blob: 744e6a9ef25d20c6c3fef74f951d1b5b49419f69 [file] [log] [blame]
Christian Heimes2c9c7a52008-05-26 13:42:13 +00001/* PyByteArray (bytearray) implementation */
2
3#define PY_SSIZE_T_CLEAN
4#include "Python.h"
5#include "structmember.h"
6#include "bytes_methods.h"
7
8static PyByteArrayObject *nullbytes = NULL;
9
10void
11PyByteArray_Fini(void)
12{
13 Py_CLEAR(nullbytes);
14}
15
16int
17PyByteArray_Init(void)
18{
19 nullbytes = PyObject_New(PyByteArrayObject, &PyByteArray_Type);
20 if (nullbytes == NULL)
21 return 0;
22 nullbytes->ob_bytes = NULL;
23 Py_SIZE(nullbytes) = nullbytes->ob_alloc = 0;
24 nullbytes->ob_exports = 0;
25 return 1;
26}
27
28/* end nullbytes support */
29
30/* Helpers */
31
32static int
33_getbytevalue(PyObject* arg, int *value)
34{
35 long face_value;
36
37 if (PyLong_Check(arg)) {
38 face_value = PyLong_AsLong(arg);
39 if (face_value < 0 || face_value >= 256) {
40 PyErr_SetString(PyExc_ValueError, "byte must be in range(0, 256)");
41 return 0;
42 }
43 } else {
44 PyErr_Format(PyExc_TypeError, "an integer is required");
45 return 0;
46 }
47
48 *value = face_value;
49 return 1;
50}
51
52static int
53bytes_getbuffer(PyByteArrayObject *obj, Py_buffer *view, int flags)
54{
55 int ret;
56 void *ptr;
57 if (view == NULL) {
58 obj->ob_exports++;
59 return 0;
60 }
61 if (obj->ob_bytes == NULL)
62 ptr = "";
63 else
64 ptr = obj->ob_bytes;
65 ret = PyBuffer_FillInfo(view, ptr, Py_SIZE(obj), 0, flags);
66 if (ret >= 0) {
67 obj->ob_exports++;
68 }
69 return ret;
70}
71
72static void
73bytes_releasebuffer(PyByteArrayObject *obj, Py_buffer *view)
74{
75 obj->ob_exports--;
76}
77
78static Py_ssize_t
79_getbuffer(PyObject *obj, Py_buffer *view)
80{
81 PyBufferProcs *buffer = Py_TYPE(obj)->tp_as_buffer;
82
83 if (buffer == NULL || buffer->bf_getbuffer == NULL)
84 {
85 PyErr_Format(PyExc_TypeError,
86 "Type %.100s doesn't support the buffer API",
87 Py_TYPE(obj)->tp_name);
88 return -1;
89 }
90
91 if (buffer->bf_getbuffer(obj, view, PyBUF_SIMPLE) < 0)
92 return -1;
93 return view->len;
94}
95
96/* Direct API functions */
97
98PyObject *
99PyByteArray_FromObject(PyObject *input)
100{
101 return PyObject_CallFunctionObjArgs((PyObject *)&PyByteArray_Type,
102 input, NULL);
103}
104
105PyObject *
106PyByteArray_FromStringAndSize(const char *bytes, Py_ssize_t size)
107{
108 PyByteArrayObject *new;
109 Py_ssize_t alloc;
110
111 if (size < 0) {
112 PyErr_SetString(PyExc_SystemError,
113 "Negative size passed to PyByteArray_FromStringAndSize");
114 return NULL;
115 }
116
117 new = PyObject_New(PyByteArrayObject, &PyByteArray_Type);
118 if (new == NULL)
119 return NULL;
120
121 if (size == 0) {
122 new->ob_bytes = NULL;
123 alloc = 0;
124 }
125 else {
126 alloc = size + 1;
127 new->ob_bytes = PyMem_Malloc(alloc);
128 if (new->ob_bytes == NULL) {
129 Py_DECREF(new);
130 return PyErr_NoMemory();
131 }
132 if (bytes != NULL)
133 memcpy(new->ob_bytes, bytes, size);
134 new->ob_bytes[size] = '\0'; /* Trailing null byte */
135 }
136 Py_SIZE(new) = size;
137 new->ob_alloc = alloc;
138 new->ob_exports = 0;
139
140 return (PyObject *)new;
141}
142
143Py_ssize_t
144PyByteArray_Size(PyObject *self)
145{
146 assert(self != NULL);
147 assert(PyByteArray_Check(self));
148
149 return PyByteArray_GET_SIZE(self);
150}
151
152char *
153PyByteArray_AsString(PyObject *self)
154{
155 assert(self != NULL);
156 assert(PyByteArray_Check(self));
157
158 return PyByteArray_AS_STRING(self);
159}
160
161int
162PyByteArray_Resize(PyObject *self, Py_ssize_t size)
163{
164 void *sval;
165 Py_ssize_t alloc = ((PyByteArrayObject *)self)->ob_alloc;
166
167 assert(self != NULL);
168 assert(PyByteArray_Check(self));
169 assert(size >= 0);
170
171 if (size < alloc / 2) {
172 /* Major downsize; resize down to exact size */
173 alloc = size + 1;
174 }
175 else if (size < alloc) {
176 /* Within allocated size; quick exit */
177 Py_SIZE(self) = size;
178 ((PyByteArrayObject *)self)->ob_bytes[size] = '\0'; /* Trailing null */
179 return 0;
180 }
181 else if (size <= alloc * 1.125) {
182 /* Moderate upsize; overallocate similar to list_resize() */
183 alloc = size + (size >> 3) + (size < 9 ? 3 : 6);
184 }
185 else {
186 /* Major upsize; resize up to exact size */
187 alloc = size + 1;
188 }
189
190 if (((PyByteArrayObject *)self)->ob_exports > 0) {
191 /*
192 fprintf(stderr, "%d: %s", ((PyByteArrayObject *)self)->ob_exports,
193 ((PyByteArrayObject *)self)->ob_bytes);
194 */
195 PyErr_SetString(PyExc_BufferError,
196 "Existing exports of data: object cannot be re-sized");
197 return -1;
198 }
199
200 sval = PyMem_Realloc(((PyByteArrayObject *)self)->ob_bytes, alloc);
201 if (sval == NULL) {
202 PyErr_NoMemory();
203 return -1;
204 }
205
206 ((PyByteArrayObject *)self)->ob_bytes = sval;
207 Py_SIZE(self) = size;
208 ((PyByteArrayObject *)self)->ob_alloc = alloc;
209 ((PyByteArrayObject *)self)->ob_bytes[size] = '\0'; /* Trailing null byte */
210
211 return 0;
212}
213
214PyObject *
215PyByteArray_Concat(PyObject *a, PyObject *b)
216{
217 Py_ssize_t size;
218 Py_buffer va, vb;
219 PyByteArrayObject *result = NULL;
220
221 va.len = -1;
222 vb.len = -1;
223 if (_getbuffer(a, &va) < 0 ||
224 _getbuffer(b, &vb) < 0) {
225 PyErr_Format(PyExc_TypeError, "can't concat %.100s to %.100s",
226 Py_TYPE(a)->tp_name, Py_TYPE(b)->tp_name);
227 goto done;
228 }
229
230 size = va.len + vb.len;
231 if (size < 0) {
232 return PyErr_NoMemory();
233 goto done;
234 }
235
236 result = (PyByteArrayObject *) PyByteArray_FromStringAndSize(NULL, size);
237 if (result != NULL) {
238 memcpy(result->ob_bytes, va.buf, va.len);
239 memcpy(result->ob_bytes + va.len, vb.buf, vb.len);
240 }
241
242 done:
243 if (va.len != -1)
244 PyObject_ReleaseBuffer(a, &va);
245 if (vb.len != -1)
246 PyObject_ReleaseBuffer(b, &vb);
247 return (PyObject *)result;
248}
249
250/* Functions stuffed into the type object */
251
252static Py_ssize_t
253bytes_length(PyByteArrayObject *self)
254{
255 return Py_SIZE(self);
256}
257
258static PyObject *
259bytes_iconcat(PyByteArrayObject *self, PyObject *other)
260{
261 Py_ssize_t mysize;
262 Py_ssize_t size;
263 Py_buffer vo;
264
265 if (_getbuffer(other, &vo) < 0) {
266 PyErr_Format(PyExc_TypeError, "can't concat %.100s to %.100s",
267 Py_TYPE(other)->tp_name, Py_TYPE(self)->tp_name);
268 return NULL;
269 }
270
271 mysize = Py_SIZE(self);
272 size = mysize + vo.len;
273 if (size < 0) {
274 PyObject_ReleaseBuffer(other, &vo);
275 return PyErr_NoMemory();
276 }
277 if (size < self->ob_alloc) {
278 Py_SIZE(self) = size;
279 self->ob_bytes[Py_SIZE(self)] = '\0'; /* Trailing null byte */
280 }
281 else if (PyByteArray_Resize((PyObject *)self, size) < 0) {
282 PyObject_ReleaseBuffer(other, &vo);
283 return NULL;
284 }
285 memcpy(self->ob_bytes + mysize, vo.buf, vo.len);
286 PyObject_ReleaseBuffer(other, &vo);
287 Py_INCREF(self);
288 return (PyObject *)self;
289}
290
291static PyObject *
292bytes_repeat(PyByteArrayObject *self, Py_ssize_t count)
293{
294 PyByteArrayObject *result;
295 Py_ssize_t mysize;
296 Py_ssize_t size;
297
298 if (count < 0)
299 count = 0;
300 mysize = Py_SIZE(self);
301 size = mysize * count;
302 if (count != 0 && size / count != mysize)
303 return PyErr_NoMemory();
304 result = (PyByteArrayObject *)PyByteArray_FromStringAndSize(NULL, size);
305 if (result != NULL && size != 0) {
306 if (mysize == 1)
307 memset(result->ob_bytes, self->ob_bytes[0], size);
308 else {
309 Py_ssize_t i;
310 for (i = 0; i < count; i++)
311 memcpy(result->ob_bytes + i*mysize, self->ob_bytes, mysize);
312 }
313 }
314 return (PyObject *)result;
315}
316
317static PyObject *
318bytes_irepeat(PyByteArrayObject *self, Py_ssize_t count)
319{
320 Py_ssize_t mysize;
321 Py_ssize_t size;
322
323 if (count < 0)
324 count = 0;
325 mysize = Py_SIZE(self);
326 size = mysize * count;
327 if (count != 0 && size / count != mysize)
328 return PyErr_NoMemory();
329 if (size < self->ob_alloc) {
330 Py_SIZE(self) = size;
331 self->ob_bytes[Py_SIZE(self)] = '\0'; /* Trailing null byte */
332 }
333 else if (PyByteArray_Resize((PyObject *)self, size) < 0)
334 return NULL;
335
336 if (mysize == 1)
337 memset(self->ob_bytes, self->ob_bytes[0], size);
338 else {
339 Py_ssize_t i;
340 for (i = 1; i < count; i++)
341 memcpy(self->ob_bytes + i*mysize, self->ob_bytes, mysize);
342 }
343
344 Py_INCREF(self);
345 return (PyObject *)self;
346}
347
348static PyObject *
349bytes_getitem(PyByteArrayObject *self, Py_ssize_t i)
350{
351 if (i < 0)
352 i += Py_SIZE(self);
353 if (i < 0 || i >= Py_SIZE(self)) {
354 PyErr_SetString(PyExc_IndexError, "bytearray index out of range");
355 return NULL;
356 }
357 return PyLong_FromLong((unsigned char)(self->ob_bytes[i]));
358}
359
360static PyObject *
361bytes_subscript(PyByteArrayObject *self, PyObject *item)
362{
363 if (PyIndex_Check(item)) {
364 Py_ssize_t i = PyNumber_AsSsize_t(item, PyExc_IndexError);
365
366 if (i == -1 && PyErr_Occurred())
367 return NULL;
368
369 if (i < 0)
370 i += PyByteArray_GET_SIZE(self);
371
372 if (i < 0 || i >= Py_SIZE(self)) {
373 PyErr_SetString(PyExc_IndexError, "bytearray index out of range");
374 return NULL;
375 }
376 return PyLong_FromLong((unsigned char)(self->ob_bytes[i]));
377 }
378 else if (PySlice_Check(item)) {
379 Py_ssize_t start, stop, step, slicelength, cur, i;
380 if (PySlice_GetIndicesEx((PySliceObject *)item,
381 PyByteArray_GET_SIZE(self),
382 &start, &stop, &step, &slicelength) < 0) {
383 return NULL;
384 }
385
386 if (slicelength <= 0)
387 return PyByteArray_FromStringAndSize("", 0);
388 else if (step == 1) {
389 return PyByteArray_FromStringAndSize(self->ob_bytes + start,
390 slicelength);
391 }
392 else {
393 char *source_buf = PyByteArray_AS_STRING(self);
394 char *result_buf = (char *)PyMem_Malloc(slicelength);
395 PyObject *result;
396
397 if (result_buf == NULL)
398 return PyErr_NoMemory();
399
400 for (cur = start, i = 0; i < slicelength;
401 cur += step, i++) {
402 result_buf[i] = source_buf[cur];
403 }
404 result = PyByteArray_FromStringAndSize(result_buf, slicelength);
405 PyMem_Free(result_buf);
406 return result;
407 }
408 }
409 else {
410 PyErr_SetString(PyExc_TypeError, "bytearray indices must be integers");
411 return NULL;
412 }
413}
414
415static int
416bytes_setslice(PyByteArrayObject *self, Py_ssize_t lo, Py_ssize_t hi,
417 PyObject *values)
418{
419 Py_ssize_t avail, needed;
420 void *bytes;
421 Py_buffer vbytes;
422 int res = 0;
423
424 vbytes.len = -1;
425 if (values == (PyObject *)self) {
426 /* Make a copy and call this function recursively */
427 int err;
428 values = PyByteArray_FromObject(values);
429 if (values == NULL)
430 return -1;
431 err = bytes_setslice(self, lo, hi, values);
432 Py_DECREF(values);
433 return err;
434 }
435 if (values == NULL) {
436 /* del b[lo:hi] */
437 bytes = NULL;
438 needed = 0;
439 }
440 else {
441 if (_getbuffer(values, &vbytes) < 0) {
442 PyErr_Format(PyExc_TypeError,
443 "can't set bytes slice from %.100s",
444 Py_TYPE(values)->tp_name);
445 return -1;
446 }
447 needed = vbytes.len;
448 bytes = vbytes.buf;
449 }
450
451 if (lo < 0)
452 lo = 0;
453 if (hi < lo)
454 hi = lo;
455 if (hi > Py_SIZE(self))
456 hi = Py_SIZE(self);
457
458 avail = hi - lo;
459 if (avail < 0)
460 lo = hi = avail = 0;
461
462 if (avail != needed) {
463 if (avail > needed) {
464 /*
465 0 lo hi old_size
466 | |<----avail----->|<-----tomove------>|
467 | |<-needed->|<-----tomove------>|
468 0 lo new_hi new_size
469 */
470 memmove(self->ob_bytes + lo + needed, self->ob_bytes + hi,
471 Py_SIZE(self) - hi);
472 }
473 /* XXX(nnorwitz): need to verify this can't overflow! */
474 if (PyByteArray_Resize((PyObject *)self,
475 Py_SIZE(self) + needed - avail) < 0) {
476 res = -1;
477 goto finish;
478 }
479 if (avail < needed) {
480 /*
481 0 lo hi old_size
482 | |<-avail->|<-----tomove------>|
483 | |<----needed---->|<-----tomove------>|
484 0 lo new_hi new_size
485 */
486 memmove(self->ob_bytes + lo + needed, self->ob_bytes + hi,
487 Py_SIZE(self) - lo - needed);
488 }
489 }
490
491 if (needed > 0)
492 memcpy(self->ob_bytes + lo, bytes, needed);
493
494
495 finish:
496 if (vbytes.len != -1)
497 PyObject_ReleaseBuffer(values, &vbytes);
498 return res;
499}
500
501static int
502bytes_setitem(PyByteArrayObject *self, Py_ssize_t i, PyObject *value)
503{
504 Py_ssize_t ival;
505
506 if (i < 0)
507 i += Py_SIZE(self);
508
509 if (i < 0 || i >= Py_SIZE(self)) {
510 PyErr_SetString(PyExc_IndexError, "bytearray index out of range");
511 return -1;
512 }
513
514 if (value == NULL)
515 return bytes_setslice(self, i, i+1, NULL);
516
517 ival = PyNumber_AsSsize_t(value, PyExc_ValueError);
518 if (ival == -1 && PyErr_Occurred())
519 return -1;
520
521 if (ival < 0 || ival >= 256) {
522 PyErr_SetString(PyExc_ValueError, "byte must be in range(0, 256)");
523 return -1;
524 }
525
526 self->ob_bytes[i] = ival;
527 return 0;
528}
529
530static int
531bytes_ass_subscript(PyByteArrayObject *self, PyObject *item, PyObject *values)
532{
533 Py_ssize_t start, stop, step, slicelen, needed;
534 char *bytes;
535
536 if (PyIndex_Check(item)) {
537 Py_ssize_t i = PyNumber_AsSsize_t(item, PyExc_IndexError);
538
539 if (i == -1 && PyErr_Occurred())
540 return -1;
541
542 if (i < 0)
543 i += PyByteArray_GET_SIZE(self);
544
545 if (i < 0 || i >= Py_SIZE(self)) {
546 PyErr_SetString(PyExc_IndexError, "bytearray index out of range");
547 return -1;
548 }
549
550 if (values == NULL) {
551 /* Fall through to slice assignment */
552 start = i;
553 stop = i + 1;
554 step = 1;
555 slicelen = 1;
556 }
557 else {
558 Py_ssize_t ival = PyNumber_AsSsize_t(values, PyExc_ValueError);
559 if (ival == -1 && PyErr_Occurred())
560 return -1;
561 if (ival < 0 || ival >= 256) {
562 PyErr_SetString(PyExc_ValueError,
563 "byte must be in range(0, 256)");
564 return -1;
565 }
566 self->ob_bytes[i] = (char)ival;
567 return 0;
568 }
569 }
570 else if (PySlice_Check(item)) {
571 if (PySlice_GetIndicesEx((PySliceObject *)item,
572 PyByteArray_GET_SIZE(self),
573 &start, &stop, &step, &slicelen) < 0) {
574 return -1;
575 }
576 }
577 else {
578 PyErr_SetString(PyExc_TypeError, "bytearray indices must be integer");
579 return -1;
580 }
581
582 if (values == NULL) {
583 bytes = NULL;
584 needed = 0;
585 }
586 else if (values == (PyObject *)self || !PyByteArray_Check(values)) {
587 /* Make a copy an call this function recursively */
588 int err;
589 values = PyByteArray_FromObject(values);
590 if (values == NULL)
591 return -1;
592 err = bytes_ass_subscript(self, item, values);
593 Py_DECREF(values);
594 return err;
595 }
596 else {
597 assert(PyByteArray_Check(values));
598 bytes = ((PyByteArrayObject *)values)->ob_bytes;
599 needed = Py_SIZE(values);
600 }
601 /* Make sure b[5:2] = ... inserts before 5, not before 2. */
602 if ((step < 0 && start < stop) ||
603 (step > 0 && start > stop))
604 stop = start;
605 if (step == 1) {
606 if (slicelen != needed) {
607 if (slicelen > needed) {
608 /*
609 0 start stop old_size
610 | |<---slicelen--->|<-----tomove------>|
611 | |<-needed->|<-----tomove------>|
612 0 lo new_hi new_size
613 */
614 memmove(self->ob_bytes + start + needed, self->ob_bytes + stop,
615 Py_SIZE(self) - stop);
616 }
617 if (PyByteArray_Resize((PyObject *)self,
618 Py_SIZE(self) + needed - slicelen) < 0)
619 return -1;
620 if (slicelen < needed) {
621 /*
622 0 lo hi old_size
623 | |<-avail->|<-----tomove------>|
624 | |<----needed---->|<-----tomove------>|
625 0 lo new_hi new_size
626 */
627 memmove(self->ob_bytes + start + needed, self->ob_bytes + stop,
628 Py_SIZE(self) - start - needed);
629 }
630 }
631
632 if (needed > 0)
633 memcpy(self->ob_bytes + start, bytes, needed);
634
635 return 0;
636 }
637 else {
638 if (needed == 0) {
639 /* Delete slice */
640 Py_ssize_t cur, i;
641
642 if (step < 0) {
643 stop = start + 1;
644 start = stop + step * (slicelen - 1) - 1;
645 step = -step;
646 }
647 for (cur = start, i = 0;
648 i < slicelen; cur += step, i++) {
649 Py_ssize_t lim = step - 1;
650
651 if (cur + step >= PyByteArray_GET_SIZE(self))
652 lim = PyByteArray_GET_SIZE(self) - cur - 1;
653
654 memmove(self->ob_bytes + cur - i,
655 self->ob_bytes + cur + 1, lim);
656 }
657 /* Move the tail of the bytes, in one chunk */
658 cur = start + slicelen*step;
659 if (cur < PyByteArray_GET_SIZE(self)) {
660 memmove(self->ob_bytes + cur - slicelen,
661 self->ob_bytes + cur,
662 PyByteArray_GET_SIZE(self) - cur);
663 }
664 if (PyByteArray_Resize((PyObject *)self,
665 PyByteArray_GET_SIZE(self) - slicelen) < 0)
666 return -1;
667
668 return 0;
669 }
670 else {
671 /* Assign slice */
672 Py_ssize_t cur, i;
673
674 if (needed != slicelen) {
675 PyErr_Format(PyExc_ValueError,
676 "attempt to assign bytes of size %zd "
677 "to extended slice of size %zd",
678 needed, slicelen);
679 return -1;
680 }
681 for (cur = start, i = 0; i < slicelen; cur += step, i++)
682 self->ob_bytes[cur] = bytes[i];
683 return 0;
684 }
685 }
686}
687
688static int
689bytes_init(PyByteArrayObject *self, PyObject *args, PyObject *kwds)
690{
691 static char *kwlist[] = {"source", "encoding", "errors", 0};
692 PyObject *arg = NULL;
693 const char *encoding = NULL;
694 const char *errors = NULL;
695 Py_ssize_t count;
696 PyObject *it;
697 PyObject *(*iternext)(PyObject *);
698
699 if (Py_SIZE(self) != 0) {
700 /* Empty previous contents (yes, do this first of all!) */
701 if (PyByteArray_Resize((PyObject *)self, 0) < 0)
702 return -1;
703 }
704
705 /* Parse arguments */
706 if (!PyArg_ParseTupleAndKeywords(args, kwds, "|Oss:bytes", kwlist,
707 &arg, &encoding, &errors))
708 return -1;
709
710 /* Make a quick exit if no first argument */
711 if (arg == NULL) {
712 if (encoding != NULL || errors != NULL) {
713 PyErr_SetString(PyExc_TypeError,
714 "encoding or errors without sequence argument");
715 return -1;
716 }
717 return 0;
718 }
719
720 if (PyUnicode_Check(arg)) {
721 /* Encode via the codec registry */
722 PyObject *encoded, *new;
723 if (encoding == NULL) {
724 PyErr_SetString(PyExc_TypeError,
725 "string argument without an encoding");
726 return -1;
727 }
728 encoded = PyCodec_Encode(arg, encoding, errors);
729 if (encoded == NULL)
730 return -1;
731 assert(PyBytes_Check(encoded));
732 new = bytes_iconcat(self, encoded);
733 Py_DECREF(encoded);
734 if (new == NULL)
735 return -1;
736 Py_DECREF(new);
737 return 0;
738 }
739
740 /* If it's not unicode, there can't be encoding or errors */
741 if (encoding != NULL || errors != NULL) {
742 PyErr_SetString(PyExc_TypeError,
743 "encoding or errors without a string argument");
744 return -1;
745 }
746
747 /* Is it an int? */
748 count = PyNumber_AsSsize_t(arg, PyExc_ValueError);
749 if (count == -1 && PyErr_Occurred())
750 PyErr_Clear();
751 else {
752 if (count < 0) {
753 PyErr_SetString(PyExc_ValueError, "negative count");
754 return -1;
755 }
756 if (count > 0) {
757 if (PyByteArray_Resize((PyObject *)self, count))
758 return -1;
759 memset(self->ob_bytes, 0, count);
760 }
761 return 0;
762 }
763
764 /* Use the buffer API */
765 if (PyObject_CheckBuffer(arg)) {
766 Py_ssize_t size;
767 Py_buffer view;
768 if (PyObject_GetBuffer(arg, &view, PyBUF_FULL_RO) < 0)
769 return -1;
770 size = view.len;
771 if (PyByteArray_Resize((PyObject *)self, size) < 0) goto fail;
772 if (PyBuffer_ToContiguous(self->ob_bytes, &view, size, 'C') < 0)
773 goto fail;
774 PyObject_ReleaseBuffer(arg, &view);
775 return 0;
776 fail:
777 PyObject_ReleaseBuffer(arg, &view);
778 return -1;
779 }
780
781 /* XXX Optimize this if the arguments is a list, tuple */
782
783 /* Get the iterator */
784 it = PyObject_GetIter(arg);
785 if (it == NULL)
786 return -1;
787 iternext = *Py_TYPE(it)->tp_iternext;
788
789 /* Run the iterator to exhaustion */
790 for (;;) {
791 PyObject *item;
792 Py_ssize_t value;
793
794 /* Get the next item */
795 item = iternext(it);
796 if (item == NULL) {
797 if (PyErr_Occurred()) {
798 if (!PyErr_ExceptionMatches(PyExc_StopIteration))
799 goto error;
800 PyErr_Clear();
801 }
802 break;
803 }
804
805 /* Interpret it as an int (__index__) */
806 value = PyNumber_AsSsize_t(item, PyExc_ValueError);
807 Py_DECREF(item);
808 if (value == -1 && PyErr_Occurred())
809 goto error;
810
811 /* Range check */
812 if (value < 0 || value >= 256) {
813 PyErr_SetString(PyExc_ValueError,
814 "bytes must be in range(0, 256)");
815 goto error;
816 }
817
818 /* Append the byte */
819 if (Py_SIZE(self) < self->ob_alloc)
820 Py_SIZE(self)++;
821 else if (PyByteArray_Resize((PyObject *)self, Py_SIZE(self)+1) < 0)
822 goto error;
823 self->ob_bytes[Py_SIZE(self)-1] = value;
824 }
825
826 /* Clean up and return success */
827 Py_DECREF(it);
828 return 0;
829
830 error:
831 /* Error handling when it != NULL */
832 Py_DECREF(it);
833 return -1;
834}
835
836/* Mostly copied from string_repr, but without the
837 "smart quote" functionality. */
838static PyObject *
839bytes_repr(PyByteArrayObject *self)
840{
841 static const char *hexdigits = "0123456789abcdef";
842 const char *quote_prefix = "bytearray(b";
843 const char *quote_postfix = ")";
844 Py_ssize_t length = Py_SIZE(self);
845 /* 14 == strlen(quote_prefix) + 2 + strlen(quote_postfix) */
846 size_t newsize = 14 + 4 * length;
847 PyObject *v;
848 if (newsize > PY_SSIZE_T_MAX || newsize / 4 - 3 != length) {
849 PyErr_SetString(PyExc_OverflowError,
850 "bytearray object is too large to make repr");
851 return NULL;
852 }
853 v = PyUnicode_FromUnicode(NULL, newsize);
854 if (v == NULL) {
855 return NULL;
856 }
857 else {
858 register Py_ssize_t i;
859 register Py_UNICODE c;
860 register Py_UNICODE *p;
861 int quote;
862
863 /* Figure out which quote to use; single is preferred */
864 quote = '\'';
865 {
866 char *test, *start;
867 start = PyByteArray_AS_STRING(self);
868 for (test = start; test < start+length; ++test) {
869 if (*test == '"') {
870 quote = '\''; /* back to single */
871 goto decided;
872 }
873 else if (*test == '\'')
874 quote = '"';
875 }
876 decided:
877 ;
878 }
879
880 p = PyUnicode_AS_UNICODE(v);
881 while (*quote_prefix)
882 *p++ = *quote_prefix++;
883 *p++ = quote;
884
885 for (i = 0; i < length; i++) {
886 /* There's at least enough room for a hex escape
887 and a closing quote. */
888 assert(newsize - (p - PyUnicode_AS_UNICODE(v)) >= 5);
889 c = self->ob_bytes[i];
890 if (c == '\'' || c == '\\')
891 *p++ = '\\', *p++ = c;
892 else if (c == '\t')
893 *p++ = '\\', *p++ = 't';
894 else if (c == '\n')
895 *p++ = '\\', *p++ = 'n';
896 else if (c == '\r')
897 *p++ = '\\', *p++ = 'r';
898 else if (c == 0)
899 *p++ = '\\', *p++ = 'x', *p++ = '0', *p++ = '0';
900 else if (c < ' ' || c >= 0x7f) {
901 *p++ = '\\';
902 *p++ = 'x';
903 *p++ = hexdigits[(c & 0xf0) >> 4];
904 *p++ = hexdigits[c & 0xf];
905 }
906 else
907 *p++ = c;
908 }
909 assert(newsize - (p - PyUnicode_AS_UNICODE(v)) >= 1);
910 *p++ = quote;
911 while (*quote_postfix) {
912 *p++ = *quote_postfix++;
913 }
914 *p = '\0';
915 if (PyUnicode_Resize(&v, (p - PyUnicode_AS_UNICODE(v)))) {
916 Py_DECREF(v);
917 return NULL;
918 }
919 return v;
920 }
921}
922
923static PyObject *
924bytes_str(PyObject *op)
925{
926 if (Py_BytesWarningFlag) {
927 if (PyErr_WarnEx(PyExc_BytesWarning,
928 "str() on a bytearray instance", 1))
929 return NULL;
930 }
931 return bytes_repr((PyByteArrayObject*)op);
932}
933
934static PyObject *
935bytes_richcompare(PyObject *self, PyObject *other, int op)
936{
937 Py_ssize_t self_size, other_size;
938 Py_buffer self_bytes, other_bytes;
939 PyObject *res;
940 Py_ssize_t minsize;
941 int cmp;
942
943 /* Bytes can be compared to anything that supports the (binary)
944 buffer API. Except that a comparison with Unicode is always an
945 error, even if the comparison is for equality. */
946 if (PyObject_IsInstance(self, (PyObject*)&PyUnicode_Type) ||
947 PyObject_IsInstance(other, (PyObject*)&PyUnicode_Type)) {
948 if (Py_BytesWarningFlag && op == Py_EQ) {
949 if (PyErr_WarnEx(PyExc_BytesWarning,
950 "Comparsion between bytearray and string", 1))
951 return NULL;
952 }
953
954 Py_INCREF(Py_NotImplemented);
955 return Py_NotImplemented;
956 }
957
958 self_size = _getbuffer(self, &self_bytes);
959 if (self_size < 0) {
960 PyErr_Clear();
961 Py_INCREF(Py_NotImplemented);
962 return Py_NotImplemented;
963 }
964
965 other_size = _getbuffer(other, &other_bytes);
966 if (other_size < 0) {
967 PyErr_Clear();
968 PyObject_ReleaseBuffer(self, &self_bytes);
969 Py_INCREF(Py_NotImplemented);
970 return Py_NotImplemented;
971 }
972
973 if (self_size != other_size && (op == Py_EQ || op == Py_NE)) {
974 /* Shortcut: if the lengths differ, the objects differ */
975 cmp = (op == Py_NE);
976 }
977 else {
978 minsize = self_size;
979 if (other_size < minsize)
980 minsize = other_size;
981
982 cmp = memcmp(self_bytes.buf, other_bytes.buf, minsize);
983 /* In ISO C, memcmp() guarantees to use unsigned bytes! */
984
985 if (cmp == 0) {
986 if (self_size < other_size)
987 cmp = -1;
988 else if (self_size > other_size)
989 cmp = 1;
990 }
991
992 switch (op) {
993 case Py_LT: cmp = cmp < 0; break;
994 case Py_LE: cmp = cmp <= 0; break;
995 case Py_EQ: cmp = cmp == 0; break;
996 case Py_NE: cmp = cmp != 0; break;
997 case Py_GT: cmp = cmp > 0; break;
998 case Py_GE: cmp = cmp >= 0; break;
999 }
1000 }
1001
1002 res = cmp ? Py_True : Py_False;
1003 PyObject_ReleaseBuffer(self, &self_bytes);
1004 PyObject_ReleaseBuffer(other, &other_bytes);
1005 Py_INCREF(res);
1006 return res;
1007}
1008
1009static void
1010bytes_dealloc(PyByteArrayObject *self)
1011{
1012 if (self->ob_bytes != 0) {
1013 PyMem_Free(self->ob_bytes);
1014 }
1015 Py_TYPE(self)->tp_free((PyObject *)self);
1016}
1017
1018
1019/* -------------------------------------------------------------------- */
1020/* Methods */
1021
1022#define STRINGLIB_CHAR char
1023#define STRINGLIB_CMP memcmp
1024#define STRINGLIB_LEN PyByteArray_GET_SIZE
1025#define STRINGLIB_STR PyByteArray_AS_STRING
1026#define STRINGLIB_NEW PyByteArray_FromStringAndSize
1027#define STRINGLIB_EMPTY nullbytes
1028#define STRINGLIB_CHECK_EXACT PyByteArray_CheckExact
1029#define STRINGLIB_MUTABLE 1
1030
1031#include "stringlib/fastsearch.h"
1032#include "stringlib/count.h"
1033#include "stringlib/find.h"
1034#include "stringlib/partition.h"
1035#include "stringlib/ctype.h"
1036#include "stringlib/transmogrify.h"
1037
1038
1039/* The following Py_LOCAL_INLINE and Py_LOCAL functions
1040were copied from the old char* style string object. */
1041
1042Py_LOCAL_INLINE(void)
1043_adjust_indices(Py_ssize_t *start, Py_ssize_t *end, Py_ssize_t len)
1044{
1045 if (*end > len)
1046 *end = len;
1047 else if (*end < 0)
1048 *end += len;
1049 if (*end < 0)
1050 *end = 0;
1051 if (*start < 0)
1052 *start += len;
1053 if (*start < 0)
1054 *start = 0;
1055}
1056
1057
1058Py_LOCAL_INLINE(Py_ssize_t)
1059bytes_find_internal(PyByteArrayObject *self, PyObject *args, int dir)
1060{
1061 PyObject *subobj;
1062 Py_buffer subbuf;
1063 Py_ssize_t start=0, end=PY_SSIZE_T_MAX;
1064 Py_ssize_t res;
1065
1066 if (!PyArg_ParseTuple(args, "O|O&O&:find/rfind/index/rindex", &subobj,
1067 _PyEval_SliceIndex, &start, _PyEval_SliceIndex, &end))
1068 return -2;
1069 if (_getbuffer(subobj, &subbuf) < 0)
1070 return -2;
1071 if (dir > 0)
1072 res = stringlib_find_slice(
1073 PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
1074 subbuf.buf, subbuf.len, start, end);
1075 else
1076 res = stringlib_rfind_slice(
1077 PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
1078 subbuf.buf, subbuf.len, start, end);
1079 PyObject_ReleaseBuffer(subobj, &subbuf);
1080 return res;
1081}
1082
1083PyDoc_STRVAR(find__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00001084"B.find(sub[, start[, end]]) -> int\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00001085\n\
1086Return the lowest index in B where subsection sub is found,\n\
1087such that sub is contained within s[start,end]. Optional\n\
1088arguments start and end are interpreted as in slice notation.\n\
1089\n\
1090Return -1 on failure.");
1091
1092static PyObject *
1093bytes_find(PyByteArrayObject *self, PyObject *args)
1094{
1095 Py_ssize_t result = bytes_find_internal(self, args, +1);
1096 if (result == -2)
1097 return NULL;
1098 return PyLong_FromSsize_t(result);
1099}
1100
1101PyDoc_STRVAR(count__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00001102"B.count(sub[, start[, end]]) -> int\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00001103\n\
1104Return the number of non-overlapping occurrences of subsection sub in\n\
1105bytes B[start:end]. Optional arguments start and end are interpreted\n\
1106as in slice notation.");
1107
1108static PyObject *
1109bytes_count(PyByteArrayObject *self, PyObject *args)
1110{
1111 PyObject *sub_obj;
1112 const char *str = PyByteArray_AS_STRING(self);
1113 Py_ssize_t start = 0, end = PY_SSIZE_T_MAX;
1114 Py_buffer vsub;
1115 PyObject *count_obj;
1116
1117 if (!PyArg_ParseTuple(args, "O|O&O&:count", &sub_obj,
1118 _PyEval_SliceIndex, &start, _PyEval_SliceIndex, &end))
1119 return NULL;
1120
1121 if (_getbuffer(sub_obj, &vsub) < 0)
1122 return NULL;
1123
1124 _adjust_indices(&start, &end, PyByteArray_GET_SIZE(self));
1125
1126 count_obj = PyLong_FromSsize_t(
1127 stringlib_count(str + start, end - start, vsub.buf, vsub.len)
1128 );
1129 PyObject_ReleaseBuffer(sub_obj, &vsub);
1130 return count_obj;
1131}
1132
1133
1134PyDoc_STRVAR(index__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00001135"B.index(sub[, start[, end]]) -> int\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00001136\n\
1137Like B.find() but raise ValueError when the subsection is not found.");
1138
1139static PyObject *
1140bytes_index(PyByteArrayObject *self, PyObject *args)
1141{
1142 Py_ssize_t result = bytes_find_internal(self, args, +1);
1143 if (result == -2)
1144 return NULL;
1145 if (result == -1) {
1146 PyErr_SetString(PyExc_ValueError,
1147 "subsection not found");
1148 return NULL;
1149 }
1150 return PyLong_FromSsize_t(result);
1151}
1152
1153
1154PyDoc_STRVAR(rfind__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00001155"B.rfind(sub[, start[, end]]) -> int\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00001156\n\
1157Return the highest index in B where subsection sub is found,\n\
1158such that sub is contained within s[start,end]. Optional\n\
1159arguments start and end are interpreted as in slice notation.\n\
1160\n\
1161Return -1 on failure.");
1162
1163static PyObject *
1164bytes_rfind(PyByteArrayObject *self, PyObject *args)
1165{
1166 Py_ssize_t result = bytes_find_internal(self, args, -1);
1167 if (result == -2)
1168 return NULL;
1169 return PyLong_FromSsize_t(result);
1170}
1171
1172
1173PyDoc_STRVAR(rindex__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00001174"B.rindex(sub[, start[, end]]) -> int\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00001175\n\
1176Like B.rfind() but raise ValueError when the subsection is not found.");
1177
1178static PyObject *
1179bytes_rindex(PyByteArrayObject *self, PyObject *args)
1180{
1181 Py_ssize_t result = bytes_find_internal(self, args, -1);
1182 if (result == -2)
1183 return NULL;
1184 if (result == -1) {
1185 PyErr_SetString(PyExc_ValueError,
1186 "subsection not found");
1187 return NULL;
1188 }
1189 return PyLong_FromSsize_t(result);
1190}
1191
1192
1193static int
1194bytes_contains(PyObject *self, PyObject *arg)
1195{
1196 Py_ssize_t ival = PyNumber_AsSsize_t(arg, PyExc_ValueError);
1197 if (ival == -1 && PyErr_Occurred()) {
1198 Py_buffer varg;
1199 int pos;
1200 PyErr_Clear();
1201 if (_getbuffer(arg, &varg) < 0)
1202 return -1;
1203 pos = stringlib_find(PyByteArray_AS_STRING(self), Py_SIZE(self),
1204 varg.buf, varg.len, 0);
1205 PyObject_ReleaseBuffer(arg, &varg);
1206 return pos >= 0;
1207 }
1208 if (ival < 0 || ival >= 256) {
1209 PyErr_SetString(PyExc_ValueError, "byte must be in range(0, 256)");
1210 return -1;
1211 }
1212
1213 return memchr(PyByteArray_AS_STRING(self), ival, Py_SIZE(self)) != NULL;
1214}
1215
1216
1217/* Matches the end (direction >= 0) or start (direction < 0) of self
1218 * against substr, using the start and end arguments. Returns
1219 * -1 on error, 0 if not found and 1 if found.
1220 */
1221Py_LOCAL(int)
1222_bytes_tailmatch(PyByteArrayObject *self, PyObject *substr, Py_ssize_t start,
1223 Py_ssize_t end, int direction)
1224{
1225 Py_ssize_t len = PyByteArray_GET_SIZE(self);
1226 const char* str;
1227 Py_buffer vsubstr;
1228 int rv = 0;
1229
1230 str = PyByteArray_AS_STRING(self);
1231
1232 if (_getbuffer(substr, &vsubstr) < 0)
1233 return -1;
1234
1235 _adjust_indices(&start, &end, len);
1236
1237 if (direction < 0) {
1238 /* startswith */
1239 if (start+vsubstr.len > len) {
1240 goto done;
1241 }
1242 } else {
1243 /* endswith */
1244 if (end-start < vsubstr.len || start > len) {
1245 goto done;
1246 }
1247
1248 if (end-vsubstr.len > start)
1249 start = end - vsubstr.len;
1250 }
1251 if (end-start >= vsubstr.len)
1252 rv = ! memcmp(str+start, vsubstr.buf, vsubstr.len);
1253
1254done:
1255 PyObject_ReleaseBuffer(substr, &vsubstr);
1256 return rv;
1257}
1258
1259
1260PyDoc_STRVAR(startswith__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00001261"B.startswith(prefix[, start[, end]]) -> bool\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00001262\n\
1263Return True if B starts with the specified prefix, False otherwise.\n\
1264With optional start, test B beginning at that position.\n\
1265With optional end, stop comparing B at that position.\n\
1266prefix can also be a tuple of strings to try.");
1267
1268static PyObject *
1269bytes_startswith(PyByteArrayObject *self, PyObject *args)
1270{
1271 Py_ssize_t start = 0;
1272 Py_ssize_t end = PY_SSIZE_T_MAX;
1273 PyObject *subobj;
1274 int result;
1275
1276 if (!PyArg_ParseTuple(args, "O|O&O&:startswith", &subobj,
1277 _PyEval_SliceIndex, &start, _PyEval_SliceIndex, &end))
1278 return NULL;
1279 if (PyTuple_Check(subobj)) {
1280 Py_ssize_t i;
1281 for (i = 0; i < PyTuple_GET_SIZE(subobj); i++) {
1282 result = _bytes_tailmatch(self,
1283 PyTuple_GET_ITEM(subobj, i),
1284 start, end, -1);
1285 if (result == -1)
1286 return NULL;
1287 else if (result) {
1288 Py_RETURN_TRUE;
1289 }
1290 }
1291 Py_RETURN_FALSE;
1292 }
1293 result = _bytes_tailmatch(self, subobj, start, end, -1);
1294 if (result == -1)
1295 return NULL;
1296 else
1297 return PyBool_FromLong(result);
1298}
1299
1300PyDoc_STRVAR(endswith__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00001301"B.endswith(suffix[, start[, end]]) -> bool\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00001302\n\
1303Return True if B ends with the specified suffix, False otherwise.\n\
1304With optional start, test B beginning at that position.\n\
1305With optional end, stop comparing B at that position.\n\
1306suffix can also be a tuple of strings to try.");
1307
1308static PyObject *
1309bytes_endswith(PyByteArrayObject *self, PyObject *args)
1310{
1311 Py_ssize_t start = 0;
1312 Py_ssize_t end = PY_SSIZE_T_MAX;
1313 PyObject *subobj;
1314 int result;
1315
1316 if (!PyArg_ParseTuple(args, "O|O&O&:endswith", &subobj,
1317 _PyEval_SliceIndex, &start, _PyEval_SliceIndex, &end))
1318 return NULL;
1319 if (PyTuple_Check(subobj)) {
1320 Py_ssize_t i;
1321 for (i = 0; i < PyTuple_GET_SIZE(subobj); i++) {
1322 result = _bytes_tailmatch(self,
1323 PyTuple_GET_ITEM(subobj, i),
1324 start, end, +1);
1325 if (result == -1)
1326 return NULL;
1327 else if (result) {
1328 Py_RETURN_TRUE;
1329 }
1330 }
1331 Py_RETURN_FALSE;
1332 }
1333 result = _bytes_tailmatch(self, subobj, start, end, +1);
1334 if (result == -1)
1335 return NULL;
1336 else
1337 return PyBool_FromLong(result);
1338}
1339
1340
1341PyDoc_STRVAR(translate__doc__,
1342"B.translate(table[, deletechars]) -> bytearray\n\
1343\n\
1344Return a copy of B, where all characters occurring in the\n\
1345optional argument deletechars are removed, and the remaining\n\
1346characters have been mapped through the given translation\n\
1347table, which must be a bytes object of length 256.");
1348
1349static PyObject *
1350bytes_translate(PyByteArrayObject *self, PyObject *args)
1351{
1352 register char *input, *output;
1353 register const char *table;
1354 register Py_ssize_t i, c, changed = 0;
1355 PyObject *input_obj = (PyObject*)self;
1356 const char *output_start;
1357 Py_ssize_t inlen;
1358 PyObject *result;
1359 int trans_table[256];
1360 PyObject *tableobj, *delobj = NULL;
1361 Py_buffer vtable, vdel;
1362
1363 if (!PyArg_UnpackTuple(args, "translate", 1, 2,
1364 &tableobj, &delobj))
1365 return NULL;
1366
1367 if (_getbuffer(tableobj, &vtable) < 0)
1368 return NULL;
1369
1370 if (vtable.len != 256) {
1371 PyErr_SetString(PyExc_ValueError,
1372 "translation table must be 256 characters long");
1373 result = NULL;
1374 goto done;
1375 }
1376
1377 if (delobj != NULL) {
1378 if (_getbuffer(delobj, &vdel) < 0) {
1379 result = NULL;
1380 goto done;
1381 }
1382 }
1383 else {
1384 vdel.buf = NULL;
1385 vdel.len = 0;
1386 }
1387
1388 table = (const char *)vtable.buf;
1389 inlen = PyByteArray_GET_SIZE(input_obj);
1390 result = PyByteArray_FromStringAndSize((char *)NULL, inlen);
1391 if (result == NULL)
1392 goto done;
1393 output_start = output = PyByteArray_AsString(result);
1394 input = PyByteArray_AS_STRING(input_obj);
1395
1396 if (vdel.len == 0) {
1397 /* If no deletions are required, use faster code */
1398 for (i = inlen; --i >= 0; ) {
1399 c = Py_CHARMASK(*input++);
1400 if (Py_CHARMASK((*output++ = table[c])) != c)
1401 changed = 1;
1402 }
1403 if (changed || !PyByteArray_CheckExact(input_obj))
1404 goto done;
1405 Py_DECREF(result);
1406 Py_INCREF(input_obj);
1407 result = input_obj;
1408 goto done;
1409 }
1410
1411 for (i = 0; i < 256; i++)
1412 trans_table[i] = Py_CHARMASK(table[i]);
1413
1414 for (i = 0; i < vdel.len; i++)
1415 trans_table[(int) Py_CHARMASK( ((unsigned char*)vdel.buf)[i] )] = -1;
1416
1417 for (i = inlen; --i >= 0; ) {
1418 c = Py_CHARMASK(*input++);
1419 if (trans_table[c] != -1)
1420 if (Py_CHARMASK(*output++ = (char)trans_table[c]) == c)
1421 continue;
1422 changed = 1;
1423 }
1424 if (!changed && PyByteArray_CheckExact(input_obj)) {
1425 Py_DECREF(result);
1426 Py_INCREF(input_obj);
1427 result = input_obj;
1428 goto done;
1429 }
1430 /* Fix the size of the resulting string */
1431 if (inlen > 0)
1432 PyByteArray_Resize(result, output - output_start);
1433
1434done:
1435 PyObject_ReleaseBuffer(tableobj, &vtable);
1436 if (delobj != NULL)
1437 PyObject_ReleaseBuffer(delobj, &vdel);
1438 return result;
1439}
1440
1441
1442#define FORWARD 1
1443#define REVERSE -1
1444
1445/* find and count characters and substrings */
1446
1447#define findchar(target, target_len, c) \
1448 ((char *)memchr((const void *)(target), c, target_len))
1449
1450/* Don't call if length < 2 */
1451#define Py_STRING_MATCH(target, offset, pattern, length) \
1452 (target[offset] == pattern[0] && \
1453 target[offset+length-1] == pattern[length-1] && \
1454 !memcmp(target+offset+1, pattern+1, length-2) )
1455
1456
1457/* Bytes ops must return a string. */
1458/* If the object is subclass of bytes, create a copy */
1459Py_LOCAL(PyByteArrayObject *)
1460return_self(PyByteArrayObject *self)
1461{
1462 if (PyByteArray_CheckExact(self)) {
1463 Py_INCREF(self);
1464 return (PyByteArrayObject *)self;
1465 }
1466 return (PyByteArrayObject *)PyByteArray_FromStringAndSize(
1467 PyByteArray_AS_STRING(self),
1468 PyByteArray_GET_SIZE(self));
1469}
1470
1471Py_LOCAL_INLINE(Py_ssize_t)
1472countchar(const char *target, Py_ssize_t target_len, char c, Py_ssize_t maxcount)
1473{
1474 Py_ssize_t count=0;
1475 const char *start=target;
1476 const char *end=target+target_len;
1477
1478 while ( (start=findchar(start, end-start, c)) != NULL ) {
1479 count++;
1480 if (count >= maxcount)
1481 break;
1482 start += 1;
1483 }
1484 return count;
1485}
1486
1487Py_LOCAL(Py_ssize_t)
1488findstring(const char *target, Py_ssize_t target_len,
1489 const char *pattern, Py_ssize_t pattern_len,
1490 Py_ssize_t start,
1491 Py_ssize_t end,
1492 int direction)
1493{
1494 if (start < 0) {
1495 start += target_len;
1496 if (start < 0)
1497 start = 0;
1498 }
1499 if (end > target_len) {
1500 end = target_len;
1501 } else if (end < 0) {
1502 end += target_len;
1503 if (end < 0)
1504 end = 0;
1505 }
1506
1507 /* zero-length substrings always match at the first attempt */
1508 if (pattern_len == 0)
1509 return (direction > 0) ? start : end;
1510
1511 end -= pattern_len;
1512
1513 if (direction < 0) {
1514 for (; end >= start; end--)
1515 if (Py_STRING_MATCH(target, end, pattern, pattern_len))
1516 return end;
1517 } else {
1518 for (; start <= end; start++)
1519 if (Py_STRING_MATCH(target, start, pattern, pattern_len))
1520 return start;
1521 }
1522 return -1;
1523}
1524
1525Py_LOCAL_INLINE(Py_ssize_t)
1526countstring(const char *target, Py_ssize_t target_len,
1527 const char *pattern, Py_ssize_t pattern_len,
1528 Py_ssize_t start,
1529 Py_ssize_t end,
1530 int direction, Py_ssize_t maxcount)
1531{
1532 Py_ssize_t count=0;
1533
1534 if (start < 0) {
1535 start += target_len;
1536 if (start < 0)
1537 start = 0;
1538 }
1539 if (end > target_len) {
1540 end = target_len;
1541 } else if (end < 0) {
1542 end += target_len;
1543 if (end < 0)
1544 end = 0;
1545 }
1546
1547 /* zero-length substrings match everywhere */
1548 if (pattern_len == 0 || maxcount == 0) {
1549 if (target_len+1 < maxcount)
1550 return target_len+1;
1551 return maxcount;
1552 }
1553
1554 end -= pattern_len;
1555 if (direction < 0) {
1556 for (; (end >= start); end--)
1557 if (Py_STRING_MATCH(target, end, pattern, pattern_len)) {
1558 count++;
1559 if (--maxcount <= 0) break;
1560 end -= pattern_len-1;
1561 }
1562 } else {
1563 for (; (start <= end); start++)
1564 if (Py_STRING_MATCH(target, start, pattern, pattern_len)) {
1565 count++;
1566 if (--maxcount <= 0)
1567 break;
1568 start += pattern_len-1;
1569 }
1570 }
1571 return count;
1572}
1573
1574
1575/* Algorithms for different cases of string replacement */
1576
1577/* len(self)>=1, from="", len(to)>=1, maxcount>=1 */
1578Py_LOCAL(PyByteArrayObject *)
1579replace_interleave(PyByteArrayObject *self,
1580 const char *to_s, Py_ssize_t to_len,
1581 Py_ssize_t maxcount)
1582{
1583 char *self_s, *result_s;
1584 Py_ssize_t self_len, result_len;
1585 Py_ssize_t count, i, product;
1586 PyByteArrayObject *result;
1587
1588 self_len = PyByteArray_GET_SIZE(self);
1589
1590 /* 1 at the end plus 1 after every character */
1591 count = self_len+1;
1592 if (maxcount < count)
1593 count = maxcount;
1594
1595 /* Check for overflow */
1596 /* result_len = count * to_len + self_len; */
1597 product = count * to_len;
1598 if (product / to_len != count) {
1599 PyErr_SetString(PyExc_OverflowError,
1600 "replace string is too long");
1601 return NULL;
1602 }
1603 result_len = product + self_len;
1604 if (result_len < 0) {
1605 PyErr_SetString(PyExc_OverflowError,
1606 "replace string is too long");
1607 return NULL;
1608 }
1609
1610 if (! (result = (PyByteArrayObject *)
1611 PyByteArray_FromStringAndSize(NULL, result_len)) )
1612 return NULL;
1613
1614 self_s = PyByteArray_AS_STRING(self);
1615 result_s = PyByteArray_AS_STRING(result);
1616
1617 /* TODO: special case single character, which doesn't need memcpy */
1618
1619 /* Lay the first one down (guaranteed this will occur) */
1620 Py_MEMCPY(result_s, to_s, to_len);
1621 result_s += to_len;
1622 count -= 1;
1623
1624 for (i=0; i<count; i++) {
1625 *result_s++ = *self_s++;
1626 Py_MEMCPY(result_s, to_s, to_len);
1627 result_s += to_len;
1628 }
1629
1630 /* Copy the rest of the original string */
1631 Py_MEMCPY(result_s, self_s, self_len-i);
1632
1633 return result;
1634}
1635
1636/* Special case for deleting a single character */
1637/* len(self)>=1, len(from)==1, to="", maxcount>=1 */
1638Py_LOCAL(PyByteArrayObject *)
1639replace_delete_single_character(PyByteArrayObject *self,
1640 char from_c, Py_ssize_t maxcount)
1641{
1642 char *self_s, *result_s;
1643 char *start, *next, *end;
1644 Py_ssize_t self_len, result_len;
1645 Py_ssize_t count;
1646 PyByteArrayObject *result;
1647
1648 self_len = PyByteArray_GET_SIZE(self);
1649 self_s = PyByteArray_AS_STRING(self);
1650
1651 count = countchar(self_s, self_len, from_c, maxcount);
1652 if (count == 0) {
1653 return return_self(self);
1654 }
1655
1656 result_len = self_len - count; /* from_len == 1 */
1657 assert(result_len>=0);
1658
1659 if ( (result = (PyByteArrayObject *)
1660 PyByteArray_FromStringAndSize(NULL, result_len)) == NULL)
1661 return NULL;
1662 result_s = PyByteArray_AS_STRING(result);
1663
1664 start = self_s;
1665 end = self_s + self_len;
1666 while (count-- > 0) {
1667 next = findchar(start, end-start, from_c);
1668 if (next == NULL)
1669 break;
1670 Py_MEMCPY(result_s, start, next-start);
1671 result_s += (next-start);
1672 start = next+1;
1673 }
1674 Py_MEMCPY(result_s, start, end-start);
1675
1676 return result;
1677}
1678
1679/* len(self)>=1, len(from)>=2, to="", maxcount>=1 */
1680
1681Py_LOCAL(PyByteArrayObject *)
1682replace_delete_substring(PyByteArrayObject *self,
1683 const char *from_s, Py_ssize_t from_len,
1684 Py_ssize_t maxcount)
1685{
1686 char *self_s, *result_s;
1687 char *start, *next, *end;
1688 Py_ssize_t self_len, result_len;
1689 Py_ssize_t count, offset;
1690 PyByteArrayObject *result;
1691
1692 self_len = PyByteArray_GET_SIZE(self);
1693 self_s = PyByteArray_AS_STRING(self);
1694
1695 count = countstring(self_s, self_len,
1696 from_s, from_len,
1697 0, self_len, 1,
1698 maxcount);
1699
1700 if (count == 0) {
1701 /* no matches */
1702 return return_self(self);
1703 }
1704
1705 result_len = self_len - (count * from_len);
1706 assert (result_len>=0);
1707
1708 if ( (result = (PyByteArrayObject *)
1709 PyByteArray_FromStringAndSize(NULL, result_len)) == NULL )
1710 return NULL;
1711
1712 result_s = PyByteArray_AS_STRING(result);
1713
1714 start = self_s;
1715 end = self_s + self_len;
1716 while (count-- > 0) {
1717 offset = findstring(start, end-start,
1718 from_s, from_len,
1719 0, end-start, FORWARD);
1720 if (offset == -1)
1721 break;
1722 next = start + offset;
1723
1724 Py_MEMCPY(result_s, start, next-start);
1725
1726 result_s += (next-start);
1727 start = next+from_len;
1728 }
1729 Py_MEMCPY(result_s, start, end-start);
1730 return result;
1731}
1732
1733/* len(self)>=1, len(from)==len(to)==1, maxcount>=1 */
1734Py_LOCAL(PyByteArrayObject *)
1735replace_single_character_in_place(PyByteArrayObject *self,
1736 char from_c, char to_c,
1737 Py_ssize_t maxcount)
1738{
1739 char *self_s, *result_s, *start, *end, *next;
1740 Py_ssize_t self_len;
1741 PyByteArrayObject *result;
1742
1743 /* The result string will be the same size */
1744 self_s = PyByteArray_AS_STRING(self);
1745 self_len = PyByteArray_GET_SIZE(self);
1746
1747 next = findchar(self_s, self_len, from_c);
1748
1749 if (next == NULL) {
1750 /* No matches; return the original bytes */
1751 return return_self(self);
1752 }
1753
1754 /* Need to make a new bytes */
1755 result = (PyByteArrayObject *) PyByteArray_FromStringAndSize(NULL, self_len);
1756 if (result == NULL)
1757 return NULL;
1758 result_s = PyByteArray_AS_STRING(result);
1759 Py_MEMCPY(result_s, self_s, self_len);
1760
1761 /* change everything in-place, starting with this one */
1762 start = result_s + (next-self_s);
1763 *start = to_c;
1764 start++;
1765 end = result_s + self_len;
1766
1767 while (--maxcount > 0) {
1768 next = findchar(start, end-start, from_c);
1769 if (next == NULL)
1770 break;
1771 *next = to_c;
1772 start = next+1;
1773 }
1774
1775 return result;
1776}
1777
1778/* len(self)>=1, len(from)==len(to)>=2, maxcount>=1 */
1779Py_LOCAL(PyByteArrayObject *)
1780replace_substring_in_place(PyByteArrayObject *self,
1781 const char *from_s, Py_ssize_t from_len,
1782 const char *to_s, Py_ssize_t to_len,
1783 Py_ssize_t maxcount)
1784{
1785 char *result_s, *start, *end;
1786 char *self_s;
1787 Py_ssize_t self_len, offset;
1788 PyByteArrayObject *result;
1789
1790 /* The result bytes will be the same size */
1791
1792 self_s = PyByteArray_AS_STRING(self);
1793 self_len = PyByteArray_GET_SIZE(self);
1794
1795 offset = findstring(self_s, self_len,
1796 from_s, from_len,
1797 0, self_len, FORWARD);
1798 if (offset == -1) {
1799 /* No matches; return the original bytes */
1800 return return_self(self);
1801 }
1802
1803 /* Need to make a new bytes */
1804 result = (PyByteArrayObject *) PyByteArray_FromStringAndSize(NULL, self_len);
1805 if (result == NULL)
1806 return NULL;
1807 result_s = PyByteArray_AS_STRING(result);
1808 Py_MEMCPY(result_s, self_s, self_len);
1809
1810 /* change everything in-place, starting with this one */
1811 start = result_s + offset;
1812 Py_MEMCPY(start, to_s, from_len);
1813 start += from_len;
1814 end = result_s + self_len;
1815
1816 while ( --maxcount > 0) {
1817 offset = findstring(start, end-start,
1818 from_s, from_len,
1819 0, end-start, FORWARD);
1820 if (offset==-1)
1821 break;
1822 Py_MEMCPY(start+offset, to_s, from_len);
1823 start += offset+from_len;
1824 }
1825
1826 return result;
1827}
1828
1829/* len(self)>=1, len(from)==1, len(to)>=2, maxcount>=1 */
1830Py_LOCAL(PyByteArrayObject *)
1831replace_single_character(PyByteArrayObject *self,
1832 char from_c,
1833 const char *to_s, Py_ssize_t to_len,
1834 Py_ssize_t maxcount)
1835{
1836 char *self_s, *result_s;
1837 char *start, *next, *end;
1838 Py_ssize_t self_len, result_len;
1839 Py_ssize_t count, product;
1840 PyByteArrayObject *result;
1841
1842 self_s = PyByteArray_AS_STRING(self);
1843 self_len = PyByteArray_GET_SIZE(self);
1844
1845 count = countchar(self_s, self_len, from_c, maxcount);
1846 if (count == 0) {
1847 /* no matches, return unchanged */
1848 return return_self(self);
1849 }
1850
1851 /* use the difference between current and new, hence the "-1" */
1852 /* result_len = self_len + count * (to_len-1) */
1853 product = count * (to_len-1);
1854 if (product / (to_len-1) != count) {
1855 PyErr_SetString(PyExc_OverflowError, "replace bytes is too long");
1856 return NULL;
1857 }
1858 result_len = self_len + product;
1859 if (result_len < 0) {
1860 PyErr_SetString(PyExc_OverflowError, "replace bytes is too long");
1861 return NULL;
1862 }
1863
1864 if ( (result = (PyByteArrayObject *)
1865 PyByteArray_FromStringAndSize(NULL, result_len)) == NULL)
1866 return NULL;
1867 result_s = PyByteArray_AS_STRING(result);
1868
1869 start = self_s;
1870 end = self_s + self_len;
1871 while (count-- > 0) {
1872 next = findchar(start, end-start, from_c);
1873 if (next == NULL)
1874 break;
1875
1876 if (next == start) {
1877 /* replace with the 'to' */
1878 Py_MEMCPY(result_s, to_s, to_len);
1879 result_s += to_len;
1880 start += 1;
1881 } else {
1882 /* copy the unchanged old then the 'to' */
1883 Py_MEMCPY(result_s, start, next-start);
1884 result_s += (next-start);
1885 Py_MEMCPY(result_s, to_s, to_len);
1886 result_s += to_len;
1887 start = next+1;
1888 }
1889 }
1890 /* Copy the remainder of the remaining bytes */
1891 Py_MEMCPY(result_s, start, end-start);
1892
1893 return result;
1894}
1895
1896/* len(self)>=1, len(from)>=2, len(to)>=2, maxcount>=1 */
1897Py_LOCAL(PyByteArrayObject *)
1898replace_substring(PyByteArrayObject *self,
1899 const char *from_s, Py_ssize_t from_len,
1900 const char *to_s, Py_ssize_t to_len,
1901 Py_ssize_t maxcount)
1902{
1903 char *self_s, *result_s;
1904 char *start, *next, *end;
1905 Py_ssize_t self_len, result_len;
1906 Py_ssize_t count, offset, product;
1907 PyByteArrayObject *result;
1908
1909 self_s = PyByteArray_AS_STRING(self);
1910 self_len = PyByteArray_GET_SIZE(self);
1911
1912 count = countstring(self_s, self_len,
1913 from_s, from_len,
1914 0, self_len, FORWARD, maxcount);
1915 if (count == 0) {
1916 /* no matches, return unchanged */
1917 return return_self(self);
1918 }
1919
1920 /* Check for overflow */
1921 /* result_len = self_len + count * (to_len-from_len) */
1922 product = count * (to_len-from_len);
1923 if (product / (to_len-from_len) != count) {
1924 PyErr_SetString(PyExc_OverflowError, "replace bytes is too long");
1925 return NULL;
1926 }
1927 result_len = self_len + product;
1928 if (result_len < 0) {
1929 PyErr_SetString(PyExc_OverflowError, "replace bytes is too long");
1930 return NULL;
1931 }
1932
1933 if ( (result = (PyByteArrayObject *)
1934 PyByteArray_FromStringAndSize(NULL, result_len)) == NULL)
1935 return NULL;
1936 result_s = PyByteArray_AS_STRING(result);
1937
1938 start = self_s;
1939 end = self_s + self_len;
1940 while (count-- > 0) {
1941 offset = findstring(start, end-start,
1942 from_s, from_len,
1943 0, end-start, FORWARD);
1944 if (offset == -1)
1945 break;
1946 next = start+offset;
1947 if (next == start) {
1948 /* replace with the 'to' */
1949 Py_MEMCPY(result_s, to_s, to_len);
1950 result_s += to_len;
1951 start += from_len;
1952 } else {
1953 /* copy the unchanged old then the 'to' */
1954 Py_MEMCPY(result_s, start, next-start);
1955 result_s += (next-start);
1956 Py_MEMCPY(result_s, to_s, to_len);
1957 result_s += to_len;
1958 start = next+from_len;
1959 }
1960 }
1961 /* Copy the remainder of the remaining bytes */
1962 Py_MEMCPY(result_s, start, end-start);
1963
1964 return result;
1965}
1966
1967
1968Py_LOCAL(PyByteArrayObject *)
1969replace(PyByteArrayObject *self,
1970 const char *from_s, Py_ssize_t from_len,
1971 const char *to_s, Py_ssize_t to_len,
1972 Py_ssize_t maxcount)
1973{
1974 if (maxcount < 0) {
1975 maxcount = PY_SSIZE_T_MAX;
1976 } else if (maxcount == 0 || PyByteArray_GET_SIZE(self) == 0) {
1977 /* nothing to do; return the original bytes */
1978 return return_self(self);
1979 }
1980
1981 if (maxcount == 0 ||
1982 (from_len == 0 && to_len == 0)) {
1983 /* nothing to do; return the original bytes */
1984 return return_self(self);
1985 }
1986
1987 /* Handle zero-length special cases */
1988
1989 if (from_len == 0) {
1990 /* insert the 'to' bytes everywhere. */
1991 /* >>> "Python".replace("", ".") */
1992 /* '.P.y.t.h.o.n.' */
1993 return replace_interleave(self, to_s, to_len, maxcount);
1994 }
1995
1996 /* Except for "".replace("", "A") == "A" there is no way beyond this */
1997 /* point for an empty self bytes to generate a non-empty bytes */
1998 /* Special case so the remaining code always gets a non-empty bytes */
1999 if (PyByteArray_GET_SIZE(self) == 0) {
2000 return return_self(self);
2001 }
2002
2003 if (to_len == 0) {
Georg Brandl17cb8a82008-05-30 08:20:09 +00002004 /* delete all occurrences of 'from' bytes */
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002005 if (from_len == 1) {
2006 return replace_delete_single_character(
2007 self, from_s[0], maxcount);
2008 } else {
2009 return replace_delete_substring(self, from_s, from_len, maxcount);
2010 }
2011 }
2012
2013 /* Handle special case where both bytes have the same length */
2014
2015 if (from_len == to_len) {
2016 if (from_len == 1) {
2017 return replace_single_character_in_place(
2018 self,
2019 from_s[0],
2020 to_s[0],
2021 maxcount);
2022 } else {
2023 return replace_substring_in_place(
2024 self, from_s, from_len, to_s, to_len, maxcount);
2025 }
2026 }
2027
2028 /* Otherwise use the more generic algorithms */
2029 if (from_len == 1) {
2030 return replace_single_character(self, from_s[0],
2031 to_s, to_len, maxcount);
2032 } else {
2033 /* len('from')>=2, len('to')>=1 */
2034 return replace_substring(self, from_s, from_len, to_s, to_len, maxcount);
2035 }
2036}
2037
2038
2039PyDoc_STRVAR(replace__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00002040"B.replace(old, new[, count]) -> bytearray\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002041\n\
2042Return a copy of B with all occurrences of subsection\n\
2043old replaced by new. If the optional argument count is\n\
2044given, only the first count occurrences are replaced.");
2045
2046static PyObject *
2047bytes_replace(PyByteArrayObject *self, PyObject *args)
2048{
2049 Py_ssize_t count = -1;
2050 PyObject *from, *to, *res;
2051 Py_buffer vfrom, vto;
2052
2053 if (!PyArg_ParseTuple(args, "OO|n:replace", &from, &to, &count))
2054 return NULL;
2055
2056 if (_getbuffer(from, &vfrom) < 0)
2057 return NULL;
2058 if (_getbuffer(to, &vto) < 0) {
2059 PyObject_ReleaseBuffer(from, &vfrom);
2060 return NULL;
2061 }
2062
2063 res = (PyObject *)replace((PyByteArrayObject *) self,
2064 vfrom.buf, vfrom.len,
2065 vto.buf, vto.len, count);
2066
2067 PyObject_ReleaseBuffer(from, &vfrom);
2068 PyObject_ReleaseBuffer(to, &vto);
2069 return res;
2070}
2071
2072
2073/* Overallocate the initial list to reduce the number of reallocs for small
2074 split sizes. Eg, "A A A A A A A A A A".split() (10 elements) has three
2075 resizes, to sizes 4, 8, then 16. Most observed string splits are for human
2076 text (roughly 11 words per line) and field delimited data (usually 1-10
2077 fields). For large strings the split algorithms are bandwidth limited
2078 so increasing the preallocation likely will not improve things.*/
2079
2080#define MAX_PREALLOC 12
2081
2082/* 5 splits gives 6 elements */
2083#define PREALLOC_SIZE(maxsplit) \
2084 (maxsplit >= MAX_PREALLOC ? MAX_PREALLOC : maxsplit+1)
2085
2086#define SPLIT_APPEND(data, left, right) \
2087 str = PyByteArray_FromStringAndSize((data) + (left), \
2088 (right) - (left)); \
2089 if (str == NULL) \
2090 goto onError; \
2091 if (PyList_Append(list, str)) { \
2092 Py_DECREF(str); \
2093 goto onError; \
2094 } \
2095 else \
2096 Py_DECREF(str);
2097
2098#define SPLIT_ADD(data, left, right) { \
2099 str = PyByteArray_FromStringAndSize((data) + (left), \
2100 (right) - (left)); \
2101 if (str == NULL) \
2102 goto onError; \
2103 if (count < MAX_PREALLOC) { \
2104 PyList_SET_ITEM(list, count, str); \
2105 } else { \
2106 if (PyList_Append(list, str)) { \
2107 Py_DECREF(str); \
2108 goto onError; \
2109 } \
2110 else \
2111 Py_DECREF(str); \
2112 } \
2113 count++; }
2114
2115/* Always force the list to the expected size. */
2116#define FIX_PREALLOC_SIZE(list) Py_SIZE(list) = count
2117
2118
2119Py_LOCAL_INLINE(PyObject *)
2120split_char(const char *s, Py_ssize_t len, char ch, Py_ssize_t maxcount)
2121{
2122 register Py_ssize_t i, j, count = 0;
2123 PyObject *str;
2124 PyObject *list = PyList_New(PREALLOC_SIZE(maxcount));
2125
2126 if (list == NULL)
2127 return NULL;
2128
2129 i = j = 0;
2130 while ((j < len) && (maxcount-- > 0)) {
2131 for(; j < len; j++) {
2132 /* I found that using memchr makes no difference */
2133 if (s[j] == ch) {
2134 SPLIT_ADD(s, i, j);
2135 i = j = j + 1;
2136 break;
2137 }
2138 }
2139 }
2140 if (i <= len) {
2141 SPLIT_ADD(s, i, len);
2142 }
2143 FIX_PREALLOC_SIZE(list);
2144 return list;
2145
2146 onError:
2147 Py_DECREF(list);
2148 return NULL;
2149}
2150
2151
2152Py_LOCAL_INLINE(PyObject *)
2153split_whitespace(const char *s, Py_ssize_t len, Py_ssize_t maxcount)
2154{
2155 register Py_ssize_t i, j, count = 0;
2156 PyObject *str;
2157 PyObject *list = PyList_New(PREALLOC_SIZE(maxcount));
2158
2159 if (list == NULL)
2160 return NULL;
2161
2162 for (i = j = 0; i < len; ) {
2163 /* find a token */
2164 while (i < len && ISSPACE(s[i]))
2165 i++;
2166 j = i;
2167 while (i < len && !ISSPACE(s[i]))
2168 i++;
2169 if (j < i) {
2170 if (maxcount-- <= 0)
2171 break;
2172 SPLIT_ADD(s, j, i);
2173 while (i < len && ISSPACE(s[i]))
2174 i++;
2175 j = i;
2176 }
2177 }
2178 if (j < len) {
2179 SPLIT_ADD(s, j, len);
2180 }
2181 FIX_PREALLOC_SIZE(list);
2182 return list;
2183
2184 onError:
2185 Py_DECREF(list);
2186 return NULL;
2187}
2188
2189PyDoc_STRVAR(split__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00002190"B.split([sep[, maxsplit]]) -> list of bytearrays\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002191\n\
2192Return a list of the sections in B, using sep as the delimiter.\n\
2193If sep is not given, B is split on ASCII whitespace characters\n\
2194(space, tab, return, newline, formfeed, vertical tab).\n\
2195If maxsplit is given, at most maxsplit splits are done.");
2196
2197static PyObject *
2198bytes_split(PyByteArrayObject *self, PyObject *args)
2199{
2200 Py_ssize_t len = PyByteArray_GET_SIZE(self), n, i, j;
2201 Py_ssize_t maxsplit = -1, count = 0;
2202 const char *s = PyByteArray_AS_STRING(self), *sub;
2203 PyObject *list, *str, *subobj = Py_None;
2204 Py_buffer vsub;
2205#ifdef USE_FAST
2206 Py_ssize_t pos;
2207#endif
2208
2209 if (!PyArg_ParseTuple(args, "|On:split", &subobj, &maxsplit))
2210 return NULL;
2211 if (maxsplit < 0)
2212 maxsplit = PY_SSIZE_T_MAX;
2213
2214 if (subobj == Py_None)
2215 return split_whitespace(s, len, maxsplit);
2216
2217 if (_getbuffer(subobj, &vsub) < 0)
2218 return NULL;
2219 sub = vsub.buf;
2220 n = vsub.len;
2221
2222 if (n == 0) {
2223 PyErr_SetString(PyExc_ValueError, "empty separator");
2224 PyObject_ReleaseBuffer(subobj, &vsub);
2225 return NULL;
2226 }
2227 if (n == 1)
2228 return split_char(s, len, sub[0], maxsplit);
2229
2230 list = PyList_New(PREALLOC_SIZE(maxsplit));
2231 if (list == NULL) {
2232 PyObject_ReleaseBuffer(subobj, &vsub);
2233 return NULL;
2234 }
2235
2236#ifdef USE_FAST
2237 i = j = 0;
2238 while (maxsplit-- > 0) {
2239 pos = fastsearch(s+i, len-i, sub, n, FAST_SEARCH);
2240 if (pos < 0)
2241 break;
2242 j = i+pos;
2243 SPLIT_ADD(s, i, j);
2244 i = j + n;
2245 }
2246#else
2247 i = j = 0;
2248 while ((j+n <= len) && (maxsplit-- > 0)) {
2249 for (; j+n <= len; j++) {
2250 if (Py_STRING_MATCH(s, j, sub, n)) {
2251 SPLIT_ADD(s, i, j);
2252 i = j = j + n;
2253 break;
2254 }
2255 }
2256 }
2257#endif
2258 SPLIT_ADD(s, i, len);
2259 FIX_PREALLOC_SIZE(list);
2260 PyObject_ReleaseBuffer(subobj, &vsub);
2261 return list;
2262
2263 onError:
2264 Py_DECREF(list);
2265 PyObject_ReleaseBuffer(subobj, &vsub);
2266 return NULL;
2267}
2268
2269/* stringlib's partition shares nullbytes in some cases.
2270 undo this, we don't want the nullbytes to be shared. */
2271static PyObject *
2272make_nullbytes_unique(PyObject *result)
2273{
2274 if (result != NULL) {
2275 int i;
2276 assert(PyTuple_Check(result));
2277 assert(PyTuple_GET_SIZE(result) == 3);
2278 for (i = 0; i < 3; i++) {
2279 if (PyTuple_GET_ITEM(result, i) == (PyObject *)nullbytes) {
2280 PyObject *new = PyByteArray_FromStringAndSize(NULL, 0);
2281 if (new == NULL) {
2282 Py_DECREF(result);
2283 result = NULL;
2284 break;
2285 }
2286 Py_DECREF(nullbytes);
2287 PyTuple_SET_ITEM(result, i, new);
2288 }
2289 }
2290 }
2291 return result;
2292}
2293
2294PyDoc_STRVAR(partition__doc__,
2295"B.partition(sep) -> (head, sep, tail)\n\
2296\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002297Search for the separator sep in B, and return the part before it,\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002298the separator itself, and the part after it. If the separator is not\n\
2299found, returns B and two empty bytearray objects.");
2300
2301static PyObject *
2302bytes_partition(PyByteArrayObject *self, PyObject *sep_obj)
2303{
2304 PyObject *bytesep, *result;
2305
2306 bytesep = PyByteArray_FromObject(sep_obj);
2307 if (! bytesep)
2308 return NULL;
2309
2310 result = stringlib_partition(
2311 (PyObject*) self,
2312 PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
2313 bytesep,
2314 PyByteArray_AS_STRING(bytesep), PyByteArray_GET_SIZE(bytesep)
2315 );
2316
2317 Py_DECREF(bytesep);
2318 return make_nullbytes_unique(result);
2319}
2320
2321PyDoc_STRVAR(rpartition__doc__,
2322"B.rpartition(sep) -> (tail, sep, head)\n\
2323\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002324Search for the separator sep in B, starting at the end of B,\n\
2325and return the part before it, the separator itself, and the\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002326part after it. If the separator is not found, returns two empty\n\
2327bytearray objects and B.");
2328
2329static PyObject *
2330bytes_rpartition(PyByteArrayObject *self, PyObject *sep_obj)
2331{
2332 PyObject *bytesep, *result;
2333
2334 bytesep = PyByteArray_FromObject(sep_obj);
2335 if (! bytesep)
2336 return NULL;
2337
2338 result = stringlib_rpartition(
2339 (PyObject*) self,
2340 PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
2341 bytesep,
2342 PyByteArray_AS_STRING(bytesep), PyByteArray_GET_SIZE(bytesep)
2343 );
2344
2345 Py_DECREF(bytesep);
2346 return make_nullbytes_unique(result);
2347}
2348
2349Py_LOCAL_INLINE(PyObject *)
2350rsplit_char(const char *s, Py_ssize_t len, char ch, Py_ssize_t maxcount)
2351{
2352 register Py_ssize_t i, j, count=0;
2353 PyObject *str;
2354 PyObject *list = PyList_New(PREALLOC_SIZE(maxcount));
2355
2356 if (list == NULL)
2357 return NULL;
2358
2359 i = j = len - 1;
2360 while ((i >= 0) && (maxcount-- > 0)) {
2361 for (; i >= 0; i--) {
2362 if (s[i] == ch) {
2363 SPLIT_ADD(s, i + 1, j + 1);
2364 j = i = i - 1;
2365 break;
2366 }
2367 }
2368 }
2369 if (j >= -1) {
2370 SPLIT_ADD(s, 0, j + 1);
2371 }
2372 FIX_PREALLOC_SIZE(list);
2373 if (PyList_Reverse(list) < 0)
2374 goto onError;
2375
2376 return list;
2377
2378 onError:
2379 Py_DECREF(list);
2380 return NULL;
2381}
2382
2383Py_LOCAL_INLINE(PyObject *)
2384rsplit_whitespace(const char *s, Py_ssize_t len, Py_ssize_t maxcount)
2385{
2386 register Py_ssize_t i, j, count = 0;
2387 PyObject *str;
2388 PyObject *list = PyList_New(PREALLOC_SIZE(maxcount));
2389
2390 if (list == NULL)
2391 return NULL;
2392
2393 for (i = j = len - 1; i >= 0; ) {
2394 /* find a token */
2395 while (i >= 0 && ISSPACE(s[i]))
2396 i--;
2397 j = i;
2398 while (i >= 0 && !ISSPACE(s[i]))
2399 i--;
2400 if (j > i) {
2401 if (maxcount-- <= 0)
2402 break;
2403 SPLIT_ADD(s, i + 1, j + 1);
2404 while (i >= 0 && ISSPACE(s[i]))
2405 i--;
2406 j = i;
2407 }
2408 }
2409 if (j >= 0) {
2410 SPLIT_ADD(s, 0, j + 1);
2411 }
2412 FIX_PREALLOC_SIZE(list);
2413 if (PyList_Reverse(list) < 0)
2414 goto onError;
2415
2416 return list;
2417
2418 onError:
2419 Py_DECREF(list);
2420 return NULL;
2421}
2422
2423PyDoc_STRVAR(rsplit__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00002424"B.rsplit(sep[, maxsplit]) -> list of bytearrays\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002425\n\
2426Return a list of the sections in B, using sep as the delimiter,\n\
2427starting at the end of B and working to the front.\n\
2428If sep is not given, B is split on ASCII whitespace characters\n\
2429(space, tab, return, newline, formfeed, vertical tab).\n\
2430If maxsplit is given, at most maxsplit splits are done.");
2431
2432static PyObject *
2433bytes_rsplit(PyByteArrayObject *self, PyObject *args)
2434{
2435 Py_ssize_t len = PyByteArray_GET_SIZE(self), n, i, j;
2436 Py_ssize_t maxsplit = -1, count = 0;
2437 const char *s = PyByteArray_AS_STRING(self), *sub;
2438 PyObject *list, *str, *subobj = Py_None;
2439 Py_buffer vsub;
2440
2441 if (!PyArg_ParseTuple(args, "|On:rsplit", &subobj, &maxsplit))
2442 return NULL;
2443 if (maxsplit < 0)
2444 maxsplit = PY_SSIZE_T_MAX;
2445
2446 if (subobj == Py_None)
2447 return rsplit_whitespace(s, len, maxsplit);
2448
2449 if (_getbuffer(subobj, &vsub) < 0)
2450 return NULL;
2451 sub = vsub.buf;
2452 n = vsub.len;
2453
2454 if (n == 0) {
2455 PyErr_SetString(PyExc_ValueError, "empty separator");
2456 PyObject_ReleaseBuffer(subobj, &vsub);
2457 return NULL;
2458 }
2459 else if (n == 1)
2460 return rsplit_char(s, len, sub[0], maxsplit);
2461
2462 list = PyList_New(PREALLOC_SIZE(maxsplit));
2463 if (list == NULL) {
2464 PyObject_ReleaseBuffer(subobj, &vsub);
2465 return NULL;
2466 }
2467
2468 j = len;
2469 i = j - n;
2470
2471 while ( (i >= 0) && (maxsplit-- > 0) ) {
2472 for (; i>=0; i--) {
2473 if (Py_STRING_MATCH(s, i, sub, n)) {
2474 SPLIT_ADD(s, i + n, j);
2475 j = i;
2476 i -= n;
2477 break;
2478 }
2479 }
2480 }
2481 SPLIT_ADD(s, 0, j);
2482 FIX_PREALLOC_SIZE(list);
2483 if (PyList_Reverse(list) < 0)
2484 goto onError;
2485 PyObject_ReleaseBuffer(subobj, &vsub);
2486 return list;
2487
2488onError:
2489 Py_DECREF(list);
2490 PyObject_ReleaseBuffer(subobj, &vsub);
2491 return NULL;
2492}
2493
2494PyDoc_STRVAR(reverse__doc__,
2495"B.reverse() -> None\n\
2496\n\
2497Reverse the order of the values in B in place.");
2498static PyObject *
2499bytes_reverse(PyByteArrayObject *self, PyObject *unused)
2500{
2501 char swap, *head, *tail;
2502 Py_ssize_t i, j, n = Py_SIZE(self);
2503
2504 j = n / 2;
2505 head = self->ob_bytes;
2506 tail = head + n - 1;
2507 for (i = 0; i < j; i++) {
2508 swap = *head;
2509 *head++ = *tail;
2510 *tail-- = swap;
2511 }
2512
2513 Py_RETURN_NONE;
2514}
2515
2516PyDoc_STRVAR(insert__doc__,
2517"B.insert(index, int) -> None\n\
2518\n\
2519Insert a single item into the bytearray before the given index.");
2520static PyObject *
2521bytes_insert(PyByteArrayObject *self, PyObject *args)
2522{
2523 int value;
2524 Py_ssize_t where, n = Py_SIZE(self);
2525
2526 if (!PyArg_ParseTuple(args, "ni:insert", &where, &value))
2527 return NULL;
2528
2529 if (n == PY_SSIZE_T_MAX) {
2530 PyErr_SetString(PyExc_OverflowError,
2531 "cannot add more objects to bytes");
2532 return NULL;
2533 }
2534 if (value < 0 || value >= 256) {
2535 PyErr_SetString(PyExc_ValueError,
2536 "byte must be in range(0, 256)");
2537 return NULL;
2538 }
2539 if (PyByteArray_Resize((PyObject *)self, n + 1) < 0)
2540 return NULL;
2541
2542 if (where < 0) {
2543 where += n;
2544 if (where < 0)
2545 where = 0;
2546 }
2547 if (where > n)
2548 where = n;
2549 memmove(self->ob_bytes + where + 1, self->ob_bytes + where, n - where);
2550 self->ob_bytes[where] = value;
2551
2552 Py_RETURN_NONE;
2553}
2554
2555PyDoc_STRVAR(append__doc__,
2556"B.append(int) -> None\n\
2557\n\
2558Append a single item to the end of B.");
2559static PyObject *
2560bytes_append(PyByteArrayObject *self, PyObject *arg)
2561{
2562 int value;
2563 Py_ssize_t n = Py_SIZE(self);
2564
2565 if (! _getbytevalue(arg, &value))
2566 return NULL;
2567 if (n == PY_SSIZE_T_MAX) {
2568 PyErr_SetString(PyExc_OverflowError,
2569 "cannot add more objects to bytes");
2570 return NULL;
2571 }
2572 if (PyByteArray_Resize((PyObject *)self, n + 1) < 0)
2573 return NULL;
2574
2575 self->ob_bytes[n] = value;
2576
2577 Py_RETURN_NONE;
2578}
2579
2580PyDoc_STRVAR(extend__doc__,
Georg Brandl17cb8a82008-05-30 08:20:09 +00002581"B.extend(iterable_of_ints) -> None\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002582\n\
2583Append all the elements from the iterator or sequence to the\n\
2584end of B.");
2585static PyObject *
2586bytes_extend(PyByteArrayObject *self, PyObject *arg)
2587{
2588 PyObject *it, *item, *bytes_obj;
2589 Py_ssize_t buf_size = 0, len = 0;
2590 int value;
2591 char *buf;
2592
2593 /* bytes_setslice code only accepts something supporting PEP 3118. */
2594 if (PyObject_CheckBuffer(arg)) {
2595 if (bytes_setslice(self, Py_SIZE(self), Py_SIZE(self), arg) == -1)
2596 return NULL;
2597
2598 Py_RETURN_NONE;
2599 }
2600
2601 it = PyObject_GetIter(arg);
2602 if (it == NULL)
2603 return NULL;
2604
2605 /* Try to determine the length of the argument. 32 is abitrary. */
2606 buf_size = _PyObject_LengthHint(arg, 32);
2607
2608 bytes_obj = PyByteArray_FromStringAndSize(NULL, buf_size);
2609 if (bytes_obj == NULL)
2610 return NULL;
2611 buf = PyByteArray_AS_STRING(bytes_obj);
2612
2613 while ((item = PyIter_Next(it)) != NULL) {
2614 if (! _getbytevalue(item, &value)) {
2615 Py_DECREF(item);
2616 Py_DECREF(it);
2617 Py_DECREF(bytes_obj);
2618 return NULL;
2619 }
2620 buf[len++] = value;
2621 Py_DECREF(item);
2622
2623 if (len >= buf_size) {
2624 buf_size = len + (len >> 1) + 1;
2625 if (PyByteArray_Resize((PyObject *)bytes_obj, buf_size) < 0) {
2626 Py_DECREF(it);
2627 Py_DECREF(bytes_obj);
2628 return NULL;
2629 }
2630 /* Recompute the `buf' pointer, since the resizing operation may
2631 have invalidated it. */
2632 buf = PyByteArray_AS_STRING(bytes_obj);
2633 }
2634 }
2635 Py_DECREF(it);
2636
2637 /* Resize down to exact size. */
2638 if (PyByteArray_Resize((PyObject *)bytes_obj, len) < 0) {
2639 Py_DECREF(bytes_obj);
2640 return NULL;
2641 }
2642
2643 if (bytes_setslice(self, Py_SIZE(self), Py_SIZE(self), bytes_obj) == -1)
2644 return NULL;
2645 Py_DECREF(bytes_obj);
2646
2647 Py_RETURN_NONE;
2648}
2649
2650PyDoc_STRVAR(pop__doc__,
2651"B.pop([index]) -> int\n\
2652\n\
2653Remove and return a single item from B. If no index\n\
2654argument is give, will pop the last value.");
2655static PyObject *
2656bytes_pop(PyByteArrayObject *self, PyObject *args)
2657{
2658 int value;
2659 Py_ssize_t where = -1, n = Py_SIZE(self);
2660
2661 if (!PyArg_ParseTuple(args, "|n:pop", &where))
2662 return NULL;
2663
2664 if (n == 0) {
2665 PyErr_SetString(PyExc_OverflowError,
2666 "cannot pop an empty bytes");
2667 return NULL;
2668 }
2669 if (where < 0)
2670 where += Py_SIZE(self);
2671 if (where < 0 || where >= Py_SIZE(self)) {
2672 PyErr_SetString(PyExc_IndexError, "pop index out of range");
2673 return NULL;
2674 }
2675
2676 value = self->ob_bytes[where];
2677 memmove(self->ob_bytes + where, self->ob_bytes + where + 1, n - where);
2678 if (PyByteArray_Resize((PyObject *)self, n - 1) < 0)
2679 return NULL;
2680
2681 return PyLong_FromLong(value);
2682}
2683
2684PyDoc_STRVAR(remove__doc__,
2685"B.remove(int) -> None\n\
2686\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002687Remove the first occurrence of a value in B.");
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002688static PyObject *
2689bytes_remove(PyByteArrayObject *self, PyObject *arg)
2690{
2691 int value;
2692 Py_ssize_t where, n = Py_SIZE(self);
2693
2694 if (! _getbytevalue(arg, &value))
2695 return NULL;
2696
2697 for (where = 0; where < n; where++) {
2698 if (self->ob_bytes[where] == value)
2699 break;
2700 }
2701 if (where == n) {
2702 PyErr_SetString(PyExc_ValueError, "value not found in bytes");
2703 return NULL;
2704 }
2705
2706 memmove(self->ob_bytes + where, self->ob_bytes + where + 1, n - where);
2707 if (PyByteArray_Resize((PyObject *)self, n - 1) < 0)
2708 return NULL;
2709
2710 Py_RETURN_NONE;
2711}
2712
2713/* XXX These two helpers could be optimized if argsize == 1 */
2714
2715static Py_ssize_t
2716lstrip_helper(unsigned char *myptr, Py_ssize_t mysize,
2717 void *argptr, Py_ssize_t argsize)
2718{
2719 Py_ssize_t i = 0;
2720 while (i < mysize && memchr(argptr, myptr[i], argsize))
2721 i++;
2722 return i;
2723}
2724
2725static Py_ssize_t
2726rstrip_helper(unsigned char *myptr, Py_ssize_t mysize,
2727 void *argptr, Py_ssize_t argsize)
2728{
2729 Py_ssize_t i = mysize - 1;
2730 while (i >= 0 && memchr(argptr, myptr[i], argsize))
2731 i--;
2732 return i + 1;
2733}
2734
2735PyDoc_STRVAR(strip__doc__,
2736"B.strip([bytes]) -> bytearray\n\
2737\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002738Strip leading and trailing bytes contained in the argument\n\
2739and return the result as a new bytearray.\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002740If the argument is omitted, strip ASCII whitespace.");
2741static PyObject *
2742bytes_strip(PyByteArrayObject *self, PyObject *args)
2743{
2744 Py_ssize_t left, right, mysize, argsize;
2745 void *myptr, *argptr;
2746 PyObject *arg = Py_None;
2747 Py_buffer varg;
2748 if (!PyArg_ParseTuple(args, "|O:strip", &arg))
2749 return NULL;
2750 if (arg == Py_None) {
2751 argptr = "\t\n\r\f\v ";
2752 argsize = 6;
2753 }
2754 else {
2755 if (_getbuffer(arg, &varg) < 0)
2756 return NULL;
2757 argptr = varg.buf;
2758 argsize = varg.len;
2759 }
2760 myptr = self->ob_bytes;
2761 mysize = Py_SIZE(self);
2762 left = lstrip_helper(myptr, mysize, argptr, argsize);
2763 if (left == mysize)
2764 right = left;
2765 else
2766 right = rstrip_helper(myptr, mysize, argptr, argsize);
2767 if (arg != Py_None)
2768 PyObject_ReleaseBuffer(arg, &varg);
2769 return PyByteArray_FromStringAndSize(self->ob_bytes + left, right - left);
2770}
2771
2772PyDoc_STRVAR(lstrip__doc__,
2773"B.lstrip([bytes]) -> bytearray\n\
2774\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002775Strip leading bytes contained in the argument\n\
2776and return the result as a new bytearray.\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002777If the argument is omitted, strip leading ASCII whitespace.");
2778static PyObject *
2779bytes_lstrip(PyByteArrayObject *self, PyObject *args)
2780{
2781 Py_ssize_t left, right, mysize, argsize;
2782 void *myptr, *argptr;
2783 PyObject *arg = Py_None;
2784 Py_buffer varg;
2785 if (!PyArg_ParseTuple(args, "|O:lstrip", &arg))
2786 return NULL;
2787 if (arg == Py_None) {
2788 argptr = "\t\n\r\f\v ";
2789 argsize = 6;
2790 }
2791 else {
2792 if (_getbuffer(arg, &varg) < 0)
2793 return NULL;
2794 argptr = varg.buf;
2795 argsize = varg.len;
2796 }
2797 myptr = self->ob_bytes;
2798 mysize = Py_SIZE(self);
2799 left = lstrip_helper(myptr, mysize, argptr, argsize);
2800 right = mysize;
2801 if (arg != Py_None)
2802 PyObject_ReleaseBuffer(arg, &varg);
2803 return PyByteArray_FromStringAndSize(self->ob_bytes + left, right - left);
2804}
2805
2806PyDoc_STRVAR(rstrip__doc__,
2807"B.rstrip([bytes]) -> bytearray\n\
2808\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002809Strip trailing bytes contained in the argument\n\
2810and return the result as a new bytearray.\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002811If the argument is omitted, strip trailing ASCII whitespace.");
2812static PyObject *
2813bytes_rstrip(PyByteArrayObject *self, PyObject *args)
2814{
2815 Py_ssize_t left, right, mysize, argsize;
2816 void *myptr, *argptr;
2817 PyObject *arg = Py_None;
2818 Py_buffer varg;
2819 if (!PyArg_ParseTuple(args, "|O:rstrip", &arg))
2820 return NULL;
2821 if (arg == Py_None) {
2822 argptr = "\t\n\r\f\v ";
2823 argsize = 6;
2824 }
2825 else {
2826 if (_getbuffer(arg, &varg) < 0)
2827 return NULL;
2828 argptr = varg.buf;
2829 argsize = varg.len;
2830 }
2831 myptr = self->ob_bytes;
2832 mysize = Py_SIZE(self);
2833 left = 0;
2834 right = rstrip_helper(myptr, mysize, argptr, argsize);
2835 if (arg != Py_None)
2836 PyObject_ReleaseBuffer(arg, &varg);
2837 return PyByteArray_FromStringAndSize(self->ob_bytes + left, right - left);
2838}
2839
2840PyDoc_STRVAR(decode_doc,
Georg Brandl17cb8a82008-05-30 08:20:09 +00002841"B.decode([encoding[, errors]]) -> str\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002842\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002843Decode B using the codec registered for encoding. encoding defaults\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002844to the default encoding. errors may be given to set a different error\n\
2845handling scheme. Default is 'strict' meaning that encoding errors raise\n\
2846a UnicodeDecodeError. Other possible values are 'ignore' and 'replace'\n\
2847as well as any other name registered with codecs.register_error that is\n\
2848able to handle UnicodeDecodeErrors.");
2849
2850static PyObject *
2851bytes_decode(PyObject *self, PyObject *args)
2852{
2853 const char *encoding = NULL;
2854 const char *errors = NULL;
2855
2856 if (!PyArg_ParseTuple(args, "|ss:decode", &encoding, &errors))
2857 return NULL;
2858 if (encoding == NULL)
2859 encoding = PyUnicode_GetDefaultEncoding();
2860 return PyCodec_Decode(self, encoding, errors);
2861}
2862
2863PyDoc_STRVAR(alloc_doc,
2864"B.__alloc__() -> int\n\
2865\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002866Return the number of bytes actually allocated.");
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002867
2868static PyObject *
2869bytes_alloc(PyByteArrayObject *self)
2870{
2871 return PyLong_FromSsize_t(self->ob_alloc);
2872}
2873
2874PyDoc_STRVAR(join_doc,
Georg Brandl17cb8a82008-05-30 08:20:09 +00002875"B.join(iterable_of_bytes) -> bytearray\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002876\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00002877Concatenate any number of bytes/bytearray objects, with B\n\
2878in between each pair, and return the result as a new bytearray.");
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002879
2880static PyObject *
2881bytes_join(PyByteArrayObject *self, PyObject *it)
2882{
2883 PyObject *seq;
2884 Py_ssize_t mysize = Py_SIZE(self);
2885 Py_ssize_t i;
2886 Py_ssize_t n;
2887 PyObject **items;
2888 Py_ssize_t totalsize = 0;
2889 PyObject *result;
2890 char *dest;
2891
2892 seq = PySequence_Fast(it, "can only join an iterable");
2893 if (seq == NULL)
2894 return NULL;
2895 n = PySequence_Fast_GET_SIZE(seq);
2896 items = PySequence_Fast_ITEMS(seq);
2897
2898 /* Compute the total size, and check that they are all bytes */
2899 /* XXX Shouldn't we use _getbuffer() on these items instead? */
2900 for (i = 0; i < n; i++) {
2901 PyObject *obj = items[i];
2902 if (!PyByteArray_Check(obj) && !PyBytes_Check(obj)) {
2903 PyErr_Format(PyExc_TypeError,
2904 "can only join an iterable of bytes "
2905 "(item %ld has type '%.100s')",
2906 /* XXX %ld isn't right on Win64 */
2907 (long)i, Py_TYPE(obj)->tp_name);
2908 goto error;
2909 }
2910 if (i > 0)
2911 totalsize += mysize;
2912 totalsize += Py_SIZE(obj);
2913 if (totalsize < 0) {
2914 PyErr_NoMemory();
2915 goto error;
2916 }
2917 }
2918
2919 /* Allocate the result, and copy the bytes */
2920 result = PyByteArray_FromStringAndSize(NULL, totalsize);
2921 if (result == NULL)
2922 goto error;
2923 dest = PyByteArray_AS_STRING(result);
2924 for (i = 0; i < n; i++) {
2925 PyObject *obj = items[i];
2926 Py_ssize_t size = Py_SIZE(obj);
2927 char *buf;
2928 if (PyByteArray_Check(obj))
2929 buf = PyByteArray_AS_STRING(obj);
2930 else
2931 buf = PyBytes_AS_STRING(obj);
2932 if (i) {
2933 memcpy(dest, self->ob_bytes, mysize);
2934 dest += mysize;
2935 }
2936 memcpy(dest, buf, size);
2937 dest += size;
2938 }
2939
2940 /* Done */
2941 Py_DECREF(seq);
2942 return result;
2943
2944 /* Error handling */
2945 error:
2946 Py_DECREF(seq);
2947 return NULL;
2948}
2949
2950PyDoc_STRVAR(fromhex_doc,
Georg Brandl17cb8a82008-05-30 08:20:09 +00002951"bytearray.fromhex(string) -> bytearray (static method)\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00002952\n\
2953Create a bytearray object from a string of hexadecimal numbers.\n\
2954Spaces between two numbers are accepted.\n\
2955Example: bytearray.fromhex('B9 01EF') -> bytearray(b'\\xb9\\x01\\xef').");
2956
2957static int
2958hex_digit_to_int(Py_UNICODE c)
2959{
2960 if (c >= 128)
2961 return -1;
2962 if (ISDIGIT(c))
2963 return c - '0';
2964 else {
2965 if (ISUPPER(c))
2966 c = TOLOWER(c);
2967 if (c >= 'a' && c <= 'f')
2968 return c - 'a' + 10;
2969 }
2970 return -1;
2971}
2972
2973static PyObject *
2974bytes_fromhex(PyObject *cls, PyObject *args)
2975{
2976 PyObject *newbytes, *hexobj;
2977 char *buf;
2978 Py_UNICODE *hex;
2979 Py_ssize_t hexlen, byteslen, i, j;
2980 int top, bot;
2981
2982 if (!PyArg_ParseTuple(args, "U:fromhex", &hexobj))
2983 return NULL;
2984 assert(PyUnicode_Check(hexobj));
2985 hexlen = PyUnicode_GET_SIZE(hexobj);
2986 hex = PyUnicode_AS_UNICODE(hexobj);
2987 byteslen = hexlen/2; /* This overestimates if there are spaces */
2988 newbytes = PyByteArray_FromStringAndSize(NULL, byteslen);
2989 if (!newbytes)
2990 return NULL;
2991 buf = PyByteArray_AS_STRING(newbytes);
2992 for (i = j = 0; i < hexlen; i += 2) {
2993 /* skip over spaces in the input */
2994 while (hex[i] == ' ')
2995 i++;
2996 if (i >= hexlen)
2997 break;
2998 top = hex_digit_to_int(hex[i]);
2999 bot = hex_digit_to_int(hex[i+1]);
3000 if (top == -1 || bot == -1) {
3001 PyErr_Format(PyExc_ValueError,
3002 "non-hexadecimal number found in "
3003 "fromhex() arg at position %zd", i);
3004 goto error;
3005 }
3006 buf[j++] = (top << 4) + bot;
3007 }
3008 if (PyByteArray_Resize(newbytes, j) < 0)
3009 goto error;
3010 return newbytes;
3011
3012 error:
3013 Py_DECREF(newbytes);
3014 return NULL;
3015}
3016
3017PyDoc_STRVAR(reduce_doc, "Return state information for pickling.");
3018
3019static PyObject *
3020bytes_reduce(PyByteArrayObject *self)
3021{
3022 PyObject *latin1, *dict;
3023 if (self->ob_bytes)
3024 latin1 = PyUnicode_DecodeLatin1(self->ob_bytes,
3025 Py_SIZE(self), NULL);
3026 else
3027 latin1 = PyUnicode_FromString("");
3028
3029 dict = PyObject_GetAttrString((PyObject *)self, "__dict__");
3030 if (dict == NULL) {
3031 PyErr_Clear();
3032 dict = Py_None;
3033 Py_INCREF(dict);
3034 }
3035
3036 return Py_BuildValue("(O(Ns)N)", Py_TYPE(self), latin1, "latin-1", dict);
3037}
3038
3039static PySequenceMethods bytes_as_sequence = {
3040 (lenfunc)bytes_length, /* sq_length */
3041 (binaryfunc)PyByteArray_Concat, /* sq_concat */
3042 (ssizeargfunc)bytes_repeat, /* sq_repeat */
3043 (ssizeargfunc)bytes_getitem, /* sq_item */
3044 0, /* sq_slice */
3045 (ssizeobjargproc)bytes_setitem, /* sq_ass_item */
3046 0, /* sq_ass_slice */
3047 (objobjproc)bytes_contains, /* sq_contains */
3048 (binaryfunc)bytes_iconcat, /* sq_inplace_concat */
3049 (ssizeargfunc)bytes_irepeat, /* sq_inplace_repeat */
3050};
3051
3052static PyMappingMethods bytes_as_mapping = {
3053 (lenfunc)bytes_length,
3054 (binaryfunc)bytes_subscript,
3055 (objobjargproc)bytes_ass_subscript,
3056};
3057
3058static PyBufferProcs bytes_as_buffer = {
3059 (getbufferproc)bytes_getbuffer,
3060 (releasebufferproc)bytes_releasebuffer,
3061};
3062
3063static PyMethodDef
3064bytes_methods[] = {
3065 {"__alloc__", (PyCFunction)bytes_alloc, METH_NOARGS, alloc_doc},
3066 {"__reduce__", (PyCFunction)bytes_reduce, METH_NOARGS, reduce_doc},
3067 {"append", (PyCFunction)bytes_append, METH_O, append__doc__},
3068 {"capitalize", (PyCFunction)stringlib_capitalize, METH_NOARGS,
3069 _Py_capitalize__doc__},
3070 {"center", (PyCFunction)stringlib_center, METH_VARARGS, center__doc__},
3071 {"count", (PyCFunction)bytes_count, METH_VARARGS, count__doc__},
3072 {"decode", (PyCFunction)bytes_decode, METH_VARARGS, decode_doc},
3073 {"endswith", (PyCFunction)bytes_endswith, METH_VARARGS, endswith__doc__},
3074 {"expandtabs", (PyCFunction)stringlib_expandtabs, METH_VARARGS,
3075 expandtabs__doc__},
3076 {"extend", (PyCFunction)bytes_extend, METH_O, extend__doc__},
3077 {"find", (PyCFunction)bytes_find, METH_VARARGS, find__doc__},
3078 {"fromhex", (PyCFunction)bytes_fromhex, METH_VARARGS|METH_CLASS,
3079 fromhex_doc},
3080 {"index", (PyCFunction)bytes_index, METH_VARARGS, index__doc__},
3081 {"insert", (PyCFunction)bytes_insert, METH_VARARGS, insert__doc__},
3082 {"isalnum", (PyCFunction)stringlib_isalnum, METH_NOARGS,
3083 _Py_isalnum__doc__},
3084 {"isalpha", (PyCFunction)stringlib_isalpha, METH_NOARGS,
3085 _Py_isalpha__doc__},
3086 {"isdigit", (PyCFunction)stringlib_isdigit, METH_NOARGS,
3087 _Py_isdigit__doc__},
3088 {"islower", (PyCFunction)stringlib_islower, METH_NOARGS,
3089 _Py_islower__doc__},
3090 {"isspace", (PyCFunction)stringlib_isspace, METH_NOARGS,
3091 _Py_isspace__doc__},
3092 {"istitle", (PyCFunction)stringlib_istitle, METH_NOARGS,
3093 _Py_istitle__doc__},
3094 {"isupper", (PyCFunction)stringlib_isupper, METH_NOARGS,
3095 _Py_isupper__doc__},
3096 {"join", (PyCFunction)bytes_join, METH_O, join_doc},
3097 {"ljust", (PyCFunction)stringlib_ljust, METH_VARARGS, ljust__doc__},
3098 {"lower", (PyCFunction)stringlib_lower, METH_NOARGS, _Py_lower__doc__},
3099 {"lstrip", (PyCFunction)bytes_lstrip, METH_VARARGS, lstrip__doc__},
3100 {"partition", (PyCFunction)bytes_partition, METH_O, partition__doc__},
3101 {"pop", (PyCFunction)bytes_pop, METH_VARARGS, pop__doc__},
3102 {"remove", (PyCFunction)bytes_remove, METH_O, remove__doc__},
3103 {"replace", (PyCFunction)bytes_replace, METH_VARARGS, replace__doc__},
3104 {"reverse", (PyCFunction)bytes_reverse, METH_NOARGS, reverse__doc__},
3105 {"rfind", (PyCFunction)bytes_rfind, METH_VARARGS, rfind__doc__},
3106 {"rindex", (PyCFunction)bytes_rindex, METH_VARARGS, rindex__doc__},
3107 {"rjust", (PyCFunction)stringlib_rjust, METH_VARARGS, rjust__doc__},
3108 {"rpartition", (PyCFunction)bytes_rpartition, METH_O, rpartition__doc__},
3109 {"rsplit", (PyCFunction)bytes_rsplit, METH_VARARGS, rsplit__doc__},
3110 {"rstrip", (PyCFunction)bytes_rstrip, METH_VARARGS, rstrip__doc__},
3111 {"split", (PyCFunction)bytes_split, METH_VARARGS, split__doc__},
3112 {"splitlines", (PyCFunction)stringlib_splitlines, METH_VARARGS,
3113 splitlines__doc__},
3114 {"startswith", (PyCFunction)bytes_startswith, METH_VARARGS ,
3115 startswith__doc__},
3116 {"strip", (PyCFunction)bytes_strip, METH_VARARGS, strip__doc__},
3117 {"swapcase", (PyCFunction)stringlib_swapcase, METH_NOARGS,
3118 _Py_swapcase__doc__},
3119 {"title", (PyCFunction)stringlib_title, METH_NOARGS, _Py_title__doc__},
3120 {"translate", (PyCFunction)bytes_translate, METH_VARARGS,
3121 translate__doc__},
3122 {"upper", (PyCFunction)stringlib_upper, METH_NOARGS, _Py_upper__doc__},
3123 {"zfill", (PyCFunction)stringlib_zfill, METH_VARARGS, zfill__doc__},
3124 {NULL}
3125};
3126
3127PyDoc_STRVAR(bytes_doc,
Georg Brandl17cb8a82008-05-30 08:20:09 +00003128"bytearray(iterable_of_ints) -> bytearray\n\
3129bytearray(string, encoding[, errors]) -> bytearray\n\
3130bytearray(bytes_or_bytearray) -> mutable copy of bytes_or_bytearray\n\
3131bytearray(memory_view) -> bytearray\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00003132\n\
3133Construct an mutable bytearray object from:\n\
3134 - an iterable yielding integers in range(256)\n\
3135 - a text string encoded using the specified encoding\n\
3136 - a bytes or a bytearray object\n\
3137 - any object implementing the buffer API.\n\
3138\n\
Georg Brandl17cb8a82008-05-30 08:20:09 +00003139bytearray(int) -> bytearray\n\
Christian Heimes2c9c7a52008-05-26 13:42:13 +00003140\n\
3141Construct a zero-initialized bytearray of the given length.");
3142
3143
3144static PyObject *bytes_iter(PyObject *seq);
3145
3146PyTypeObject PyByteArray_Type = {
3147 PyVarObject_HEAD_INIT(&PyType_Type, 0)
3148 "bytearray",
3149 sizeof(PyByteArrayObject),
3150 0,
3151 (destructor)bytes_dealloc, /* tp_dealloc */
3152 0, /* tp_print */
3153 0, /* tp_getattr */
3154 0, /* tp_setattr */
3155 0, /* tp_compare */
3156 (reprfunc)bytes_repr, /* tp_repr */
3157 0, /* tp_as_number */
3158 &bytes_as_sequence, /* tp_as_sequence */
3159 &bytes_as_mapping, /* tp_as_mapping */
3160 0, /* tp_hash */
3161 0, /* tp_call */
3162 bytes_str, /* tp_str */
3163 PyObject_GenericGetAttr, /* tp_getattro */
3164 0, /* tp_setattro */
3165 &bytes_as_buffer, /* tp_as_buffer */
3166 Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /* tp_flags */
3167 bytes_doc, /* tp_doc */
3168 0, /* tp_traverse */
3169 0, /* tp_clear */
3170 (richcmpfunc)bytes_richcompare, /* tp_richcompare */
3171 0, /* tp_weaklistoffset */
3172 bytes_iter, /* tp_iter */
3173 0, /* tp_iternext */
3174 bytes_methods, /* tp_methods */
3175 0, /* tp_members */
3176 0, /* tp_getset */
3177 0, /* tp_base */
3178 0, /* tp_dict */
3179 0, /* tp_descr_get */
3180 0, /* tp_descr_set */
3181 0, /* tp_dictoffset */
3182 (initproc)bytes_init, /* tp_init */
3183 PyType_GenericAlloc, /* tp_alloc */
3184 PyType_GenericNew, /* tp_new */
3185 PyObject_Del, /* tp_free */
3186};
3187
3188/*********************** Bytes Iterator ****************************/
3189
3190typedef struct {
3191 PyObject_HEAD
3192 Py_ssize_t it_index;
3193 PyByteArrayObject *it_seq; /* Set to NULL when iterator is exhausted */
3194} bytesiterobject;
3195
3196static void
3197bytesiter_dealloc(bytesiterobject *it)
3198{
3199 _PyObject_GC_UNTRACK(it);
3200 Py_XDECREF(it->it_seq);
3201 PyObject_GC_Del(it);
3202}
3203
3204static int
3205bytesiter_traverse(bytesiterobject *it, visitproc visit, void *arg)
3206{
3207 Py_VISIT(it->it_seq);
3208 return 0;
3209}
3210
3211static PyObject *
3212bytesiter_next(bytesiterobject *it)
3213{
3214 PyByteArrayObject *seq;
3215 PyObject *item;
3216
3217 assert(it != NULL);
3218 seq = it->it_seq;
3219 if (seq == NULL)
3220 return NULL;
3221 assert(PyByteArray_Check(seq));
3222
3223 if (it->it_index < PyByteArray_GET_SIZE(seq)) {
3224 item = PyLong_FromLong(
3225 (unsigned char)seq->ob_bytes[it->it_index]);
3226 if (item != NULL)
3227 ++it->it_index;
3228 return item;
3229 }
3230
3231 Py_DECREF(seq);
3232 it->it_seq = NULL;
3233 return NULL;
3234}
3235
3236static PyObject *
3237bytesiter_length_hint(bytesiterobject *it)
3238{
3239 Py_ssize_t len = 0;
3240 if (it->it_seq)
3241 len = PyByteArray_GET_SIZE(it->it_seq) - it->it_index;
3242 return PyLong_FromSsize_t(len);
3243}
3244
3245PyDoc_STRVAR(length_hint_doc,
3246 "Private method returning an estimate of len(list(it)).");
3247
3248static PyMethodDef bytesiter_methods[] = {
3249 {"__length_hint__", (PyCFunction)bytesiter_length_hint, METH_NOARGS,
3250 length_hint_doc},
3251 {NULL, NULL} /* sentinel */
3252};
3253
3254PyTypeObject PyByteArrayIter_Type = {
3255 PyVarObject_HEAD_INIT(&PyType_Type, 0)
3256 "bytearray_iterator", /* tp_name */
3257 sizeof(bytesiterobject), /* tp_basicsize */
3258 0, /* tp_itemsize */
3259 /* methods */
3260 (destructor)bytesiter_dealloc, /* tp_dealloc */
3261 0, /* tp_print */
3262 0, /* tp_getattr */
3263 0, /* tp_setattr */
3264 0, /* tp_compare */
3265 0, /* tp_repr */
3266 0, /* tp_as_number */
3267 0, /* tp_as_sequence */
3268 0, /* tp_as_mapping */
3269 0, /* tp_hash */
3270 0, /* tp_call */
3271 0, /* tp_str */
3272 PyObject_GenericGetAttr, /* tp_getattro */
3273 0, /* tp_setattro */
3274 0, /* tp_as_buffer */
3275 Py_TPFLAGS_DEFAULT | Py_TPFLAGS_HAVE_GC, /* tp_flags */
3276 0, /* tp_doc */
3277 (traverseproc)bytesiter_traverse, /* tp_traverse */
3278 0, /* tp_clear */
3279 0, /* tp_richcompare */
3280 0, /* tp_weaklistoffset */
3281 PyObject_SelfIter, /* tp_iter */
3282 (iternextfunc)bytesiter_next, /* tp_iternext */
3283 bytesiter_methods, /* tp_methods */
3284 0,
3285};
3286
3287static PyObject *
3288bytes_iter(PyObject *seq)
3289{
3290 bytesiterobject *it;
3291
3292 if (!PyByteArray_Check(seq)) {
3293 PyErr_BadInternalCall();
3294 return NULL;
3295 }
3296 it = PyObject_GC_New(bytesiterobject, &PyByteArrayIter_Type);
3297 if (it == NULL)
3298 return NULL;
3299 it->it_index = 0;
3300 Py_INCREF(seq);
3301 it->it_seq = (PyByteArrayObject *)seq;
3302 _PyObject_GC_TRACK(it);
3303 return (PyObject *)it;
3304}