresslevel\n\
1343 is given, must be a number between 1 and 9.\n\
1344 Data read is always returned in bytes; data written ought to be bytes.\n\
1347 static PyTypeObject BZ2File_Type = {
1348 PyVarObject_HEAD_INIT(NULL, 0)
1349 "bz2.BZ2File", /*tp_name*/
1350 sizeof(BZ2FileObject), /*tp_basicsize*/
1351 0, /*tp_itemsize*/
1352 (destructor)BZ2File_dealloc, /*tp_dealloc*/
1353 0, /*tp_print*/
1354 0, /*tp_getattr*/
1355 0, /*tp_setattr*/
1356 0, /*tp_reserved*/
1357 0, /*tp_repr*/
1358 0, /*tp_as_number*/
1359 0, /*tp_as_sequence*/
1360 0, /*tp_as_mapping*/
1361 0, /*tp_hash*/
1362 0, /*tp_call*/
1363 0, /*tp_str*/
1364 PyObject_GenericGetAttr,/*tp_getattro*/
1365 PyObject_GenericSetAttr,/*tp_setattro*/
1366 0, /*tp_as_buffer*/
1367 Py_TPFLAGS_DEFAULT|Py_TPFLAGS_BASETYPE, /*tp_flags*/
1368 BZ2File__doc__, /*tp_doc*/
1369 0, /*tp_traverse*/
1370 0, /*tp_clear*/
1371 0, /*tp_richcompare*/
1372 0, /*tp_weaklistoffset*/
1373 (getiterfunc)BZ2File_getiter, /*tp_iter*/
1374 (iternextfunc)BZ2File_iternext, /*tp_iternext*/
1375 BZ2File_methods, /*tp_methods*/
1376 0, /*tp_members*/
1377 BZ2File_getset, /*tp_getset*/
1378 0, /*tp_base*/
1379 0, /*tp_dict*/
1380 0, /*tp_descr_get*/
1381 0, /*tp_descr_set*/
1382 0, /*tp_dictoffset*/
1383 (initproc)BZ2File_init, /*tp_init*/
1384 PyType_GenericAlloc, /*tp_alloc*/
1385 PyType_GenericNew, /*tp_new*/
1386 PyObject_Free, /*tp_free*/
1387 0, /*tp_is_gc*/
1391 /* ===================================================================== */
1392 /* Methods of BZ2Comp. */
1394 PyDoc_STRVAR(BZ2Comp_compress__doc__,
1395 "compress(data) -> string\n\
1397 Provide more data to the compressor object. It will return chunks of\n\
1398 compressed data whenever possible. When you've finished providing data\n\
1399 to compress, call the flush() method to finish the compression process,\n\
1400 and return what is left in the internal buffers.\n\
1403 static PyObject *
1404 BZ2Comp_compress(BZ2CompObject *self, PyObject *args)
1406 Py_buffer pdata;
1407 char *data;
1408 int datasize;
1409 int bufsize = SMALLCHUNK;
1410 PY_LONG_LONG totalout;
1411 PyObject *ret = NULL;
1412 bz_stream *bzs = &self->bzs;
1413 int bzerror;
1415 if (!PyArg_ParseTuple(args, "y*:compress", &pdata))
1416 return NULL;
1417 data = pdata.buf;
1418 datasize = pdata.len;
1420 if (datasize == 0) {
1421 PyBuffer_Release(&pdata);
1422 return PyBytes_FromStringAndSize("", 0);
1425 ACQUIRE_LOCK(self);
1426 if (!self->running) {
1427 PyErr_SetString(PyExc_ValueError,
1428 "this object was already flushed");
1429 goto error;
1432 ret = PyBytes_FromStringAndSize(NULL, bufsize);
1433 if (!ret)
1434 goto error;
1436 bzs->next_in = data;
1437 bzs->avail_in = datasize;
1438 bzs->next_out = BUF(ret);
1439 bzs->avail_out = bufsize;
1441 totalout = BZS_TOTAL_OUT(bzs);
1443 for (;;) {
1444 Py_BEGIN_ALLOW_THREADS
1445 bzerror = BZ2_bzCompress(bzs, BZ_RUN);
1446 Py_END_ALLOW_THREADS
1447 if (bzerror != BZ_RUN_OK) {
1448 Util_CatchBZ2Error(bzerror);
1449 goto error;
1451 if (bzs->avail_in == 0)
1452 break; /* no more input data */
1453 if (bzs->avail_out == 0) {
1454 bufsize = Util_NewBufferSize(bufsize);
1455 if (_PyBytes_Resize(&ret, bufsize) < 0) {
1456 BZ2_bzCompressEnd(bzs);
1457 goto error;
1459 bzs->next_out = BUF(ret) + (BZS_TOTAL_OUT(bzs)
1460 - totalout);
1461 bzs->avail_out = bufsize - (bzs->next_out - BUF(ret));
1465 if (_PyBytes_Resize(&ret,
1466 (Py_ssize_t)(BZS_TOTAL_OUT(bzs) - totalout)) < 0)
1467 goto error;
1469 RELEASE_LOCK(self);
1470 PyBuffer_Release(&pdata);
1471 return ret;
1473 error:
1474 RELEASE_LOCK(self);
1475 PyBuffer_Release(&pdata);
1476 Py_XDECREF(ret);
1477 return NULL;
1480 PyDoc_STRVAR(BZ2Comp_flush__doc__,
1481 "flush() -> string\n\
1483 Finish the compression process and return what is left in internal buffers.\n\
1484 You must not use the compressor object after calling this method.\n\
1487 static PyObject *
1488 BZ2Comp_flush(BZ2CompObject *self)
1490 int bufsize = SMALLCHUNK;
1491 PyObject *ret = NULL;
1492 bz_stream *bzs = &self->bzs;
1493 PY_LONG_LONG totalout;
1494 int bzerror;
1496 ACQUIRE_LOCK(self);
1497 if (!self->running) {
1498 PyErr_SetString(PyExc_ValueError, "object was already "
1499 "flushed");
1500 goto error;
1502 self->running = 0;
1504 ret = PyBytes_FromStringAndSize(NULL, bufsize);
1505 if (!ret)
1506 goto error;
1508 bzs->next_out = BUF(ret);
1509 bzs->avail_out = bufsize;
1511 totalout = BZS_TOTAL_OUT(bzs);
1513 for (;;) {
1514 Py_BEGIN_ALLOW_THREADS
1515 bzerror = BZ2_bzCompress(bzs, BZ_FINISH);
1516 Py_END_ALLOW_THREADS
1517 if (bzerror == BZ_STREAM_END) {
1518 break;
1519 } else if (bzerror != BZ_FINISH_OK) {
1520 Util_CatchBZ2Error(bzerror);
1521 goto error;
1523 if (bzs->avail_out == 0) {
1524 bufsize = Util_NewBufferSize(bufsize);
1525 if (_PyBytes_Resize(&ret, bufsize) < 0)
1526 goto error;
1527 bzs->next_out = BUF(ret);
1528 bzs->next_out = BUF(ret) + (BZS_TOTAL_OUT(bzs)
1529 - totalout);
1530 bzs->avail_out = bufsize - (bzs->next_out - BUF(ret));
1534 if (bzs->avail_out != 0) {
1535 if (_PyBytes_Resize(&ret,
1536 (Py_ssize_t)(BZS_TOTAL_OUT(bzs) - totalout)) < 0)
1537 goto error;
1540 RELEASE_LOCK(self);
1541 return ret;
1543 error:
1544 RELEASE_LOCK(self);
1545 Py_XDECREF(ret);
1546 return NULL;
1549 static PyMethodDef BZ2Comp_methods[] = {
1550 {"compress", (PyCFunction)BZ2Comp_compress, METH_VARARGS,
1551 BZ2Comp_compress__doc__},
1552 {"flush", (PyCFunction)BZ2Comp_flush, METH_NOARGS,
1553 BZ2Comp_flush__doc__},
1554 {NULL, NULL} /* sentinel */
1558 /* ===================================================================== */
1559 /* Slot definitions for BZ2Comp_Type. */
1561 static int
1562 BZ2Comp_init(BZ2CompObject *self, PyObject *args, PyObject *kwargs)
1564 int compresslevel = 9;
1565 int bzerror;
1566 static char *kwlist[] = {"compresslevel", 0};
1568 if (!PyArg_ParseTupleAndKeywords(args, kwargs, "|i:BZ2Compressor",
1569 kwlist, &compresslevel))
1570 return -1;
1572 if (compresslevel < 1 || compresslevel > 9) {
1573 PyErr_SetString(PyExc_ValueError,
1574 "compresslevel must be between 1 and 9");
1575 goto error;
1578 #ifdef WITH_THREAD
1579 self->lock = PyThread_allocate_lock();
1580 if (!self->lock) {
1581 PyErr_SetString(PyExc_MemoryError, "unable to allocate lock");
1582 goto error;
1584 #endif
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;
1593 self->running = 1;
1595 return 0;
1596 error:
1597 #ifdef WITH_THREAD
1598 if (self->lock) {
1599 PyThread_free_lock(self->lock);
1600 self->lock = NULL;
1602 #endif
1603 return -1;
1606 static void
1607 BZ2Comp_dealloc(BZ2CompObject *self)
1609 #ifdef WITH_THREAD
1610 if (self->lock)
1611 PyThread_free_lock(self->lock);
1612 #endif
1613 BZ2_bzCompressEnd(&self->bzs);
1614 Py_TYPE(self)->tp_free((PyObject *)self);
1618 /* ===================================================================== */
1619 /* BZ2Comp_Type definition. */
1621 PyDoc_STRVAR(BZ2Comp__doc__,
1622 "BZ2Compressor([compresslevel=9]) -> compressor object\n\
1624 Create a new compressor object. This object may be used to compress\n\
1625 data sequentially. If you want to compress data in one shot, use the\n\
1626 compress() function instead. The compresslevel parameter, if given,\n\
1627 must be a number between 1 and 9.\n\
1630 static PyTypeObject BZ2Comp_Type = {
1631 PyVarObject_HEAD_INIT(NULL, 0)
1632 "bz2.BZ2Compressor", /*tp_name*/
1633 sizeof(BZ2CompObject), /*tp_basicsize*/
1634 0, /*tp_itemsize*/
1635 (destructor)BZ2Comp_dealloc, /*tp_dealloc*/
1636 0, /*tp_print*/
1637 0, /*tp_getattr*/
1638 0, /*tp_setattr*/
1639 0, /*tp_reserved*/
1640 0, /*tp_repr*/
1641 0, /*tp_as_number*/
1642 0, /*tp_as_sequence*/
1643 0, /*tp_as_mapping*/
1644 0, /*tp_hash*/
1645 0, /*tp_call*/
1646 0, /*tp_str*/
1647 PyObject_GenericGetAttr,/*tp_getattro*/
1648 PyObject_GenericSetAttr,/*tp_setattro*/
1649 0, /*tp_as_buffer*/
1650 Py_TPFLAGS_DEFAULT|Py_TPFLAGS_BASETYPE, /*tp_flags*/
1651 BZ2Comp__doc__, /*tp_doc*/
1652 0, /*tp_traverse*/
1653 0, /*tp_clear*/
1654 0, /*tp_richcompare*/
1655 0, /*tp_weaklistoffset*/
1656 0, /*tp_iter*/
1657 0, /*tp_iternext*/
1658 BZ2Comp_methods, /*tp_methods*/
1659 0, /*tp_members*/
1660 0, /*tp_getset*/
1661 0, /*tp_base*/
1662 0, /*tp_dict*/
1663 0, /*tp_descr_get*/
1664 0, /*tp_descr_set*/
1665 0, /*tp_dictoffset*/
1666 (initproc)BZ2Comp_init, /*tp_init*/
1667 PyType_GenericAlloc, /*tp_alloc*/
1668 PyType_GenericNew, /*tp_new*/
1669 PyObject_Free, /*tp_free*/
1670 0, /*tp_is_gc*/
1674 /* ===================================================================== */
1675 /* Members of BZ2Decomp. */
1677 #undef OFF
1678 #define OFF(x) offsetof(BZ2DecompObject, x)
1680 static PyMemberDef BZ2Decomp_members[] = {
1681 {"unused_data", T_OBJECT, OFF(unused_data), READONLY},
1682 {NULL} /* Sentinel */
1686 /* ===================================================================== */
1687 /* Methods of BZ2Decomp. */
1689 PyDoc_STRVAR(BZ2Decomp_decompress__doc__,
1690 "decompress(data) -> string\n\
1692 Provide more data to the decompressor object. It will return chunks\n\
1693 of decompressed data whenever possible. If you try to decompress data\n\
1694 after the end of stream is found, EOFError will be raised. If any data\n\
1695 was found after the end of stream, it'll be ignored and saved in\n\
1696 unused_data attribute.\n\
1699 static PyObject *
1700 BZ2Decomp_decompress(BZ2DecompObject *self, PyObject *args)
1702 Py_buffer pdata;
1703 char *data;
1704 int datasize;
1705 int bufsize = SMALLCHUNK;
1706 PY_LONG_LONG totalout;
1707 PyObject *ret = NULL;
1708 bz_stream *bzs = &self->bzs;
1709 int bzerror;
1711 if (!PyArg_ParseTuple(args, "y*:decompress", &pdata))
1712 return NULL;
1713 data = pdata.buf;
1714 datasize = pdata.len;
1716 ACQUIRE_LOCK(self);
1717 if (!self->running) {
1718 PyErr_SetString(PyExc_EOFError, "end of stream was "
1719 "already found");
1720 goto error;
1723 ret = PyBytes_FromStringAndSize(NULL, bufsize);
1724 if (!ret)
1725 goto error;
1727 bzs->next_in = data;
1728 bzs->avail_in = datasize;
1729 bzs->next_out = BUF(ret);
1730 bzs->avail_out = bufsize;
1732 totalout = BZS_TOTAL_OUT(bzs);
1734 for (;;) {
1735 Py_BEGIN_ALLOW_THREADS
1736 bzerror = BZ2_bzDecompress(bzs);
1737 Py_END_ALLOW_THREADS
1738 if (bzerror == BZ_STREAM_END) {
1739 if (bzs->avail_in != 0) {
1740 Py_DECREF(self->unused_data);
1741 self->unused_data =
1742 PyBytes_FromStringAndSize(bzs->next_in,
1743 bzs->avail_in);
1745 self->running = 0;
1746 break;
1748 if (bzerror != BZ_OK) {
1749 Util_CatchBZ2Error(bzerror);
1750 goto error;
1752 if (bzs->avail_in == 0)
1753 break; /* no more input data */
1754 if (bzs->avail_out == 0) {
1755 bufsize = Util_NewBufferSize(bufsize);
1756 if (_PyBytes_Resize(&ret, bufsize) < 0) {
1757 BZ2_bzDecompressEnd(bzs);
1758 goto error;
1760 bzs->next_out = BUF(ret);
1761 bzs->next_out = BUF(ret) + (BZS_TOTAL_OUT(bzs)
1762 - totalout);
1763 bzs->avail_out = bufsize - (bzs->next_out - BUF(ret));
1767 if (bzs->avail_out != 0) {
1768 if (_PyBytes_Resize(&ret,
1769 (Py_ssize_t)(BZS_TOTAL_OUT(bzs) - totalout)) < 0)
1770 goto error;
1773 RELEASE_LOCK(self);
1774 PyBuffer_Release(&pdata);
1775 return ret;
1777 error:
1778 RELEASE_LOCK(self);
1779 PyBuffer_Release(&pdata);
1780 Py_XDECREF(ret);
1781 return NULL;
1784 static PyMethodDef BZ2Decomp_methods[] = {
1785 {"decompress", (PyCFunction)BZ2Decomp_decompress, METH_VARARGS, BZ2Decomp_decompress__doc__},
1786 {NULL, NULL} /* sentinel */
1790 /* ===================================================================== */
1791 /* Slot definitions for BZ2Decomp_Type. */
1793 static int
1794 BZ2Decomp_init(BZ2DecompObject *self, PyObject *args, PyObject *kwargs)
1796 int bzerror;
1798 if (!PyArg_ParseTuple(args, ":BZ2Decompressor"))
1799 return -1;
1801 #ifdef WITH_THREAD
1802 self->lock = PyThread_allocate_lock();
1803 if (!self->lock) {
1804 PyErr_SetString(PyExc_MemoryError, "unable to allocate lock");
1805 goto error;
1807 #endif
1809 self->unused_data = PyBytes_FromStringAndSize("", 0);
1810 if (!self->unused_data)
1811 goto error;
1813 memset(&self->bzs, 0, sizeof(bz_stream));
1814 bzerror = BZ2_bzDecompressInit(&self->bzs, 0, 0);
1815 if (bzerror != BZ_OK) {
1816 Util_CatchBZ2Error(bzerror);
1817 goto error;
1820 self->running = 1;
1822 return 0;
1824 error:
1825 #ifdef WITH_THREAD
1826 if (self->lock) {
1827 PyThread_free_lock(self->lock);
1828 self->lock = NULL;
1830 #endif
1831 Py_CLEAR(self->unused_data);
1832 return -1;
1835 static void
1836 BZ2Decomp_dealloc(BZ2DecompObject *self)
1838 #ifdef WITH_THREAD
1839 if (self->lock)
1840 PyThread_free_lock(self->lock);
1841 #endif
1842 Py_XDECREF(self->unused_data);
1843 BZ2_bzDecompressEnd(&self->bzs);
1844 Py_TYPE(self)->tp_free((PyObject *)self);
1848 /* ===================================================================== */
1849 /* BZ2Decomp_Type definition. */
1851 PyDoc_STRVAR(BZ2Decomp__doc__,
1852 "BZ2Decompressor() -> decompressor object\n\
1854 Create a new decompressor object. This object may be used to decompress\n\
1855 data sequentially. If you want to decompress data in one shot, use the\n\
1856 decompress() function instead.\n\
1859 static PyTypeObject BZ2Decomp_Type = {
1860 PyVarObject_HEAD_INIT(NULL, 0)
1861 "bz2.BZ2Decompressor", /*tp_name*/
1862 sizeof(BZ2DecompObject), /*tp_basicsize*/
1863 0, /*tp_itemsize*/
1864 (destructor)BZ2Decomp_dealloc, /*tp_dealloc*/
1865 0, /*tp_print*/
1866 0, /*tp_getattr*/
1867 0, /*tp_setattr*/
1868 0, /*tp_reserved*/
1869 0, /*tp_repr*/
1870 0, /*tp_as_number*/
1871 0, /*tp_as_sequence*/
1872 0, /*tp_as_mapping*/
1873 0, /*tp_hash*/
1874 0, /*tp_call*/
1875 0, /*tp_str*/
1876 PyObject_GenericGetAttr,/*tp_getattro*/
1877 PyObject_GenericSetAttr,/*tp_setattro*/
1878 0, /*tp_as_buffer*/
1879 Py_TPFLAGS_DEFAULT|Py_TPFLAGS_BASETYPE, /*tp_flags*/
1880 BZ2Decomp__doc__, /*tp_doc*/
1881 0, /*tp_traverse*/
1882 0, /*tp_clear*/
1883 0, /*tp_richcompare*/
1884 0, /*tp_weaklistoffset*/
1885 0, /*tp_iter*/
1886 0, /*tp_iternext*/
1887 BZ2Decomp_methods, /*tp_methods*/
1888 BZ2Decomp_members, /*tp_members*/
1889 0, /*tp_getset*/
1890 0, /*tp_base*/
1891 0, /*tp_dict*/
1892 0, /*tp_descr_get*/
1893 0, /*tp_descr_set*/
1894 0, /*tp_dictoffset*/
1895 (initproc)BZ2Decomp_init, /*tp_init*/
1896 PyType_GenericAlloc, /*tp_alloc*/
1897 PyType_GenericNew, /*tp_new*/
1898 PyObject_Free, /*tp_free*/
1899 0, /*tp_is_gc*/
1903 /* ===================================================================== */
1904 /* Module functions. */
1906 PyDoc_STRVAR(bz2_compress__doc__,
1907 "compress(data [, compresslevel=9]) -> string\n\
1909 Compress data in one shot. If you want to compress data sequentially,\n\
1910 use an instance of BZ2Compressor instead. The compresslevel parameter, if\n\
1911 given, must be a number between 1 and 9.\n\
1914 static PyObject *
1915 bz2_compress(PyObject *self, PyObject *args, PyObject *kwargs)
1917 int compresslevel=9;
1918 Py_buffer pdata;
1919 char *data;
1920 int datasize;
1921 int bufsize;
1922 PyObject *ret = NULL;
1923 bz_stream _bzs;
1924 bz_stream *bzs = &_bzs;
1925 int bzerror;
1926 static char *kwlist[] = {"data", "compresslevel", 0};
1928 if (!PyArg_ParseTupleAndKeywords(args, kwargs, "y*|i",
1929 kwlist, &pdata,
1930 &compresslevel))
1931 return NULL;
1932 data = pdata.buf;
1933 datasize = pdata.len;
1935 if (compresslevel < 1 || compresslevel > 9) {
1936 PyErr_SetString(PyExc_ValueError,
1937 "compresslevel must be between 1 and 9");
1938 PyBuffer_Release(&pdata);
1939 return NULL;
1942 /* Conforming to bz2 manual, this is large enough to fit compressed
1943 * data in one shot. We will check it later anyway. */
1944 bufsize = datasize + (datasize/100+1) + 600;
1946 ret = PyBytes_FromStringAndSize(NULL, bufsize);
1947 if (!ret) {
1948 PyBuffer_Release(&pdata);
1949 return NULL;
1952 memset(bzs, 0, sizeof(bz_stream));
1954 bzs->next_in = data;
1955 bzs->avail_in = datasize;
1956 bzs->next_out = BUF(ret);
1957 bzs->avail_out = bufsize;
1959 bzerror = BZ2_bzCompressInit(bzs, compresslevel, 0, 0);
1960 if (bzerror != BZ_OK) {
1961 Util_CatchBZ2Error(bzerror);
1962 PyBuffer_Release(&pdata);
1963 Py_DECREF(ret);
1964 return NULL;
1967 for (;;) {
1968 Py_BEGIN_ALLOW_THREADS
1969 bzerror = BZ2_bzCompress(bzs, BZ_FINISH);
1970 Py_END_ALLOW_THREADS
1971 if (bzerror == BZ_STREAM_END) {
1972 break;
1973 } else if (bzerror != BZ_FINISH_OK) {
1974 BZ2_bzCompressEnd(bzs);
1975 Util_CatchBZ2Error(bzerror);
1976 PyBuffer_Release(&pdata);
1977 Py_DECREF(ret);
1978 return NULL;
1980 if (bzs->avail_out == 0) {
1981 bufsize = Util_NewBufferSize(bufsize);
1982 if (_PyBytes_Resize(&ret, bufsize) < 0) {
1983 BZ2_bzCompressEnd(bzs);
1984 PyBuffer_Release(&pdata);
1985 return NULL;
1987 bzs->next_out = BUF(ret) + BZS_TOTAL_OUT(bzs);
1988 bzs->avail_out = bufsize - (bzs->next_out - BUF(ret));
1992 if (bzs->avail_out != 0) {
1993 if (_PyBytes_Resize(&ret, (Py_ssize_t)BZS_TOTAL_OUT(bzs)) < 0) {
1994 ret = NULL;
1997 BZ2_bzCompressEnd(bzs);
1999 PyBuffer_Release(&pdata);
2000 return ret;
2003 PyDoc_STRVAR(bz2_decompress__doc__,
2004 "decompress(data) -> decompressed data\n\
2006 Decompress data in one shot. If you want to decompress data sequentially,\n\
2007 use an instance of BZ2Decompressor instead.\n\
2010 static PyObject *
2011 bz2_decompress(PyObject *self, PyObject *args)
2013 Py_buffer pdata;
2014 char *data;
2015 int datasize;
2016 int bufsize = SMALLCHUNK;
2017 PyObject *ret;
2018 bz_stream _bzs;
2019 bz_stream *bzs = &_bzs;
2020 int bzerror;
2022 if (!PyArg_ParseTuple(args, "y*:decompress", &pdata))
2023 return NULL;
2024 data = pdata.buf;
2025 datasize = pdata.len;
2027 if (datasize == 0) {
2028 PyBuffer_Release(&pdata);
2029 return PyBytes_FromStringAndSize("", 0);
2032 ret = PyBytes_FromStringAndSize(NULL, bufsize);
2033 if (!ret) {
2034 PyBuffer_Release(&pdata);
2035 return NULL;
2038 memset(bzs, 0, sizeof(bz_stream));
2040 bzs->next_in = data;
2041 bzs->avail_in = datasize;
2042 bzs->next_out = BUF(ret);
2043 bzs->avail_out = bufsize;
2045 bzerror = BZ2_bzDecompressInit(bzs, 0, 0);
2046 if (bzerror != BZ_OK) {
2047 Util_CatchBZ2Error(bzerror);
2048 Py_DECREF(ret);
2049 PyBuffer_Release(&pdata);
2050 return NULL;
2053 for (;;) {
2054 Py_BEGIN_ALLOW_THREADS
2055 bzerror = BZ2_bzDecompress(bzs);
2056 Py_END_ALLOW_THREADS
2057 if (bzerror == BZ_STREAM_END) {
2058 break;
2059 } else if (bzerror != BZ_OK) {
2060 BZ2_bzDecompressEnd(bzs);
2061 Util_CatchBZ2Error(bzerror);
2062 PyBuffer_Release(&pdata);
2063 Py_DECREF(ret);
2064 return NULL;
2066 if (bzs->avail_in == 0) {
2067 BZ2_bzDecompressEnd(bzs);
2068 PyErr_SetString(PyExc_ValueError,
2069 "couldn't find end of stream");
2070 PyBuffer_Release(&pdata);
2071 Py_DECREF(ret);
2072 return NULL;
2074 if (bzs->avail_out == 0) {
2075 bufsize = Util_NewBufferSize(bufsize);
2076 if (_PyBytes_Resize(&ret, bufsize) < 0) {
2077 BZ2_bzDecompressEnd(bzs);
2078 PyBuffer_Release(&pdata);
2079 return NULL;
2081 bzs->next_out = BUF(ret) + BZS_TOTAL_OUT(bzs);
2082 bzs->avail_out = bufsize - (bzs->next_out - BUF(ret));
2086 if (bzs->avail_out != 0) {
2087 if (_PyBytes_Resize(&ret, (Py_ssize_t)BZS_TOTAL_OUT(bzs)) < 0) {
2088 ret = NULL;
2091 BZ2_bzDecompressEnd(bzs);
2092 PyBuffer_Release(&pdata);
2094 return ret;
2097 static PyMethodDef bz2_methods[] = {
2098 {"compress", (PyCFunction) bz2_compress, METH_VARARGS|METH_KEYWORDS,
2099 bz2_compress__doc__},
2100 {"decompress", (PyCFunction) bz2_decompress, METH_VARARGS,
2101 bz2_decompress__doc__},
2102 {NULL, NULL} /* sentinel */
2105 /* ===================================================================== */
2106 /* Initialization function. */
2108 PyDoc_STRVAR(bz2__doc__,
2109 "The python bz2 module provides a comprehensive interface for\n\
2110 the bz2 compression library. It implements a complete file\n\
2111 interface, one shot (de)compression functions, and types for\n\
2112 sequential (de)compression.\n\
2116 static struct PyModuleDef bz2module = {
2117 PyModuleDef_HEAD_INIT,
2118 "bz2",
2119 bz2__doc__,
2121 bz2_methods,
2122 NULL,
2123 NULL,
2124 NULL,
2125 NULL
2128 PyMODINIT_FUNC
2129 PyInit_bz2(void)
2131 PyObject *m;
2133 Py_TYPE(&BZ2File_Type) = &PyType_Type;
2134 Py_TYPE(&BZ2Comp_Type) = &PyType_Type;
2135 Py_TYPE(&BZ2Decomp_Type) = &PyType_Type;
2137 m = PyModule_Create(&bz2module);
2138 if (m == NULL)
2139 return NULL;
2141 PyModule_AddObject(m, "__author__", PyUnicode_FromString(__author__));
2143 Py_INCREF(&BZ2File_Type);
2144 PyModule_AddObject(m, "BZ2File", (PyObject *)&BZ2File_Type);
2146 Py_INCREF(&BZ2Comp_Type);
2147 PyModule_AddObject(m, "BZ2Compressor", (PyObject *)&BZ2Comp_Type);
2149 Py_INCREF(&BZ2Decomp_Type);
2150 PyModule_AddObject(m, "BZ2Decompressor", (PyObject *)&BZ2Decomp_Type);
2151 return m;