1 | n/a | /* |
---|
2 | n/a | An implementation of the new I/O lib as defined by PEP 3116 - "New I/O" |
---|
3 | n/a | |
---|
4 | n/a | Classes defined here: UnsupportedOperation, BlockingIOError. |
---|
5 | n/a | Functions defined here: open(). |
---|
6 | n/a | |
---|
7 | n/a | Mostly written by Amaury Forgeot d'Arc |
---|
8 | n/a | */ |
---|
9 | n/a | |
---|
10 | n/a | #define PY_SSIZE_T_CLEAN |
---|
11 | n/a | #include "Python.h" |
---|
12 | n/a | #include "structmember.h" |
---|
13 | n/a | #include "_iomodule.h" |
---|
14 | n/a | |
---|
15 | n/a | #ifdef HAVE_SYS_TYPES_H |
---|
16 | n/a | #include <sys/types.h> |
---|
17 | n/a | #endif /* HAVE_SYS_TYPES_H */ |
---|
18 | n/a | |
---|
19 | n/a | #ifdef HAVE_SYS_STAT_H |
---|
20 | n/a | #include <sys/stat.h> |
---|
21 | n/a | #endif /* HAVE_SYS_STAT_H */ |
---|
22 | n/a | |
---|
23 | n/a | #ifdef MS_WINDOWS |
---|
24 | n/a | #include <consoleapi.h> |
---|
25 | n/a | #endif |
---|
26 | n/a | |
---|
27 | n/a | /* Various interned strings */ |
---|
28 | n/a | |
---|
29 | n/a | PyObject *_PyIO_str_close; |
---|
30 | n/a | PyObject *_PyIO_str_closed; |
---|
31 | n/a | PyObject *_PyIO_str_decode; |
---|
32 | n/a | PyObject *_PyIO_str_encode; |
---|
33 | n/a | PyObject *_PyIO_str_fileno; |
---|
34 | n/a | PyObject *_PyIO_str_flush; |
---|
35 | n/a | PyObject *_PyIO_str_getstate; |
---|
36 | n/a | PyObject *_PyIO_str_isatty; |
---|
37 | n/a | PyObject *_PyIO_str_newlines; |
---|
38 | n/a | PyObject *_PyIO_str_nl; |
---|
39 | n/a | PyObject *_PyIO_str_read; |
---|
40 | n/a | PyObject *_PyIO_str_read1; |
---|
41 | n/a | PyObject *_PyIO_str_readable; |
---|
42 | n/a | PyObject *_PyIO_str_readall; |
---|
43 | n/a | PyObject *_PyIO_str_readinto; |
---|
44 | n/a | PyObject *_PyIO_str_readline; |
---|
45 | n/a | PyObject *_PyIO_str_reset; |
---|
46 | n/a | PyObject *_PyIO_str_seek; |
---|
47 | n/a | PyObject *_PyIO_str_seekable; |
---|
48 | n/a | PyObject *_PyIO_str_setstate; |
---|
49 | n/a | PyObject *_PyIO_str_tell; |
---|
50 | n/a | PyObject *_PyIO_str_truncate; |
---|
51 | n/a | PyObject *_PyIO_str_writable; |
---|
52 | n/a | PyObject *_PyIO_str_write; |
---|
53 | n/a | |
---|
54 | n/a | PyObject *_PyIO_empty_str; |
---|
55 | n/a | PyObject *_PyIO_empty_bytes; |
---|
56 | n/a | PyObject *_PyIO_zero; |
---|
57 | n/a | |
---|
58 | n/a | PyDoc_STRVAR(module_doc, |
---|
59 | n/a | "The io module provides the Python interfaces to stream handling. The\n" |
---|
60 | n/a | "builtin open function is defined in this module.\n" |
---|
61 | n/a | "\n" |
---|
62 | n/a | "At the top of the I/O hierarchy is the abstract base class IOBase. It\n" |
---|
63 | n/a | "defines the basic interface to a stream. Note, however, that there is no\n" |
---|
64 | n/a | "separation between reading and writing to streams; implementations are\n" |
---|
65 | n/a | "allowed to raise an IOError if they do not support a given operation.\n" |
---|
66 | n/a | "\n" |
---|
67 | n/a | "Extending IOBase is RawIOBase which deals simply with the reading and\n" |
---|
68 | n/a | "writing of raw bytes to a stream. FileIO subclasses RawIOBase to provide\n" |
---|
69 | n/a | "an interface to OS files.\n" |
---|
70 | n/a | "\n" |
---|
71 | n/a | "BufferedIOBase deals with buffering on a raw byte stream (RawIOBase). Its\n" |
---|
72 | n/a | "subclasses, BufferedWriter, BufferedReader, and BufferedRWPair buffer\n" |
---|
73 | n/a | "streams that are readable, writable, and both respectively.\n" |
---|
74 | n/a | "BufferedRandom provides a buffered interface to random access\n" |
---|
75 | n/a | "streams. BytesIO is a simple stream of in-memory bytes.\n" |
---|
76 | n/a | "\n" |
---|
77 | n/a | "Another IOBase subclass, TextIOBase, deals with the encoding and decoding\n" |
---|
78 | n/a | "of streams into text. TextIOWrapper, which extends it, is a buffered text\n" |
---|
79 | n/a | "interface to a buffered raw stream (`BufferedIOBase`). Finally, StringIO\n" |
---|
80 | n/a | "is an in-memory stream for text.\n" |
---|
81 | n/a | "\n" |
---|
82 | n/a | "Argument names are not part of the specification, and only the arguments\n" |
---|
83 | n/a | "of open() are intended to be used as keyword arguments.\n" |
---|
84 | n/a | "\n" |
---|
85 | n/a | "data:\n" |
---|
86 | n/a | "\n" |
---|
87 | n/a | "DEFAULT_BUFFER_SIZE\n" |
---|
88 | n/a | "\n" |
---|
89 | n/a | " An int containing the default buffer size used by the module's buffered\n" |
---|
90 | n/a | " I/O classes. open() uses the file's blksize (as obtained by os.stat) if\n" |
---|
91 | n/a | " possible.\n" |
---|
92 | n/a | ); |
---|
93 | n/a | |
---|
94 | n/a | |
---|
95 | n/a | /* |
---|
96 | n/a | * The main open() function |
---|
97 | n/a | */ |
---|
98 | n/a | /*[clinic input] |
---|
99 | n/a | module _io |
---|
100 | n/a | |
---|
101 | n/a | _io.open |
---|
102 | n/a | file: object |
---|
103 | n/a | mode: str = "r" |
---|
104 | n/a | buffering: int = -1 |
---|
105 | n/a | encoding: str(accept={str, NoneType}) = NULL |
---|
106 | n/a | errors: str(accept={str, NoneType}) = NULL |
---|
107 | n/a | newline: str(accept={str, NoneType}) = NULL |
---|
108 | n/a | closefd: int(c_default="1") = True |
---|
109 | n/a | opener: object = None |
---|
110 | n/a | |
---|
111 | n/a | Open file and return a stream. Raise IOError upon failure. |
---|
112 | n/a | |
---|
113 | n/a | file is either a text or byte string giving the name (and the path |
---|
114 | n/a | if the file isn't in the current working directory) of the file to |
---|
115 | n/a | be opened or an integer file descriptor of the file to be |
---|
116 | n/a | wrapped. (If a file descriptor is given, it is closed when the |
---|
117 | n/a | returned I/O object is closed, unless closefd is set to False.) |
---|
118 | n/a | |
---|
119 | n/a | mode is an optional string that specifies the mode in which the file |
---|
120 | n/a | is opened. It defaults to 'r' which means open for reading in text |
---|
121 | n/a | mode. Other common values are 'w' for writing (truncating the file if |
---|
122 | n/a | it already exists), 'x' for creating and writing to a new file, and |
---|
123 | n/a | 'a' for appending (which on some Unix systems, means that all writes |
---|
124 | n/a | append to the end of the file regardless of the current seek position). |
---|
125 | n/a | In text mode, if encoding is not specified the encoding used is platform |
---|
126 | n/a | dependent: locale.getpreferredencoding(False) is called to get the |
---|
127 | n/a | current locale encoding. (For reading and writing raw bytes use binary |
---|
128 | n/a | mode and leave encoding unspecified.) The available modes are: |
---|
129 | n/a | |
---|
130 | n/a | ========= =============================================================== |
---|
131 | n/a | Character Meaning |
---|
132 | n/a | --------- --------------------------------------------------------------- |
---|
133 | n/a | 'r' open for reading (default) |
---|
134 | n/a | 'w' open for writing, truncating the file first |
---|
135 | n/a | 'x' create a new file and open it for writing |
---|
136 | n/a | 'a' open for writing, appending to the end of the file if it exists |
---|
137 | n/a | 'b' binary mode |
---|
138 | n/a | 't' text mode (default) |
---|
139 | n/a | '+' open a disk file for updating (reading and writing) |
---|
140 | n/a | 'U' universal newline mode (deprecated) |
---|
141 | n/a | ========= =============================================================== |
---|
142 | n/a | |
---|
143 | n/a | The default mode is 'rt' (open for reading text). For binary random |
---|
144 | n/a | access, the mode 'w+b' opens and truncates the file to 0 bytes, while |
---|
145 | n/a | 'r+b' opens the file without truncation. The 'x' mode implies 'w' and |
---|
146 | n/a | raises an `FileExistsError` if the file already exists. |
---|
147 | n/a | |
---|
148 | n/a | Python distinguishes between files opened in binary and text modes, |
---|
149 | n/a | even when the underlying operating system doesn't. Files opened in |
---|
150 | n/a | binary mode (appending 'b' to the mode argument) return contents as |
---|
151 | n/a | bytes objects without any decoding. In text mode (the default, or when |
---|
152 | n/a | 't' is appended to the mode argument), the contents of the file are |
---|
153 | n/a | returned as strings, the bytes having been first decoded using a |
---|
154 | n/a | platform-dependent encoding or using the specified encoding if given. |
---|
155 | n/a | |
---|
156 | n/a | 'U' mode is deprecated and will raise an exception in future versions |
---|
157 | n/a | of Python. It has no effect in Python 3. Use newline to control |
---|
158 | n/a | universal newlines mode. |
---|
159 | n/a | |
---|
160 | n/a | buffering is an optional integer used to set the buffering policy. |
---|
161 | n/a | Pass 0 to switch buffering off (only allowed in binary mode), 1 to select |
---|
162 | n/a | line buffering (only usable in text mode), and an integer > 1 to indicate |
---|
163 | n/a | the size of a fixed-size chunk buffer. When no buffering argument is |
---|
164 | n/a | given, the default buffering policy works as follows: |
---|
165 | n/a | |
---|
166 | n/a | * Binary files are buffered in fixed-size chunks; the size of the buffer |
---|
167 | n/a | is chosen using a heuristic trying to determine the underlying device's |
---|
168 | n/a | "block size" and falling back on `io.DEFAULT_BUFFER_SIZE`. |
---|
169 | n/a | On many systems, the buffer will typically be 4096 or 8192 bytes long. |
---|
170 | n/a | |
---|
171 | n/a | * "Interactive" text files (files for which isatty() returns True) |
---|
172 | n/a | use line buffering. Other text files use the policy described above |
---|
173 | n/a | for binary files. |
---|
174 | n/a | |
---|
175 | n/a | encoding is the name of the encoding used to decode or encode the |
---|
176 | n/a | file. This should only be used in text mode. The default encoding is |
---|
177 | n/a | platform dependent, but any encoding supported by Python can be |
---|
178 | n/a | passed. See the codecs module for the list of supported encodings. |
---|
179 | n/a | |
---|
180 | n/a | errors is an optional string that specifies how encoding errors are to |
---|
181 | n/a | be handled---this argument should not be used in binary mode. Pass |
---|
182 | n/a | 'strict' to raise a ValueError exception if there is an encoding error |
---|
183 | n/a | (the default of None has the same effect), or pass 'ignore' to ignore |
---|
184 | n/a | errors. (Note that ignoring encoding errors can lead to data loss.) |
---|
185 | n/a | See the documentation for codecs.register or run 'help(codecs.Codec)' |
---|
186 | n/a | for a list of the permitted encoding error strings. |
---|
187 | n/a | |
---|
188 | n/a | newline controls how universal newlines works (it only applies to text |
---|
189 | n/a | mode). It can be None, '', '\n', '\r', and '\r\n'. It works as |
---|
190 | n/a | follows: |
---|
191 | n/a | |
---|
192 | n/a | * On input, if newline is None, universal newlines mode is |
---|
193 | n/a | enabled. Lines in the input can end in '\n', '\r', or '\r\n', and |
---|
194 | n/a | these are translated into '\n' before being returned to the |
---|
195 | n/a | caller. If it is '', universal newline mode is enabled, but line |
---|
196 | n/a | endings are returned to the caller untranslated. If it has any of |
---|
197 | n/a | the other legal values, input lines are only terminated by the given |
---|
198 | n/a | string, and the line ending is returned to the caller untranslated. |
---|
199 | n/a | |
---|
200 | n/a | * On output, if newline is None, any '\n' characters written are |
---|
201 | n/a | translated to the system default line separator, os.linesep. If |
---|
202 | n/a | newline is '' or '\n', no translation takes place. If newline is any |
---|
203 | n/a | of the other legal values, any '\n' characters written are translated |
---|
204 | n/a | to the given string. |
---|
205 | n/a | |
---|
206 | n/a | If closefd is False, the underlying file descriptor will be kept open |
---|
207 | n/a | when the file is closed. This does not work when a file name is given |
---|
208 | n/a | and must be True in that case. |
---|
209 | n/a | |
---|
210 | n/a | A custom opener can be used by passing a callable as *opener*. The |
---|
211 | n/a | underlying file descriptor for the file object is then obtained by |
---|
212 | n/a | calling *opener* with (*file*, *flags*). *opener* must return an open |
---|
213 | n/a | file descriptor (passing os.open as *opener* results in functionality |
---|
214 | n/a | similar to passing None). |
---|
215 | n/a | |
---|
216 | n/a | open() returns a file object whose type depends on the mode, and |
---|
217 | n/a | through which the standard file operations such as reading and writing |
---|
218 | n/a | are performed. When open() is used to open a file in a text mode ('w', |
---|
219 | n/a | 'r', 'wt', 'rt', etc.), it returns a TextIOWrapper. When used to open |
---|
220 | n/a | a file in a binary mode, the returned class varies: in read binary |
---|
221 | n/a | mode, it returns a BufferedReader; in write binary and append binary |
---|
222 | n/a | modes, it returns a BufferedWriter, and in read/write mode, it returns |
---|
223 | n/a | a BufferedRandom. |
---|
224 | n/a | |
---|
225 | n/a | It is also possible to use a string or bytearray as a file for both |
---|
226 | n/a | reading and writing. For strings StringIO can be used like a file |
---|
227 | n/a | opened in a text mode, and for bytes a BytesIO can be used like a file |
---|
228 | n/a | opened in a binary mode. |
---|
229 | n/a | [clinic start generated code]*/ |
---|
230 | n/a | |
---|
231 | n/a | static PyObject * |
---|
232 | n/a | _io_open_impl(PyObject *module, PyObject *file, const char *mode, |
---|
233 | n/a | int buffering, const char *encoding, const char *errors, |
---|
234 | n/a | const char *newline, int closefd, PyObject *opener) |
---|
235 | n/a | /*[clinic end generated code: output=aefafc4ce2b46dc0 input=f4e1ca75223987bc]*/ |
---|
236 | n/a | { |
---|
237 | n/a | unsigned i; |
---|
238 | n/a | |
---|
239 | n/a | int creating = 0, reading = 0, writing = 0, appending = 0, updating = 0; |
---|
240 | n/a | int text = 0, binary = 0, universal = 0; |
---|
241 | n/a | |
---|
242 | n/a | char rawmode[6], *m; |
---|
243 | n/a | int line_buffering, is_number; |
---|
244 | n/a | long isatty; |
---|
245 | n/a | |
---|
246 | n/a | PyObject *raw, *modeobj = NULL, *buffer, *wrapper, *result = NULL, *path_or_fd = NULL; |
---|
247 | n/a | |
---|
248 | n/a | _Py_IDENTIFIER(_blksize); |
---|
249 | n/a | _Py_IDENTIFIER(isatty); |
---|
250 | n/a | _Py_IDENTIFIER(mode); |
---|
251 | n/a | _Py_IDENTIFIER(close); |
---|
252 | n/a | |
---|
253 | n/a | is_number = PyNumber_Check(file); |
---|
254 | n/a | |
---|
255 | n/a | if (is_number) { |
---|
256 | n/a | path_or_fd = file; |
---|
257 | n/a | Py_INCREF(path_or_fd); |
---|
258 | n/a | } else { |
---|
259 | n/a | path_or_fd = PyOS_FSPath(file); |
---|
260 | n/a | if (path_or_fd == NULL) { |
---|
261 | n/a | return NULL; |
---|
262 | n/a | } |
---|
263 | n/a | } |
---|
264 | n/a | |
---|
265 | n/a | if (!is_number && |
---|
266 | n/a | !PyUnicode_Check(path_or_fd) && |
---|
267 | n/a | !PyBytes_Check(path_or_fd)) { |
---|
268 | n/a | PyErr_Format(PyExc_TypeError, "invalid file: %R", file); |
---|
269 | n/a | goto error; |
---|
270 | n/a | } |
---|
271 | n/a | |
---|
272 | n/a | /* Decode mode */ |
---|
273 | n/a | for (i = 0; i < strlen(mode); i++) { |
---|
274 | n/a | char c = mode[i]; |
---|
275 | n/a | |
---|
276 | n/a | switch (c) { |
---|
277 | n/a | case 'x': |
---|
278 | n/a | creating = 1; |
---|
279 | n/a | break; |
---|
280 | n/a | case 'r': |
---|
281 | n/a | reading = 1; |
---|
282 | n/a | break; |
---|
283 | n/a | case 'w': |
---|
284 | n/a | writing = 1; |
---|
285 | n/a | break; |
---|
286 | n/a | case 'a': |
---|
287 | n/a | appending = 1; |
---|
288 | n/a | break; |
---|
289 | n/a | case '+': |
---|
290 | n/a | updating = 1; |
---|
291 | n/a | break; |
---|
292 | n/a | case 't': |
---|
293 | n/a | text = 1; |
---|
294 | n/a | break; |
---|
295 | n/a | case 'b': |
---|
296 | n/a | binary = 1; |
---|
297 | n/a | break; |
---|
298 | n/a | case 'U': |
---|
299 | n/a | universal = 1; |
---|
300 | n/a | reading = 1; |
---|
301 | n/a | break; |
---|
302 | n/a | default: |
---|
303 | n/a | goto invalid_mode; |
---|
304 | n/a | } |
---|
305 | n/a | |
---|
306 | n/a | /* c must not be duplicated */ |
---|
307 | n/a | if (strchr(mode+i+1, c)) { |
---|
308 | n/a | invalid_mode: |
---|
309 | n/a | PyErr_Format(PyExc_ValueError, "invalid mode: '%s'", mode); |
---|
310 | n/a | goto error; |
---|
311 | n/a | } |
---|
312 | n/a | |
---|
313 | n/a | } |
---|
314 | n/a | |
---|
315 | n/a | m = rawmode; |
---|
316 | n/a | if (creating) *(m++) = 'x'; |
---|
317 | n/a | if (reading) *(m++) = 'r'; |
---|
318 | n/a | if (writing) *(m++) = 'w'; |
---|
319 | n/a | if (appending) *(m++) = 'a'; |
---|
320 | n/a | if (updating) *(m++) = '+'; |
---|
321 | n/a | *m = '\0'; |
---|
322 | n/a | |
---|
323 | n/a | /* Parameters validation */ |
---|
324 | n/a | if (universal) { |
---|
325 | n/a | if (creating || writing || appending || updating) { |
---|
326 | n/a | PyErr_SetString(PyExc_ValueError, |
---|
327 | n/a | "mode U cannot be combined with x', 'w', 'a', or '+'"); |
---|
328 | n/a | goto error; |
---|
329 | n/a | } |
---|
330 | n/a | if (PyErr_WarnEx(PyExc_DeprecationWarning, |
---|
331 | n/a | "'U' mode is deprecated", 1) < 0) |
---|
332 | n/a | goto error; |
---|
333 | n/a | reading = 1; |
---|
334 | n/a | } |
---|
335 | n/a | |
---|
336 | n/a | if (text && binary) { |
---|
337 | n/a | PyErr_SetString(PyExc_ValueError, |
---|
338 | n/a | "can't have text and binary mode at once"); |
---|
339 | n/a | goto error; |
---|
340 | n/a | } |
---|
341 | n/a | |
---|
342 | n/a | if (creating + reading + writing + appending > 1) { |
---|
343 | n/a | PyErr_SetString(PyExc_ValueError, |
---|
344 | n/a | "must have exactly one of create/read/write/append mode"); |
---|
345 | n/a | goto error; |
---|
346 | n/a | } |
---|
347 | n/a | |
---|
348 | n/a | if (binary && encoding != NULL) { |
---|
349 | n/a | PyErr_SetString(PyExc_ValueError, |
---|
350 | n/a | "binary mode doesn't take an encoding argument"); |
---|
351 | n/a | goto error; |
---|
352 | n/a | } |
---|
353 | n/a | |
---|
354 | n/a | if (binary && errors != NULL) { |
---|
355 | n/a | PyErr_SetString(PyExc_ValueError, |
---|
356 | n/a | "binary mode doesn't take an errors argument"); |
---|
357 | n/a | goto error; |
---|
358 | n/a | } |
---|
359 | n/a | |
---|
360 | n/a | if (binary && newline != NULL) { |
---|
361 | n/a | PyErr_SetString(PyExc_ValueError, |
---|
362 | n/a | "binary mode doesn't take a newline argument"); |
---|
363 | n/a | goto error; |
---|
364 | n/a | } |
---|
365 | n/a | |
---|
366 | n/a | /* Create the Raw file stream */ |
---|
367 | n/a | { |
---|
368 | n/a | PyObject *RawIO_class = (PyObject *)&PyFileIO_Type; |
---|
369 | n/a | #ifdef MS_WINDOWS |
---|
370 | n/a | if (!Py_LegacyWindowsStdioFlag && _PyIO_get_console_type(path_or_fd) != '\0') { |
---|
371 | n/a | RawIO_class = (PyObject *)&PyWindowsConsoleIO_Type; |
---|
372 | n/a | encoding = "utf-8"; |
---|
373 | n/a | } |
---|
374 | n/a | #endif |
---|
375 | n/a | raw = PyObject_CallFunction(RawIO_class, |
---|
376 | n/a | "OsiO", path_or_fd, rawmode, closefd, opener); |
---|
377 | n/a | } |
---|
378 | n/a | |
---|
379 | n/a | if (raw == NULL) |
---|
380 | n/a | goto error; |
---|
381 | n/a | result = raw; |
---|
382 | n/a | |
---|
383 | n/a | Py_DECREF(path_or_fd); |
---|
384 | n/a | path_or_fd = NULL; |
---|
385 | n/a | |
---|
386 | n/a | modeobj = PyUnicode_FromString(mode); |
---|
387 | n/a | if (modeobj == NULL) |
---|
388 | n/a | goto error; |
---|
389 | n/a | |
---|
390 | n/a | /* buffering */ |
---|
391 | n/a | { |
---|
392 | n/a | PyObject *res = _PyObject_CallMethodId(raw, &PyId_isatty, NULL); |
---|
393 | n/a | if (res == NULL) |
---|
394 | n/a | goto error; |
---|
395 | n/a | isatty = PyLong_AsLong(res); |
---|
396 | n/a | Py_DECREF(res); |
---|
397 | n/a | if (isatty == -1 && PyErr_Occurred()) |
---|
398 | n/a | goto error; |
---|
399 | n/a | } |
---|
400 | n/a | |
---|
401 | n/a | if (buffering == 1 || (buffering < 0 && isatty)) { |
---|
402 | n/a | buffering = -1; |
---|
403 | n/a | line_buffering = 1; |
---|
404 | n/a | } |
---|
405 | n/a | else |
---|
406 | n/a | line_buffering = 0; |
---|
407 | n/a | |
---|
408 | n/a | if (buffering < 0) { |
---|
409 | n/a | PyObject *blksize_obj; |
---|
410 | n/a | blksize_obj = _PyObject_GetAttrId(raw, &PyId__blksize); |
---|
411 | n/a | if (blksize_obj == NULL) |
---|
412 | n/a | goto error; |
---|
413 | n/a | buffering = PyLong_AsLong(blksize_obj); |
---|
414 | n/a | Py_DECREF(blksize_obj); |
---|
415 | n/a | if (buffering == -1 && PyErr_Occurred()) |
---|
416 | n/a | goto error; |
---|
417 | n/a | } |
---|
418 | n/a | if (buffering < 0) { |
---|
419 | n/a | PyErr_SetString(PyExc_ValueError, |
---|
420 | n/a | "invalid buffering size"); |
---|
421 | n/a | goto error; |
---|
422 | n/a | } |
---|
423 | n/a | |
---|
424 | n/a | /* if not buffering, returns the raw file object */ |
---|
425 | n/a | if (buffering == 0) { |
---|
426 | n/a | if (!binary) { |
---|
427 | n/a | PyErr_SetString(PyExc_ValueError, |
---|
428 | n/a | "can't have unbuffered text I/O"); |
---|
429 | n/a | goto error; |
---|
430 | n/a | } |
---|
431 | n/a | |
---|
432 | n/a | Py_DECREF(modeobj); |
---|
433 | n/a | return result; |
---|
434 | n/a | } |
---|
435 | n/a | |
---|
436 | n/a | /* wraps into a buffered file */ |
---|
437 | n/a | { |
---|
438 | n/a | PyObject *Buffered_class; |
---|
439 | n/a | |
---|
440 | n/a | if (updating) |
---|
441 | n/a | Buffered_class = (PyObject *)&PyBufferedRandom_Type; |
---|
442 | n/a | else if (creating || writing || appending) |
---|
443 | n/a | Buffered_class = (PyObject *)&PyBufferedWriter_Type; |
---|
444 | n/a | else if (reading) |
---|
445 | n/a | Buffered_class = (PyObject *)&PyBufferedReader_Type; |
---|
446 | n/a | else { |
---|
447 | n/a | PyErr_Format(PyExc_ValueError, |
---|
448 | n/a | "unknown mode: '%s'", mode); |
---|
449 | n/a | goto error; |
---|
450 | n/a | } |
---|
451 | n/a | |
---|
452 | n/a | buffer = PyObject_CallFunction(Buffered_class, "Oi", raw, buffering); |
---|
453 | n/a | } |
---|
454 | n/a | if (buffer == NULL) |
---|
455 | n/a | goto error; |
---|
456 | n/a | result = buffer; |
---|
457 | n/a | Py_DECREF(raw); |
---|
458 | n/a | |
---|
459 | n/a | |
---|
460 | n/a | /* if binary, returns the buffered file */ |
---|
461 | n/a | if (binary) { |
---|
462 | n/a | Py_DECREF(modeobj); |
---|
463 | n/a | return result; |
---|
464 | n/a | } |
---|
465 | n/a | |
---|
466 | n/a | /* wraps into a TextIOWrapper */ |
---|
467 | n/a | wrapper = PyObject_CallFunction((PyObject *)&PyTextIOWrapper_Type, |
---|
468 | n/a | "Osssi", |
---|
469 | n/a | buffer, |
---|
470 | n/a | encoding, errors, newline, |
---|
471 | n/a | line_buffering); |
---|
472 | n/a | if (wrapper == NULL) |
---|
473 | n/a | goto error; |
---|
474 | n/a | result = wrapper; |
---|
475 | n/a | Py_DECREF(buffer); |
---|
476 | n/a | |
---|
477 | n/a | if (_PyObject_SetAttrId(wrapper, &PyId_mode, modeobj) < 0) |
---|
478 | n/a | goto error; |
---|
479 | n/a | Py_DECREF(modeobj); |
---|
480 | n/a | return result; |
---|
481 | n/a | |
---|
482 | n/a | error: |
---|
483 | n/a | if (result != NULL) { |
---|
484 | n/a | PyObject *exc, *val, *tb, *close_result; |
---|
485 | n/a | PyErr_Fetch(&exc, &val, &tb); |
---|
486 | n/a | close_result = _PyObject_CallMethodId(result, &PyId_close, NULL); |
---|
487 | n/a | _PyErr_ChainExceptions(exc, val, tb); |
---|
488 | n/a | Py_XDECREF(close_result); |
---|
489 | n/a | Py_DECREF(result); |
---|
490 | n/a | } |
---|
491 | n/a | Py_XDECREF(path_or_fd); |
---|
492 | n/a | Py_XDECREF(modeobj); |
---|
493 | n/a | return NULL; |
---|
494 | n/a | } |
---|
495 | n/a | |
---|
496 | n/a | /* |
---|
497 | n/a | * Private helpers for the io module. |
---|
498 | n/a | */ |
---|
499 | n/a | |
---|
500 | n/a | Py_off_t |
---|
501 | n/a | PyNumber_AsOff_t(PyObject *item, PyObject *err) |
---|
502 | n/a | { |
---|
503 | n/a | Py_off_t result; |
---|
504 | n/a | PyObject *runerr; |
---|
505 | n/a | PyObject *value = PyNumber_Index(item); |
---|
506 | n/a | if (value == NULL) |
---|
507 | n/a | return -1; |
---|
508 | n/a | |
---|
509 | n/a | /* We're done if PyLong_AsSsize_t() returns without error. */ |
---|
510 | n/a | result = PyLong_AsOff_t(value); |
---|
511 | n/a | if (result != -1 || !(runerr = PyErr_Occurred())) |
---|
512 | n/a | goto finish; |
---|
513 | n/a | |
---|
514 | n/a | /* Error handling code -- only manage OverflowError differently */ |
---|
515 | n/a | if (!PyErr_GivenExceptionMatches(runerr, PyExc_OverflowError)) |
---|
516 | n/a | goto finish; |
---|
517 | n/a | |
---|
518 | n/a | PyErr_Clear(); |
---|
519 | n/a | /* If no error-handling desired then the default clipping |
---|
520 | n/a | is sufficient. |
---|
521 | n/a | */ |
---|
522 | n/a | if (!err) { |
---|
523 | n/a | assert(PyLong_Check(value)); |
---|
524 | n/a | /* Whether or not it is less than or equal to |
---|
525 | n/a | zero is determined by the sign of ob_size |
---|
526 | n/a | */ |
---|
527 | n/a | if (_PyLong_Sign(value) < 0) |
---|
528 | n/a | result = PY_OFF_T_MIN; |
---|
529 | n/a | else |
---|
530 | n/a | result = PY_OFF_T_MAX; |
---|
531 | n/a | } |
---|
532 | n/a | else { |
---|
533 | n/a | /* Otherwise replace the error with caller's error object. */ |
---|
534 | n/a | PyErr_Format(err, |
---|
535 | n/a | "cannot fit '%.200s' into an offset-sized integer", |
---|
536 | n/a | item->ob_type->tp_name); |
---|
537 | n/a | } |
---|
538 | n/a | |
---|
539 | n/a | finish: |
---|
540 | n/a | Py_DECREF(value); |
---|
541 | n/a | return result; |
---|
542 | n/a | } |
---|
543 | n/a | |
---|
544 | n/a | |
---|
545 | n/a | /* Basically the "n" format code with the ability to turn None into -1. */ |
---|
546 | n/a | int |
---|
547 | n/a | _PyIO_ConvertSsize_t(PyObject *obj, void *result) { |
---|
548 | n/a | Py_ssize_t limit; |
---|
549 | n/a | if (obj == Py_None) { |
---|
550 | n/a | limit = -1; |
---|
551 | n/a | } |
---|
552 | n/a | else if (PyNumber_Check(obj)) { |
---|
553 | n/a | limit = PyNumber_AsSsize_t(obj, PyExc_OverflowError); |
---|
554 | n/a | if (limit == -1 && PyErr_Occurred()) |
---|
555 | n/a | return 0; |
---|
556 | n/a | } |
---|
557 | n/a | else { |
---|
558 | n/a | PyErr_Format(PyExc_TypeError, |
---|
559 | n/a | "integer argument expected, got '%.200s'", |
---|
560 | n/a | Py_TYPE(obj)->tp_name); |
---|
561 | n/a | return 0; |
---|
562 | n/a | } |
---|
563 | n/a | *((Py_ssize_t *)result) = limit; |
---|
564 | n/a | return 1; |
---|
565 | n/a | } |
---|
566 | n/a | |
---|
567 | n/a | |
---|
568 | n/a | _PyIO_State * |
---|
569 | n/a | _PyIO_get_module_state(void) |
---|
570 | n/a | { |
---|
571 | n/a | PyObject *mod = PyState_FindModule(&_PyIO_Module); |
---|
572 | n/a | _PyIO_State *state; |
---|
573 | n/a | if (mod == NULL || (state = IO_MOD_STATE(mod)) == NULL) { |
---|
574 | n/a | PyErr_SetString(PyExc_RuntimeError, |
---|
575 | n/a | "could not find io module state " |
---|
576 | n/a | "(interpreter shutdown?)"); |
---|
577 | n/a | return NULL; |
---|
578 | n/a | } |
---|
579 | n/a | return state; |
---|
580 | n/a | } |
---|
581 | n/a | |
---|
582 | n/a | PyObject * |
---|
583 | n/a | _PyIO_get_locale_module(_PyIO_State *state) |
---|
584 | n/a | { |
---|
585 | n/a | PyObject *mod; |
---|
586 | n/a | if (state->locale_module != NULL) { |
---|
587 | n/a | assert(PyWeakref_CheckRef(state->locale_module)); |
---|
588 | n/a | mod = PyWeakref_GET_OBJECT(state->locale_module); |
---|
589 | n/a | if (mod != Py_None) { |
---|
590 | n/a | Py_INCREF(mod); |
---|
591 | n/a | return mod; |
---|
592 | n/a | } |
---|
593 | n/a | Py_CLEAR(state->locale_module); |
---|
594 | n/a | } |
---|
595 | n/a | mod = PyImport_ImportModule("_bootlocale"); |
---|
596 | n/a | if (mod == NULL) |
---|
597 | n/a | return NULL; |
---|
598 | n/a | state->locale_module = PyWeakref_NewRef(mod, NULL); |
---|
599 | n/a | if (state->locale_module == NULL) { |
---|
600 | n/a | Py_DECREF(mod); |
---|
601 | n/a | return NULL; |
---|
602 | n/a | } |
---|
603 | n/a | return mod; |
---|
604 | n/a | } |
---|
605 | n/a | |
---|
606 | n/a | |
---|
607 | n/a | static int |
---|
608 | n/a | iomodule_traverse(PyObject *mod, visitproc visit, void *arg) { |
---|
609 | n/a | _PyIO_State *state = IO_MOD_STATE(mod); |
---|
610 | n/a | if (!state->initialized) |
---|
611 | n/a | return 0; |
---|
612 | n/a | if (state->locale_module != NULL) { |
---|
613 | n/a | Py_VISIT(state->locale_module); |
---|
614 | n/a | } |
---|
615 | n/a | Py_VISIT(state->unsupported_operation); |
---|
616 | n/a | return 0; |
---|
617 | n/a | } |
---|
618 | n/a | |
---|
619 | n/a | |
---|
620 | n/a | static int |
---|
621 | n/a | iomodule_clear(PyObject *mod) { |
---|
622 | n/a | _PyIO_State *state = IO_MOD_STATE(mod); |
---|
623 | n/a | if (!state->initialized) |
---|
624 | n/a | return 0; |
---|
625 | n/a | if (state->locale_module != NULL) |
---|
626 | n/a | Py_CLEAR(state->locale_module); |
---|
627 | n/a | Py_CLEAR(state->unsupported_operation); |
---|
628 | n/a | return 0; |
---|
629 | n/a | } |
---|
630 | n/a | |
---|
631 | n/a | static void |
---|
632 | n/a | iomodule_free(PyObject *mod) { |
---|
633 | n/a | iomodule_clear(mod); |
---|
634 | n/a | } |
---|
635 | n/a | |
---|
636 | n/a | |
---|
637 | n/a | /* |
---|
638 | n/a | * Module definition |
---|
639 | n/a | */ |
---|
640 | n/a | |
---|
641 | n/a | #include "clinic/_iomodule.c.h" |
---|
642 | n/a | |
---|
643 | n/a | static PyMethodDef module_methods[] = { |
---|
644 | n/a | _IO_OPEN_METHODDEF |
---|
645 | n/a | {NULL, NULL} |
---|
646 | n/a | }; |
---|
647 | n/a | |
---|
648 | n/a | struct PyModuleDef _PyIO_Module = { |
---|
649 | n/a | PyModuleDef_HEAD_INIT, |
---|
650 | n/a | "io", |
---|
651 | n/a | module_doc, |
---|
652 | n/a | sizeof(_PyIO_State), |
---|
653 | n/a | module_methods, |
---|
654 | n/a | NULL, |
---|
655 | n/a | iomodule_traverse, |
---|
656 | n/a | iomodule_clear, |
---|
657 | n/a | (freefunc)iomodule_free, |
---|
658 | n/a | }; |
---|
659 | n/a | |
---|
660 | n/a | PyMODINIT_FUNC |
---|
661 | n/a | PyInit__io(void) |
---|
662 | n/a | { |
---|
663 | n/a | PyObject *m = PyModule_Create(&_PyIO_Module); |
---|
664 | n/a | _PyIO_State *state = NULL; |
---|
665 | n/a | if (m == NULL) |
---|
666 | n/a | return NULL; |
---|
667 | n/a | state = IO_MOD_STATE(m); |
---|
668 | n/a | state->initialized = 0; |
---|
669 | n/a | |
---|
670 | n/a | #define ADD_TYPE(type, name) \ |
---|
671 | n/a | if (PyType_Ready(type) < 0) \ |
---|
672 | n/a | goto fail; \ |
---|
673 | n/a | Py_INCREF(type); \ |
---|
674 | n/a | if (PyModule_AddObject(m, name, (PyObject *)type) < 0) { \ |
---|
675 | n/a | Py_DECREF(type); \ |
---|
676 | n/a | goto fail; \ |
---|
677 | n/a | } |
---|
678 | n/a | |
---|
679 | n/a | /* DEFAULT_BUFFER_SIZE */ |
---|
680 | n/a | if (PyModule_AddIntMacro(m, DEFAULT_BUFFER_SIZE) < 0) |
---|
681 | n/a | goto fail; |
---|
682 | n/a | |
---|
683 | n/a | /* UnsupportedOperation inherits from ValueError and IOError */ |
---|
684 | n/a | state->unsupported_operation = PyObject_CallFunction( |
---|
685 | n/a | (PyObject *)&PyType_Type, "s(OO){}", |
---|
686 | n/a | "UnsupportedOperation", PyExc_OSError, PyExc_ValueError); |
---|
687 | n/a | if (state->unsupported_operation == NULL) |
---|
688 | n/a | goto fail; |
---|
689 | n/a | Py_INCREF(state->unsupported_operation); |
---|
690 | n/a | if (PyModule_AddObject(m, "UnsupportedOperation", |
---|
691 | n/a | state->unsupported_operation) < 0) |
---|
692 | n/a | goto fail; |
---|
693 | n/a | |
---|
694 | n/a | /* BlockingIOError, for compatibility */ |
---|
695 | n/a | Py_INCREF(PyExc_BlockingIOError); |
---|
696 | n/a | if (PyModule_AddObject(m, "BlockingIOError", |
---|
697 | n/a | (PyObject *) PyExc_BlockingIOError) < 0) |
---|
698 | n/a | goto fail; |
---|
699 | n/a | |
---|
700 | n/a | /* Concrete base types of the IO ABCs. |
---|
701 | n/a | (the ABCs themselves are declared through inheritance in io.py) |
---|
702 | n/a | */ |
---|
703 | n/a | ADD_TYPE(&PyIOBase_Type, "_IOBase"); |
---|
704 | n/a | ADD_TYPE(&PyRawIOBase_Type, "_RawIOBase"); |
---|
705 | n/a | ADD_TYPE(&PyBufferedIOBase_Type, "_BufferedIOBase"); |
---|
706 | n/a | ADD_TYPE(&PyTextIOBase_Type, "_TextIOBase"); |
---|
707 | n/a | |
---|
708 | n/a | /* Implementation of concrete IO objects. */ |
---|
709 | n/a | /* FileIO */ |
---|
710 | n/a | PyFileIO_Type.tp_base = &PyRawIOBase_Type; |
---|
711 | n/a | ADD_TYPE(&PyFileIO_Type, "FileIO"); |
---|
712 | n/a | |
---|
713 | n/a | /* BytesIO */ |
---|
714 | n/a | PyBytesIO_Type.tp_base = &PyBufferedIOBase_Type; |
---|
715 | n/a | ADD_TYPE(&PyBytesIO_Type, "BytesIO"); |
---|
716 | n/a | if (PyType_Ready(&_PyBytesIOBuffer_Type) < 0) |
---|
717 | n/a | goto fail; |
---|
718 | n/a | |
---|
719 | n/a | /* StringIO */ |
---|
720 | n/a | PyStringIO_Type.tp_base = &PyTextIOBase_Type; |
---|
721 | n/a | ADD_TYPE(&PyStringIO_Type, "StringIO"); |
---|
722 | n/a | |
---|
723 | n/a | #ifdef MS_WINDOWS |
---|
724 | n/a | /* WindowsConsoleIO */ |
---|
725 | n/a | PyWindowsConsoleIO_Type.tp_base = &PyRawIOBase_Type; |
---|
726 | n/a | ADD_TYPE(&PyWindowsConsoleIO_Type, "_WindowsConsoleIO"); |
---|
727 | n/a | #endif |
---|
728 | n/a | |
---|
729 | n/a | /* BufferedReader */ |
---|
730 | n/a | PyBufferedReader_Type.tp_base = &PyBufferedIOBase_Type; |
---|
731 | n/a | ADD_TYPE(&PyBufferedReader_Type, "BufferedReader"); |
---|
732 | n/a | |
---|
733 | n/a | /* BufferedWriter */ |
---|
734 | n/a | PyBufferedWriter_Type.tp_base = &PyBufferedIOBase_Type; |
---|
735 | n/a | ADD_TYPE(&PyBufferedWriter_Type, "BufferedWriter"); |
---|
736 | n/a | |
---|
737 | n/a | /* BufferedRWPair */ |
---|
738 | n/a | PyBufferedRWPair_Type.tp_base = &PyBufferedIOBase_Type; |
---|
739 | n/a | ADD_TYPE(&PyBufferedRWPair_Type, "BufferedRWPair"); |
---|
740 | n/a | |
---|
741 | n/a | /* BufferedRandom */ |
---|
742 | n/a | PyBufferedRandom_Type.tp_base = &PyBufferedIOBase_Type; |
---|
743 | n/a | ADD_TYPE(&PyBufferedRandom_Type, "BufferedRandom"); |
---|
744 | n/a | |
---|
745 | n/a | /* TextIOWrapper */ |
---|
746 | n/a | PyTextIOWrapper_Type.tp_base = &PyTextIOBase_Type; |
---|
747 | n/a | ADD_TYPE(&PyTextIOWrapper_Type, "TextIOWrapper"); |
---|
748 | n/a | |
---|
749 | n/a | /* IncrementalNewlineDecoder */ |
---|
750 | n/a | ADD_TYPE(&PyIncrementalNewlineDecoder_Type, "IncrementalNewlineDecoder"); |
---|
751 | n/a | |
---|
752 | n/a | /* Interned strings */ |
---|
753 | n/a | #define ADD_INTERNED(name) \ |
---|
754 | n/a | if (!_PyIO_str_ ## name && \ |
---|
755 | n/a | !(_PyIO_str_ ## name = PyUnicode_InternFromString(# name))) \ |
---|
756 | n/a | goto fail; |
---|
757 | n/a | |
---|
758 | n/a | ADD_INTERNED(close) |
---|
759 | n/a | ADD_INTERNED(closed) |
---|
760 | n/a | ADD_INTERNED(decode) |
---|
761 | n/a | ADD_INTERNED(encode) |
---|
762 | n/a | ADD_INTERNED(fileno) |
---|
763 | n/a | ADD_INTERNED(flush) |
---|
764 | n/a | ADD_INTERNED(getstate) |
---|
765 | n/a | ADD_INTERNED(isatty) |
---|
766 | n/a | ADD_INTERNED(newlines) |
---|
767 | n/a | ADD_INTERNED(read) |
---|
768 | n/a | ADD_INTERNED(read1) |
---|
769 | n/a | ADD_INTERNED(readable) |
---|
770 | n/a | ADD_INTERNED(readall) |
---|
771 | n/a | ADD_INTERNED(readinto) |
---|
772 | n/a | ADD_INTERNED(readline) |
---|
773 | n/a | ADD_INTERNED(reset) |
---|
774 | n/a | ADD_INTERNED(seek) |
---|
775 | n/a | ADD_INTERNED(seekable) |
---|
776 | n/a | ADD_INTERNED(setstate) |
---|
777 | n/a | ADD_INTERNED(tell) |
---|
778 | n/a | ADD_INTERNED(truncate) |
---|
779 | n/a | ADD_INTERNED(write) |
---|
780 | n/a | ADD_INTERNED(writable) |
---|
781 | n/a | |
---|
782 | n/a | if (!_PyIO_str_nl && |
---|
783 | n/a | !(_PyIO_str_nl = PyUnicode_InternFromString("\n"))) |
---|
784 | n/a | goto fail; |
---|
785 | n/a | |
---|
786 | n/a | if (!_PyIO_empty_str && |
---|
787 | n/a | !(_PyIO_empty_str = PyUnicode_FromStringAndSize(NULL, 0))) |
---|
788 | n/a | goto fail; |
---|
789 | n/a | if (!_PyIO_empty_bytes && |
---|
790 | n/a | !(_PyIO_empty_bytes = PyBytes_FromStringAndSize(NULL, 0))) |
---|
791 | n/a | goto fail; |
---|
792 | n/a | if (!_PyIO_zero && |
---|
793 | n/a | !(_PyIO_zero = PyLong_FromLong(0L))) |
---|
794 | n/a | goto fail; |
---|
795 | n/a | |
---|
796 | n/a | state->initialized = 1; |
---|
797 | n/a | |
---|
798 | n/a | return m; |
---|
799 | n/a | |
---|
800 | n/a | fail: |
---|
801 | n/a | Py_XDECREF(state->unsupported_operation); |
---|
802 | n/a | Py_DECREF(m); |
---|
803 | n/a | return NULL; |
---|
804 | n/a | } |
---|