]>
Commit | Line | Data |
---|---|---|
7eb75bcc DM |
1 | /*\r |
2 | An implementation of Text I/O as defined by PEP 3116 - "New I/O"\r | |
3 | \r | |
4 | Classes defined here: TextIOBase, IncrementalNewlineDecoder, TextIOWrapper.\r | |
5 | \r | |
6 | Written by Amaury Forgeot d'Arc and Antoine Pitrou\r | |
7 | */\r | |
8 | \r | |
9 | #define PY_SSIZE_T_CLEAN\r | |
10 | #include "Python.h"\r | |
11 | #include "structmember.h"\r | |
12 | #include "_iomodule.h"\r | |
13 | \r | |
14 | /* TextIOBase */\r | |
15 | \r | |
16 | PyDoc_STRVAR(textiobase_doc,\r | |
17 | "Base class for text I/O.\n"\r | |
18 | "\n"\r | |
19 | "This class provides a character and line based interface to stream\n"\r | |
20 | "I/O. There is no readinto method because Python's character strings\n"\r | |
21 | "are immutable. There is no public constructor.\n"\r | |
22 | );\r | |
23 | \r | |
24 | static PyObject *\r | |
25 | _unsupported(const char *message)\r | |
26 | {\r | |
27 | PyErr_SetString(_PyIO_unsupported_operation, message);\r | |
28 | return NULL;\r | |
29 | }\r | |
30 | \r | |
31 | PyDoc_STRVAR(textiobase_detach_doc,\r | |
32 | "Separate the underlying buffer from the TextIOBase and return it.\n"\r | |
33 | "\n"\r | |
34 | "After the underlying buffer has been detached, the TextIO is in an\n"\r | |
35 | "unusable state.\n"\r | |
36 | );\r | |
37 | \r | |
38 | static PyObject *\r | |
39 | textiobase_detach(PyObject *self)\r | |
40 | {\r | |
41 | return _unsupported("detach");\r | |
42 | }\r | |
43 | \r | |
44 | PyDoc_STRVAR(textiobase_read_doc,\r | |
45 | "Read at most n characters from stream.\n"\r | |
46 | "\n"\r | |
47 | "Read from underlying buffer until we have n characters or we hit EOF.\n"\r | |
48 | "If n is negative or omitted, read until EOF.\n"\r | |
49 | );\r | |
50 | \r | |
51 | static PyObject *\r | |
52 | textiobase_read(PyObject *self, PyObject *args)\r | |
53 | {\r | |
54 | return _unsupported("read");\r | |
55 | }\r | |
56 | \r | |
57 | PyDoc_STRVAR(textiobase_readline_doc,\r | |
58 | "Read until newline or EOF.\n"\r | |
59 | "\n"\r | |
60 | "Returns an empty string if EOF is hit immediately.\n"\r | |
61 | );\r | |
62 | \r | |
63 | static PyObject *\r | |
64 | textiobase_readline(PyObject *self, PyObject *args)\r | |
65 | {\r | |
66 | return _unsupported("readline");\r | |
67 | }\r | |
68 | \r | |
69 | PyDoc_STRVAR(textiobase_write_doc,\r | |
70 | "Write string to stream.\n"\r | |
71 | "Returns the number of characters written (which is always equal to\n"\r | |
72 | "the length of the string).\n"\r | |
73 | );\r | |
74 | \r | |
75 | static PyObject *\r | |
76 | textiobase_write(PyObject *self, PyObject *args)\r | |
77 | {\r | |
78 | return _unsupported("write");\r | |
79 | }\r | |
80 | \r | |
81 | PyDoc_STRVAR(textiobase_encoding_doc,\r | |
82 | "Encoding of the text stream.\n"\r | |
83 | "\n"\r | |
84 | "Subclasses should override.\n"\r | |
85 | );\r | |
86 | \r | |
87 | static PyObject *\r | |
88 | textiobase_encoding_get(PyObject *self, void *context)\r | |
89 | {\r | |
90 | Py_RETURN_NONE;\r | |
91 | }\r | |
92 | \r | |
93 | PyDoc_STRVAR(textiobase_newlines_doc,\r | |
94 | "Line endings translated so far.\n"\r | |
95 | "\n"\r | |
96 | "Only line endings translated during reading are considered.\n"\r | |
97 | "\n"\r | |
98 | "Subclasses should override.\n"\r | |
99 | );\r | |
100 | \r | |
101 | static PyObject *\r | |
102 | textiobase_newlines_get(PyObject *self, void *context)\r | |
103 | {\r | |
104 | Py_RETURN_NONE;\r | |
105 | }\r | |
106 | \r | |
107 | PyDoc_STRVAR(textiobase_errors_doc,\r | |
108 | "The error setting of the decoder or encoder.\n"\r | |
109 | "\n"\r | |
110 | "Subclasses should override.\n"\r | |
111 | );\r | |
112 | \r | |
113 | static PyObject *\r | |
114 | textiobase_errors_get(PyObject *self, void *context)\r | |
115 | {\r | |
116 | Py_RETURN_NONE;\r | |
117 | }\r | |
118 | \r | |
119 | \r | |
120 | static PyMethodDef textiobase_methods[] = {\r | |
121 | {"detach", (PyCFunction)textiobase_detach, METH_NOARGS, textiobase_detach_doc},\r | |
122 | {"read", textiobase_read, METH_VARARGS, textiobase_read_doc},\r | |
123 | {"readline", textiobase_readline, METH_VARARGS, textiobase_readline_doc},\r | |
124 | {"write", textiobase_write, METH_VARARGS, textiobase_write_doc},\r | |
125 | {NULL, NULL}\r | |
126 | };\r | |
127 | \r | |
128 | static PyGetSetDef textiobase_getset[] = {\r | |
129 | {"encoding", (getter)textiobase_encoding_get, NULL, textiobase_encoding_doc},\r | |
130 | {"newlines", (getter)textiobase_newlines_get, NULL, textiobase_newlines_doc},\r | |
131 | {"errors", (getter)textiobase_errors_get, NULL, textiobase_errors_doc},\r | |
132 | {NULL}\r | |
133 | };\r | |
134 | \r | |
135 | PyTypeObject PyTextIOBase_Type = {\r | |
136 | PyVarObject_HEAD_INIT(NULL, 0)\r | |
137 | "_io._TextIOBase", /*tp_name*/\r | |
138 | 0, /*tp_basicsize*/\r | |
139 | 0, /*tp_itemsize*/\r | |
140 | 0, /*tp_dealloc*/\r | |
141 | 0, /*tp_print*/\r | |
142 | 0, /*tp_getattr*/\r | |
143 | 0, /*tp_setattr*/\r | |
144 | 0, /*tp_compare */\r | |
145 | 0, /*tp_repr*/\r | |
146 | 0, /*tp_as_number*/\r | |
147 | 0, /*tp_as_sequence*/\r | |
148 | 0, /*tp_as_mapping*/\r | |
149 | 0, /*tp_hash */\r | |
150 | 0, /*tp_call*/\r | |
151 | 0, /*tp_str*/\r | |
152 | 0, /*tp_getattro*/\r | |
153 | 0, /*tp_setattro*/\r | |
154 | 0, /*tp_as_buffer*/\r | |
155 | Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /*tp_flags*/\r | |
156 | textiobase_doc, /* tp_doc */\r | |
157 | 0, /* tp_traverse */\r | |
158 | 0, /* tp_clear */\r | |
159 | 0, /* tp_richcompare */\r | |
160 | 0, /* tp_weaklistoffset */\r | |
161 | 0, /* tp_iter */\r | |
162 | 0, /* tp_iternext */\r | |
163 | textiobase_methods, /* tp_methods */\r | |
164 | 0, /* tp_members */\r | |
165 | textiobase_getset, /* tp_getset */\r | |
166 | &PyIOBase_Type, /* tp_base */\r | |
167 | 0, /* tp_dict */\r | |
168 | 0, /* tp_descr_get */\r | |
169 | 0, /* tp_descr_set */\r | |
170 | 0, /* tp_dictoffset */\r | |
171 | 0, /* tp_init */\r | |
172 | 0, /* tp_alloc */\r | |
173 | 0, /* tp_new */\r | |
174 | };\r | |
175 | \r | |
176 | \r | |
177 | /* IncrementalNewlineDecoder */\r | |
178 | \r | |
179 | PyDoc_STRVAR(incrementalnewlinedecoder_doc,\r | |
180 | "Codec used when reading a file in universal newlines mode. It wraps\n"\r | |
181 | "another incremental decoder, translating \\r\\n and \\r into \\n. It also\n"\r | |
182 | "records the types of newlines encountered. When used with\n"\r | |
183 | "translate=False, it ensures that the newline sequence is returned in\n"\r | |
184 | "one piece. When used with decoder=None, it expects unicode strings as\n"\r | |
185 | "decode input and translates newlines without first invoking an external\n"\r | |
186 | "decoder.\n"\r | |
187 | );\r | |
188 | \r | |
189 | typedef struct {\r | |
190 | PyObject_HEAD\r | |
191 | PyObject *decoder;\r | |
192 | PyObject *errors;\r | |
193 | signed int pendingcr: 1;\r | |
194 | signed int translate: 1;\r | |
195 | unsigned int seennl: 3;\r | |
196 | } nldecoder_object;\r | |
197 | \r | |
198 | static int\r | |
199 | incrementalnewlinedecoder_init(nldecoder_object *self,\r | |
200 | PyObject *args, PyObject *kwds)\r | |
201 | {\r | |
202 | PyObject *decoder;\r | |
203 | int translate;\r | |
204 | PyObject *errors = NULL;\r | |
205 | char *kwlist[] = {"decoder", "translate", "errors", NULL};\r | |
206 | \r | |
207 | if (!PyArg_ParseTupleAndKeywords(args, kwds, "Oi|O:IncrementalNewlineDecoder",\r | |
208 | kwlist, &decoder, &translate, &errors))\r | |
209 | return -1;\r | |
210 | \r | |
211 | self->decoder = decoder;\r | |
212 | Py_INCREF(decoder);\r | |
213 | \r | |
214 | if (errors == NULL) {\r | |
215 | self->errors = PyUnicode_FromString("strict");\r | |
216 | if (self->errors == NULL)\r | |
217 | return -1;\r | |
218 | }\r | |
219 | else {\r | |
220 | Py_INCREF(errors);\r | |
221 | self->errors = errors;\r | |
222 | }\r | |
223 | \r | |
224 | self->translate = translate;\r | |
225 | self->seennl = 0;\r | |
226 | self->pendingcr = 0;\r | |
227 | \r | |
228 | return 0;\r | |
229 | }\r | |
230 | \r | |
231 | static void\r | |
232 | incrementalnewlinedecoder_dealloc(nldecoder_object *self)\r | |
233 | {\r | |
234 | Py_CLEAR(self->decoder);\r | |
235 | Py_CLEAR(self->errors);\r | |
236 | Py_TYPE(self)->tp_free((PyObject *)self);\r | |
237 | }\r | |
238 | \r | |
239 | static int\r | |
240 | check_decoded(PyObject *decoded)\r | |
241 | {\r | |
242 | if (decoded == NULL)\r | |
243 | return -1;\r | |
244 | if (!PyUnicode_Check(decoded)) {\r | |
245 | PyErr_Format(PyExc_TypeError,\r | |
246 | "decoder should return a string result, not '%.200s'",\r | |
247 | Py_TYPE(decoded)->tp_name);\r | |
248 | Py_DECREF(decoded);\r | |
249 | return -1;\r | |
250 | }\r | |
251 | return 0;\r | |
252 | }\r | |
253 | \r | |
254 | #define SEEN_CR 1\r | |
255 | #define SEEN_LF 2\r | |
256 | #define SEEN_CRLF 4\r | |
257 | #define SEEN_ALL (SEEN_CR | SEEN_LF | SEEN_CRLF)\r | |
258 | \r | |
259 | PyObject *\r | |
260 | _PyIncrementalNewlineDecoder_decode(PyObject *_self,\r | |
261 | PyObject *input, int final)\r | |
262 | {\r | |
263 | PyObject *output;\r | |
264 | Py_ssize_t output_len;\r | |
265 | nldecoder_object *self = (nldecoder_object *) _self;\r | |
266 | \r | |
267 | if (self->decoder == NULL) {\r | |
268 | PyErr_SetString(PyExc_ValueError,\r | |
269 | "IncrementalNewlineDecoder.__init__ not called");\r | |
270 | return NULL;\r | |
271 | }\r | |
272 | \r | |
273 | /* decode input (with the eventual \r from a previous pass) */\r | |
274 | if (self->decoder != Py_None) {\r | |
275 | output = PyObject_CallMethodObjArgs(self->decoder,\r | |
276 | _PyIO_str_decode, input, final ? Py_True : Py_False, NULL);\r | |
277 | }\r | |
278 | else {\r | |
279 | output = input;\r | |
280 | Py_INCREF(output);\r | |
281 | }\r | |
282 | \r | |
283 | if (check_decoded(output) < 0)\r | |
284 | return NULL;\r | |
285 | \r | |
286 | output_len = PyUnicode_GET_SIZE(output);\r | |
287 | if (self->pendingcr && (final || output_len > 0)) {\r | |
288 | Py_UNICODE *out;\r | |
289 | PyObject *modified = PyUnicode_FromUnicode(NULL, output_len + 1);\r | |
290 | if (modified == NULL)\r | |
291 | goto error;\r | |
292 | out = PyUnicode_AS_UNICODE(modified);\r | |
293 | out[0] = '\r';\r | |
294 | memcpy(out + 1, PyUnicode_AS_UNICODE(output),\r | |
295 | output_len * sizeof(Py_UNICODE));\r | |
296 | Py_DECREF(output);\r | |
297 | output = modified;\r | |
298 | self->pendingcr = 0;\r | |
299 | output_len++;\r | |
300 | }\r | |
301 | \r | |
302 | /* retain last \r even when not translating data:\r | |
303 | * then readline() is sure to get \r\n in one pass\r | |
304 | */\r | |
305 | if (!final) {\r | |
306 | if (output_len > 0\r | |
307 | && PyUnicode_AS_UNICODE(output)[output_len - 1] == '\r') {\r | |
308 | \r | |
309 | if (Py_REFCNT(output) == 1) {\r | |
310 | if (PyUnicode_Resize(&output, output_len - 1) < 0)\r | |
311 | goto error;\r | |
312 | }\r | |
313 | else {\r | |
314 | PyObject *modified = PyUnicode_FromUnicode(\r | |
315 | PyUnicode_AS_UNICODE(output),\r | |
316 | output_len - 1);\r | |
317 | if (modified == NULL)\r | |
318 | goto error;\r | |
319 | Py_DECREF(output);\r | |
320 | output = modified;\r | |
321 | }\r | |
322 | self->pendingcr = 1;\r | |
323 | }\r | |
324 | }\r | |
325 | \r | |
326 | /* Record which newlines are read and do newline translation if desired,\r | |
327 | all in one pass. */\r | |
328 | {\r | |
329 | Py_UNICODE *in_str;\r | |
330 | Py_ssize_t len;\r | |
331 | int seennl = self->seennl;\r | |
332 | int only_lf = 0;\r | |
333 | \r | |
334 | in_str = PyUnicode_AS_UNICODE(output);\r | |
335 | len = PyUnicode_GET_SIZE(output);\r | |
336 | \r | |
337 | if (len == 0)\r | |
338 | return output;\r | |
339 | \r | |
340 | /* If, up to now, newlines are consistently \n, do a quick check\r | |
341 | for the \r *byte* with the libc's optimized memchr.\r | |
342 | */\r | |
343 | if (seennl == SEEN_LF || seennl == 0) {\r | |
344 | only_lf = (memchr(in_str, '\r', len * sizeof(Py_UNICODE)) == NULL);\r | |
345 | }\r | |
346 | \r | |
347 | if (only_lf) {\r | |
348 | /* If not already seen, quick scan for a possible "\n" character.\r | |
349 | (there's nothing else to be done, even when in translation mode)\r | |
350 | */\r | |
351 | if (seennl == 0 &&\r | |
352 | memchr(in_str, '\n', len * sizeof(Py_UNICODE)) != NULL) {\r | |
353 | Py_UNICODE *s, *end;\r | |
354 | s = in_str;\r | |
355 | end = in_str + len;\r | |
356 | for (;;) {\r | |
357 | Py_UNICODE c;\r | |
358 | /* Fast loop for non-control characters */\r | |
359 | while (*s > '\n')\r | |
360 | s++;\r | |
361 | c = *s++;\r | |
362 | if (c == '\n') {\r | |
363 | seennl |= SEEN_LF;\r | |
364 | break;\r | |
365 | }\r | |
366 | if (s > end)\r | |
367 | break;\r | |
368 | }\r | |
369 | }\r | |
370 | /* Finished: we have scanned for newlines, and none of them\r | |
371 | need translating */\r | |
372 | }\r | |
373 | else if (!self->translate) {\r | |
374 | Py_UNICODE *s, *end;\r | |
375 | /* We have already seen all newline types, no need to scan again */\r | |
376 | if (seennl == SEEN_ALL)\r | |
377 | goto endscan;\r | |
378 | s = in_str;\r | |
379 | end = in_str + len;\r | |
380 | for (;;) {\r | |
381 | Py_UNICODE c;\r | |
382 | /* Fast loop for non-control characters */\r | |
383 | while (*s > '\r')\r | |
384 | s++;\r | |
385 | c = *s++;\r | |
386 | if (c == '\n')\r | |
387 | seennl |= SEEN_LF;\r | |
388 | else if (c == '\r') {\r | |
389 | if (*s == '\n') {\r | |
390 | seennl |= SEEN_CRLF;\r | |
391 | s++;\r | |
392 | }\r | |
393 | else\r | |
394 | seennl |= SEEN_CR;\r | |
395 | }\r | |
396 | if (s > end)\r | |
397 | break;\r | |
398 | if (seennl == SEEN_ALL)\r | |
399 | break;\r | |
400 | }\r | |
401 | endscan:\r | |
402 | ;\r | |
403 | }\r | |
404 | else {\r | |
405 | PyObject *translated = NULL;\r | |
406 | Py_UNICODE *out_str;\r | |
407 | Py_UNICODE *in, *out, *end;\r | |
408 | if (Py_REFCNT(output) != 1) {\r | |
409 | /* We could try to optimize this so that we only do a copy\r | |
410 | when there is something to translate. On the other hand,\r | |
411 | most decoders should only output non-shared strings, i.e.\r | |
412 | translation is done in place. */\r | |
413 | translated = PyUnicode_FromUnicode(NULL, len);\r | |
414 | if (translated == NULL)\r | |
415 | goto error;\r | |
416 | assert(Py_REFCNT(translated) == 1);\r | |
417 | memcpy(PyUnicode_AS_UNICODE(translated),\r | |
418 | PyUnicode_AS_UNICODE(output),\r | |
419 | len * sizeof(Py_UNICODE));\r | |
420 | }\r | |
421 | else {\r | |
422 | translated = output;\r | |
423 | }\r | |
424 | out_str = PyUnicode_AS_UNICODE(translated);\r | |
425 | in = in_str;\r | |
426 | out = out_str;\r | |
427 | end = in_str + len;\r | |
428 | for (;;) {\r | |
429 | Py_UNICODE c;\r | |
430 | /* Fast loop for non-control characters */\r | |
431 | while ((c = *in++) > '\r')\r | |
432 | *out++ = c;\r | |
433 | if (c == '\n') {\r | |
434 | *out++ = c;\r | |
435 | seennl |= SEEN_LF;\r | |
436 | continue;\r | |
437 | }\r | |
438 | if (c == '\r') {\r | |
439 | if (*in == '\n') {\r | |
440 | in++;\r | |
441 | seennl |= SEEN_CRLF;\r | |
442 | }\r | |
443 | else\r | |
444 | seennl |= SEEN_CR;\r | |
445 | *out++ = '\n';\r | |
446 | continue;\r | |
447 | }\r | |
448 | if (in > end)\r | |
449 | break;\r | |
450 | *out++ = c;\r | |
451 | }\r | |
452 | if (translated != output) {\r | |
453 | Py_DECREF(output);\r | |
454 | output = translated;\r | |
455 | }\r | |
456 | if (out - out_str != len) {\r | |
457 | if (PyUnicode_Resize(&output, out - out_str) < 0)\r | |
458 | goto error;\r | |
459 | }\r | |
460 | }\r | |
461 | self->seennl |= seennl;\r | |
462 | }\r | |
463 | \r | |
464 | return output;\r | |
465 | \r | |
466 | error:\r | |
467 | Py_DECREF(output);\r | |
468 | return NULL;\r | |
469 | }\r | |
470 | \r | |
471 | static PyObject *\r | |
472 | incrementalnewlinedecoder_decode(nldecoder_object *self,\r | |
473 | PyObject *args, PyObject *kwds)\r | |
474 | {\r | |
475 | char *kwlist[] = {"input", "final", NULL};\r | |
476 | PyObject *input;\r | |
477 | int final = 0;\r | |
478 | \r | |
479 | if (!PyArg_ParseTupleAndKeywords(args, kwds, "O|i:IncrementalNewlineDecoder",\r | |
480 | kwlist, &input, &final))\r | |
481 | return NULL;\r | |
482 | return _PyIncrementalNewlineDecoder_decode((PyObject *) self, input, final);\r | |
483 | }\r | |
484 | \r | |
485 | static PyObject *\r | |
486 | incrementalnewlinedecoder_getstate(nldecoder_object *self, PyObject *args)\r | |
487 | {\r | |
488 | PyObject *buffer;\r | |
489 | unsigned PY_LONG_LONG flag;\r | |
490 | \r | |
491 | if (self->decoder != Py_None) {\r | |
492 | PyObject *state = PyObject_CallMethodObjArgs(self->decoder,\r | |
493 | _PyIO_str_getstate, NULL);\r | |
494 | if (state == NULL)\r | |
495 | return NULL;\r | |
496 | if (!PyArg_Parse(state, "(OK)", &buffer, &flag)) {\r | |
497 | Py_DECREF(state);\r | |
498 | return NULL;\r | |
499 | }\r | |
500 | Py_INCREF(buffer);\r | |
501 | Py_DECREF(state);\r | |
502 | }\r | |
503 | else {\r | |
504 | buffer = PyBytes_FromString("");\r | |
505 | flag = 0;\r | |
506 | }\r | |
507 | flag <<= 1;\r | |
508 | if (self->pendingcr)\r | |
509 | flag |= 1;\r | |
510 | return Py_BuildValue("NK", buffer, flag);\r | |
511 | }\r | |
512 | \r | |
513 | static PyObject *\r | |
514 | incrementalnewlinedecoder_setstate(nldecoder_object *self, PyObject *state)\r | |
515 | {\r | |
516 | PyObject *buffer;\r | |
517 | unsigned PY_LONG_LONG flag;\r | |
518 | \r | |
519 | if (!PyArg_Parse(state, "(OK)", &buffer, &flag))\r | |
520 | return NULL;\r | |
521 | \r | |
522 | self->pendingcr = (int) flag & 1;\r | |
523 | flag >>= 1;\r | |
524 | \r | |
525 | if (self->decoder != Py_None)\r | |
526 | return PyObject_CallMethod(self->decoder,\r | |
527 | "setstate", "((OK))", buffer, flag);\r | |
528 | else\r | |
529 | Py_RETURN_NONE;\r | |
530 | }\r | |
531 | \r | |
532 | static PyObject *\r | |
533 | incrementalnewlinedecoder_reset(nldecoder_object *self, PyObject *args)\r | |
534 | {\r | |
535 | self->seennl = 0;\r | |
536 | self->pendingcr = 0;\r | |
537 | if (self->decoder != Py_None)\r | |
538 | return PyObject_CallMethodObjArgs(self->decoder, _PyIO_str_reset, NULL);\r | |
539 | else\r | |
540 | Py_RETURN_NONE;\r | |
541 | }\r | |
542 | \r | |
543 | static PyObject *\r | |
544 | incrementalnewlinedecoder_newlines_get(nldecoder_object *self, void *context)\r | |
545 | {\r | |
546 | switch (self->seennl) {\r | |
547 | case SEEN_CR:\r | |
548 | return PyUnicode_FromString("\r");\r | |
549 | case SEEN_LF:\r | |
550 | return PyUnicode_FromString("\n");\r | |
551 | case SEEN_CRLF:\r | |
552 | return PyUnicode_FromString("\r\n");\r | |
553 | case SEEN_CR | SEEN_LF:\r | |
554 | return Py_BuildValue("ss", "\r", "\n");\r | |
555 | case SEEN_CR | SEEN_CRLF:\r | |
556 | return Py_BuildValue("ss", "\r", "\r\n");\r | |
557 | case SEEN_LF | SEEN_CRLF:\r | |
558 | return Py_BuildValue("ss", "\n", "\r\n");\r | |
559 | case SEEN_CR | SEEN_LF | SEEN_CRLF:\r | |
560 | return Py_BuildValue("sss", "\r", "\n", "\r\n");\r | |
561 | default:\r | |
562 | Py_RETURN_NONE;\r | |
563 | }\r | |
564 | \r | |
565 | }\r | |
566 | \r | |
567 | \r | |
568 | static PyMethodDef incrementalnewlinedecoder_methods[] = {\r | |
569 | {"decode", (PyCFunction)incrementalnewlinedecoder_decode, METH_VARARGS|METH_KEYWORDS},\r | |
570 | {"getstate", (PyCFunction)incrementalnewlinedecoder_getstate, METH_NOARGS},\r | |
571 | {"setstate", (PyCFunction)incrementalnewlinedecoder_setstate, METH_O},\r | |
572 | {"reset", (PyCFunction)incrementalnewlinedecoder_reset, METH_NOARGS},\r | |
573 | {NULL}\r | |
574 | };\r | |
575 | \r | |
576 | static PyGetSetDef incrementalnewlinedecoder_getset[] = {\r | |
577 | {"newlines", (getter)incrementalnewlinedecoder_newlines_get, NULL, NULL},\r | |
578 | {NULL}\r | |
579 | };\r | |
580 | \r | |
581 | PyTypeObject PyIncrementalNewlineDecoder_Type = {\r | |
582 | PyVarObject_HEAD_INIT(NULL, 0)\r | |
583 | "_io.IncrementalNewlineDecoder", /*tp_name*/\r | |
584 | sizeof(nldecoder_object), /*tp_basicsize*/\r | |
585 | 0, /*tp_itemsize*/\r | |
586 | (destructor)incrementalnewlinedecoder_dealloc, /*tp_dealloc*/\r | |
587 | 0, /*tp_print*/\r | |
588 | 0, /*tp_getattr*/\r | |
589 | 0, /*tp_setattr*/\r | |
590 | 0, /*tp_compare */\r | |
591 | 0, /*tp_repr*/\r | |
592 | 0, /*tp_as_number*/\r | |
593 | 0, /*tp_as_sequence*/\r | |
594 | 0, /*tp_as_mapping*/\r | |
595 | 0, /*tp_hash */\r | |
596 | 0, /*tp_call*/\r | |
597 | 0, /*tp_str*/\r | |
598 | 0, /*tp_getattro*/\r | |
599 | 0, /*tp_setattro*/\r | |
600 | 0, /*tp_as_buffer*/\r | |
601 | Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /*tp_flags*/\r | |
602 | incrementalnewlinedecoder_doc, /* tp_doc */\r | |
603 | 0, /* tp_traverse */\r | |
604 | 0, /* tp_clear */\r | |
605 | 0, /* tp_richcompare */\r | |
606 | 0, /*tp_weaklistoffset*/\r | |
607 | 0, /* tp_iter */\r | |
608 | 0, /* tp_iternext */\r | |
609 | incrementalnewlinedecoder_methods, /* tp_methods */\r | |
610 | 0, /* tp_members */\r | |
611 | incrementalnewlinedecoder_getset, /* tp_getset */\r | |
612 | 0, /* tp_base */\r | |
613 | 0, /* tp_dict */\r | |
614 | 0, /* tp_descr_get */\r | |
615 | 0, /* tp_descr_set */\r | |
616 | 0, /* tp_dictoffset */\r | |
617 | (initproc)incrementalnewlinedecoder_init, /* tp_init */\r | |
618 | 0, /* tp_alloc */\r | |
619 | PyType_GenericNew, /* tp_new */\r | |
620 | };\r | |
621 | \r | |
622 | \r | |
623 | /* TextIOWrapper */\r | |
624 | \r | |
625 | PyDoc_STRVAR(textiowrapper_doc,\r | |
626 | "Character and line based layer over a BufferedIOBase object, buffer.\n"\r | |
627 | "\n"\r | |
628 | "encoding gives the name of the encoding that the stream will be\n"\r | |
629 | "decoded or encoded with. It defaults to locale.getpreferredencoding.\n"\r | |
630 | "\n"\r | |
631 | "errors determines the strictness of encoding and decoding (see the\n"\r | |
632 | "codecs.register) and defaults to \"strict\".\n"\r | |
633 | "\n"\r | |
634 | "newline controls how line endings are handled. It can be None, '',\n"\r | |
635 | "'\\n', '\\r', and '\\r\\n'. It works as follows:\n"\r | |
636 | "\n"\r | |
637 | "* On input, if newline is None, universal newlines mode is\n"\r | |
638 | " enabled. Lines in the input can end in '\\n', '\\r', or '\\r\\n', and\n"\r | |
639 | " these are translated into '\\n' before being returned to the\n"\r | |
640 | " caller. If it is '', universal newline mode is enabled, but line\n"\r | |
641 | " endings are returned to the caller untranslated. If it has any of\n"\r | |
642 | " the other legal values, input lines are only terminated by the given\n"\r | |
643 | " string, and the line ending is returned to the caller untranslated.\n"\r | |
644 | "\n"\r | |
645 | "* On output, if newline is None, any '\\n' characters written are\n"\r | |
646 | " translated to the system default line separator, os.linesep. If\n"\r | |
647 | " newline is '', no translation takes place. If newline is any of the\n"\r | |
648 | " other legal values, any '\\n' characters written are translated to\n"\r | |
649 | " the given string.\n"\r | |
650 | "\n"\r | |
651 | "If line_buffering is True, a call to flush is implied when a call to\n"\r | |
652 | "write contains a newline character."\r | |
653 | );\r | |
654 | \r | |
655 | typedef PyObject *\r | |
656 | (*encodefunc_t)(PyObject *, PyObject *);\r | |
657 | \r | |
658 | typedef struct\r | |
659 | {\r | |
660 | PyObject_HEAD\r | |
661 | int ok; /* initialized? */\r | |
662 | int detached;\r | |
663 | Py_ssize_t chunk_size;\r | |
664 | PyObject *buffer;\r | |
665 | PyObject *encoding;\r | |
666 | PyObject *encoder;\r | |
667 | PyObject *decoder;\r | |
668 | PyObject *readnl;\r | |
669 | PyObject *errors;\r | |
670 | const char *writenl; /* utf-8 encoded, NULL stands for \n */\r | |
671 | char line_buffering;\r | |
672 | char readuniversal;\r | |
673 | char readtranslate;\r | |
674 | char writetranslate;\r | |
675 | char seekable;\r | |
676 | char telling;\r | |
677 | /* Specialized encoding func (see below) */\r | |
678 | encodefunc_t encodefunc;\r | |
679 | /* Whether or not it's the start of the stream */\r | |
680 | char encoding_start_of_stream;\r | |
681 | \r | |
682 | /* Reads and writes are internally buffered in order to speed things up.\r | |
683 | However, any read will first flush the write buffer if itsn't empty.\r | |
684 | \r | |
685 | Please also note that text to be written is first encoded before being\r | |
686 | buffered. This is necessary so that encoding errors are immediately\r | |
687 | reported to the caller, but it unfortunately means that the\r | |
688 | IncrementalEncoder (whose encode() method is always written in Python)\r | |
689 | becomes a bottleneck for small writes.\r | |
690 | */\r | |
691 | PyObject *decoded_chars; /* buffer for text returned from decoder */\r | |
692 | Py_ssize_t decoded_chars_used; /* offset into _decoded_chars for read() */\r | |
693 | PyObject *pending_bytes; /* list of bytes objects waiting to be\r | |
694 | written, or NULL */\r | |
695 | Py_ssize_t pending_bytes_count;\r | |
696 | PyObject *snapshot;\r | |
697 | /* snapshot is either None, or a tuple (dec_flags, next_input) where\r | |
698 | * dec_flags is the second (integer) item of the decoder state and\r | |
699 | * next_input is the chunk of input bytes that comes next after the\r | |
700 | * snapshot point. We use this to reconstruct decoder states in tell().\r | |
701 | */\r | |
702 | \r | |
703 | /* Cache raw object if it's a FileIO object */\r | |
704 | PyObject *raw;\r | |
705 | \r | |
706 | PyObject *weakreflist;\r | |
707 | PyObject *dict;\r | |
708 | } textio;\r | |
709 | \r | |
710 | \r | |
711 | /* A couple of specialized cases in order to bypass the slow incremental\r | |
712 | encoding methods for the most popular encodings. */\r | |
713 | \r | |
714 | static PyObject *\r | |
715 | ascii_encode(textio *self, PyObject *text)\r | |
716 | {\r | |
717 | return PyUnicode_EncodeASCII(PyUnicode_AS_UNICODE(text),\r | |
718 | PyUnicode_GET_SIZE(text),\r | |
719 | PyBytes_AS_STRING(self->errors));\r | |
720 | }\r | |
721 | \r | |
722 | static PyObject *\r | |
723 | utf16be_encode(textio *self, PyObject *text)\r | |
724 | {\r | |
725 | return PyUnicode_EncodeUTF16(PyUnicode_AS_UNICODE(text),\r | |
726 | PyUnicode_GET_SIZE(text),\r | |
727 | PyBytes_AS_STRING(self->errors), 1);\r | |
728 | }\r | |
729 | \r | |
730 | static PyObject *\r | |
731 | utf16le_encode(textio *self, PyObject *text)\r | |
732 | {\r | |
733 | return PyUnicode_EncodeUTF16(PyUnicode_AS_UNICODE(text),\r | |
734 | PyUnicode_GET_SIZE(text),\r | |
735 | PyBytes_AS_STRING(self->errors), -1);\r | |
736 | }\r | |
737 | \r | |
738 | static PyObject *\r | |
739 | utf16_encode(textio *self, PyObject *text)\r | |
740 | {\r | |
741 | if (!self->encoding_start_of_stream) {\r | |
742 | /* Skip the BOM and use native byte ordering */\r | |
743 | #if defined(WORDS_BIGENDIAN)\r | |
744 | return utf16be_encode(self, text);\r | |
745 | #else\r | |
746 | return utf16le_encode(self, text);\r | |
747 | #endif\r | |
748 | }\r | |
749 | return PyUnicode_EncodeUTF16(PyUnicode_AS_UNICODE(text),\r | |
750 | PyUnicode_GET_SIZE(text),\r | |
751 | PyBytes_AS_STRING(self->errors), 0);\r | |
752 | }\r | |
753 | \r | |
754 | static PyObject *\r | |
755 | utf32be_encode(textio *self, PyObject *text)\r | |
756 | {\r | |
757 | return PyUnicode_EncodeUTF32(PyUnicode_AS_UNICODE(text),\r | |
758 | PyUnicode_GET_SIZE(text),\r | |
759 | PyBytes_AS_STRING(self->errors), 1);\r | |
760 | }\r | |
761 | \r | |
762 | static PyObject *\r | |
763 | utf32le_encode(textio *self, PyObject *text)\r | |
764 | {\r | |
765 | return PyUnicode_EncodeUTF32(PyUnicode_AS_UNICODE(text),\r | |
766 | PyUnicode_GET_SIZE(text),\r | |
767 | PyBytes_AS_STRING(self->errors), -1);\r | |
768 | }\r | |
769 | \r | |
770 | static PyObject *\r | |
771 | utf32_encode(textio *self, PyObject *text)\r | |
772 | {\r | |
773 | if (!self->encoding_start_of_stream) {\r | |
774 | /* Skip the BOM and use native byte ordering */\r | |
775 | #if defined(WORDS_BIGENDIAN)\r | |
776 | return utf32be_encode(self, text);\r | |
777 | #else\r | |
778 | return utf32le_encode(self, text);\r | |
779 | #endif\r | |
780 | }\r | |
781 | return PyUnicode_EncodeUTF32(PyUnicode_AS_UNICODE(text),\r | |
782 | PyUnicode_GET_SIZE(text),\r | |
783 | PyBytes_AS_STRING(self->errors), 0);\r | |
784 | }\r | |
785 | \r | |
786 | static PyObject *\r | |
787 | utf8_encode(textio *self, PyObject *text)\r | |
788 | {\r | |
789 | return PyUnicode_EncodeUTF8(PyUnicode_AS_UNICODE(text),\r | |
790 | PyUnicode_GET_SIZE(text),\r | |
791 | PyBytes_AS_STRING(self->errors));\r | |
792 | }\r | |
793 | \r | |
794 | static PyObject *\r | |
795 | latin1_encode(textio *self, PyObject *text)\r | |
796 | {\r | |
797 | return PyUnicode_EncodeLatin1(PyUnicode_AS_UNICODE(text),\r | |
798 | PyUnicode_GET_SIZE(text),\r | |
799 | PyBytes_AS_STRING(self->errors));\r | |
800 | }\r | |
801 | \r | |
802 | /* Map normalized encoding names onto the specialized encoding funcs */\r | |
803 | \r | |
804 | typedef struct {\r | |
805 | const char *name;\r | |
806 | encodefunc_t encodefunc;\r | |
807 | } encodefuncentry;\r | |
808 | \r | |
809 | static encodefuncentry encodefuncs[] = {\r | |
810 | {"ascii", (encodefunc_t) ascii_encode},\r | |
811 | {"iso8859-1", (encodefunc_t) latin1_encode},\r | |
812 | {"utf-8", (encodefunc_t) utf8_encode},\r | |
813 | {"utf-16-be", (encodefunc_t) utf16be_encode},\r | |
814 | {"utf-16-le", (encodefunc_t) utf16le_encode},\r | |
815 | {"utf-16", (encodefunc_t) utf16_encode},\r | |
816 | {"utf-32-be", (encodefunc_t) utf32be_encode},\r | |
817 | {"utf-32-le", (encodefunc_t) utf32le_encode},\r | |
818 | {"utf-32", (encodefunc_t) utf32_encode},\r | |
819 | {NULL, NULL}\r | |
820 | };\r | |
821 | \r | |
822 | \r | |
823 | static int\r | |
824 | textiowrapper_init(textio *self, PyObject *args, PyObject *kwds)\r | |
825 | {\r | |
826 | char *kwlist[] = {"buffer", "encoding", "errors",\r | |
827 | "newline", "line_buffering",\r | |
828 | NULL};\r | |
829 | PyObject *buffer, *raw;\r | |
830 | char *encoding = NULL;\r | |
831 | char *errors = NULL;\r | |
832 | char *newline = NULL;\r | |
833 | int line_buffering = 0;\r | |
834 | \r | |
835 | PyObject *res;\r | |
836 | int r;\r | |
837 | \r | |
838 | self->ok = 0;\r | |
839 | self->detached = 0;\r | |
840 | if (!PyArg_ParseTupleAndKeywords(args, kwds, "O|zzzi:fileio",\r | |
841 | kwlist, &buffer, &encoding, &errors,\r | |
842 | &newline, &line_buffering))\r | |
843 | return -1;\r | |
844 | \r | |
845 | if (newline && newline[0] != '\0'\r | |
846 | && !(newline[0] == '\n' && newline[1] == '\0')\r | |
847 | && !(newline[0] == '\r' && newline[1] == '\0')\r | |
848 | && !(newline[0] == '\r' && newline[1] == '\n' && newline[2] == '\0')) {\r | |
849 | PyErr_Format(PyExc_ValueError,\r | |
850 | "illegal newline value: %s", newline);\r | |
851 | return -1;\r | |
852 | }\r | |
853 | \r | |
854 | Py_CLEAR(self->buffer);\r | |
855 | Py_CLEAR(self->encoding);\r | |
856 | Py_CLEAR(self->encoder);\r | |
857 | Py_CLEAR(self->decoder);\r | |
858 | Py_CLEAR(self->readnl);\r | |
859 | Py_CLEAR(self->decoded_chars);\r | |
860 | Py_CLEAR(self->pending_bytes);\r | |
861 | Py_CLEAR(self->snapshot);\r | |
862 | Py_CLEAR(self->errors);\r | |
863 | Py_CLEAR(self->raw);\r | |
864 | self->decoded_chars_used = 0;\r | |
865 | self->pending_bytes_count = 0;\r | |
866 | self->encodefunc = NULL;\r | |
867 | self->writenl = NULL;\r | |
868 | \r | |
869 | if (encoding == NULL && self->encoding == NULL) {\r | |
870 | if (_PyIO_locale_module == NULL) {\r | |
871 | _PyIO_locale_module = PyImport_ImportModule("locale");\r | |
872 | if (_PyIO_locale_module == NULL)\r | |
873 | goto catch_ImportError;\r | |
874 | else\r | |
875 | goto use_locale;\r | |
876 | }\r | |
877 | else {\r | |
878 | use_locale:\r | |
879 | self->encoding = PyObject_CallMethod(\r | |
880 | _PyIO_locale_module, "getpreferredencoding", NULL);\r | |
881 | if (self->encoding == NULL) {\r | |
882 | catch_ImportError:\r | |
883 | /*\r | |
884 | Importing locale can raise a ImportError because of\r | |
885 | _functools, and locale.getpreferredencoding can raise a\r | |
886 | ImportError if _locale is not available. These will happen\r | |
887 | during module building.\r | |
888 | */\r | |
889 | if (PyErr_ExceptionMatches(PyExc_ImportError)) {\r | |
890 | PyErr_Clear();\r | |
891 | self->encoding = PyString_FromString("ascii");\r | |
892 | }\r | |
893 | else\r | |
894 | goto error;\r | |
895 | }\r | |
896 | else if (!PyString_Check(self->encoding))\r | |
897 | Py_CLEAR(self->encoding);\r | |
898 | }\r | |
899 | }\r | |
900 | if (self->encoding != NULL)\r | |
901 | encoding = PyString_AsString(self->encoding);\r | |
902 | else if (encoding != NULL) {\r | |
903 | self->encoding = PyString_FromString(encoding);\r | |
904 | if (self->encoding == NULL)\r | |
905 | goto error;\r | |
906 | }\r | |
907 | else {\r | |
908 | PyErr_SetString(PyExc_IOError,\r | |
909 | "could not determine default encoding");\r | |
910 | }\r | |
911 | \r | |
912 | if (errors == NULL)\r | |
913 | errors = "strict";\r | |
914 | self->errors = PyBytes_FromString(errors);\r | |
915 | if (self->errors == NULL)\r | |
916 | goto error;\r | |
917 | \r | |
918 | self->chunk_size = 8192;\r | |
919 | self->readuniversal = (newline == NULL || newline[0] == '\0');\r | |
920 | self->line_buffering = line_buffering;\r | |
921 | self->readtranslate = (newline == NULL);\r | |
922 | if (newline) {\r | |
923 | self->readnl = PyString_FromString(newline);\r | |
924 | if (self->readnl == NULL)\r | |
925 | return -1;\r | |
926 | }\r | |
927 | self->writetranslate = (newline == NULL || newline[0] != '\0');\r | |
928 | if (!self->readuniversal && self->writetranslate) {\r | |
929 | self->writenl = PyString_AsString(self->readnl);\r | |
930 | if (!strcmp(self->writenl, "\n"))\r | |
931 | self->writenl = NULL;\r | |
932 | }\r | |
933 | #ifdef MS_WINDOWS\r | |
934 | else\r | |
935 | self->writenl = "\r\n";\r | |
936 | #endif\r | |
937 | \r | |
938 | /* Build the decoder object */\r | |
939 | res = PyObject_CallMethod(buffer, "readable", NULL);\r | |
940 | if (res == NULL)\r | |
941 | goto error;\r | |
942 | r = PyObject_IsTrue(res);\r | |
943 | Py_DECREF(res);\r | |
944 | if (r == -1)\r | |
945 | goto error;\r | |
946 | if (r == 1) {\r | |
947 | self->decoder = PyCodec_IncrementalDecoder(\r | |
948 | encoding, errors);\r | |
949 | if (self->decoder == NULL)\r | |
950 | goto error;\r | |
951 | \r | |
952 | if (self->readuniversal) {\r | |
953 | PyObject *incrementalDecoder = PyObject_CallFunction(\r | |
954 | (PyObject *)&PyIncrementalNewlineDecoder_Type,\r | |
955 | "Oi", self->decoder, (int)self->readtranslate);\r | |
956 | if (incrementalDecoder == NULL)\r | |
957 | goto error;\r | |
958 | Py_CLEAR(self->decoder);\r | |
959 | self->decoder = incrementalDecoder;\r | |
960 | }\r | |
961 | }\r | |
962 | \r | |
963 | /* Build the encoder object */\r | |
964 | res = PyObject_CallMethod(buffer, "writable", NULL);\r | |
965 | if (res == NULL)\r | |
966 | goto error;\r | |
967 | r = PyObject_IsTrue(res);\r | |
968 | Py_DECREF(res);\r | |
969 | if (r == -1)\r | |
970 | goto error;\r | |
971 | if (r == 1) {\r | |
972 | PyObject *ci;\r | |
973 | self->encoder = PyCodec_IncrementalEncoder(\r | |
974 | encoding, errors);\r | |
975 | if (self->encoder == NULL)\r | |
976 | goto error;\r | |
977 | /* Get the normalized named of the codec */\r | |
978 | ci = _PyCodec_Lookup(encoding);\r | |
979 | if (ci == NULL)\r | |
980 | goto error;\r | |
981 | res = PyObject_GetAttrString(ci, "name");\r | |
982 | Py_DECREF(ci);\r | |
983 | if (res == NULL) {\r | |
984 | if (PyErr_ExceptionMatches(PyExc_AttributeError))\r | |
985 | PyErr_Clear();\r | |
986 | else\r | |
987 | goto error;\r | |
988 | }\r | |
989 | else if (PyString_Check(res)) {\r | |
990 | encodefuncentry *e = encodefuncs;\r | |
991 | while (e->name != NULL) {\r | |
992 | if (!strcmp(PyString_AS_STRING(res), e->name)) {\r | |
993 | self->encodefunc = e->encodefunc;\r | |
994 | break;\r | |
995 | }\r | |
996 | e++;\r | |
997 | }\r | |
998 | }\r | |
999 | Py_XDECREF(res);\r | |
1000 | }\r | |
1001 | \r | |
1002 | self->buffer = buffer;\r | |
1003 | Py_INCREF(buffer);\r | |
1004 | \r | |
1005 | if (Py_TYPE(buffer) == &PyBufferedReader_Type ||\r | |
1006 | Py_TYPE(buffer) == &PyBufferedWriter_Type ||\r | |
1007 | Py_TYPE(buffer) == &PyBufferedRandom_Type) {\r | |
1008 | raw = PyObject_GetAttrString(buffer, "raw");\r | |
1009 | /* Cache the raw FileIO object to speed up 'closed' checks */\r | |
1010 | if (raw == NULL) {\r | |
1011 | if (PyErr_ExceptionMatches(PyExc_AttributeError))\r | |
1012 | PyErr_Clear();\r | |
1013 | else\r | |
1014 | goto error;\r | |
1015 | }\r | |
1016 | else if (Py_TYPE(raw) == &PyFileIO_Type)\r | |
1017 | self->raw = raw;\r | |
1018 | else\r | |
1019 | Py_DECREF(raw);\r | |
1020 | }\r | |
1021 | \r | |
1022 | res = PyObject_CallMethod(buffer, "seekable", NULL);\r | |
1023 | if (res == NULL)\r | |
1024 | goto error;\r | |
1025 | r = PyObject_IsTrue(res);\r | |
1026 | Py_DECREF(res);\r | |
1027 | if (r < 0)\r | |
1028 | goto error;\r | |
1029 | self->seekable = self->telling = r;\r | |
1030 | \r | |
1031 | self->encoding_start_of_stream = 0;\r | |
1032 | if (self->seekable && self->encoder) {\r | |
1033 | PyObject *cookieObj;\r | |
1034 | int cmp;\r | |
1035 | \r | |
1036 | self->encoding_start_of_stream = 1;\r | |
1037 | \r | |
1038 | cookieObj = PyObject_CallMethodObjArgs(buffer, _PyIO_str_tell, NULL);\r | |
1039 | if (cookieObj == NULL)\r | |
1040 | goto error;\r | |
1041 | \r | |
1042 | cmp = PyObject_RichCompareBool(cookieObj, _PyIO_zero, Py_EQ);\r | |
1043 | Py_DECREF(cookieObj);\r | |
1044 | if (cmp < 0) {\r | |
1045 | goto error;\r | |
1046 | }\r | |
1047 | \r | |
1048 | if (cmp == 0) {\r | |
1049 | self->encoding_start_of_stream = 0;\r | |
1050 | res = PyObject_CallMethodObjArgs(self->encoder, _PyIO_str_setstate,\r | |
1051 | _PyIO_zero, NULL);\r | |
1052 | if (res == NULL)\r | |
1053 | goto error;\r | |
1054 | Py_DECREF(res);\r | |
1055 | }\r | |
1056 | }\r | |
1057 | \r | |
1058 | self->ok = 1;\r | |
1059 | return 0;\r | |
1060 | \r | |
1061 | error:\r | |
1062 | return -1;\r | |
1063 | }\r | |
1064 | \r | |
1065 | static int\r | |
1066 | _textiowrapper_clear(textio *self)\r | |
1067 | {\r | |
1068 | if (self->ok && _PyIOBase_finalize((PyObject *) self) < 0)\r | |
1069 | return -1;\r | |
1070 | self->ok = 0;\r | |
1071 | Py_CLEAR(self->buffer);\r | |
1072 | Py_CLEAR(self->encoding);\r | |
1073 | Py_CLEAR(self->encoder);\r | |
1074 | Py_CLEAR(self->decoder);\r | |
1075 | Py_CLEAR(self->readnl);\r | |
1076 | Py_CLEAR(self->decoded_chars);\r | |
1077 | Py_CLEAR(self->pending_bytes);\r | |
1078 | Py_CLEAR(self->snapshot);\r | |
1079 | Py_CLEAR(self->errors);\r | |
1080 | Py_CLEAR(self->raw);\r | |
1081 | return 0;\r | |
1082 | }\r | |
1083 | \r | |
1084 | static void\r | |
1085 | textiowrapper_dealloc(textio *self)\r | |
1086 | {\r | |
1087 | if (_textiowrapper_clear(self) < 0)\r | |
1088 | return;\r | |
1089 | _PyObject_GC_UNTRACK(self);\r | |
1090 | if (self->weakreflist != NULL)\r | |
1091 | PyObject_ClearWeakRefs((PyObject *)self);\r | |
1092 | Py_CLEAR(self->dict);\r | |
1093 | Py_TYPE(self)->tp_free((PyObject *)self);\r | |
1094 | }\r | |
1095 | \r | |
1096 | static int\r | |
1097 | textiowrapper_traverse(textio *self, visitproc visit, void *arg)\r | |
1098 | {\r | |
1099 | Py_VISIT(self->buffer);\r | |
1100 | Py_VISIT(self->encoding);\r | |
1101 | Py_VISIT(self->encoder);\r | |
1102 | Py_VISIT(self->decoder);\r | |
1103 | Py_VISIT(self->readnl);\r | |
1104 | Py_VISIT(self->decoded_chars);\r | |
1105 | Py_VISIT(self->pending_bytes);\r | |
1106 | Py_VISIT(self->snapshot);\r | |
1107 | Py_VISIT(self->errors);\r | |
1108 | Py_VISIT(self->raw);\r | |
1109 | \r | |
1110 | Py_VISIT(self->dict);\r | |
1111 | return 0;\r | |
1112 | }\r | |
1113 | \r | |
1114 | static int\r | |
1115 | textiowrapper_clear(textio *self)\r | |
1116 | {\r | |
1117 | if (_textiowrapper_clear(self) < 0)\r | |
1118 | return -1;\r | |
1119 | Py_CLEAR(self->dict);\r | |
1120 | return 0;\r | |
1121 | }\r | |
1122 | \r | |
1123 | static PyObject *\r | |
1124 | textiowrapper_closed_get(textio *self, void *context);\r | |
1125 | \r | |
1126 | /* This macro takes some shortcuts to make the common case faster. */\r | |
1127 | #define CHECK_CLOSED(self) \\r | |
1128 | do { \\r | |
1129 | int r; \\r | |
1130 | PyObject *_res; \\r | |
1131 | if (Py_TYPE(self) == &PyTextIOWrapper_Type) { \\r | |
1132 | if (self->raw != NULL) \\r | |
1133 | r = _PyFileIO_closed(self->raw); \\r | |
1134 | else { \\r | |
1135 | _res = textiowrapper_closed_get(self, NULL); \\r | |
1136 | if (_res == NULL) \\r | |
1137 | return NULL; \\r | |
1138 | r = PyObject_IsTrue(_res); \\r | |
1139 | Py_DECREF(_res); \\r | |
1140 | if (r < 0) \\r | |
1141 | return NULL; \\r | |
1142 | } \\r | |
1143 | if (r > 0) { \\r | |
1144 | PyErr_SetString(PyExc_ValueError, \\r | |
1145 | "I/O operation on closed file."); \\r | |
1146 | return NULL; \\r | |
1147 | } \\r | |
1148 | } \\r | |
1149 | else if (_PyIOBase_check_closed((PyObject *)self, Py_True) == NULL) \\r | |
1150 | return NULL; \\r | |
1151 | } while (0)\r | |
1152 | \r | |
1153 | #define CHECK_INITIALIZED(self) \\r | |
1154 | if (self->ok <= 0) { \\r | |
1155 | PyErr_SetString(PyExc_ValueError, \\r | |
1156 | "I/O operation on uninitialized object"); \\r | |
1157 | return NULL; \\r | |
1158 | }\r | |
1159 | \r | |
1160 | #define CHECK_ATTACHED(self) \\r | |
1161 | CHECK_INITIALIZED(self); \\r | |
1162 | if (self->detached) { \\r | |
1163 | PyErr_SetString(PyExc_ValueError, \\r | |
1164 | "underlying buffer has been detached"); \\r | |
1165 | return NULL; \\r | |
1166 | }\r | |
1167 | \r | |
1168 | #define CHECK_ATTACHED_INT(self) \\r | |
1169 | if (self->ok <= 0) { \\r | |
1170 | PyErr_SetString(PyExc_ValueError, \\r | |
1171 | "I/O operation on uninitialized object"); \\r | |
1172 | return -1; \\r | |
1173 | } else if (self->detached) { \\r | |
1174 | PyErr_SetString(PyExc_ValueError, \\r | |
1175 | "underlying buffer has been detached"); \\r | |
1176 | return -1; \\r | |
1177 | }\r | |
1178 | \r | |
1179 | \r | |
1180 | static PyObject *\r | |
1181 | textiowrapper_detach(textio *self)\r | |
1182 | {\r | |
1183 | PyObject *buffer, *res;\r | |
1184 | CHECK_ATTACHED(self);\r | |
1185 | res = PyObject_CallMethodObjArgs((PyObject *)self, _PyIO_str_flush, NULL);\r | |
1186 | if (res == NULL)\r | |
1187 | return NULL;\r | |
1188 | Py_DECREF(res);\r | |
1189 | buffer = self->buffer;\r | |
1190 | self->buffer = NULL;\r | |
1191 | self->detached = 1;\r | |
1192 | return buffer;\r | |
1193 | }\r | |
1194 | \r | |
1195 | Py_LOCAL_INLINE(const Py_UNICODE *)\r | |
1196 | findchar(const Py_UNICODE *s, Py_ssize_t size, Py_UNICODE ch)\r | |
1197 | {\r | |
1198 | /* like wcschr, but doesn't stop at NULL characters */\r | |
1199 | while (size-- > 0) {\r | |
1200 | if (*s == ch)\r | |
1201 | return s;\r | |
1202 | s++;\r | |
1203 | }\r | |
1204 | return NULL;\r | |
1205 | }\r | |
1206 | \r | |
1207 | /* Flush the internal write buffer. This doesn't explicitly flush the\r | |
1208 | underlying buffered object, though. */\r | |
1209 | static int\r | |
1210 | _textiowrapper_writeflush(textio *self)\r | |
1211 | {\r | |
1212 | PyObject *pending, *b, *ret;\r | |
1213 | \r | |
1214 | if (self->pending_bytes == NULL)\r | |
1215 | return 0;\r | |
1216 | \r | |
1217 | pending = self->pending_bytes;\r | |
1218 | Py_INCREF(pending);\r | |
1219 | self->pending_bytes_count = 0;\r | |
1220 | Py_CLEAR(self->pending_bytes);\r | |
1221 | \r | |
1222 | b = _PyBytes_Join(_PyIO_empty_bytes, pending);\r | |
1223 | Py_DECREF(pending);\r | |
1224 | if (b == NULL)\r | |
1225 | return -1;\r | |
1226 | ret = NULL;\r | |
1227 | do {\r | |
1228 | ret = PyObject_CallMethodObjArgs(self->buffer,\r | |
1229 | _PyIO_str_write, b, NULL);\r | |
1230 | } while (ret == NULL && _PyIO_trap_eintr());\r | |
1231 | Py_DECREF(b);\r | |
1232 | if (ret == NULL)\r | |
1233 | return -1;\r | |
1234 | Py_DECREF(ret);\r | |
1235 | return 0;\r | |
1236 | }\r | |
1237 | \r | |
1238 | static PyObject *\r | |
1239 | textiowrapper_write(textio *self, PyObject *args)\r | |
1240 | {\r | |
1241 | PyObject *ret;\r | |
1242 | PyObject *text; /* owned reference */\r | |
1243 | PyObject *b;\r | |
1244 | Py_ssize_t textlen;\r | |
1245 | int haslf = 0;\r | |
1246 | int needflush = 0;\r | |
1247 | \r | |
1248 | CHECK_ATTACHED(self);\r | |
1249 | \r | |
1250 | if (!PyArg_ParseTuple(args, "U:write", &text)) {\r | |
1251 | return NULL;\r | |
1252 | }\r | |
1253 | \r | |
1254 | CHECK_CLOSED(self);\r | |
1255 | \r | |
1256 | if (self->encoder == NULL) {\r | |
1257 | PyErr_SetString(PyExc_IOError, "not writable");\r | |
1258 | return NULL;\r | |
1259 | }\r | |
1260 | \r | |
1261 | Py_INCREF(text);\r | |
1262 | \r | |
1263 | textlen = PyUnicode_GetSize(text);\r | |
1264 | \r | |
1265 | if ((self->writetranslate && self->writenl != NULL) || self->line_buffering)\r | |
1266 | if (findchar(PyUnicode_AS_UNICODE(text),\r | |
1267 | PyUnicode_GET_SIZE(text), '\n'))\r | |
1268 | haslf = 1;\r | |
1269 | \r | |
1270 | if (haslf && self->writetranslate && self->writenl != NULL) {\r | |
1271 | PyObject *newtext = PyObject_CallMethod(\r | |
1272 | text, "replace", "ss", "\n", self->writenl);\r | |
1273 | Py_DECREF(text);\r | |
1274 | if (newtext == NULL)\r | |
1275 | return NULL;\r | |
1276 | text = newtext;\r | |
1277 | }\r | |
1278 | \r | |
1279 | if (self->line_buffering &&\r | |
1280 | (haslf ||\r | |
1281 | findchar(PyUnicode_AS_UNICODE(text),\r | |
1282 | PyUnicode_GET_SIZE(text), '\r')))\r | |
1283 | needflush = 1;\r | |
1284 | \r | |
1285 | /* XXX What if we were just reading? */\r | |
1286 | if (self->encodefunc != NULL) {\r | |
1287 | b = (*self->encodefunc)((PyObject *) self, text);\r | |
1288 | self->encoding_start_of_stream = 0;\r | |
1289 | }\r | |
1290 | else\r | |
1291 | b = PyObject_CallMethodObjArgs(self->encoder,\r | |
1292 | _PyIO_str_encode, text, NULL);\r | |
1293 | Py_DECREF(text);\r | |
1294 | if (b == NULL)\r | |
1295 | return NULL;\r | |
1296 | \r | |
1297 | if (self->pending_bytes == NULL) {\r | |
1298 | self->pending_bytes = PyList_New(0);\r | |
1299 | if (self->pending_bytes == NULL) {\r | |
1300 | Py_DECREF(b);\r | |
1301 | return NULL;\r | |
1302 | }\r | |
1303 | self->pending_bytes_count = 0;\r | |
1304 | }\r | |
1305 | if (PyList_Append(self->pending_bytes, b) < 0) {\r | |
1306 | Py_DECREF(b);\r | |
1307 | return NULL;\r | |
1308 | }\r | |
1309 | self->pending_bytes_count += PyBytes_GET_SIZE(b);\r | |
1310 | Py_DECREF(b);\r | |
1311 | if (self->pending_bytes_count > self->chunk_size || needflush) {\r | |
1312 | if (_textiowrapper_writeflush(self) < 0)\r | |
1313 | return NULL;\r | |
1314 | }\r | |
1315 | \r | |
1316 | if (needflush) {\r | |
1317 | ret = PyObject_CallMethodObjArgs(self->buffer, _PyIO_str_flush, NULL);\r | |
1318 | if (ret == NULL)\r | |
1319 | return NULL;\r | |
1320 | Py_DECREF(ret);\r | |
1321 | }\r | |
1322 | \r | |
1323 | Py_CLEAR(self->snapshot);\r | |
1324 | \r | |
1325 | if (self->decoder) {\r | |
1326 | ret = PyObject_CallMethod(self->decoder, "reset", NULL);\r | |
1327 | if (ret == NULL)\r | |
1328 | return NULL;\r | |
1329 | Py_DECREF(ret);\r | |
1330 | }\r | |
1331 | \r | |
1332 | return PyLong_FromSsize_t(textlen);\r | |
1333 | }\r | |
1334 | \r | |
1335 | /* Steal a reference to chars and store it in the decoded_char buffer;\r | |
1336 | */\r | |
1337 | static void\r | |
1338 | textiowrapper_set_decoded_chars(textio *self, PyObject *chars)\r | |
1339 | {\r | |
1340 | Py_CLEAR(self->decoded_chars);\r | |
1341 | self->decoded_chars = chars;\r | |
1342 | self->decoded_chars_used = 0;\r | |
1343 | }\r | |
1344 | \r | |
1345 | static PyObject *\r | |
1346 | textiowrapper_get_decoded_chars(textio *self, Py_ssize_t n)\r | |
1347 | {\r | |
1348 | PyObject *chars;\r | |
1349 | Py_ssize_t avail;\r | |
1350 | \r | |
1351 | if (self->decoded_chars == NULL)\r | |
1352 | return PyUnicode_FromStringAndSize(NULL, 0);\r | |
1353 | \r | |
1354 | avail = (PyUnicode_GET_SIZE(self->decoded_chars)\r | |
1355 | - self->decoded_chars_used);\r | |
1356 | \r | |
1357 | assert(avail >= 0);\r | |
1358 | \r | |
1359 | if (n < 0 || n > avail)\r | |
1360 | n = avail;\r | |
1361 | \r | |
1362 | if (self->decoded_chars_used > 0 || n < avail) {\r | |
1363 | chars = PyUnicode_FromUnicode(\r | |
1364 | PyUnicode_AS_UNICODE(self->decoded_chars)\r | |
1365 | + self->decoded_chars_used, n);\r | |
1366 | if (chars == NULL)\r | |
1367 | return NULL;\r | |
1368 | }\r | |
1369 | else {\r | |
1370 | chars = self->decoded_chars;\r | |
1371 | Py_INCREF(chars);\r | |
1372 | }\r | |
1373 | \r | |
1374 | self->decoded_chars_used += n;\r | |
1375 | return chars;\r | |
1376 | }\r | |
1377 | \r | |
1378 | /* Read and decode the next chunk of data from the BufferedReader.\r | |
1379 | */\r | |
1380 | static int\r | |
1381 | textiowrapper_read_chunk(textio *self)\r | |
1382 | {\r | |
1383 | PyObject *dec_buffer = NULL;\r | |
1384 | PyObject *dec_flags = NULL;\r | |
1385 | PyObject *input_chunk = NULL;\r | |
1386 | PyObject *decoded_chars, *chunk_size;\r | |
1387 | int eof;\r | |
1388 | \r | |
1389 | /* The return value is True unless EOF was reached. The decoded string is\r | |
1390 | * placed in self._decoded_chars (replacing its previous value). The\r | |
1391 | * entire input chunk is sent to the decoder, though some of it may remain\r | |
1392 | * buffered in the decoder, yet to be converted.\r | |
1393 | */\r | |
1394 | \r | |
1395 | if (self->decoder == NULL) {\r | |
1396 | PyErr_SetString(PyExc_IOError, "not readable");\r | |
1397 | return -1;\r | |
1398 | }\r | |
1399 | \r | |
1400 | if (self->telling) {\r | |
1401 | /* To prepare for tell(), we need to snapshot a point in the file\r | |
1402 | * where the decoder's input buffer is empty.\r | |
1403 | */\r | |
1404 | \r | |
1405 | PyObject *state = PyObject_CallMethodObjArgs(self->decoder,\r | |
1406 | _PyIO_str_getstate, NULL);\r | |
1407 | if (state == NULL)\r | |
1408 | return -1;\r | |
1409 | /* Given this, we know there was a valid snapshot point\r | |
1410 | * len(dec_buffer) bytes ago with decoder state (b'', dec_flags).\r | |
1411 | */\r | |
1412 | if (PyArg_Parse(state, "(OO)", &dec_buffer, &dec_flags) < 0) {\r | |
1413 | Py_DECREF(state);\r | |
1414 | return -1;\r | |
1415 | }\r | |
1416 | Py_INCREF(dec_buffer);\r | |
1417 | Py_INCREF(dec_flags);\r | |
1418 | Py_DECREF(state);\r | |
1419 | }\r | |
1420 | \r | |
1421 | /* Read a chunk, decode it, and put the result in self._decoded_chars. */\r | |
1422 | chunk_size = PyLong_FromSsize_t(self->chunk_size);\r | |
1423 | if (chunk_size == NULL)\r | |
1424 | goto fail;\r | |
1425 | input_chunk = PyObject_CallMethodObjArgs(self->buffer,\r | |
1426 | _PyIO_str_read1, chunk_size, NULL);\r | |
1427 | Py_DECREF(chunk_size);\r | |
1428 | if (input_chunk == NULL)\r | |
1429 | goto fail;\r | |
1430 | if (!PyBytes_Check(input_chunk)) {\r | |
1431 | PyErr_Format(PyExc_TypeError,\r | |
1432 | "underlying read1() should have returned a bytes object, "\r | |
1433 | "not '%.200s'", Py_TYPE(input_chunk)->tp_name);\r | |
1434 | goto fail;\r | |
1435 | }\r | |
1436 | \r | |
1437 | eof = (PyBytes_Size(input_chunk) == 0);\r | |
1438 | \r | |
1439 | if (Py_TYPE(self->decoder) == &PyIncrementalNewlineDecoder_Type) {\r | |
1440 | decoded_chars = _PyIncrementalNewlineDecoder_decode(\r | |
1441 | self->decoder, input_chunk, eof);\r | |
1442 | }\r | |
1443 | else {\r | |
1444 | decoded_chars = PyObject_CallMethodObjArgs(self->decoder,\r | |
1445 | _PyIO_str_decode, input_chunk, eof ? Py_True : Py_False, NULL);\r | |
1446 | }\r | |
1447 | \r | |
1448 | if (check_decoded(decoded_chars) < 0)\r | |
1449 | goto fail;\r | |
1450 | textiowrapper_set_decoded_chars(self, decoded_chars);\r | |
1451 | if (PyUnicode_GET_SIZE(decoded_chars) > 0)\r | |
1452 | eof = 0;\r | |
1453 | \r | |
1454 | if (self->telling) {\r | |
1455 | /* At the snapshot point, len(dec_buffer) bytes before the read, the\r | |
1456 | * next input to be decoded is dec_buffer + input_chunk.\r | |
1457 | */\r | |
1458 | PyObject *next_input = PyNumber_Add(dec_buffer, input_chunk);\r | |
1459 | if (next_input == NULL)\r | |
1460 | goto fail;\r | |
1461 | if (!PyBytes_Check(next_input)) {\r | |
1462 | PyErr_Format(PyExc_TypeError,\r | |
1463 | "decoder getstate() should have returned a bytes "\r | |
1464 | "object, not '%.200s'",\r | |
1465 | Py_TYPE(next_input)->tp_name);\r | |
1466 | Py_DECREF(next_input);\r | |
1467 | goto fail;\r | |
1468 | }\r | |
1469 | Py_DECREF(dec_buffer);\r | |
1470 | Py_CLEAR(self->snapshot);\r | |
1471 | self->snapshot = Py_BuildValue("NN", dec_flags, next_input);\r | |
1472 | }\r | |
1473 | Py_DECREF(input_chunk);\r | |
1474 | \r | |
1475 | return (eof == 0);\r | |
1476 | \r | |
1477 | fail:\r | |
1478 | Py_XDECREF(dec_buffer);\r | |
1479 | Py_XDECREF(dec_flags);\r | |
1480 | Py_XDECREF(input_chunk);\r | |
1481 | return -1;\r | |
1482 | }\r | |
1483 | \r | |
1484 | static PyObject *\r | |
1485 | textiowrapper_read(textio *self, PyObject *args)\r | |
1486 | {\r | |
1487 | Py_ssize_t n = -1;\r | |
1488 | PyObject *result = NULL, *chunks = NULL;\r | |
1489 | \r | |
1490 | CHECK_ATTACHED(self);\r | |
1491 | \r | |
1492 | if (!PyArg_ParseTuple(args, "|O&:read", &_PyIO_ConvertSsize_t, &n))\r | |
1493 | return NULL;\r | |
1494 | \r | |
1495 | CHECK_CLOSED(self);\r | |
1496 | \r | |
1497 | if (self->decoder == NULL) {\r | |
1498 | PyErr_SetString(PyExc_IOError, "not readable");\r | |
1499 | return NULL;\r | |
1500 | }\r | |
1501 | \r | |
1502 | if (_textiowrapper_writeflush(self) < 0)\r | |
1503 | return NULL;\r | |
1504 | \r | |
1505 | if (n < 0) {\r | |
1506 | /* Read everything */\r | |
1507 | PyObject *bytes = PyObject_CallMethod(self->buffer, "read", NULL);\r | |
1508 | PyObject *decoded, *final;\r | |
1509 | if (bytes == NULL)\r | |
1510 | goto fail;\r | |
1511 | decoded = PyObject_CallMethodObjArgs(self->decoder, _PyIO_str_decode,\r | |
1512 | bytes, Py_True, NULL);\r | |
1513 | Py_DECREF(bytes);\r | |
1514 | if (check_decoded(decoded) < 0)\r | |
1515 | goto fail;\r | |
1516 | \r | |
1517 | result = textiowrapper_get_decoded_chars(self, -1);\r | |
1518 | \r | |
1519 | if (result == NULL) {\r | |
1520 | Py_DECREF(decoded);\r | |
1521 | return NULL;\r | |
1522 | }\r | |
1523 | \r | |
1524 | final = PyUnicode_Concat(result, decoded);\r | |
1525 | Py_DECREF(result);\r | |
1526 | Py_DECREF(decoded);\r | |
1527 | if (final == NULL)\r | |
1528 | goto fail;\r | |
1529 | \r | |
1530 | Py_CLEAR(self->snapshot);\r | |
1531 | return final;\r | |
1532 | }\r | |
1533 | else {\r | |
1534 | int res = 1;\r | |
1535 | Py_ssize_t remaining = n;\r | |
1536 | \r | |
1537 | result = textiowrapper_get_decoded_chars(self, n);\r | |
1538 | if (result == NULL)\r | |
1539 | goto fail;\r | |
1540 | remaining -= PyUnicode_GET_SIZE(result);\r | |
1541 | \r | |
1542 | /* Keep reading chunks until we have n characters to return */\r | |
1543 | while (remaining > 0) {\r | |
1544 | res = textiowrapper_read_chunk(self);\r | |
1545 | if (res < 0) {\r | |
1546 | /* NOTE: PyErr_SetFromErrno() calls PyErr_CheckSignals()\r | |
1547 | when EINTR occurs so we needn't do it ourselves. */\r | |
1548 | if (_PyIO_trap_eintr()) {\r | |
1549 | continue;\r | |
1550 | }\r | |
1551 | goto fail;\r | |
1552 | }\r | |
1553 | if (res == 0) /* EOF */\r | |
1554 | break;\r | |
1555 | if (chunks == NULL) {\r | |
1556 | chunks = PyList_New(0);\r | |
1557 | if (chunks == NULL)\r | |
1558 | goto fail;\r | |
1559 | }\r | |
1560 | if (PyList_Append(chunks, result) < 0)\r | |
1561 | goto fail;\r | |
1562 | Py_DECREF(result);\r | |
1563 | result = textiowrapper_get_decoded_chars(self, remaining);\r | |
1564 | if (result == NULL)\r | |
1565 | goto fail;\r | |
1566 | remaining -= PyUnicode_GET_SIZE(result);\r | |
1567 | }\r | |
1568 | if (chunks != NULL) {\r | |
1569 | if (result != NULL && PyList_Append(chunks, result) < 0)\r | |
1570 | goto fail;\r | |
1571 | Py_CLEAR(result);\r | |
1572 | result = PyUnicode_Join(_PyIO_empty_str, chunks);\r | |
1573 | if (result == NULL)\r | |
1574 | goto fail;\r | |
1575 | Py_CLEAR(chunks);\r | |
1576 | }\r | |
1577 | return result;\r | |
1578 | }\r | |
1579 | fail:\r | |
1580 | Py_XDECREF(result);\r | |
1581 | Py_XDECREF(chunks);\r | |
1582 | return NULL;\r | |
1583 | }\r | |
1584 | \r | |
1585 | \r | |
1586 | /* NOTE: `end` must point to the real end of the Py_UNICODE storage,\r | |
1587 | that is to the NUL character. Otherwise the function will produce\r | |
1588 | incorrect results. */\r | |
1589 | static Py_UNICODE *\r | |
1590 | find_control_char(Py_UNICODE *start, Py_UNICODE *end, Py_UNICODE ch)\r | |
1591 | {\r | |
1592 | Py_UNICODE *s = start;\r | |
1593 | for (;;) {\r | |
1594 | while (*s > ch)\r | |
1595 | s++;\r | |
1596 | if (*s == ch)\r | |
1597 | return s;\r | |
1598 | if (s == end)\r | |
1599 | return NULL;\r | |
1600 | s++;\r | |
1601 | }\r | |
1602 | }\r | |
1603 | \r | |
1604 | Py_ssize_t\r | |
1605 | _PyIO_find_line_ending(\r | |
1606 | int translated, int universal, PyObject *readnl,\r | |
1607 | Py_UNICODE *start, Py_UNICODE *end, Py_ssize_t *consumed)\r | |
1608 | {\r | |
1609 | Py_ssize_t len = end - start;\r | |
1610 | \r | |
1611 | if (translated) {\r | |
1612 | /* Newlines are already translated, only search for \n */\r | |
1613 | Py_UNICODE *pos = find_control_char(start, end, '\n');\r | |
1614 | if (pos != NULL)\r | |
1615 | return pos - start + 1;\r | |
1616 | else {\r | |
1617 | *consumed = len;\r | |
1618 | return -1;\r | |
1619 | }\r | |
1620 | }\r | |
1621 | else if (universal) {\r | |
1622 | /* Universal newline search. Find any of \r, \r\n, \n\r | |
1623 | * The decoder ensures that \r\n are not split in two pieces\r | |
1624 | */\r | |
1625 | Py_UNICODE *s = start;\r | |
1626 | for (;;) {\r | |
1627 | Py_UNICODE ch;\r | |
1628 | /* Fast path for non-control chars. The loop always ends\r | |
1629 | since the Py_UNICODE storage is NUL-terminated. */\r | |
1630 | while (*s > '\r')\r | |
1631 | s++;\r | |
1632 | if (s >= end) {\r | |
1633 | *consumed = len;\r | |
1634 | return -1;\r | |
1635 | }\r | |
1636 | ch = *s++;\r | |
1637 | if (ch == '\n')\r | |
1638 | return s - start;\r | |
1639 | if (ch == '\r') {\r | |
1640 | if (*s == '\n')\r | |
1641 | return s - start + 1;\r | |
1642 | else\r | |
1643 | return s - start;\r | |
1644 | }\r | |
1645 | }\r | |
1646 | }\r | |
1647 | else {\r | |
1648 | /* Non-universal mode. */\r | |
1649 | Py_ssize_t readnl_len = PyString_GET_SIZE(readnl);\r | |
1650 | unsigned char *nl = (unsigned char *) PyString_AS_STRING(readnl);\r | |
1651 | if (readnl_len == 1) {\r | |
1652 | Py_UNICODE *pos = find_control_char(start, end, nl[0]);\r | |
1653 | if (pos != NULL)\r | |
1654 | return pos - start + 1;\r | |
1655 | *consumed = len;\r | |
1656 | return -1;\r | |
1657 | }\r | |
1658 | else {\r | |
1659 | Py_UNICODE *s = start;\r | |
1660 | Py_UNICODE *e = end - readnl_len + 1;\r | |
1661 | Py_UNICODE *pos;\r | |
1662 | if (e < s)\r | |
1663 | e = s;\r | |
1664 | while (s < e) {\r | |
1665 | Py_ssize_t i;\r | |
1666 | Py_UNICODE *pos = find_control_char(s, end, nl[0]);\r | |
1667 | if (pos == NULL || pos >= e)\r | |
1668 | break;\r | |
1669 | for (i = 1; i < readnl_len; i++) {\r | |
1670 | if (pos[i] != nl[i])\r | |
1671 | break;\r | |
1672 | }\r | |
1673 | if (i == readnl_len)\r | |
1674 | return pos - start + readnl_len;\r | |
1675 | s = pos + 1;\r | |
1676 | }\r | |
1677 | pos = find_control_char(e, end, nl[0]);\r | |
1678 | if (pos == NULL)\r | |
1679 | *consumed = len;\r | |
1680 | else\r | |
1681 | *consumed = pos - start;\r | |
1682 | return -1;\r | |
1683 | }\r | |
1684 | }\r | |
1685 | }\r | |
1686 | \r | |
1687 | static PyObject *\r | |
1688 | _textiowrapper_readline(textio *self, Py_ssize_t limit)\r | |
1689 | {\r | |
1690 | PyObject *line = NULL, *chunks = NULL, *remaining = NULL;\r | |
1691 | Py_ssize_t start, endpos, chunked, offset_to_buffer;\r | |
1692 | int res;\r | |
1693 | \r | |
1694 | CHECK_CLOSED(self);\r | |
1695 | \r | |
1696 | if (_textiowrapper_writeflush(self) < 0)\r | |
1697 | return NULL;\r | |
1698 | \r | |
1699 | chunked = 0;\r | |
1700 | \r | |
1701 | while (1) {\r | |
1702 | Py_UNICODE *ptr;\r | |
1703 | Py_ssize_t line_len;\r | |
1704 | Py_ssize_t consumed = 0;\r | |
1705 | \r | |
1706 | /* First, get some data if necessary */\r | |
1707 | res = 1;\r | |
1708 | while (!self->decoded_chars ||\r | |
1709 | !PyUnicode_GET_SIZE(self->decoded_chars)) {\r | |
1710 | res = textiowrapper_read_chunk(self);\r | |
1711 | if (res < 0) {\r | |
1712 | /* NOTE: PyErr_SetFromErrno() calls PyErr_CheckSignals()\r | |
1713 | when EINTR occurs so we needn't do it ourselves. */\r | |
1714 | if (_PyIO_trap_eintr()) {\r | |
1715 | continue;\r | |
1716 | }\r | |
1717 | goto error;\r | |
1718 | }\r | |
1719 | if (res == 0)\r | |
1720 | break;\r | |
1721 | }\r | |
1722 | if (res == 0) {\r | |
1723 | /* end of file */\r | |
1724 | textiowrapper_set_decoded_chars(self, NULL);\r | |
1725 | Py_CLEAR(self->snapshot);\r | |
1726 | start = endpos = offset_to_buffer = 0;\r | |
1727 | break;\r | |
1728 | }\r | |
1729 | \r | |
1730 | if (remaining == NULL) {\r | |
1731 | line = self->decoded_chars;\r | |
1732 | start = self->decoded_chars_used;\r | |
1733 | offset_to_buffer = 0;\r | |
1734 | Py_INCREF(line);\r | |
1735 | }\r | |
1736 | else {\r | |
1737 | assert(self->decoded_chars_used == 0);\r | |
1738 | line = PyUnicode_Concat(remaining, self->decoded_chars);\r | |
1739 | start = 0;\r | |
1740 | offset_to_buffer = PyUnicode_GET_SIZE(remaining);\r | |
1741 | Py_CLEAR(remaining);\r | |
1742 | if (line == NULL)\r | |
1743 | goto error;\r | |
1744 | }\r | |
1745 | \r | |
1746 | ptr = PyUnicode_AS_UNICODE(line);\r | |
1747 | line_len = PyUnicode_GET_SIZE(line);\r | |
1748 | \r | |
1749 | endpos = _PyIO_find_line_ending(\r | |
1750 | self->readtranslate, self->readuniversal, self->readnl,\r | |
1751 | ptr + start, ptr + line_len, &consumed);\r | |
1752 | if (endpos >= 0) {\r | |
1753 | endpos += start;\r | |
1754 | if (limit >= 0 && (endpos - start) + chunked >= limit)\r | |
1755 | endpos = start + limit - chunked;\r | |
1756 | break;\r | |
1757 | }\r | |
1758 | \r | |
1759 | /* We can put aside up to `endpos` */\r | |
1760 | endpos = consumed + start;\r | |
1761 | if (limit >= 0 && (endpos - start) + chunked >= limit) {\r | |
1762 | /* Didn't find line ending, but reached length limit */\r | |
1763 | endpos = start + limit - chunked;\r | |
1764 | break;\r | |
1765 | }\r | |
1766 | \r | |
1767 | if (endpos > start) {\r | |
1768 | /* No line ending seen yet - put aside current data */\r | |
1769 | PyObject *s;\r | |
1770 | if (chunks == NULL) {\r | |
1771 | chunks = PyList_New(0);\r | |
1772 | if (chunks == NULL)\r | |
1773 | goto error;\r | |
1774 | }\r | |
1775 | s = PyUnicode_FromUnicode(ptr + start, endpos - start);\r | |
1776 | if (s == NULL)\r | |
1777 | goto error;\r | |
1778 | if (PyList_Append(chunks, s) < 0) {\r | |
1779 | Py_DECREF(s);\r | |
1780 | goto error;\r | |
1781 | }\r | |
1782 | chunked += PyUnicode_GET_SIZE(s);\r | |
1783 | Py_DECREF(s);\r | |
1784 | }\r | |
1785 | /* There may be some remaining bytes we'll have to prepend to the\r | |
1786 | next chunk of data */\r | |
1787 | if (endpos < line_len) {\r | |
1788 | remaining = PyUnicode_FromUnicode(\r | |
1789 | ptr + endpos, line_len - endpos);\r | |
1790 | if (remaining == NULL)\r | |
1791 | goto error;\r | |
1792 | }\r | |
1793 | Py_CLEAR(line);\r | |
1794 | /* We have consumed the buffer */\r | |
1795 | textiowrapper_set_decoded_chars(self, NULL);\r | |
1796 | }\r | |
1797 | \r | |
1798 | if (line != NULL) {\r | |
1799 | /* Our line ends in the current buffer */\r | |
1800 | self->decoded_chars_used = endpos - offset_to_buffer;\r | |
1801 | if (start > 0 || endpos < PyUnicode_GET_SIZE(line)) {\r | |
1802 | if (start == 0 && Py_REFCNT(line) == 1) {\r | |
1803 | if (PyUnicode_Resize(&line, endpos) < 0)\r | |
1804 | goto error;\r | |
1805 | }\r | |
1806 | else {\r | |
1807 | PyObject *s = PyUnicode_FromUnicode(\r | |
1808 | PyUnicode_AS_UNICODE(line) + start, endpos - start);\r | |
1809 | Py_CLEAR(line);\r | |
1810 | if (s == NULL)\r | |
1811 | goto error;\r | |
1812 | line = s;\r | |
1813 | }\r | |
1814 | }\r | |
1815 | }\r | |
1816 | if (remaining != NULL) {\r | |
1817 | if (chunks == NULL) {\r | |
1818 | chunks = PyList_New(0);\r | |
1819 | if (chunks == NULL)\r | |
1820 | goto error;\r | |
1821 | }\r | |
1822 | if (PyList_Append(chunks, remaining) < 0)\r | |
1823 | goto error;\r | |
1824 | Py_CLEAR(remaining);\r | |
1825 | }\r | |
1826 | if (chunks != NULL) {\r | |
1827 | if (line != NULL && PyList_Append(chunks, line) < 0)\r | |
1828 | goto error;\r | |
1829 | Py_CLEAR(line);\r | |
1830 | line = PyUnicode_Join(_PyIO_empty_str, chunks);\r | |
1831 | if (line == NULL)\r | |
1832 | goto error;\r | |
1833 | Py_DECREF(chunks);\r | |
1834 | }\r | |
1835 | if (line == NULL)\r | |
1836 | line = PyUnicode_FromStringAndSize(NULL, 0);\r | |
1837 | \r | |
1838 | return line;\r | |
1839 | \r | |
1840 | error:\r | |
1841 | Py_XDECREF(chunks);\r | |
1842 | Py_XDECREF(remaining);\r | |
1843 | Py_XDECREF(line);\r | |
1844 | return NULL;\r | |
1845 | }\r | |
1846 | \r | |
1847 | static PyObject *\r | |
1848 | textiowrapper_readline(textio *self, PyObject *args)\r | |
1849 | {\r | |
1850 | PyObject *limitobj = NULL;\r | |
1851 | Py_ssize_t limit = -1;\r | |
1852 | \r | |
1853 | CHECK_ATTACHED(self);\r | |
1854 | if (!PyArg_ParseTuple(args, "|O:readline", &limitobj)) {\r | |
1855 | return NULL;\r | |
1856 | }\r | |
1857 | if (limitobj) {\r | |
1858 | if (!PyNumber_Check(limitobj)) {\r | |
1859 | PyErr_Format(PyExc_TypeError,\r | |
1860 | "integer argument expected, got '%.200s'",\r | |
1861 | Py_TYPE(limitobj)->tp_name);\r | |
1862 | return NULL;\r | |
1863 | }\r | |
1864 | limit = PyNumber_AsSsize_t(limitobj, PyExc_OverflowError);\r | |
1865 | if (limit == -1 && PyErr_Occurred())\r | |
1866 | return NULL;\r | |
1867 | }\r | |
1868 | return _textiowrapper_readline(self, limit);\r | |
1869 | }\r | |
1870 | \r | |
1871 | /* Seek and Tell */\r | |
1872 | \r | |
1873 | typedef struct {\r | |
1874 | Py_off_t start_pos;\r | |
1875 | int dec_flags;\r | |
1876 | int bytes_to_feed;\r | |
1877 | int chars_to_skip;\r | |
1878 | char need_eof;\r | |
1879 | } cookie_type;\r | |
1880 | \r | |
1881 | /*\r | |
1882 | To speed up cookie packing/unpacking, we store the fields in a temporary\r | |
1883 | string and call _PyLong_FromByteArray() or _PyLong_AsByteArray (resp.).\r | |
1884 | The following macros define at which offsets in the intermediary byte\r | |
1885 | string the various CookieStruct fields will be stored.\r | |
1886 | */\r | |
1887 | \r | |
1888 | #define COOKIE_BUF_LEN (sizeof(Py_off_t) + 3 * sizeof(int) + sizeof(char))\r | |
1889 | \r | |
1890 | #if defined(WORDS_BIGENDIAN)\r | |
1891 | \r | |
1892 | # define IS_LITTLE_ENDIAN 0\r | |
1893 | \r | |
1894 | /* We want the least significant byte of start_pos to also be the least\r | |
1895 | significant byte of the cookie, which means that in big-endian mode we\r | |
1896 | must copy the fields in reverse order. */\r | |
1897 | \r | |
1898 | # define OFF_START_POS (sizeof(char) + 3 * sizeof(int))\r | |
1899 | # define OFF_DEC_FLAGS (sizeof(char) + 2 * sizeof(int))\r | |
1900 | # define OFF_BYTES_TO_FEED (sizeof(char) + sizeof(int))\r | |
1901 | # define OFF_CHARS_TO_SKIP (sizeof(char))\r | |
1902 | # define OFF_NEED_EOF 0\r | |
1903 | \r | |
1904 | #else\r | |
1905 | \r | |
1906 | # define IS_LITTLE_ENDIAN 1\r | |
1907 | \r | |
1908 | /* Little-endian mode: the least significant byte of start_pos will\r | |
1909 | naturally end up the least significant byte of the cookie. */\r | |
1910 | \r | |
1911 | # define OFF_START_POS 0\r | |
1912 | # define OFF_DEC_FLAGS (sizeof(Py_off_t))\r | |
1913 | # define OFF_BYTES_TO_FEED (sizeof(Py_off_t) + sizeof(int))\r | |
1914 | # define OFF_CHARS_TO_SKIP (sizeof(Py_off_t) + 2 * sizeof(int))\r | |
1915 | # define OFF_NEED_EOF (sizeof(Py_off_t) + 3 * sizeof(int))\r | |
1916 | \r | |
1917 | #endif\r | |
1918 | \r | |
1919 | static int\r | |
1920 | textiowrapper_parse_cookie(cookie_type *cookie, PyObject *cookieObj)\r | |
1921 | {\r | |
1922 | unsigned char buffer[COOKIE_BUF_LEN];\r | |
1923 | PyLongObject *cookieLong = (PyLongObject *)PyNumber_Long(cookieObj);\r | |
1924 | if (cookieLong == NULL)\r | |
1925 | return -1;\r | |
1926 | \r | |
1927 | if (_PyLong_AsByteArray(cookieLong, buffer, sizeof(buffer),\r | |
1928 | IS_LITTLE_ENDIAN, 0) < 0) {\r | |
1929 | Py_DECREF(cookieLong);\r | |
1930 | return -1;\r | |
1931 | }\r | |
1932 | Py_DECREF(cookieLong);\r | |
1933 | \r | |
1934 | memcpy(&cookie->start_pos, buffer + OFF_START_POS, sizeof(cookie->start_pos));\r | |
1935 | memcpy(&cookie->dec_flags, buffer + OFF_DEC_FLAGS, sizeof(cookie->dec_flags));\r | |
1936 | memcpy(&cookie->bytes_to_feed, buffer + OFF_BYTES_TO_FEED, sizeof(cookie->bytes_to_feed));\r | |
1937 | memcpy(&cookie->chars_to_skip, buffer + OFF_CHARS_TO_SKIP, sizeof(cookie->chars_to_skip));\r | |
1938 | memcpy(&cookie->need_eof, buffer + OFF_NEED_EOF, sizeof(cookie->need_eof));\r | |
1939 | \r | |
1940 | return 0;\r | |
1941 | }\r | |
1942 | \r | |
1943 | static PyObject *\r | |
1944 | textiowrapper_build_cookie(cookie_type *cookie)\r | |
1945 | {\r | |
1946 | unsigned char buffer[COOKIE_BUF_LEN];\r | |
1947 | \r | |
1948 | memcpy(buffer + OFF_START_POS, &cookie->start_pos, sizeof(cookie->start_pos));\r | |
1949 | memcpy(buffer + OFF_DEC_FLAGS, &cookie->dec_flags, sizeof(cookie->dec_flags));\r | |
1950 | memcpy(buffer + OFF_BYTES_TO_FEED, &cookie->bytes_to_feed, sizeof(cookie->bytes_to_feed));\r | |
1951 | memcpy(buffer + OFF_CHARS_TO_SKIP, &cookie->chars_to_skip, sizeof(cookie->chars_to_skip));\r | |
1952 | memcpy(buffer + OFF_NEED_EOF, &cookie->need_eof, sizeof(cookie->need_eof));\r | |
1953 | \r | |
1954 | return _PyLong_FromByteArray(buffer, sizeof(buffer), IS_LITTLE_ENDIAN, 0);\r | |
1955 | }\r | |
1956 | #undef IS_LITTLE_ENDIAN\r | |
1957 | \r | |
1958 | static int\r | |
1959 | _textiowrapper_decoder_setstate(textio *self, cookie_type *cookie)\r | |
1960 | {\r | |
1961 | PyObject *res;\r | |
1962 | /* When seeking to the start of the stream, we call decoder.reset()\r | |
1963 | rather than decoder.getstate().\r | |
1964 | This is for a few decoders such as utf-16 for which the state value\r | |
1965 | at start is not (b"", 0) but e.g. (b"", 2) (meaning, in the case of\r | |
1966 | utf-16, that we are expecting a BOM).\r | |
1967 | */\r | |
1968 | if (cookie->start_pos == 0 && cookie->dec_flags == 0)\r | |
1969 | res = PyObject_CallMethodObjArgs(self->decoder, _PyIO_str_reset, NULL);\r | |
1970 | else\r | |
1971 | res = PyObject_CallMethod(self->decoder, "setstate",\r | |
1972 | "((si))", "", cookie->dec_flags);\r | |
1973 | if (res == NULL)\r | |
1974 | return -1;\r | |
1975 | Py_DECREF(res);\r | |
1976 | return 0;\r | |
1977 | }\r | |
1978 | \r | |
1979 | static int\r | |
1980 | _textiowrapper_encoder_setstate(textio *self, cookie_type *cookie)\r | |
1981 | {\r | |
1982 | PyObject *res;\r | |
1983 | /* Same as _textiowrapper_decoder_setstate() above. */\r | |
1984 | if (cookie->start_pos == 0 && cookie->dec_flags == 0) {\r | |
1985 | res = PyObject_CallMethodObjArgs(self->encoder, _PyIO_str_reset, NULL);\r | |
1986 | self->encoding_start_of_stream = 1;\r | |
1987 | }\r | |
1988 | else {\r | |
1989 | res = PyObject_CallMethodObjArgs(self->encoder, _PyIO_str_setstate,\r | |
1990 | _PyIO_zero, NULL);\r | |
1991 | self->encoding_start_of_stream = 0;\r | |
1992 | }\r | |
1993 | if (res == NULL)\r | |
1994 | return -1;\r | |
1995 | Py_DECREF(res);\r | |
1996 | return 0;\r | |
1997 | }\r | |
1998 | \r | |
1999 | static PyObject *\r | |
2000 | textiowrapper_seek(textio *self, PyObject *args)\r | |
2001 | {\r | |
2002 | PyObject *cookieObj, *posobj;\r | |
2003 | cookie_type cookie;\r | |
2004 | int whence = 0;\r | |
2005 | PyObject *res;\r | |
2006 | int cmp;\r | |
2007 | \r | |
2008 | CHECK_ATTACHED(self);\r | |
2009 | \r | |
2010 | if (!PyArg_ParseTuple(args, "O|i:seek", &cookieObj, &whence))\r | |
2011 | return NULL;\r | |
2012 | CHECK_CLOSED(self);\r | |
2013 | \r | |
2014 | Py_INCREF(cookieObj);\r | |
2015 | \r | |
2016 | if (!self->seekable) {\r | |
2017 | PyErr_SetString(PyExc_IOError,\r | |
2018 | "underlying stream is not seekable");\r | |
2019 | goto fail;\r | |
2020 | }\r | |
2021 | \r | |
2022 | if (whence == 1) {\r | |
2023 | /* seek relative to current position */\r | |
2024 | cmp = PyObject_RichCompareBool(cookieObj, _PyIO_zero, Py_EQ);\r | |
2025 | if (cmp < 0)\r | |
2026 | goto fail;\r | |
2027 | \r | |
2028 | if (cmp == 0) {\r | |
2029 | PyErr_SetString(PyExc_IOError,\r | |
2030 | "can't do nonzero cur-relative seeks");\r | |
2031 | goto fail;\r | |
2032 | }\r | |
2033 | \r | |
2034 | /* Seeking to the current position should attempt to\r | |
2035 | * sync the underlying buffer with the current position.\r | |
2036 | */\r | |
2037 | Py_DECREF(cookieObj);\r | |
2038 | cookieObj = PyObject_CallMethod((PyObject *)self, "tell", NULL);\r | |
2039 | if (cookieObj == NULL)\r | |
2040 | goto fail;\r | |
2041 | }\r | |
2042 | else if (whence == 2) {\r | |
2043 | /* seek relative to end of file */\r | |
2044 | \r | |
2045 | cmp = PyObject_RichCompareBool(cookieObj, _PyIO_zero, Py_EQ);\r | |
2046 | if (cmp < 0)\r | |
2047 | goto fail;\r | |
2048 | \r | |
2049 | if (cmp == 0) {\r | |
2050 | PyErr_SetString(PyExc_IOError,\r | |
2051 | "can't do nonzero end-relative seeks");\r | |
2052 | goto fail;\r | |
2053 | }\r | |
2054 | \r | |
2055 | res = PyObject_CallMethod((PyObject *)self, "flush", NULL);\r | |
2056 | if (res == NULL)\r | |
2057 | goto fail;\r | |
2058 | Py_DECREF(res);\r | |
2059 | \r | |
2060 | textiowrapper_set_decoded_chars(self, NULL);\r | |
2061 | Py_CLEAR(self->snapshot);\r | |
2062 | if (self->decoder) {\r | |
2063 | res = PyObject_CallMethod(self->decoder, "reset", NULL);\r | |
2064 | if (res == NULL)\r | |
2065 | goto fail;\r | |
2066 | Py_DECREF(res);\r | |
2067 | }\r | |
2068 | \r | |
2069 | res = PyObject_CallMethod(self->buffer, "seek", "ii", 0, 2);\r | |
2070 | Py_XDECREF(cookieObj);\r | |
2071 | return res;\r | |
2072 | }\r | |
2073 | else if (whence != 0) {\r | |
2074 | PyErr_Format(PyExc_ValueError,\r | |
2075 | "invalid whence (%d, should be 0, 1 or 2)", whence);\r | |
2076 | goto fail;\r | |
2077 | }\r | |
2078 | \r | |
2079 | cmp = PyObject_RichCompareBool(cookieObj, _PyIO_zero, Py_LT);\r | |
2080 | if (cmp < 0)\r | |
2081 | goto fail;\r | |
2082 | \r | |
2083 | if (cmp == 1) {\r | |
2084 | PyObject *repr = PyObject_Repr(cookieObj);\r | |
2085 | if (repr != NULL) {\r | |
2086 | PyErr_Format(PyExc_ValueError,\r | |
2087 | "negative seek position %s",\r | |
2088 | PyString_AS_STRING(repr));\r | |
2089 | Py_DECREF(repr);\r | |
2090 | }\r | |
2091 | goto fail;\r | |
2092 | }\r | |
2093 | \r | |
2094 | res = PyObject_CallMethodObjArgs((PyObject *)self, _PyIO_str_flush, NULL);\r | |
2095 | if (res == NULL)\r | |
2096 | goto fail;\r | |
2097 | Py_DECREF(res);\r | |
2098 | \r | |
2099 | /* The strategy of seek() is to go back to the safe start point\r | |
2100 | * and replay the effect of read(chars_to_skip) from there.\r | |
2101 | */\r | |
2102 | if (textiowrapper_parse_cookie(&cookie, cookieObj) < 0)\r | |
2103 | goto fail;\r | |
2104 | \r | |
2105 | /* Seek back to the safe start point. */\r | |
2106 | posobj = PyLong_FromOff_t(cookie.start_pos);\r | |
2107 | if (posobj == NULL)\r | |
2108 | goto fail;\r | |
2109 | res = PyObject_CallMethodObjArgs(self->buffer,\r | |
2110 | _PyIO_str_seek, posobj, NULL);\r | |
2111 | Py_DECREF(posobj);\r | |
2112 | if (res == NULL)\r | |
2113 | goto fail;\r | |
2114 | Py_DECREF(res);\r | |
2115 | \r | |
2116 | textiowrapper_set_decoded_chars(self, NULL);\r | |
2117 | Py_CLEAR(self->snapshot);\r | |
2118 | \r | |
2119 | /* Restore the decoder to its state from the safe start point. */\r | |
2120 | if (self->decoder) {\r | |
2121 | if (_textiowrapper_decoder_setstate(self, &cookie) < 0)\r | |
2122 | goto fail;\r | |
2123 | }\r | |
2124 | \r | |
2125 | if (cookie.chars_to_skip) {\r | |
2126 | /* Just like _read_chunk, feed the decoder and save a snapshot. */\r | |
2127 | PyObject *input_chunk = PyObject_CallMethod(\r | |
2128 | self->buffer, "read", "i", cookie.bytes_to_feed);\r | |
2129 | PyObject *decoded;\r | |
2130 | \r | |
2131 | if (input_chunk == NULL)\r | |
2132 | goto fail;\r | |
2133 | \r | |
2134 | if (!PyBytes_Check(input_chunk)) {\r | |
2135 | PyErr_Format(PyExc_TypeError,\r | |
2136 | "underlying read() should have returned a bytes "\r | |
2137 | "object, not '%.200s'",\r | |
2138 | Py_TYPE(input_chunk)->tp_name);\r | |
2139 | Py_DECREF(input_chunk);\r | |
2140 | goto fail;\r | |
2141 | }\r | |
2142 | \r | |
2143 | self->snapshot = Py_BuildValue("iN", cookie.dec_flags, input_chunk);\r | |
2144 | if (self->snapshot == NULL) {\r | |
2145 | Py_DECREF(input_chunk);\r | |
2146 | goto fail;\r | |
2147 | }\r | |
2148 | \r | |
2149 | decoded = PyObject_CallMethod(self->decoder, "decode",\r | |
2150 | "Oi", input_chunk, (int)cookie.need_eof);\r | |
2151 | \r | |
2152 | if (check_decoded(decoded) < 0)\r | |
2153 | goto fail;\r | |
2154 | \r | |
2155 | textiowrapper_set_decoded_chars(self, decoded);\r | |
2156 | \r | |
2157 | /* Skip chars_to_skip of the decoded characters. */\r | |
2158 | if (PyUnicode_GetSize(self->decoded_chars) < cookie.chars_to_skip) {\r | |
2159 | PyErr_SetString(PyExc_IOError, "can't restore logical file position");\r | |
2160 | goto fail;\r | |
2161 | }\r | |
2162 | self->decoded_chars_used = cookie.chars_to_skip;\r | |
2163 | }\r | |
2164 | else {\r | |
2165 | self->snapshot = Py_BuildValue("is", cookie.dec_flags, "");\r | |
2166 | if (self->snapshot == NULL)\r | |
2167 | goto fail;\r | |
2168 | }\r | |
2169 | \r | |
2170 | /* Finally, reset the encoder (merely useful for proper BOM handling) */\r | |
2171 | if (self->encoder) {\r | |
2172 | if (_textiowrapper_encoder_setstate(self, &cookie) < 0)\r | |
2173 | goto fail;\r | |
2174 | }\r | |
2175 | return cookieObj;\r | |
2176 | fail:\r | |
2177 | Py_XDECREF(cookieObj);\r | |
2178 | return NULL;\r | |
2179 | \r | |
2180 | }\r | |
2181 | \r | |
2182 | static PyObject *\r | |
2183 | textiowrapper_tell(textio *self, PyObject *args)\r | |
2184 | {\r | |
2185 | PyObject *res;\r | |
2186 | PyObject *posobj = NULL;\r | |
2187 | cookie_type cookie = {0,0,0,0,0};\r | |
2188 | PyObject *next_input;\r | |
2189 | Py_ssize_t chars_to_skip, chars_decoded;\r | |
2190 | PyObject *saved_state = NULL;\r | |
2191 | char *input, *input_end;\r | |
2192 | \r | |
2193 | CHECK_ATTACHED(self);\r | |
2194 | CHECK_CLOSED(self);\r | |
2195 | \r | |
2196 | if (!self->seekable) {\r | |
2197 | PyErr_SetString(PyExc_IOError,\r | |
2198 | "underlying stream is not seekable");\r | |
2199 | goto fail;\r | |
2200 | }\r | |
2201 | if (!self->telling) {\r | |
2202 | PyErr_SetString(PyExc_IOError,\r | |
2203 | "telling position disabled by next() call");\r | |
2204 | goto fail;\r | |
2205 | }\r | |
2206 | \r | |
2207 | if (_textiowrapper_writeflush(self) < 0)\r | |
2208 | return NULL;\r | |
2209 | res = PyObject_CallMethod((PyObject *)self, "flush", NULL);\r | |
2210 | if (res == NULL)\r | |
2211 | goto fail;\r | |
2212 | Py_DECREF(res);\r | |
2213 | \r | |
2214 | posobj = PyObject_CallMethod(self->buffer, "tell", NULL);\r | |
2215 | if (posobj == NULL)\r | |
2216 | goto fail;\r | |
2217 | \r | |
2218 | if (self->decoder == NULL || self->snapshot == NULL) {\r | |
2219 | assert (self->decoded_chars == NULL || PyUnicode_GetSize(self->decoded_chars) == 0);\r | |
2220 | return posobj;\r | |
2221 | }\r | |
2222 | \r | |
2223 | #if defined(HAVE_LARGEFILE_SUPPORT)\r | |
2224 | cookie.start_pos = PyLong_AsLongLong(posobj);\r | |
2225 | #else\r | |
2226 | cookie.start_pos = PyLong_AsLong(posobj);\r | |
2227 | #endif\r | |
2228 | if (PyErr_Occurred())\r | |
2229 | goto fail;\r | |
2230 | \r | |
2231 | /* Skip backward to the snapshot point (see _read_chunk). */\r | |
2232 | if (!PyArg_Parse(self->snapshot, "(iO)", &cookie.dec_flags, &next_input))\r | |
2233 | goto fail;\r | |
2234 | \r | |
2235 | assert (PyBytes_Check(next_input));\r | |
2236 | \r | |
2237 | cookie.start_pos -= PyBytes_GET_SIZE(next_input);\r | |
2238 | \r | |
2239 | /* How many decoded characters have been used up since the snapshot? */\r | |
2240 | if (self->decoded_chars_used == 0) {\r | |
2241 | /* We haven't moved from the snapshot point. */\r | |
2242 | Py_DECREF(posobj);\r | |
2243 | return textiowrapper_build_cookie(&cookie);\r | |
2244 | }\r | |
2245 | \r | |
2246 | chars_to_skip = self->decoded_chars_used;\r | |
2247 | \r | |
2248 | /* Starting from the snapshot position, we will walk the decoder\r | |
2249 | * forward until it gives us enough decoded characters.\r | |
2250 | */\r | |
2251 | saved_state = PyObject_CallMethodObjArgs(self->decoder,\r | |
2252 | _PyIO_str_getstate, NULL);\r | |
2253 | if (saved_state == NULL)\r | |
2254 | goto fail;\r | |
2255 | \r | |
2256 | /* Note our initial start point. */\r | |
2257 | if (_textiowrapper_decoder_setstate(self, &cookie) < 0)\r | |
2258 | goto fail;\r | |
2259 | \r | |
2260 | /* Feed the decoder one byte at a time. As we go, note the\r | |
2261 | * nearest "safe start point" before the current location\r | |
2262 | * (a point where the decoder has nothing buffered, so seek()\r | |
2263 | * can safely start from there and advance to this location).\r | |
2264 | */\r | |
2265 | chars_decoded = 0;\r | |
2266 | input = PyBytes_AS_STRING(next_input);\r | |
2267 | input_end = input + PyBytes_GET_SIZE(next_input);\r | |
2268 | while (input < input_end) {\r | |
2269 | PyObject *state;\r | |
2270 | char *dec_buffer;\r | |
2271 | Py_ssize_t dec_buffer_len;\r | |
2272 | int dec_flags;\r | |
2273 | \r | |
2274 | PyObject *decoded = PyObject_CallMethod(\r | |
2275 | self->decoder, "decode", "s#", input, (Py_ssize_t)1);\r | |
2276 | if (check_decoded(decoded) < 0)\r | |
2277 | goto fail;\r | |
2278 | chars_decoded += PyUnicode_GET_SIZE(decoded);\r | |
2279 | Py_DECREF(decoded);\r | |
2280 | \r | |
2281 | cookie.bytes_to_feed += 1;\r | |
2282 | \r | |
2283 | state = PyObject_CallMethodObjArgs(self->decoder,\r | |
2284 | _PyIO_str_getstate, NULL);\r | |
2285 | if (state == NULL)\r | |
2286 | goto fail;\r | |
2287 | if (!PyArg_Parse(state, "(s#i)", &dec_buffer, &dec_buffer_len, &dec_flags)) {\r | |
2288 | Py_DECREF(state);\r | |
2289 | goto fail;\r | |
2290 | }\r | |
2291 | Py_DECREF(state);\r | |
2292 | \r | |
2293 | if (dec_buffer_len == 0 && chars_decoded <= chars_to_skip) {\r | |
2294 | /* Decoder buffer is empty, so this is a safe start point. */\r | |
2295 | cookie.start_pos += cookie.bytes_to_feed;\r | |
2296 | chars_to_skip -= chars_decoded;\r | |
2297 | cookie.dec_flags = dec_flags;\r | |
2298 | cookie.bytes_to_feed = 0;\r | |
2299 | chars_decoded = 0;\r | |
2300 | }\r | |
2301 | if (chars_decoded >= chars_to_skip)\r | |
2302 | break;\r | |
2303 | input++;\r | |
2304 | }\r | |
2305 | if (input == input_end) {\r | |
2306 | /* We didn't get enough decoded data; signal EOF to get more. */\r | |
2307 | PyObject *decoded = PyObject_CallMethod(\r | |
2308 | self->decoder, "decode", "si", "", /* final = */ 1);\r | |
2309 | if (check_decoded(decoded) < 0)\r | |
2310 | goto fail;\r | |
2311 | chars_decoded += PyUnicode_GET_SIZE(decoded);\r | |
2312 | Py_DECREF(decoded);\r | |
2313 | cookie.need_eof = 1;\r | |
2314 | \r | |
2315 | if (chars_decoded < chars_to_skip) {\r | |
2316 | PyErr_SetString(PyExc_IOError,\r | |
2317 | "can't reconstruct logical file position");\r | |
2318 | goto fail;\r | |
2319 | }\r | |
2320 | }\r | |
2321 | \r | |
2322 | /* finally */\r | |
2323 | Py_XDECREF(posobj);\r | |
2324 | res = PyObject_CallMethod(self->decoder, "setstate", "(O)", saved_state);\r | |
2325 | Py_DECREF(saved_state);\r | |
2326 | if (res == NULL)\r | |
2327 | return NULL;\r | |
2328 | Py_DECREF(res);\r | |
2329 | \r | |
2330 | /* The returned cookie corresponds to the last safe start point. */\r | |
2331 | cookie.chars_to_skip = Py_SAFE_DOWNCAST(chars_to_skip, Py_ssize_t, int);\r | |
2332 | return textiowrapper_build_cookie(&cookie);\r | |
2333 | \r | |
2334 | fail:\r | |
2335 | Py_XDECREF(posobj);\r | |
2336 | if (saved_state) {\r | |
2337 | PyObject *type, *value, *traceback;\r | |
2338 | PyErr_Fetch(&type, &value, &traceback);\r | |
2339 | \r | |
2340 | res = PyObject_CallMethod(self->decoder, "setstate", "(O)", saved_state);\r | |
2341 | _PyErr_ReplaceException(type, value, traceback);\r | |
2342 | Py_DECREF(saved_state);\r | |
2343 | Py_XDECREF(res);\r | |
2344 | }\r | |
2345 | return NULL;\r | |
2346 | }\r | |
2347 | \r | |
2348 | static PyObject *\r | |
2349 | textiowrapper_truncate(textio *self, PyObject *args)\r | |
2350 | {\r | |
2351 | PyObject *pos = Py_None;\r | |
2352 | PyObject *res;\r | |
2353 | \r | |
2354 | CHECK_ATTACHED(self)\r | |
2355 | if (!PyArg_ParseTuple(args, "|O:truncate", &pos)) {\r | |
2356 | return NULL;\r | |
2357 | }\r | |
2358 | \r | |
2359 | res = PyObject_CallMethodObjArgs((PyObject *) self, _PyIO_str_flush, NULL);\r | |
2360 | if (res == NULL)\r | |
2361 | return NULL;\r | |
2362 | Py_DECREF(res);\r | |
2363 | \r | |
2364 | return PyObject_CallMethodObjArgs(self->buffer, _PyIO_str_truncate, pos, NULL);\r | |
2365 | }\r | |
2366 | \r | |
2367 | static PyObject *\r | |
2368 | textiowrapper_repr(textio *self)\r | |
2369 | {\r | |
2370 | PyObject *nameobj, *res;\r | |
2371 | PyObject *namerepr = NULL, *encrepr = NULL;\r | |
2372 | \r | |
2373 | CHECK_INITIALIZED(self);\r | |
2374 | \r | |
2375 | nameobj = PyObject_GetAttrString((PyObject *) self, "name");\r | |
2376 | if (nameobj == NULL) {\r | |
2377 | if (PyErr_ExceptionMatches(PyExc_Exception))\r | |
2378 | PyErr_Clear();\r | |
2379 | else\r | |
2380 | goto error;\r | |
2381 | encrepr = PyObject_Repr(self->encoding);\r | |
2382 | res = PyString_FromFormat("<_io.TextIOWrapper encoding=%s>",\r | |
2383 | PyString_AS_STRING(encrepr));\r | |
2384 | }\r | |
2385 | else {\r | |
2386 | encrepr = PyObject_Repr(self->encoding);\r | |
2387 | namerepr = PyObject_Repr(nameobj);\r | |
2388 | res = PyString_FromFormat("<_io.TextIOWrapper name=%s encoding=%s>",\r | |
2389 | PyString_AS_STRING(namerepr),\r | |
2390 | PyString_AS_STRING(encrepr));\r | |
2391 | Py_DECREF(nameobj);\r | |
2392 | }\r | |
2393 | Py_XDECREF(namerepr);\r | |
2394 | Py_XDECREF(encrepr);\r | |
2395 | return res;\r | |
2396 | \r | |
2397 | error:\r | |
2398 | Py_XDECREF(namerepr);\r | |
2399 | Py_XDECREF(encrepr);\r | |
2400 | return NULL;\r | |
2401 | }\r | |
2402 | \r | |
2403 | \r | |
2404 | /* Inquiries */\r | |
2405 | \r | |
2406 | static PyObject *\r | |
2407 | textiowrapper_fileno(textio *self, PyObject *args)\r | |
2408 | {\r | |
2409 | CHECK_ATTACHED(self);\r | |
2410 | return PyObject_CallMethod(self->buffer, "fileno", NULL);\r | |
2411 | }\r | |
2412 | \r | |
2413 | static PyObject *\r | |
2414 | textiowrapper_seekable(textio *self, PyObject *args)\r | |
2415 | {\r | |
2416 | CHECK_ATTACHED(self);\r | |
2417 | return PyObject_CallMethod(self->buffer, "seekable", NULL);\r | |
2418 | }\r | |
2419 | \r | |
2420 | static PyObject *\r | |
2421 | textiowrapper_readable(textio *self, PyObject *args)\r | |
2422 | {\r | |
2423 | CHECK_ATTACHED(self);\r | |
2424 | return PyObject_CallMethod(self->buffer, "readable", NULL);\r | |
2425 | }\r | |
2426 | \r | |
2427 | static PyObject *\r | |
2428 | textiowrapper_writable(textio *self, PyObject *args)\r | |
2429 | {\r | |
2430 | CHECK_ATTACHED(self);\r | |
2431 | return PyObject_CallMethod(self->buffer, "writable", NULL);\r | |
2432 | }\r | |
2433 | \r | |
2434 | static PyObject *\r | |
2435 | textiowrapper_isatty(textio *self, PyObject *args)\r | |
2436 | {\r | |
2437 | CHECK_ATTACHED(self);\r | |
2438 | return PyObject_CallMethod(self->buffer, "isatty", NULL);\r | |
2439 | }\r | |
2440 | \r | |
2441 | static PyObject *\r | |
2442 | textiowrapper_flush(textio *self, PyObject *args)\r | |
2443 | {\r | |
2444 | CHECK_ATTACHED(self);\r | |
2445 | CHECK_CLOSED(self);\r | |
2446 | self->telling = self->seekable;\r | |
2447 | if (_textiowrapper_writeflush(self) < 0)\r | |
2448 | return NULL;\r | |
2449 | return PyObject_CallMethod(self->buffer, "flush", NULL);\r | |
2450 | }\r | |
2451 | \r | |
2452 | static PyObject *\r | |
2453 | textiowrapper_close(textio *self, PyObject *args)\r | |
2454 | {\r | |
2455 | PyObject *res;\r | |
2456 | int r;\r | |
2457 | CHECK_ATTACHED(self);\r | |
2458 | \r | |
2459 | res = textiowrapper_closed_get(self, NULL);\r | |
2460 | if (res == NULL)\r | |
2461 | return NULL;\r | |
2462 | r = PyObject_IsTrue(res);\r | |
2463 | Py_DECREF(res);\r | |
2464 | if (r < 0)\r | |
2465 | return NULL;\r | |
2466 | \r | |
2467 | if (r > 0) {\r | |
2468 | Py_RETURN_NONE; /* stream already closed */\r | |
2469 | }\r | |
2470 | else {\r | |
2471 | PyObject *exc = NULL, *val, *tb;\r | |
2472 | res = PyObject_CallMethod((PyObject *)self, "flush", NULL);\r | |
2473 | if (res == NULL)\r | |
2474 | PyErr_Fetch(&exc, &val, &tb);\r | |
2475 | else\r | |
2476 | Py_DECREF(res);\r | |
2477 | \r | |
2478 | res = PyObject_CallMethod(self->buffer, "close", NULL);\r | |
2479 | if (exc != NULL) {\r | |
2480 | _PyErr_ReplaceException(exc, val, tb);\r | |
2481 | Py_CLEAR(res);\r | |
2482 | }\r | |
2483 | return res;\r | |
2484 | }\r | |
2485 | }\r | |
2486 | \r | |
2487 | static PyObject *\r | |
2488 | textiowrapper_iternext(textio *self)\r | |
2489 | {\r | |
2490 | PyObject *line;\r | |
2491 | \r | |
2492 | CHECK_ATTACHED(self);\r | |
2493 | \r | |
2494 | self->telling = 0;\r | |
2495 | if (Py_TYPE(self) == &PyTextIOWrapper_Type) {\r | |
2496 | /* Skip method call overhead for speed */\r | |
2497 | line = _textiowrapper_readline(self, -1);\r | |
2498 | }\r | |
2499 | else {\r | |
2500 | line = PyObject_CallMethodObjArgs((PyObject *)self,\r | |
2501 | _PyIO_str_readline, NULL);\r | |
2502 | if (line && !PyUnicode_Check(line)) {\r | |
2503 | PyErr_Format(PyExc_IOError,\r | |
2504 | "readline() should have returned an str object, "\r | |
2505 | "not '%.200s'", Py_TYPE(line)->tp_name);\r | |
2506 | Py_DECREF(line);\r | |
2507 | return NULL;\r | |
2508 | }\r | |
2509 | }\r | |
2510 | \r | |
2511 | if (line == NULL)\r | |
2512 | return NULL;\r | |
2513 | \r | |
2514 | if (PyUnicode_GET_SIZE(line) == 0) {\r | |
2515 | /* Reached EOF or would have blocked */\r | |
2516 | Py_DECREF(line);\r | |
2517 | Py_CLEAR(self->snapshot);\r | |
2518 | self->telling = self->seekable;\r | |
2519 | return NULL;\r | |
2520 | }\r | |
2521 | \r | |
2522 | return line;\r | |
2523 | }\r | |
2524 | \r | |
2525 | static PyObject *\r | |
2526 | textiowrapper_name_get(textio *self, void *context)\r | |
2527 | {\r | |
2528 | CHECK_ATTACHED(self);\r | |
2529 | return PyObject_GetAttrString(self->buffer, "name");\r | |
2530 | }\r | |
2531 | \r | |
2532 | static PyObject *\r | |
2533 | textiowrapper_closed_get(textio *self, void *context)\r | |
2534 | {\r | |
2535 | CHECK_ATTACHED(self);\r | |
2536 | return PyObject_GetAttr(self->buffer, _PyIO_str_closed);\r | |
2537 | }\r | |
2538 | \r | |
2539 | static PyObject *\r | |
2540 | textiowrapper_newlines_get(textio *self, void *context)\r | |
2541 | {\r | |
2542 | PyObject *res;\r | |
2543 | CHECK_ATTACHED(self);\r | |
2544 | if (self->decoder == NULL)\r | |
2545 | Py_RETURN_NONE;\r | |
2546 | res = PyObject_GetAttr(self->decoder, _PyIO_str_newlines);\r | |
2547 | if (res == NULL) {\r | |
2548 | if (PyErr_ExceptionMatches(PyExc_AttributeError)) {\r | |
2549 | PyErr_Clear();\r | |
2550 | Py_RETURN_NONE;\r | |
2551 | }\r | |
2552 | else {\r | |
2553 | return NULL;\r | |
2554 | }\r | |
2555 | }\r | |
2556 | return res;\r | |
2557 | }\r | |
2558 | \r | |
2559 | static PyObject *\r | |
2560 | textiowrapper_errors_get(textio *self, void *context)\r | |
2561 | {\r | |
2562 | CHECK_INITIALIZED(self);\r | |
2563 | Py_INCREF(self->errors);\r | |
2564 | return self->errors;\r | |
2565 | }\r | |
2566 | \r | |
2567 | static PyObject *\r | |
2568 | textiowrapper_chunk_size_get(textio *self, void *context)\r | |
2569 | {\r | |
2570 | CHECK_ATTACHED(self);\r | |
2571 | return PyLong_FromSsize_t(self->chunk_size);\r | |
2572 | }\r | |
2573 | \r | |
2574 | static int\r | |
2575 | textiowrapper_chunk_size_set(textio *self, PyObject *arg, void *context)\r | |
2576 | {\r | |
2577 | Py_ssize_t n;\r | |
2578 | CHECK_ATTACHED_INT(self);\r | |
2579 | n = PyNumber_AsSsize_t(arg, PyExc_TypeError);\r | |
2580 | if (n == -1 && PyErr_Occurred())\r | |
2581 | return -1;\r | |
2582 | if (n <= 0) {\r | |
2583 | PyErr_SetString(PyExc_ValueError,\r | |
2584 | "a strictly positive integer is required");\r | |
2585 | return -1;\r | |
2586 | }\r | |
2587 | self->chunk_size = n;\r | |
2588 | return 0;\r | |
2589 | }\r | |
2590 | \r | |
2591 | static PyMethodDef textiowrapper_methods[] = {\r | |
2592 | {"detach", (PyCFunction)textiowrapper_detach, METH_NOARGS},\r | |
2593 | {"write", (PyCFunction)textiowrapper_write, METH_VARARGS},\r | |
2594 | {"read", (PyCFunction)textiowrapper_read, METH_VARARGS},\r | |
2595 | {"readline", (PyCFunction)textiowrapper_readline, METH_VARARGS},\r | |
2596 | {"flush", (PyCFunction)textiowrapper_flush, METH_NOARGS},\r | |
2597 | {"close", (PyCFunction)textiowrapper_close, METH_NOARGS},\r | |
2598 | \r | |
2599 | {"fileno", (PyCFunction)textiowrapper_fileno, METH_NOARGS},\r | |
2600 | {"seekable", (PyCFunction)textiowrapper_seekable, METH_NOARGS},\r | |
2601 | {"readable", (PyCFunction)textiowrapper_readable, METH_NOARGS},\r | |
2602 | {"writable", (PyCFunction)textiowrapper_writable, METH_NOARGS},\r | |
2603 | {"isatty", (PyCFunction)textiowrapper_isatty, METH_NOARGS},\r | |
2604 | \r | |
2605 | {"seek", (PyCFunction)textiowrapper_seek, METH_VARARGS},\r | |
2606 | {"tell", (PyCFunction)textiowrapper_tell, METH_NOARGS},\r | |
2607 | {"truncate", (PyCFunction)textiowrapper_truncate, METH_VARARGS},\r | |
2608 | {NULL, NULL}\r | |
2609 | };\r | |
2610 | \r | |
2611 | static PyMemberDef textiowrapper_members[] = {\r | |
2612 | {"encoding", T_OBJECT, offsetof(textio, encoding), READONLY},\r | |
2613 | {"buffer", T_OBJECT, offsetof(textio, buffer), READONLY},\r | |
2614 | {"line_buffering", T_BOOL, offsetof(textio, line_buffering), READONLY},\r | |
2615 | {NULL}\r | |
2616 | };\r | |
2617 | \r | |
2618 | static PyGetSetDef textiowrapper_getset[] = {\r | |
2619 | {"name", (getter)textiowrapper_name_get, NULL, NULL},\r | |
2620 | {"closed", (getter)textiowrapper_closed_get, NULL, NULL},\r | |
2621 | /* {"mode", (getter)TextIOWrapper_mode_get, NULL, NULL},\r | |
2622 | */\r | |
2623 | {"newlines", (getter)textiowrapper_newlines_get, NULL, NULL},\r | |
2624 | {"errors", (getter)textiowrapper_errors_get, NULL, NULL},\r | |
2625 | {"_CHUNK_SIZE", (getter)textiowrapper_chunk_size_get,\r | |
2626 | (setter)textiowrapper_chunk_size_set, NULL},\r | |
2627 | {NULL}\r | |
2628 | };\r | |
2629 | \r | |
2630 | PyTypeObject PyTextIOWrapper_Type = {\r | |
2631 | PyVarObject_HEAD_INIT(NULL, 0)\r | |
2632 | "_io.TextIOWrapper", /*tp_name*/\r | |
2633 | sizeof(textio), /*tp_basicsize*/\r | |
2634 | 0, /*tp_itemsize*/\r | |
2635 | (destructor)textiowrapper_dealloc, /*tp_dealloc*/\r | |
2636 | 0, /*tp_print*/\r | |
2637 | 0, /*tp_getattr*/\r | |
2638 | 0, /*tps_etattr*/\r | |
2639 | 0, /*tp_compare */\r | |
2640 | (reprfunc)textiowrapper_repr,/*tp_repr*/\r | |
2641 | 0, /*tp_as_number*/\r | |
2642 | 0, /*tp_as_sequence*/\r | |
2643 | 0, /*tp_as_mapping*/\r | |
2644 | 0, /*tp_hash */\r | |
2645 | 0, /*tp_call*/\r | |
2646 | 0, /*tp_str*/\r | |
2647 | 0, /*tp_getattro*/\r | |
2648 | 0, /*tp_setattro*/\r | |
2649 | 0, /*tp_as_buffer*/\r | |
2650 | Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE\r | |
2651 | | Py_TPFLAGS_HAVE_GC, /*tp_flags*/\r | |
2652 | textiowrapper_doc, /* tp_doc */\r | |
2653 | (traverseproc)textiowrapper_traverse, /* tp_traverse */\r | |
2654 | (inquiry)textiowrapper_clear, /* tp_clear */\r | |
2655 | 0, /* tp_richcompare */\r | |
2656 | offsetof(textio, weakreflist), /*tp_weaklistoffset*/\r | |
2657 | 0, /* tp_iter */\r | |
2658 | (iternextfunc)textiowrapper_iternext, /* tp_iternext */\r | |
2659 | textiowrapper_methods, /* tp_methods */\r | |
2660 | textiowrapper_members, /* tp_members */\r | |
2661 | textiowrapper_getset, /* tp_getset */\r | |
2662 | 0, /* tp_base */\r | |
2663 | 0, /* tp_dict */\r | |
2664 | 0, /* tp_descr_get */\r | |
2665 | 0, /* tp_descr_set */\r | |
2666 | offsetof(textio, dict), /*tp_dictoffset*/\r | |
2667 | (initproc)textiowrapper_init, /* tp_init */\r | |
2668 | 0, /* tp_alloc */\r | |
2669 | PyType_GenericNew, /* tp_new */\r | |
2670 | };\r |