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 "seperation between reading and writing to streams; implementations are\n"
62 "allowed to throw 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;
303 int line_buffering
, isatty
;
305 PyObject
*raw
, *modeobj
= NULL
, *buffer
= NULL
, *wrapper
= NULL
;
307 if (!PyArg_ParseTupleAndKeywords(args
, kwds
, "O|sizzzi:open", kwlist
,
308 &file
, &mode
, &buffering
,
309 &encoding
, &errors
, &newline
,
314 if (!PyUnicode_Check(file
) &&
315 !PyBytes_Check(file
) &&
316 !PyNumber_Check(file
)) {
317 PyObject
*repr
= PyObject_Repr(file
);
319 PyErr_Format(PyExc_TypeError
, "invalid file: %s",
320 PyString_AS_STRING(repr
));
327 for (i
= 0; i
< strlen(mode
); i
++) {
357 /* c must not be duplicated */
358 if (strchr(mode
+i
+1, c
)) {
360 PyErr_Format(PyExc_ValueError
, "invalid mode: '%s'", mode
);
367 if (reading
) *(m
++) = 'r';
368 if (writing
) *(m
++) = 'w';
369 if (appending
) *(m
++) = 'a';
370 if (updating
) *(m
++) = '+';
373 /* Parameters validation */
375 if (writing
|| appending
) {
376 PyErr_SetString(PyExc_ValueError
,
377 "can't use U and writing mode at once");
383 if (text
&& binary
) {
384 PyErr_SetString(PyExc_ValueError
,
385 "can't have text and binary mode at once");
389 if (reading
+ writing
+ appending
> 1) {
390 PyErr_SetString(PyExc_ValueError
,
391 "must have exactly one of read/write/append mode");
395 if (binary
&& encoding
!= NULL
) {
396 PyErr_SetString(PyExc_ValueError
,
397 "binary mode doesn't take an encoding argument");
401 if (binary
&& errors
!= NULL
) {
402 PyErr_SetString(PyExc_ValueError
,
403 "binary mode doesn't take an errors argument");
407 if (binary
&& newline
!= NULL
) {
408 PyErr_SetString(PyExc_ValueError
,
409 "binary mode doesn't take a newline argument");
413 /* Create the Raw file stream */
414 raw
= PyObject_CallFunction((PyObject
*)&PyFileIO_Type
,
415 "Osi", file
, rawmode
, closefd
);
419 modeobj
= PyUnicode_FromString(mode
);
425 PyObject
*res
= PyObject_CallMethod(raw
, "isatty", NULL
);
428 isatty
= PyLong_AsLong(res
);
430 if (isatty
== -1 && PyErr_Occurred())
434 if (buffering
== 1 || (buffering
< 0 && isatty
)) {
442 buffering
= DEFAULT_BUFFER_SIZE
;
443 #ifdef HAVE_STRUCT_STAT_ST_BLKSIZE
447 PyObject
*res
= PyObject_CallMethod(raw
, "fileno", NULL
);
451 fileno
= PyInt_AsLong(res
);
453 if (fileno
== -1 && PyErr_Occurred())
456 if (fstat(fileno
, &st
) >= 0)
457 buffering
= st
.st_blksize
;
462 PyErr_SetString(PyExc_ValueError
,
463 "invalid buffering size");
467 /* if not buffering, returns the raw file object */
468 if (buffering
== 0) {
470 PyErr_SetString(PyExc_ValueError
,
471 "can't have unbuffered text I/O");
479 /* wraps into a buffered file */
481 PyObject
*Buffered_class
;
484 Buffered_class
= (PyObject
*)&PyBufferedRandom_Type
;
485 else if (writing
|| appending
)
486 Buffered_class
= (PyObject
*)&PyBufferedWriter_Type
;
488 Buffered_class
= (PyObject
*)&PyBufferedReader_Type
;
490 PyErr_Format(PyExc_ValueError
,
491 "unknown mode: '%s'", mode
);
495 buffer
= PyObject_CallFunction(Buffered_class
, "Oi", raw
, buffering
);
502 /* if binary, returns the buffered file */
508 /* wraps into a TextIOWrapper */
509 wrapper
= PyObject_CallFunction((PyObject
*)&PyTextIOWrapper_Type
,
512 encoding
, errors
, newline
,
518 if (PyObject_SetAttrString(wrapper
, "mode", modeobj
) < 0)
532 * Private helpers for the io module.
536 PyNumber_AsOff_t(PyObject
*item
, PyObject
*err
)
540 PyObject
*value
= PyNumber_Index(item
);
544 if (PyInt_Check(value
)) {
545 /* We assume a long always fits in a Py_off_t... */
546 result
= (Py_off_t
) PyInt_AS_LONG(value
);
550 /* We're done if PyLong_AsSsize_t() returns without error. */
551 result
= PyLong_AsOff_t(value
);
552 if (result
!= -1 || !(runerr
= PyErr_Occurred()))
555 /* Error handling code -- only manage OverflowError differently */
556 if (!PyErr_GivenExceptionMatches(runerr
, PyExc_OverflowError
))
560 /* If no error-handling desired then the default clipping
564 assert(PyLong_Check(value
));
565 /* Whether or not it is less than or equal to
566 zero is determined by the sign of ob_size
568 if (_PyLong_Sign(value
) < 0)
569 result
= PY_OFF_T_MIN
;
571 result
= PY_OFF_T_MAX
;
574 /* Otherwise replace the error with caller's error object. */
576 "cannot fit '%.200s' into an offset-sized integer",
577 item
->ob_type
->tp_name
);
586 /* Basically the "n" format code with the ability to turn None into -1. */
588 _PyIO_ConvertSsize_t(PyObject
*obj
, void *result
) {
590 if (obj
== Py_None
) {
593 else if (PyNumber_Check(obj
)) {
594 limit
= PyNumber_AsSsize_t(obj
, PyExc_OverflowError
);
595 if (limit
== -1 && PyErr_Occurred())
599 PyErr_Format(PyExc_TypeError
,
600 "integer argument expected, got '%.200s'",
601 Py_TYPE(obj
)->tp_name
);
604 *((Py_ssize_t
*)result
) = limit
;
613 PyObject
*_PyIO_os_module
= NULL
;
614 PyObject
*_PyIO_locale_module
= NULL
;
615 PyObject
*_PyIO_unsupported_operation
= NULL
;
617 static PyMethodDef module_methods
[] = {
618 {"open", (PyCFunction
)io_open
, METH_VARARGS
|METH_KEYWORDS
, open_doc
},
625 PyObject
*m
= Py_InitModule4("_io", module_methods
,
626 module_doc
, NULL
, PYTHON_API_VERSION
);
630 /* put os in the module state */
631 _PyIO_os_module
= PyImport_ImportModule("os");
632 if (_PyIO_os_module
== NULL
)
635 #define ADD_TYPE(type, name) \
636 if (PyType_Ready(type) < 0) \
639 if (PyModule_AddObject(m, name, (PyObject *)type) < 0) { \
644 /* DEFAULT_BUFFER_SIZE */
645 if (PyModule_AddIntMacro(m
, DEFAULT_BUFFER_SIZE
) < 0)
648 /* UnsupportedOperation inherits from ValueError and IOError */
649 _PyIO_unsupported_operation
= PyObject_CallFunction(
650 (PyObject
*)&PyType_Type
, "s(OO){}",
651 "UnsupportedOperation", PyExc_ValueError
, PyExc_IOError
);
652 if (_PyIO_unsupported_operation
== NULL
)
654 Py_INCREF(_PyIO_unsupported_operation
);
655 if (PyModule_AddObject(m
, "UnsupportedOperation",
656 _PyIO_unsupported_operation
) < 0)
659 /* BlockingIOError */
660 _PyExc_BlockingIOError
.tp_base
= (PyTypeObject
*) PyExc_IOError
;
661 ADD_TYPE(&_PyExc_BlockingIOError
, "BlockingIOError");
663 /* Concrete base types of the IO ABCs.
664 (the ABCs themselves are declared through inheritance in io.py)
666 ADD_TYPE(&PyIOBase_Type
, "_IOBase");
667 ADD_TYPE(&PyRawIOBase_Type
, "_RawIOBase");
668 ADD_TYPE(&PyBufferedIOBase_Type
, "_BufferedIOBase");
669 ADD_TYPE(&PyTextIOBase_Type
, "_TextIOBase");
671 /* Implementation of concrete IO objects. */
673 PyFileIO_Type
.tp_base
= &PyRawIOBase_Type
;
674 ADD_TYPE(&PyFileIO_Type
, "FileIO");
677 PyBytesIO_Type
.tp_base
= &PyBufferedIOBase_Type
;
678 ADD_TYPE(&PyBytesIO_Type
, "BytesIO");
681 PyStringIO_Type
.tp_base
= &PyTextIOBase_Type
;
682 ADD_TYPE(&PyStringIO_Type
, "StringIO");
685 PyBufferedReader_Type
.tp_base
= &PyBufferedIOBase_Type
;
686 ADD_TYPE(&PyBufferedReader_Type
, "BufferedReader");
689 PyBufferedWriter_Type
.tp_base
= &PyBufferedIOBase_Type
;
690 ADD_TYPE(&PyBufferedWriter_Type
, "BufferedWriter");
693 PyBufferedRWPair_Type
.tp_base
= &PyBufferedIOBase_Type
;
694 ADD_TYPE(&PyBufferedRWPair_Type
, "BufferedRWPair");
697 PyBufferedRandom_Type
.tp_base
= &PyBufferedIOBase_Type
;
698 ADD_TYPE(&PyBufferedRandom_Type
, "BufferedRandom");
701 PyTextIOWrapper_Type
.tp_base
= &PyTextIOBase_Type
;
702 ADD_TYPE(&PyTextIOWrapper_Type
, "TextIOWrapper");
704 /* IncrementalNewlineDecoder */
705 ADD_TYPE(&PyIncrementalNewlineDecoder_Type
, "IncrementalNewlineDecoder");
707 /* Interned strings */
708 if (!(_PyIO_str_close
= PyString_InternFromString("close")))
710 if (!(_PyIO_str_closed
= PyString_InternFromString("closed")))
712 if (!(_PyIO_str_decode
= PyString_InternFromString("decode")))
714 if (!(_PyIO_str_encode
= PyString_InternFromString("encode")))
716 if (!(_PyIO_str_fileno
= PyString_InternFromString("fileno")))
718 if (!(_PyIO_str_flush
= PyString_InternFromString("flush")))
720 if (!(_PyIO_str_getstate
= PyString_InternFromString("getstate")))
722 if (!(_PyIO_str_isatty
= PyString_InternFromString("isatty")))
724 if (!(_PyIO_str_newlines
= PyString_InternFromString("newlines")))
726 if (!(_PyIO_str_nl
= PyString_InternFromString("\n")))
728 if (!(_PyIO_str_read
= PyString_InternFromString("read")))
730 if (!(_PyIO_str_read1
= PyString_InternFromString("read1")))
732 if (!(_PyIO_str_readable
= PyString_InternFromString("readable")))
734 if (!(_PyIO_str_readinto
= PyString_InternFromString("readinto")))
736 if (!(_PyIO_str_readline
= PyString_InternFromString("readline")))
738 if (!(_PyIO_str_reset
= PyString_InternFromString("reset")))
740 if (!(_PyIO_str_seek
= PyString_InternFromString("seek")))
742 if (!(_PyIO_str_seekable
= PyString_InternFromString("seekable")))
744 if (!(_PyIO_str_setstate
= PyString_InternFromString("setstate")))
746 if (!(_PyIO_str_tell
= PyString_InternFromString("tell")))
748 if (!(_PyIO_str_truncate
= PyString_InternFromString("truncate")))
750 if (!(_PyIO_str_write
= PyString_InternFromString("write")))
752 if (!(_PyIO_str_writable
= PyString_InternFromString("writable")))
755 if (!(_PyIO_empty_str
= PyUnicode_FromStringAndSize(NULL
, 0)))
757 if (!(_PyIO_empty_bytes
= PyBytes_FromStringAndSize(NULL
, 0)))
759 if (!(_PyIO_zero
= PyLong_FromLong(0L)))
765 Py_CLEAR(_PyIO_os_module
);
766 Py_CLEAR(_PyIO_unsupported_operation
);