2 An implementation of the new I/O lib as defined by PEP 3116 - "New I/O"
4 Classes defined here: UnsupportedOperation, BlockingIOError.
5 Functions defined here: open().
7 Mostly written by Amaury Forgeot d'Arc
10 #define PY_SSIZE_T_CLEAN
12 #include "structmember.h"
13 #include "_iomodule.h"
15 #ifdef HAVE_SYS_TYPES_H
16 #include <sys/types.h>
17 #endif /* HAVE_SYS_TYPES_H */
19 #ifdef HAVE_SYS_STAT_H
21 #endif /* HAVE_SYS_STAT_H */
24 /* Various interned strings */
26 PyObject
*_PyIO_str_close
;
27 PyObject
*_PyIO_str_closed
;
28 PyObject
*_PyIO_str_decode
;
29 PyObject
*_PyIO_str_encode
;
30 PyObject
*_PyIO_str_fileno
;
31 PyObject
*_PyIO_str_flush
;
32 PyObject
*_PyIO_str_getstate
;
33 PyObject
*_PyIO_str_isatty
;
34 PyObject
*_PyIO_str_newlines
;
35 PyObject
*_PyIO_str_nl
;
36 PyObject
*_PyIO_str_read
;
37 PyObject
*_PyIO_str_read1
;
38 PyObject
*_PyIO_str_readable
;
39 PyObject
*_PyIO_str_readinto
;
40 PyObject
*_PyIO_str_readline
;
41 PyObject
*_PyIO_str_reset
;
42 PyObject
*_PyIO_str_seek
;
43 PyObject
*_PyIO_str_seekable
;
44 PyObject
*_PyIO_str_setstate
;
45 PyObject
*_PyIO_str_tell
;
46 PyObject
*_PyIO_str_truncate
;
47 PyObject
*_PyIO_str_writable
;
48 PyObject
*_PyIO_str_write
;
50 PyObject
*_PyIO_empty_str
;
51 PyObject
*_PyIO_empty_bytes
;
55 PyDoc_STRVAR(module_doc
,
56 "The io module provides the Python interfaces to stream handling. The\n"
57 "builtin open function is defined in this module.\n"
59 "At the top of the I/O hierarchy is the abstract base class IOBase. It\n"
60 "defines the basic interface to a stream. Note, however, that there is no\n"
61 "separation between reading and writing to streams; implementations are\n"
62 "allowed to raise an IOError if they do not support a given operation.\n"
64 "Extending IOBase is RawIOBase which deals simply with the reading and\n"
65 "writing of raw bytes to a stream. FileIO subclasses RawIOBase to provide\n"
66 "an interface to OS files.\n"
68 "BufferedIOBase deals with buffering on a raw byte stream (RawIOBase). Its\n"
69 "subclasses, BufferedWriter, BufferedReader, and BufferedRWPair buffer\n"
70 "streams that are readable, writable, and both respectively.\n"
71 "BufferedRandom provides a buffered interface to random access\n"
72 "streams. BytesIO is a simple stream of in-memory bytes.\n"
74 "Another IOBase subclass, TextIOBase, deals with the encoding and decoding\n"
75 "of streams into text. TextIOWrapper, which extends it, is a buffered text\n"
76 "interface to a buffered raw stream (`BufferedIOBase`). Finally, StringIO\n"
77 "is a in-memory stream for text.\n"
79 "Argument names are not part of the specification, and only the arguments\n"
80 "of open() are intended to be used as keyword arguments.\n"
84 "DEFAULT_BUFFER_SIZE\n"
86 " An int containing the default buffer size used by the module's buffered\n"
87 " I/O classes. open() uses the file's blksize (as obtained by os.stat) if\n"
93 * BlockingIOError extends IOError
97 blockingioerror_init(PyBlockingIOErrorObject
*self
, PyObject
*args
,
100 PyObject
*myerrno
= NULL
, *strerror
= NULL
;
101 PyObject
*baseargs
= NULL
;
102 Py_ssize_t written
= 0;
104 assert(PyTuple_Check(args
));
107 if (!PyArg_ParseTuple(args
, "OO|n:BlockingIOError",
108 &myerrno
, &strerror
, &written
))
111 baseargs
= PyTuple_Pack(2, myerrno
, strerror
);
112 if (baseargs
== NULL
)
114 /* This will take care of initializing of myerrno and strerror members */
115 if (((PyTypeObject
*)PyExc_IOError
)->tp_init(
116 (PyObject
*)self
, baseargs
, kwds
) == -1) {
122 self
->written
= written
;
126 static PyMemberDef blockingioerror_members
[] = {
127 {"characters_written", T_PYSSIZET
, offsetof(PyBlockingIOErrorObject
, written
), 0},
128 {NULL
} /* Sentinel */
131 static PyTypeObject _PyExc_BlockingIOError
= {
132 PyVarObject_HEAD_INIT(NULL
, 0)
133 "BlockingIOError", /*tp_name*/
134 sizeof(PyBlockingIOErrorObject
), /*tp_basicsize*/
143 0, /*tp_as_sequence*/
151 Py_TPFLAGS_DEFAULT
| Py_TPFLAGS_BASETYPE
, /*tp_flags*/
152 PyDoc_STR("Exception raised when I/O would block "
153 "on a non-blocking I/O stream"), /* tp_doc */
156 0, /* tp_richcompare */
157 0, /* tp_weaklistoffset */
161 blockingioerror_members
, /* tp_members */
165 0, /* tp_descr_get */
166 0, /* tp_descr_set */
167 0, /* tp_dictoffset */
168 (initproc
)blockingioerror_init
, /* tp_init */
172 PyObject
*PyExc_BlockingIOError
= (PyObject
*)&_PyExc_BlockingIOError
;
176 * The main open() function
178 PyDoc_STRVAR(open_doc
,
179 "Open file and return a stream. Raise IOError upon failure.\n"
181 "file is either a text or byte string giving the name (and the path\n"
182 "if the file isn't in the current working directory) of the file to\n"
183 "be opened or an integer file descriptor of the file to be\n"
184 "wrapped. (If a file descriptor is given, it is closed when the\n"
185 "returned I/O object is closed, unless closefd is set to False.)\n"
187 "mode is an optional string that specifies the mode in which the file\n"
188 "is opened. It defaults to 'r' which means open for reading in text\n"
189 "mode. Other common values are 'w' for writing (truncating the file if\n"
190 "it already exists), and 'a' for appending (which on some Unix systems,\n"
191 "means that all writes append to the end of the file regardless of the\n"
192 "current seek position). In text mode, if encoding is not specified the\n"
193 "encoding used is platform dependent. (For reading and writing raw\n"
194 "bytes use binary mode and leave encoding unspecified.) The available\n"
197 "========= ===============================================================\n"
198 "Character Meaning\n"
199 "--------- ---------------------------------------------------------------\n"
200 "'r' open for reading (default)\n"
201 "'w' open for writing, truncating the file first\n"
202 "'a' open for writing, appending to the end of the file if it exists\n"
204 "'t' text mode (default)\n"
205 "'+' open a disk file for updating (reading and writing)\n"
206 "'U' universal newline mode (for backwards compatibility; unneeded\n"
208 "========= ===============================================================\n"
210 "The default mode is 'rt' (open for reading text). For binary random\n"
211 "access, the mode 'w+b' opens and truncates the file to 0 bytes, while\n"
212 "'r+b' opens the file without truncation.\n"
214 "Python distinguishes between files opened in binary and text modes,\n"
215 "even when the underlying operating system doesn't. Files opened in\n"
216 "binary mode (appending 'b' to the mode argument) return contents as\n"
217 "bytes objects without any decoding. In text mode (the default, or when\n"
218 "'t' is appended to the mode argument), the contents of the file are\n"
219 "returned as strings, the bytes having been first decoded using a\n"
220 "platform-dependent encoding or using the specified encoding if given.\n"
222 "buffering is an optional integer used to set the buffering policy.\n"
223 "Pass 0 to switch buffering off (only allowed in binary mode), 1 to select\n"
224 "line buffering (only usable in text mode), and an integer > 1 to indicate\n"
225 "the size of a fixed-size chunk buffer. When no buffering argument is\n"
226 "given, the default buffering policy works as follows:\n"
228 "* Binary files are buffered in fixed-size chunks; the size of the buffer\n"
229 " is chosen using a heuristic trying to determine the underlying device's\n"
230 " \"block size\" and falling back on `io.DEFAULT_BUFFER_SIZE`.\n"
231 " On many systems, the buffer will typically be 4096 or 8192 bytes long.\n"
233 "* \"Interactive\" text files (files for which isatty() returns True)\n"
234 " use line buffering. Other text files use the policy described above\n"
235 " for binary files.\n"
237 "encoding is the name of the encoding used to decode or encode the\n"
238 "file. This should only be used in text mode. The default encoding is\n"
239 "platform dependent, but any encoding supported by Python can be\n"
240 "passed. See the codecs module for the list of supported encodings.\n"
242 "errors is an optional string that specifies how encoding errors are to\n"
243 "be handled---this argument should not be used in binary mode. Pass\n"
244 "'strict' to raise a ValueError exception if there is an encoding error\n"
245 "(the default of None has the same effect), or pass 'ignore' to ignore\n"
246 "errors. (Note that ignoring encoding errors can lead to data loss.)\n"
247 "See the documentation for codecs.register for a list of the permitted\n"
248 "encoding error strings.\n"
250 "newline controls how universal newlines works (it only applies to text\n"
251 "mode). It can be None, '', '\\n', '\\r', and '\\r\\n'. It works as\n"
254 "* On input, if newline is None, universal newlines mode is\n"
255 " enabled. Lines in the input can end in '\\n', '\\r', or '\\r\\n', and\n"
256 " these are translated into '\\n' before being returned to the\n"
257 " caller. If it is '', universal newline mode is enabled, but line\n"
258 " endings are returned to the caller untranslated. If it has any of\n"
259 " the other legal values, input lines are only terminated by the given\n"
260 " string, and the line ending is returned to the caller untranslated.\n"
262 "* On output, if newline is None, any '\\n' characters written are\n"
263 " translated to the system default line separator, os.linesep. If\n"
264 " newline is '', no translation takes place. If newline is any of the\n"
265 " other legal values, any '\\n' characters written are translated to\n"
266 " the given string.\n"
268 "If closefd is False, the underlying file descriptor will be kept open\n"
269 "when the file is closed. This does not work when a file name is given\n"
270 "and must be True in that case.\n"
272 "open() returns a file object whose type depends on the mode, and\n"
273 "through which the standard file operations such as reading and writing\n"
274 "are performed. When open() is used to open a file in a text mode ('w',\n"
275 "'r', 'wt', 'rt', etc.), it returns a TextIOWrapper. When used to open\n"
276 "a file in a binary mode, the returned class varies: in read binary\n"
277 "mode, it returns a BufferedReader; in write binary and append binary\n"
278 "modes, it returns a BufferedWriter, and in read/write mode, it returns\n"
279 "a BufferedRandom.\n"
281 "It is also possible to use a string or bytearray as a file for both\n"
282 "reading and writing. For strings StringIO can be used like a file\n"
283 "opened in a text mode, and for bytes a BytesIO can be used like a file\n"
284 "opened in a binary mode.\n"
288 io_open(PyObject
*self
, PyObject
*args
, PyObject
*kwds
)
290 char *kwlist
[] = {"file", "mode", "buffering",
291 "encoding", "errors", "newline",
295 int buffering
= -1, closefd
= 1;
296 char *encoding
= NULL
, *errors
= NULL
, *newline
= NULL
;
299 int reading
= 0, writing
= 0, appending
= 0, updating
= 0;
300 int text
= 0, binary
= 0, universal
= 0;
306 PyObject
*raw
, *modeobj
= NULL
, *buffer
, *wrapper
, *result
= NULL
;
308 if (!PyArg_ParseTupleAndKeywords(args
, kwds
, "O|sizzzi:open", kwlist
,
309 &file
, &mode
, &buffering
,
310 &encoding
, &errors
, &newline
,
315 if (!PyUnicode_Check(file
) &&
316 !PyBytes_Check(file
) &&
317 !PyNumber_Check(file
)) {
318 PyObject
*repr
= PyObject_Repr(file
);
320 PyErr_Format(PyExc_TypeError
, "invalid file: %s",
321 PyString_AS_STRING(repr
));
328 for (i
= 0; i
< strlen(mode
); i
++) {
358 /* c must not be duplicated */
359 if (strchr(mode
+i
+1, c
)) {
361 PyErr_Format(PyExc_ValueError
, "invalid mode: '%s'", mode
);
368 if (reading
) *(m
++) = 'r';
369 if (writing
) *(m
++) = 'w';
370 if (appending
) *(m
++) = 'a';
371 if (updating
) *(m
++) = '+';
374 /* Parameters validation */
376 if (writing
|| appending
) {
377 PyErr_SetString(PyExc_ValueError
,
378 "can't use U and writing mode at once");
384 if (text
&& binary
) {
385 PyErr_SetString(PyExc_ValueError
,
386 "can't have text and binary mode at once");
390 if (reading
+ writing
+ appending
> 1) {
391 PyErr_SetString(PyExc_ValueError
,
392 "must have exactly one of read/write/append mode");
396 if (binary
&& encoding
!= NULL
) {
397 PyErr_SetString(PyExc_ValueError
,
398 "binary mode doesn't take an encoding argument");
402 if (binary
&& errors
!= NULL
) {
403 PyErr_SetString(PyExc_ValueError
,
404 "binary mode doesn't take an errors argument");
408 if (binary
&& newline
!= NULL
) {
409 PyErr_SetString(PyExc_ValueError
,
410 "binary mode doesn't take a newline argument");
414 /* Create the Raw file stream */
415 raw
= PyObject_CallFunction((PyObject
*)&PyFileIO_Type
,
416 "Osi", file
, rawmode
, closefd
);
421 modeobj
= PyUnicode_FromString(mode
);
427 PyObject
*res
= PyObject_CallMethod(raw
, "isatty", NULL
);
430 isatty
= PyLong_AsLong(res
);
432 if (isatty
== -1 && PyErr_Occurred())
436 if (buffering
== 1 || (buffering
< 0 && isatty
)) {
444 buffering
= DEFAULT_BUFFER_SIZE
;
445 #ifdef HAVE_STRUCT_STAT_ST_BLKSIZE
449 PyObject
*res
= PyObject_CallMethod(raw
, "fileno", NULL
);
453 fileno
= _PyInt_AsInt(res
);
455 if (fileno
== -1 && PyErr_Occurred())
458 if (fstat(fileno
, &st
) >= 0 && st
.st_blksize
> 1)
459 buffering
= st
.st_blksize
;
464 PyErr_SetString(PyExc_ValueError
,
465 "invalid buffering size");
469 /* if not buffering, returns the raw file object */
470 if (buffering
== 0) {
472 PyErr_SetString(PyExc_ValueError
,
473 "can't have unbuffered text I/O");
481 /* wraps into a buffered file */
483 PyObject
*Buffered_class
;
486 Buffered_class
= (PyObject
*)&PyBufferedRandom_Type
;
487 else if (writing
|| appending
)
488 Buffered_class
= (PyObject
*)&PyBufferedWriter_Type
;
490 Buffered_class
= (PyObject
*)&PyBufferedReader_Type
;
492 PyErr_Format(PyExc_ValueError
,
493 "unknown mode: '%s'", mode
);
497 buffer
= PyObject_CallFunction(Buffered_class
, "Oi", raw
, buffering
);
505 /* if binary, returns the buffered file */
511 /* wraps into a TextIOWrapper */
512 wrapper
= PyObject_CallFunction((PyObject
*)&PyTextIOWrapper_Type
,
515 encoding
, errors
, newline
,
522 if (PyObject_SetAttrString(wrapper
, "mode", modeobj
) < 0)
528 if (result
!= NULL
) {
529 PyObject
*exc
, *val
, *tb
, *close_result
;
530 PyErr_Fetch(&exc
, &val
, &tb
);
531 close_result
= PyObject_CallMethod(result
, "close", NULL
);
532 _PyErr_ReplaceException(exc
, val
, tb
);
533 Py_XDECREF(close_result
);
541 * Private helpers for the io module.
545 PyNumber_AsOff_t(PyObject
*item
, PyObject
*err
)
549 PyObject
*value
= PyNumber_Index(item
);
553 if (PyInt_Check(value
)) {
554 /* We assume a long always fits in a Py_off_t... */
555 result
= (Py_off_t
) PyInt_AS_LONG(value
);
559 /* We're done if PyLong_AsSsize_t() returns without error. */
560 result
= PyLong_AsOff_t(value
);
561 if (result
!= -1 || !(runerr
= PyErr_Occurred()))
564 /* Error handling code -- only manage OverflowError differently */
565 if (!PyErr_GivenExceptionMatches(runerr
, PyExc_OverflowError
))
569 /* If no error-handling desired then the default clipping
573 assert(PyLong_Check(value
));
574 /* Whether or not it is less than or equal to
575 zero is determined by the sign of ob_size
577 if (_PyLong_Sign(value
) < 0)
578 result
= PY_OFF_T_MIN
;
580 result
= PY_OFF_T_MAX
;
583 /* Otherwise replace the error with caller's error object. */
585 "cannot fit '%.200s' into an offset-sized integer",
586 item
->ob_type
->tp_name
);
595 /* Basically the "n" format code with the ability to turn None into -1. */
597 _PyIO_ConvertSsize_t(PyObject
*obj
, void *result
) {
599 if (obj
== Py_None
) {
602 else if (PyNumber_Check(obj
)) {
603 limit
= PyNumber_AsSsize_t(obj
, PyExc_OverflowError
);
604 if (limit
== -1 && PyErr_Occurred())
608 PyErr_Format(PyExc_TypeError
,
609 "integer argument expected, got '%.200s'",
610 Py_TYPE(obj
)->tp_name
);
613 *((Py_ssize_t
*)result
) = limit
;
622 PyObject
*_PyIO_os_module
= NULL
;
623 PyObject
*_PyIO_locale_module
= NULL
;
624 PyObject
*_PyIO_unsupported_operation
= NULL
;
626 static PyMethodDef module_methods
[] = {
627 {"open", (PyCFunction
)io_open
, METH_VARARGS
|METH_KEYWORDS
, open_doc
},
634 PyObject
*m
= Py_InitModule4("_io", module_methods
,
635 module_doc
, NULL
, PYTHON_API_VERSION
);
639 /* put os in the module state */
640 _PyIO_os_module
= PyImport_ImportModule("os");
641 if (_PyIO_os_module
== NULL
)
644 #define ADD_TYPE(type, name) \
645 if (PyType_Ready(type) < 0) \
648 if (PyModule_AddObject(m, name, (PyObject *)type) < 0) { \
653 /* DEFAULT_BUFFER_SIZE */
654 if (PyModule_AddIntMacro(m
, DEFAULT_BUFFER_SIZE
) < 0)
657 /* UnsupportedOperation inherits from ValueError and IOError */
658 _PyIO_unsupported_operation
= PyObject_CallFunction(
659 (PyObject
*)&PyType_Type
, "s(OO){}",
660 "UnsupportedOperation", PyExc_ValueError
, PyExc_IOError
);
661 if (_PyIO_unsupported_operation
== NULL
)
663 Py_INCREF(_PyIO_unsupported_operation
);
664 if (PyModule_AddObject(m
, "UnsupportedOperation",
665 _PyIO_unsupported_operation
) < 0)
668 /* BlockingIOError */
669 _PyExc_BlockingIOError
.tp_base
= (PyTypeObject
*) PyExc_IOError
;
670 ADD_TYPE(&_PyExc_BlockingIOError
, "BlockingIOError");
672 /* Concrete base types of the IO ABCs.
673 (the ABCs themselves are declared through inheritance in io.py)
675 ADD_TYPE(&PyIOBase_Type
, "_IOBase");
676 ADD_TYPE(&PyRawIOBase_Type
, "_RawIOBase");
677 ADD_TYPE(&PyBufferedIOBase_Type
, "_BufferedIOBase");
678 ADD_TYPE(&PyTextIOBase_Type
, "_TextIOBase");
680 /* Implementation of concrete IO objects. */
682 PyFileIO_Type
.tp_base
= &PyRawIOBase_Type
;
683 ADD_TYPE(&PyFileIO_Type
, "FileIO");
686 PyBytesIO_Type
.tp_base
= &PyBufferedIOBase_Type
;
687 ADD_TYPE(&PyBytesIO_Type
, "BytesIO");
690 PyStringIO_Type
.tp_base
= &PyTextIOBase_Type
;
691 ADD_TYPE(&PyStringIO_Type
, "StringIO");
694 PyBufferedReader_Type
.tp_base
= &PyBufferedIOBase_Type
;
695 ADD_TYPE(&PyBufferedReader_Type
, "BufferedReader");
698 PyBufferedWriter_Type
.tp_base
= &PyBufferedIOBase_Type
;
699 ADD_TYPE(&PyBufferedWriter_Type
, "BufferedWriter");
702 PyBufferedRWPair_Type
.tp_base
= &PyBufferedIOBase_Type
;
703 ADD_TYPE(&PyBufferedRWPair_Type
, "BufferedRWPair");
706 PyBufferedRandom_Type
.tp_base
= &PyBufferedIOBase_Type
;
707 ADD_TYPE(&PyBufferedRandom_Type
, "BufferedRandom");
710 PyTextIOWrapper_Type
.tp_base
= &PyTextIOBase_Type
;
711 ADD_TYPE(&PyTextIOWrapper_Type
, "TextIOWrapper");
713 /* IncrementalNewlineDecoder */
714 ADD_TYPE(&PyIncrementalNewlineDecoder_Type
, "IncrementalNewlineDecoder");
716 /* Interned strings */
717 if (!(_PyIO_str_close
= PyString_InternFromString("close")))
719 if (!(_PyIO_str_closed
= PyString_InternFromString("closed")))
721 if (!(_PyIO_str_decode
= PyString_InternFromString("decode")))
723 if (!(_PyIO_str_encode
= PyString_InternFromString("encode")))
725 if (!(_PyIO_str_fileno
= PyString_InternFromString("fileno")))
727 if (!(_PyIO_str_flush
= PyString_InternFromString("flush")))
729 if (!(_PyIO_str_getstate
= PyString_InternFromString("getstate")))
731 if (!(_PyIO_str_isatty
= PyString_InternFromString("isatty")))
733 if (!(_PyIO_str_newlines
= PyString_InternFromString("newlines")))
735 if (!(_PyIO_str_nl
= PyString_InternFromString("\n")))
737 if (!(_PyIO_str_read
= PyString_InternFromString("read")))
739 if (!(_PyIO_str_read1
= PyString_InternFromString("read1")))
741 if (!(_PyIO_str_readable
= PyString_InternFromString("readable")))
743 if (!(_PyIO_str_readinto
= PyString_InternFromString("readinto")))
745 if (!(_PyIO_str_readline
= PyString_InternFromString("readline")))
747 if (!(_PyIO_str_reset
= PyString_InternFromString("reset")))
749 if (!(_PyIO_str_seek
= PyString_InternFromString("seek")))
751 if (!(_PyIO_str_seekable
= PyString_InternFromString("seekable")))
753 if (!(_PyIO_str_setstate
= PyString_InternFromString("setstate")))
755 if (!(_PyIO_str_tell
= PyString_InternFromString("tell")))
757 if (!(_PyIO_str_truncate
= PyString_InternFromString("truncate")))
759 if (!(_PyIO_str_write
= PyString_InternFromString("write")))
761 if (!(_PyIO_str_writable
= PyString_InternFromString("writable")))
764 if (!(_PyIO_empty_str
= PyUnicode_FromStringAndSize(NULL
, 0)))
766 if (!(_PyIO_empty_bytes
= PyBytes_FromStringAndSize(NULL
, 0)))
768 if (!(_PyIO_zero
= PyLong_FromLong(0L)))
774 Py_CLEAR(_PyIO_os_module
);
775 Py_CLEAR(_PyIO_unsupported_operation
);