Google Git
Sign in
chromium / external / github.com / python / cpython / b225c08de889d2bf070e6c981c5f386cf06d961c / . / Objects / unicodeobject.c
blob: ec5684b1d0950265931ea063770bffffb75b2952 [file] [log] [blame]
/*
Unicode implementation based on original code by Fredrik Lundh,
modified by Marc-Andre Lemburg <mal@lemburg.com>.
Major speed upgrades to the method implementations at the Reykjavik
NeedForSpeed sprint, by Fredrik Lundh and Andrew Dalke.
Copyright (c) Corporation for National Research Initiatives.
--------------------------------------------------------------------
The original string type implementation is:
Copyright (c) 1999 by Secret Labs AB
Copyright (c) 1999 by Fredrik Lundh
By obtaining, using, and/or copying this software and/or its
associated documentation, you agree that you have read, understood,
and will comply with the following terms and conditions:
Permission to use, copy, modify, and distribute this software and its
associated documentation for any purpose and without fee is hereby
granted, provided that the above copyright notice appears in all
copies, and that both that copyright notice and this permission notice
appear in supporting documentation, and that the name of Secret Labs
AB or the author not be used in advertising or publicity pertaining to
distribution of the software without specific, written prior
permission.
SECRET LABS AB AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO
THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS. IN NO EVENT SHALL SECRET LABS AB OR THE AUTHOR BE LIABLE FOR
ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
--------------------------------------------------------------------
*/
#define PY_SSIZE_T_CLEAN
#include "Python.h"
#include "pycore_abstract.h" // _PyIndex_Check()
#include "pycore_atomic_funcs.h" // _Py_atomic_size_get()
#include "pycore_bytesobject.h" // _PyBytes_Repeat()
#include "pycore_bytes_methods.h" // _Py_bytes_lower()
#include "pycore_format.h" // F_LJUST
#include "pycore_initconfig.h" // _PyStatus_OK()
#include "pycore_interp.h" // PyInterpreterState.fs_codec
#include "pycore_long.h" // _PyLong_FormatWriter()
#include "pycore_object.h" // _PyObject_GC_TRACK(), _Py_FatalRefcountError()
#include "pycore_pathconfig.h" // _Py_DumpPathConfig()
#include "pycore_pylifecycle.h" // _Py_SetFileSystemEncoding()
#include "pycore_pystate.h" // _PyInterpreterState_GET()
#include "pycore_ucnhash.h" // _PyUnicode_Name_CAPI
#include "pycore_unicodeobject.h" // struct _Py_unicode_state
#include "pycore_unicodeobject_generated.h" // _PyUnicode_InitStaticStrings()
#include "stringlib/eq.h" // unicode_eq()
#include <stddef.h> // ptrdiff_t
#ifdef MS_WINDOWS
#include <windows.h>
#endif
#ifdef HAVE_NON_UNICODE_WCHAR_T_REPRESENTATION
# include "pycore_fileutils.h" // _Py_LocaleUsesNonUnicodeWchar()
#endif
/* Uncomment to display statistics on interned strings at exit
in _PyUnicode_ClearInterned(). */
/* #define INTERNED_STATS 1 */
/*[clinic input]
class str "PyObject *" "&PyUnicode_Type"
[clinic start generated code]*/
/*[clinic end generated code: output=da39a3ee5e6b4b0d input=4884c934de622cf6]*/
/*[python input]
class Py_UCS4_converter(CConverter):
type = 'Py_UCS4'
converter = 'convert_uc'
def converter_init(self):
if self.default is not unspecified:
self.c_default = ascii(self.default)
if len(self.c_default) > 4 or self.c_default[0] != "'":
self.c_default = hex(ord(self.default))
[python start generated code]*/
/*[python end generated code: output=da39a3ee5e6b4b0d input=88f5dd06cd8e7a61]*/
/* --- Globals ------------------------------------------------------------
NOTE: In the interpreter's initialization phase, some globals are currently
initialized dynamically as needed. In the process Unicode objects may
be created before the Unicode type is ready.
*/
#ifdef __cplusplus
extern "C" {
#endif
// Maximum code point of Unicode 6.0: 0x10ffff (1,114,111).
// The value must be the same in fileutils.c.
#define MAX_UNICODE 0x10ffff
#ifdef Py_DEBUG
# define _PyUnicode_CHECK(op) _PyUnicode_CheckConsistency(op, 0)
#else
# define _PyUnicode_CHECK(op) PyUnicode_Check(op)
#endif
#define _PyUnicode_UTF8(op) \
(_PyCompactUnicodeObject_CAST(op)->utf8)
#define PyUnicode_UTF8(op) \
(assert(_PyUnicode_CHECK(op)), \
PyUnicode_IS_COMPACT_ASCII(op) ? \
((char*)(_PyASCIIObject_CAST(op) + 1)) : \
_PyUnicode_UTF8(op))
#define _PyUnicode_UTF8_LENGTH(op) \
(_PyCompactUnicodeObject_CAST(op)->utf8_length)
#define PyUnicode_UTF8_LENGTH(op) \
(assert(_PyUnicode_CHECK(op)), \
PyUnicode_IS_COMPACT_ASCII(op) ? \
_PyASCIIObject_CAST(op)->length : \
_PyUnicode_UTF8_LENGTH(op))
#define _PyUnicode_LENGTH(op) \
(_PyASCIIObject_CAST(op)->length)
#define _PyUnicode_STATE(op) \
(_PyASCIIObject_CAST(op)->state)
#define _PyUnicode_HASH(op) \
(_PyASCIIObject_CAST(op)->hash)
#define _PyUnicode_KIND(op) \
(assert(_PyUnicode_CHECK(op)), \
_PyASCIIObject_CAST(op)->state.kind)
#define _PyUnicode_GET_LENGTH(op) \
(assert(_PyUnicode_CHECK(op)), \
_PyASCIIObject_CAST(op)->length)
#define _PyUnicode_DATA_ANY(op) \
(_PyUnicodeObject_CAST(op)->data.any)
#define _PyUnicode_SHARE_UTF8(op) \
(assert(_PyUnicode_CHECK(op)), \
assert(!PyUnicode_IS_COMPACT_ASCII(op)), \
(_PyUnicode_UTF8(op) == PyUnicode_DATA(op)))
/* true if the Unicode object has an allocated UTF-8 memory block
(not shared with other data) */
#define _PyUnicode_HAS_UTF8_MEMORY(op) \
((!PyUnicode_IS_COMPACT_ASCII(op) \
&& _PyUnicode_UTF8(op) \
&& _PyUnicode_UTF8(op) != PyUnicode_DATA(op)))
/* Generic helper macro to convert characters of different types.
from_type and to_type have to be valid type names, begin and end
are pointers to the source characters which should be of type
"from_type *". to is a pointer of type "to_type *" and points to the
buffer where the result characters are written to. */
#define _PyUnicode_CONVERT_BYTES(from_type, to_type, begin, end, to) \
do { \
to_type *_to = (to_type *)(to); \
const from_type *_iter = (const from_type *)(begin);\
const from_type *_end = (const from_type *)(end);\
Py_ssize_t n = (_end) - (_iter); \
const from_type *_unrolled_end = \
_iter + _Py_SIZE_ROUND_DOWN(n, 4); \
while (_iter < (_unrolled_end)) { \
_to[0] = (to_type) _iter[0]; \
_to[1] = (to_type) _iter[1]; \
_to[2] = (to_type) _iter[2]; \
_to[3] = (to_type) _iter[3]; \
_iter += 4; _to += 4; \
} \
while (_iter < (_end)) \
*_to++ = (to_type) *_iter++; \
} while (0)
#define LATIN1(ch) \
(ch < 128 \
? (PyObject*)&_Py_SINGLETON(strings).ascii[ch] \
: (PyObject*)&_Py_SINGLETON(strings).latin1[ch - 128])
#ifdef MS_WINDOWS
/* On Windows, overallocate by 50% is the best factor */
# define OVERALLOCATE_FACTOR 2
#else
/* On Linux, overallocate by 25% is the best factor */
# define OVERALLOCATE_FACTOR 4
#endif
/* Forward declaration */
static inline int
_PyUnicodeWriter_WriteCharInline(_PyUnicodeWriter *writer, Py_UCS4 ch);
static inline void
_PyUnicodeWriter_InitWithBuffer(_PyUnicodeWriter *writer, PyObject *buffer);
static PyObject *
unicode_encode_utf8(PyObject *unicode, _Py_error_handler error_handler,
const char *errors);
static PyObject *
unicode_decode_utf8(const char *s, Py_ssize_t size,
_Py_error_handler error_handler, const char *errors,
Py_ssize_t *consumed);
#ifdef Py_DEBUG
static inline int unicode_is_finalizing(void);
static int unicode_is_singleton(PyObject *unicode);
#endif
// Return a borrowed reference to the empty string singleton.
static inline PyObject* unicode_get_empty(void)
{
_Py_DECLARE_STR(empty, "");
return &_Py_STR(empty);
}
// Return a strong reference to the empty string singleton.
static inline PyObject* unicode_new_empty(void)
{
PyObject *empty = unicode_get_empty();
return Py_NewRef(empty);
}
/* This dictionary holds all interned unicode strings. Note that references
to strings in this dictionary are *not* counted in the string's ob_refcnt.
When the interned string reaches a refcnt of 0 the string deallocation
function will delete the reference from this dictionary.
*/
static inline PyObject *get_interned_dict(PyInterpreterState *interp)
{
return _Py_INTERP_CACHED_OBJECT(interp, interned_strings);
}
Py_ssize_t
_PyUnicode_InternedSize()
{
return PyObject_Length(get_interned_dict(_PyInterpreterState_GET()));
}
static int
init_interned_dict(PyInterpreterState *interp)
{
assert(get_interned_dict(interp) == NULL);
PyObject *interned = interned = PyDict_New();
if (interned == NULL) {
return -1;
}
_Py_INTERP_CACHED_OBJECT(interp, interned_strings) = interned;
return 0;
}
static void
clear_interned_dict(PyInterpreterState *interp)
{
PyObject *interned = get_interned_dict(interp);
if (interned != NULL) {
PyDict_Clear(interned);
Py_DECREF(interned);
_Py_INTERP_CACHED_OBJECT(interp, interned_strings) = NULL;
}
}
#define _Py_RETURN_UNICODE_EMPTY() \
do { \
return unicode_new_empty(); \
} while (0)
static inline void
unicode_fill(int kind, void *data, Py_UCS4 value,
Py_ssize_t start, Py_ssize_t length)
{
assert(0 <= start);
switch (kind) {
case PyUnicode_1BYTE_KIND: {
assert(value <= 0xff);
Py_UCS1 ch = (unsigned char)value;
Py_UCS1 *to = (Py_UCS1 *)data + start;
memset(to, ch, length);
break;
}
case PyUnicode_2BYTE_KIND: {
assert(value <= 0xffff);
Py_UCS2 ch = (Py_UCS2)value;
Py_UCS2 *to = (Py_UCS2 *)data + start;
const Py_UCS2 *end = to + length;
for (; to < end; ++to) *to = ch;
break;
}
case PyUnicode_4BYTE_KIND: {
assert(value <= MAX_UNICODE);
Py_UCS4 ch = value;
Py_UCS4 * to = (Py_UCS4 *)data + start;
const Py_UCS4 *end = to + length;
for (; to < end; ++to) *to = ch;
break;
}
default: Py_UNREACHABLE();
}
}
/* Fast detection of the most frequent whitespace characters */
const unsigned char _Py_ascii_whitespace[] = {
0, 0, 0, 0, 0, 0, 0, 0,
/* case 0x0009: * CHARACTER TABULATION */
/* case 0x000A: * LINE FEED */
/* case 0x000B: * LINE TABULATION */
/* case 0x000C: * FORM FEED */
/* case 0x000D: * CARRIAGE RETURN */
0, 1, 1, 1, 1, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
/* case 0x001C: * FILE SEPARATOR */
/* case 0x001D: * GROUP SEPARATOR */
/* case 0x001E: * RECORD SEPARATOR */
/* case 0x001F: * UNIT SEPARATOR */
0, 0, 0, 0, 1, 1, 1, 1,
/* case 0x0020: * SPACE */
1, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0
};
/* forward */
static PyObject* get_latin1_char(unsigned char ch);
static int unicode_modifiable(PyObject *unicode);
static PyObject *
_PyUnicode_FromUCS1(const Py_UCS1 *s, Py_ssize_t size);
static PyObject *
_PyUnicode_FromUCS2(const Py_UCS2 *s, Py_ssize_t size);
static PyObject *
_PyUnicode_FromUCS4(const Py_UCS4 *s, Py_ssize_t size);
static PyObject *
unicode_encode_call_errorhandler(const char *errors,
PyObject **errorHandler,const char *encoding, const char *reason,
PyObject *unicode, PyObject **exceptionObject,
Py_ssize_t startpos, Py_ssize_t endpos, Py_ssize_t *newpos);
static void
raise_encode_exception(PyObject **exceptionObject,
const char *encoding,
PyObject *unicode,
Py_ssize_t startpos, Py_ssize_t endpos,
const char *reason);
/* Same for linebreaks */
static const unsigned char ascii_linebreak[] = {
0, 0, 0, 0, 0, 0, 0, 0,
/* 0x000A, * LINE FEED */
/* 0x000B, * LINE TABULATION */
/* 0x000C, * FORM FEED */
/* 0x000D, * CARRIAGE RETURN */
0, 0, 1, 1, 1, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
/* 0x001C, * FILE SEPARATOR */
/* 0x001D, * GROUP SEPARATOR */
/* 0x001E, * RECORD SEPARATOR */
0, 0, 0, 0, 1, 1, 1, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0
};
static int convert_uc(PyObject *obj, void *addr);
struct encoding_map;
#include "clinic/unicodeobject.c.h"
_Py_error_handler
_Py_GetErrorHandler(const char *errors)
{
if (errors == NULL || strcmp(errors, "strict") == 0) {
return _Py_ERROR_STRICT;
}
if (strcmp(errors, "surrogateescape") == 0) {
return _Py_ERROR_SURROGATEESCAPE;
}
if (strcmp(errors, "replace") == 0) {
return _Py_ERROR_REPLACE;
}
if (strcmp(errors, "ignore") == 0) {
return _Py_ERROR_IGNORE;
}
if (strcmp(errors, "backslashreplace") == 0) {
return _Py_ERROR_BACKSLASHREPLACE;
}
if (strcmp(errors, "surrogatepass") == 0) {
return _Py_ERROR_SURROGATEPASS;
}
if (strcmp(errors, "xmlcharrefreplace") == 0) {
return _Py_ERROR_XMLCHARREFREPLACE;
}
return _Py_ERROR_OTHER;
}
static _Py_error_handler
get_error_handler_wide(const wchar_t *errors)
{
if (errors == NULL || wcscmp(errors, L"strict") == 0) {
return _Py_ERROR_STRICT;
}
if (wcscmp(errors, L"surrogateescape") == 0) {
return _Py_ERROR_SURROGATEESCAPE;
}
if (wcscmp(errors, L"replace") == 0) {
return _Py_ERROR_REPLACE;
}
if (wcscmp(errors, L"ignore") == 0) {
return _Py_ERROR_IGNORE;
}
if (wcscmp(errors, L"backslashreplace") == 0) {
return _Py_ERROR_BACKSLASHREPLACE;
}
if (wcscmp(errors, L"surrogatepass") == 0) {
return _Py_ERROR_SURROGATEPASS;
}
if (wcscmp(errors, L"xmlcharrefreplace") == 0) {
return _Py_ERROR_XMLCHARREFREPLACE;
}
return _Py_ERROR_OTHER;
}
static inline int
unicode_check_encoding_errors(const char *encoding, const char *errors)
{
if (encoding == NULL && errors == NULL) {
return 0;
}
PyInterpreterState *interp = _PyInterpreterState_GET();
#ifndef Py_DEBUG
/* In release mode, only check in development mode (-X dev) */
if (!_PyInterpreterState_GetConfig(interp)->dev_mode) {
return 0;
}
#else
/* Always check in debug mode */
#endif
/* Avoid calling _PyCodec_Lookup() and PyCodec_LookupError() before the
codec registry is ready: before_PyUnicode_InitEncodings() is called. */
if (!interp->unicode.fs_codec.encoding) {
return 0;
}
/* Disable checks during Python finalization. For example, it allows to
call _PyObject_Dump() during finalization for debugging purpose. */
if (interp->finalizing) {
return 0;
}
if (encoding != NULL
// Fast path for the most common built-in encodings. Even if the codec
// is cached, _PyCodec_Lookup() decodes the bytes string from UTF-8 to
// create a temporary Unicode string (the key in the cache).
&& strcmp(encoding, "utf-8") != 0
&& strcmp(encoding, "utf8") != 0
&& strcmp(encoding, "ascii") != 0)
{
PyObject *handler = _PyCodec_Lookup(encoding);
if (handler == NULL) {
return -1;
}
Py_DECREF(handler);
}
if (errors != NULL
// Fast path for the most common built-in error handlers.
&& strcmp(errors, "strict") != 0
&& strcmp(errors, "ignore") != 0
&& strcmp(errors, "replace") != 0
&& strcmp(errors, "surrogateescape") != 0
&& strcmp(errors, "surrogatepass") != 0)
{
PyObject *handler = PyCodec_LookupError(errors);
if (handler == NULL) {
return -1;
}
Py_DECREF(handler);
}
return 0;
}
int
_PyUnicode_CheckConsistency(PyObject *op, int check_content)
{
#define CHECK(expr) \
do { if (!(expr)) { _PyObject_ASSERT_FAILED_MSG(op, Py_STRINGIFY(expr)); } } while (0)
assert(op != NULL);
CHECK(PyUnicode_Check(op));
PyASCIIObject *ascii = _PyASCIIObject_CAST(op);
int kind = ascii->state.kind;
if (ascii->state.ascii == 1 && ascii->state.compact == 1) {
CHECK(kind == PyUnicode_1BYTE_KIND);
}
else {
PyCompactUnicodeObject *compact = _PyCompactUnicodeObject_CAST(op);
void *data;
if (ascii->state.compact == 1) {
data = compact + 1;
CHECK(kind == PyUnicode_1BYTE_KIND
|| kind == PyUnicode_2BYTE_KIND
|| kind == PyUnicode_4BYTE_KIND);
CHECK(ascii->state.ascii == 0);
CHECK(compact->utf8 != data);
}
else {
PyUnicodeObject *unicode = _PyUnicodeObject_CAST(op);
data = unicode->data.any;
CHECK(kind == PyUnicode_1BYTE_KIND
|| kind == PyUnicode_2BYTE_KIND
|| kind == PyUnicode_4BYTE_KIND);
CHECK(ascii->state.compact == 0);
CHECK(data != NULL);
if (ascii->state.ascii) {
CHECK(compact->utf8 == data);
CHECK(compact->utf8_length == ascii->length);
}
else {
CHECK(compact->utf8 != data);
}
}
if (compact->utf8 == NULL)
CHECK(compact->utf8_length == 0);
}
/* check that the best kind is used: O(n) operation */
if (check_content) {
Py_ssize_t i;
Py_UCS4 maxchar = 0;
const void *data;
Py_UCS4 ch;
data = PyUnicode_DATA(ascii);
for (i=0; i < ascii->length; i++)
{
ch = PyUnicode_READ(kind, data, i);
if (ch > maxchar)
maxchar = ch;
}
if (kind == PyUnicode_1BYTE_KIND) {
if (ascii->state.ascii == 0) {
CHECK(maxchar >= 128);
CHECK(maxchar <= 255);
}
else
CHECK(maxchar < 128);
}
else if (kind == PyUnicode_2BYTE_KIND) {
CHECK(maxchar >= 0x100);
CHECK(maxchar <= 0xFFFF);
}
else {
CHECK(maxchar >= 0x10000);
CHECK(maxchar <= MAX_UNICODE);
}
CHECK(PyUnicode_READ(kind, data, ascii->length) == 0);
}
return 1;
#undef CHECK
}
static PyObject*
unicode_result(PyObject *unicode)
{
assert(_PyUnicode_CHECK(unicode));
Py_ssize_t length = PyUnicode_GET_LENGTH(unicode);
if (length == 0) {
PyObject *empty = unicode_get_empty();
if (unicode != empty) {
Py_DECREF(unicode);
Py_INCREF(empty);
}
return empty;
}
if (length == 1) {
int kind = PyUnicode_KIND(unicode);
if (kind == PyUnicode_1BYTE_KIND) {
const Py_UCS1 *data = PyUnicode_1BYTE_DATA(unicode);
Py_UCS1 ch = data[0];
PyObject *latin1_char = LATIN1(ch);
if (unicode != latin1_char) {
Py_INCREF(latin1_char);
Py_DECREF(unicode);
}
return latin1_char;
}
}
assert(_PyUnicode_CheckConsistency(unicode, 1));
return unicode;
}
static PyObject*
unicode_result_unchanged(PyObject *unicode)
{
if (PyUnicode_CheckExact(unicode)) {
return Py_NewRef(unicode);
}
else
/* Subtype -- return genuine unicode string with the same value. */
return _PyUnicode_Copy(unicode);
}
/* Implementation of the "backslashreplace" error handler for 8-bit encodings:
ASCII, Latin1, UTF-8, etc. */
static char*
backslashreplace(_PyBytesWriter *writer, char *str,
PyObject *unicode, Py_ssize_t collstart, Py_ssize_t collend)
{
Py_ssize_t size, i;
Py_UCS4 ch;
int kind;
const void *data;
kind = PyUnicode_KIND(unicode);
data = PyUnicode_DATA(unicode);
size = 0;
/* determine replacement size */
for (i = collstart; i < collend; ++i) {
Py_ssize_t incr;
ch = PyUnicode_READ(kind, data, i);
if (ch < 0x100)
incr = 2+2;
else if (ch < 0x10000)
incr = 2+4;
else {
assert(ch <= MAX_UNICODE);
incr = 2+8;
}
if (size > PY_SSIZE_T_MAX - incr) {
PyErr_SetString(PyExc_OverflowError,
"encoded result is too long for a Python string");
return NULL;
}
size += incr;
}
str = _PyBytesWriter_Prepare(writer, str, size);
if (str == NULL)
return NULL;
/* generate replacement */
for (i = collstart; i < collend; ++i) {
ch = PyUnicode_READ(kind, data, i);
*str++ = '\\';
if (ch >= 0x00010000) {
*str++ = 'U';
*str++ = Py_hexdigits[(ch>>28)&0xf];
*str++ = Py_hexdigits[(ch>>24)&0xf];
*str++ = Py_hexdigits[(ch>>20)&0xf];
*str++ = Py_hexdigits[(ch>>16)&0xf];
*str++ = Py_hexdigits[(ch>>12)&0xf];
*str++ = Py_hexdigits[(ch>>8)&0xf];
}
else if (ch >= 0x100) {
*str++ = 'u';
*str++ = Py_hexdigits[(ch>>12)&0xf];
*str++ = Py_hexdigits[(ch>>8)&0xf];
}
else
*str++ = 'x';
*str++ = Py_hexdigits[(ch>>4)&0xf];
*str++ = Py_hexdigits[ch&0xf];
}
return str;
}
/* Implementation of the "xmlcharrefreplace" error handler for 8-bit encodings:
ASCII, Latin1, UTF-8, etc. */
static char*
xmlcharrefreplace(_PyBytesWriter *writer, char *str,
PyObject *unicode, Py_ssize_t collstart, Py_ssize_t collend)
{
Py_ssize_t size, i;
Py_UCS4 ch;
int kind;
const void *data;
kind = PyUnicode_KIND(unicode);
data = PyUnicode_DATA(unicode);
size = 0;
/* determine replacement size */
for (i = collstart; i < collend; ++i) {
Py_ssize_t incr;
ch = PyUnicode_READ(kind, data, i);
if (ch < 10)
incr = 2+1+1;
else if (ch < 100)
incr = 2+2+1;
else if (ch < 1000)
incr = 2+3+1;
else if (ch < 10000)
incr = 2+4+1;
else if (ch < 100000)
incr = 2+5+1;
else if (ch < 1000000)
incr = 2+6+1;
else {
assert(ch <= MAX_UNICODE);
incr = 2+7+1;
}
if (size > PY_SSIZE_T_MAX - incr) {
PyErr_SetString(PyExc_OverflowError,
"encoded result is too long for a Python string");
return NULL;
}
size += incr;
}
str = _PyBytesWriter_Prepare(writer, str, size);
if (str == NULL)
return NULL;
/* generate replacement */
for (i = collstart; i < collend; ++i) {
size = sprintf(str, "&#%d;", PyUnicode_READ(kind, data, i));
if (size < 0) {
return NULL;
}
str += size;
}
return str;
}
/* --- Bloom Filters ----------------------------------------------------- */
/* stuff to implement simple "bloom filters" for Unicode characters.
to keep things simple, we use a single bitmask, using the least 5
bits from each unicode characters as the bit index. */
/* the linebreak mask is set up by _PyUnicode_Init() below */
#if LONG_BIT >= 128
#define BLOOM_WIDTH 128
#elif LONG_BIT >= 64
#define BLOOM_WIDTH 64
#elif LONG_BIT >= 32
#define BLOOM_WIDTH 32
#else
#error "LONG_BIT is smaller than 32"
#endif
#define BLOOM_MASK unsigned long
static BLOOM_MASK bloom_linebreak = ~(BLOOM_MASK)0;
#define BLOOM(mask, ch) ((mask & (1UL << ((ch) & (BLOOM_WIDTH - 1)))))
#define BLOOM_LINEBREAK(ch) \
((ch) < 128U ? ascii_linebreak[(ch)] : \
(BLOOM(bloom_linebreak, (ch)) && Py_UNICODE_ISLINEBREAK(ch)))
static inline BLOOM_MASK
make_bloom_mask(int kind, const void* ptr, Py_ssize_t len)
{
#define BLOOM_UPDATE(TYPE, MASK, PTR, LEN) \
do { \
TYPE *data = (TYPE *)PTR; \
TYPE *end = data + LEN; \
Py_UCS4 ch; \
for (; data != end; data++) { \
ch = *data; \
MASK |= (1UL << (ch & (BLOOM_WIDTH - 1))); \
} \
break; \
} while (0)
/* calculate simple bloom-style bitmask for a given unicode string */
BLOOM_MASK mask;
mask = 0;
switch (kind) {
case PyUnicode_1BYTE_KIND:
BLOOM_UPDATE(Py_UCS1, mask, ptr, len);
break;
case PyUnicode_2BYTE_KIND:
BLOOM_UPDATE(Py_UCS2, mask, ptr, len);
break;
case PyUnicode_4BYTE_KIND:
BLOOM_UPDATE(Py_UCS4, mask, ptr, len);
break;
default:
Py_UNREACHABLE();
}
return mask;
#undef BLOOM_UPDATE
}
static int
ensure_unicode(PyObject *obj)
{
if (!PyUnicode_Check(obj)) {
PyErr_Format(PyExc_TypeError,
"must be str, not %.100s",
Py_TYPE(obj)->tp_name);
return -1;
}
return 0;
}
/* Compilation of templated routines */
#define STRINGLIB_GET_EMPTY() unicode_get_empty()
#include "stringlib/asciilib.h"
#include "stringlib/fastsearch.h"
#include "stringlib/partition.h"
#include "stringlib/split.h"
#include "stringlib/count.h"
#include "stringlib/find.h"
#include "stringlib/find_max_char.h"
#include "stringlib/undef.h"
#include "stringlib/ucs1lib.h"
#include "stringlib/fastsearch.h"
#include "stringlib/partition.h"
#include "stringlib/split.h"
#include "stringlib/count.h"
#include "stringlib/find.h"
#include "stringlib/replace.h"
#include "stringlib/find_max_char.h"
#include "stringlib/undef.h"
#include "stringlib/ucs2lib.h"
#include "stringlib/fastsearch.h"
#include "stringlib/partition.h"
#include "stringlib/split.h"
#include "stringlib/count.h"
#include "stringlib/find.h"
#include "stringlib/replace.h"
#include "stringlib/find_max_char.h"
#include "stringlib/undef.h"
#include "stringlib/ucs4lib.h"
#include "stringlib/fastsearch.h"
#include "stringlib/partition.h"
#include "stringlib/split.h"
#include "stringlib/count.h"
#include "stringlib/find.h"
#include "stringlib/replace.h"
#include "stringlib/find_max_char.h"
#include "stringlib/undef.h"
#undef STRINGLIB_GET_EMPTY
/* --- Unicode Object ----------------------------------------------------- */
static inline Py_ssize_t
findchar(const void *s, int kind,
Py_ssize_t size, Py_UCS4 ch,
int direction)
{
switch (kind) {
case PyUnicode_1BYTE_KIND:
if ((Py_UCS1) ch != ch)
return -1;
if (direction > 0)
return ucs1lib_find_char((const Py_UCS1 *) s, size, (Py_UCS1) ch);
else
return ucs1lib_rfind_char((const Py_UCS1 *) s, size, (Py_UCS1) ch);
case PyUnicode_2BYTE_KIND:
if ((Py_UCS2) ch != ch)
return -1;
if (direction > 0)
return ucs2lib_find_char((const Py_UCS2 *) s, size, (Py_UCS2) ch);
else
return ucs2lib_rfind_char((const Py_UCS2 *) s, size, (Py_UCS2) ch);
case PyUnicode_4BYTE_KIND:
if (direction > 0)
return ucs4lib_find_char((const Py_UCS4 *) s, size, ch);
else
return ucs4lib_rfind_char((const Py_UCS4 *) s, size, ch);
default:
Py_UNREACHABLE();
}
}
#ifdef Py_DEBUG
/* Fill the data of a Unicode string with invalid characters to detect bugs
earlier.
_PyUnicode_CheckConsistency(str, 1) detects invalid characters, at least for
ASCII and UCS-4 strings. U+00FF is invalid in ASCII and U+FFFFFFFF is an
invalid character in Unicode 6.0. */
static void
unicode_fill_invalid(PyObject *unicode, Py_ssize_t old_length)
{
int kind = PyUnicode_KIND(unicode);
Py_UCS1 *data = PyUnicode_1BYTE_DATA(unicode);
Py_ssize_t length = _PyUnicode_LENGTH(unicode);
if (length <= old_length)
return;
memset(data + old_length * kind, 0xff, (length - old_length) * kind);
}
#endif
static PyObject*
resize_compact(PyObject *unicode, Py_ssize_t length)
{
Py_ssize_t char_size;
Py_ssize_t struct_size;
Py_ssize_t new_size;
PyObject *new_unicode;
#ifdef Py_DEBUG
Py_ssize_t old_length = _PyUnicode_LENGTH(unicode);
#endif
assert(unicode_modifiable(unicode));
assert(PyUnicode_IS_COMPACT(unicode));
char_size = PyUnicode_KIND(unicode);
if (PyUnicode_IS_ASCII(unicode))
struct_size = sizeof(PyASCIIObject);
else
struct_size = sizeof(PyCompactUnicodeObject);
if (length > ((PY_SSIZE_T_MAX - struct_size) / char_size - 1)) {
PyErr_NoMemory();
return NULL;
}
new_size = (struct_size + (length + 1) * char_size);
if (_PyUnicode_HAS_UTF8_MEMORY(unicode)) {
PyObject_Free(_PyUnicode_UTF8(unicode));
_PyUnicode_UTF8(unicode) = NULL;
_PyUnicode_UTF8_LENGTH(unicode) = 0;
}
#ifdef Py_TRACE_REFS
_Py_ForgetReference(unicode);
#endif
new_unicode = (PyObject *)PyObject_Realloc(unicode, new_size);
if (new_unicode == NULL) {
_Py_NewReferenceNoTotal(unicode);
PyErr_NoMemory();
return NULL;
}
unicode = new_unicode;
_Py_NewReferenceNoTotal(unicode);
_PyUnicode_LENGTH(unicode) = length;
#ifdef Py_DEBUG
unicode_fill_invalid(unicode, old_length);
#endif
PyUnicode_WRITE(PyUnicode_KIND(unicode), PyUnicode_DATA(unicode),
length, 0);
assert(_PyUnicode_CheckConsistency(unicode, 0));
return unicode;
}
static int
resize_inplace(PyObject *unicode, Py_ssize_t length)
{
assert(!PyUnicode_IS_COMPACT(unicode));
assert(Py_REFCNT(unicode) == 1);
Py_ssize_t new_size;
Py_ssize_t char_size;
int share_utf8;
void *data;
#ifdef Py_DEBUG
Py_ssize_t old_length = _PyUnicode_LENGTH(unicode);
#endif
data = _PyUnicode_DATA_ANY(unicode);
char_size = PyUnicode_KIND(unicode);
share_utf8 = _PyUnicode_SHARE_UTF8(unicode);
if (length > (PY_SSIZE_T_MAX / char_size - 1)) {
PyErr_NoMemory();
return -1;
}
new_size = (length + 1) * char_size;
if (!share_utf8 && _PyUnicode_HAS_UTF8_MEMORY(unicode))
{
PyObject_Free(_PyUnicode_UTF8(unicode));
_PyUnicode_UTF8(unicode) = NULL;
_PyUnicode_UTF8_LENGTH(unicode) = 0;
}
data = (PyObject *)PyObject_Realloc(data, new_size);
if (data == NULL) {
PyErr_NoMemory();
return -1;
}
_PyUnicode_DATA_ANY(unicode) = data;
if (share_utf8) {
_PyUnicode_UTF8(unicode) = data;
_PyUnicode_UTF8_LENGTH(unicode) = length;
}
_PyUnicode_LENGTH(unicode) = length;
PyUnicode_WRITE(PyUnicode_KIND(unicode), data, length, 0);
#ifdef Py_DEBUG
unicode_fill_invalid(unicode, old_length);
#endif
/* check for integer overflow */
if (length > PY_SSIZE_T_MAX / (Py_ssize_t)sizeof(wchar_t) - 1) {
PyErr_NoMemory();
return -1;
}
assert(_PyUnicode_CheckConsistency(unicode, 0));
return 0;
}
static PyObject*
resize_copy(PyObject *unicode, Py_ssize_t length)
{
Py_ssize_t copy_length;
PyObject *copy;
copy = PyUnicode_New(length, PyUnicode_MAX_CHAR_VALUE(unicode));
if (copy == NULL)
return NULL;
copy_length = Py_MIN(length, PyUnicode_GET_LENGTH(unicode));
_PyUnicode_FastCopyCharacters(copy, 0, unicode, 0, copy_length);
return copy;
}
static const char*
unicode_kind_name(PyObject *unicode)
{
/* don't check consistency: unicode_kind_name() is called from
_PyUnicode_Dump() */
if (!PyUnicode_IS_COMPACT(unicode))
{
switch (PyUnicode_KIND(unicode))
{
case PyUnicode_1BYTE_KIND:
if (PyUnicode_IS_ASCII(unicode))
return "legacy ascii";
else
return "legacy latin1";
case PyUnicode_2BYTE_KIND:
return "legacy UCS2";
case PyUnicode_4BYTE_KIND:
return "legacy UCS4";
default:
return "<legacy invalid kind>";
}
}
switch (PyUnicode_KIND(unicode)) {
case PyUnicode_1BYTE_KIND:
if (PyUnicode_IS_ASCII(unicode))
return "ascii";
else
return "latin1";
case PyUnicode_2BYTE_KIND:
return "UCS2";
case PyUnicode_4BYTE_KIND:
return "UCS4";
default:
return "<invalid compact kind>";
}
}
#ifdef Py_DEBUG
/* Functions wrapping macros for use in debugger */
const char *_PyUnicode_utf8(void *unicode_raw){
PyObject *unicode = _PyObject_CAST(unicode_raw);
return PyUnicode_UTF8(unicode);
}
const void *_PyUnicode_compact_data(void *unicode_raw) {
PyObject *unicode = _PyObject_CAST(unicode_raw);
return _PyUnicode_COMPACT_DATA(unicode);
}
const void *_PyUnicode_data(void *unicode_raw) {
PyObject *unicode = _PyObject_CAST(unicode_raw);
printf("obj %p\n", (void*)unicode);
printf("compact %d\n", PyUnicode_IS_COMPACT(unicode));
printf("compact ascii %d\n", PyUnicode_IS_COMPACT_ASCII(unicode));
printf("ascii op %p\n", (void*)(_PyASCIIObject_CAST(unicode) + 1));
printf("compact op %p\n", (void*)(_PyCompactUnicodeObject_CAST(unicode) + 1));
printf("compact data %p\n", _PyUnicode_COMPACT_DATA(unicode));
return PyUnicode_DATA(unicode);
}
void
_PyUnicode_Dump(PyObject *op)
{
PyASCIIObject *ascii = _PyASCIIObject_CAST(op);
PyCompactUnicodeObject *compact = _PyCompactUnicodeObject_CAST(op);
PyUnicodeObject *unicode = _PyUnicodeObject_CAST(op);
const void *data;
if (ascii->state.compact)
{
if (ascii->state.ascii)
data = (ascii + 1);
else
data = (compact + 1);
}
else
data = unicode->data.any;
printf("%s: len=%zu, ", unicode_kind_name(op), ascii->length);
if (!ascii->state.ascii) {
printf("utf8=%p (%zu)", (void *)compact->utf8, compact->utf8_length);
}
printf(", data=%p\n", data);
}
#endif
PyObject *
PyUnicode_New(Py_ssize_t size, Py_UCS4 maxchar)
{
/* Optimization for empty strings */
if (size == 0) {
return unicode_new_empty();
}
PyObject *obj;
PyCompactUnicodeObject *unicode;
void *data;
int kind;
int is_ascii;
Py_ssize_t char_size;
Py_ssize_t struct_size;
is_ascii = 0;
struct_size = sizeof(PyCompactUnicodeObject);
if (maxchar < 128) {
kind = PyUnicode_1BYTE_KIND;
char_size = 1;
is_ascii = 1;
struct_size = sizeof(PyASCIIObject);
}
else if (maxchar < 256) {
kind = PyUnicode_1BYTE_KIND;
char_size = 1;
}
else if (maxchar < 65536) {
kind = PyUnicode_2BYTE_KIND;
char_size = 2;
}
else {
if (maxchar > MAX_UNICODE) {
PyErr_SetString(PyExc_SystemError,
"invalid maximum character passed to PyUnicode_New");
return NULL;
}
kind = PyUnicode_4BYTE_KIND;
char_size = 4;
}
/* Ensure we won't overflow the size. */
if (size < 0) {
PyErr_SetString(PyExc_SystemError,
"Negative size passed to PyUnicode_New");
return NULL;
}
if (size > ((PY_SSIZE_T_MAX - struct_size) / char_size - 1))
return PyErr_NoMemory();
/* Duplicated allocation code from _PyObject_New() instead of a call to
* PyObject_New() so we are able to allocate space for the object and
* it's data buffer.
*/
obj = (PyObject *) PyObject_Malloc(struct_size + (size + 1) * char_size);
if (obj == NULL) {
return PyErr_NoMemory();
}
_PyObject_Init(obj, &PyUnicode_Type);
unicode = (PyCompactUnicodeObject *)obj;
if (is_ascii)
data = ((PyASCIIObject*)obj) + 1;
else
data = unicode + 1;
_PyUnicode_LENGTH(unicode) = size;
_PyUnicode_HASH(unicode) = -1;
_PyUnicode_STATE(unicode).interned = 0;
_PyUnicode_STATE(unicode).kind = kind;
_PyUnicode_STATE(unicode).compact = 1;
_PyUnicode_STATE(unicode).ascii = is_ascii;
if (is_ascii) {
((char*)data)[size] = 0;
}
else if (kind == PyUnicode_1BYTE_KIND) {
((char*)data)[size] = 0;
unicode->utf8 = NULL;
unicode->utf8_length = 0;
}
else {
unicode->utf8 = NULL;
unicode->utf8_length = 0;
if (kind == PyUnicode_2BYTE_KIND)
((Py_UCS2*)data)[size] = 0;
else /* kind == PyUnicode_4BYTE_KIND */
((Py_UCS4*)data)[size] = 0;
}
#ifdef Py_DEBUG
unicode_fill_invalid((PyObject*)unicode, 0);
#endif
assert(_PyUnicode_CheckConsistency((PyObject*)unicode, 0));
return obj;
}
#if SIZEOF_WCHAR_T == 2
/* Helper function to convert a 16-bits wchar_t representation to UCS4, this
will decode surrogate pairs, the other conversions are implemented as macros
for efficiency.
This function assumes that unicode can hold one more code point than wstr
characters for a terminating null character. */
static void
unicode_convert_wchar_to_ucs4(const wchar_t *begin, const wchar_t *end,
PyObject *unicode)
{
const wchar_t *iter;
Py_UCS4 *ucs4_out;
assert(unicode != NULL);
assert(_PyUnicode_CHECK(unicode));
assert(_PyUnicode_KIND(unicode) == PyUnicode_4BYTE_KIND);
ucs4_out = PyUnicode_4BYTE_DATA(unicode);
for (iter = begin; iter < end; ) {
assert(ucs4_out < (PyUnicode_4BYTE_DATA(unicode) +
_PyUnicode_GET_LENGTH(unicode)));
if (Py_UNICODE_IS_HIGH_SURROGATE(iter[0])
&& (iter+1) < end
&& Py_UNICODE_IS_LOW_SURROGATE(iter[1]))
{
*ucs4_out++ = Py_UNICODE_JOIN_SURROGATES(iter[0], iter[1]);
iter += 2;
}
else {
*ucs4_out++ = *iter;
iter++;
}
}
assert(ucs4_out == (PyUnicode_4BYTE_DATA(unicode) +
_PyUnicode_GET_LENGTH(unicode)));
}
#endif
static int
unicode_check_modifiable(PyObject *unicode)
{
if (!unicode_modifiable(unicode)) {
PyErr_SetString(PyExc_SystemError,
"Cannot modify a string currently used");
return -1;
}
return 0;
}
static int
_copy_characters(PyObject *to, Py_ssize_t to_start,
PyObject *from, Py_ssize_t from_start,
Py_ssize_t how_many, int check_maxchar)
{
int from_kind, to_kind;
const void *from_data;
void *to_data;
assert(0 <= how_many);
assert(0 <= from_start);
assert(0 <= to_start);
assert(PyUnicode_Check(from));
assert(from_start + how_many <= PyUnicode_GET_LENGTH(from));
assert(PyUnicode_Check(to));
assert(to_start + how_many <= PyUnicode_GET_LENGTH(to));
if (how_many == 0)
return 0;
from_kind = PyUnicode_KIND(from);
from_data = PyUnicode_DATA(from);
to_kind = PyUnicode_KIND(to);
to_data = PyUnicode_DATA(to);
#ifdef Py_DEBUG
if (!check_maxchar
&& PyUnicode_MAX_CHAR_VALUE(from) > PyUnicode_MAX_CHAR_VALUE(to))
{
Py_UCS4 to_maxchar = PyUnicode_MAX_CHAR_VALUE(to);
Py_UCS4 ch;
Py_ssize_t i;
for (i=0; i < how_many; i++) {
ch = PyUnicode_READ(from_kind, from_data, from_start + i);
assert(ch <= to_maxchar);
}
}
#endif
if (from_kind == to_kind) {
if (check_maxchar
&& !PyUnicode_IS_ASCII(from) && PyUnicode_IS_ASCII(to))
{
/* Writing Latin-1 characters into an ASCII string requires to
check that all written characters are pure ASCII */
Py_UCS4 max_char;
max_char = ucs1lib_find_max_char(from_data,
(const Py_UCS1*)from_data + how_many);
if (max_char >= 128)
return -1;
}
memcpy((char*)to_data + to_kind * to_start,
(const char*)from_data + from_kind * from_start,
to_kind * how_many);
}
else if (from_kind == PyUnicode_1BYTE_KIND
&& to_kind == PyUnicode_2BYTE_KIND)
{
_PyUnicode_CONVERT_BYTES(
Py_UCS1, Py_UCS2,
PyUnicode_1BYTE_DATA(from) + from_start,
PyUnicode_1BYTE_DATA(from) + from_start + how_many,
PyUnicode_2BYTE_DATA(to) + to_start
);
}
else if (from_kind == PyUnicode_1BYTE_KIND
&& to_kind == PyUnicode_4BYTE_KIND)
{
_PyUnicode_CONVERT_BYTES(
Py_UCS1, Py_UCS4,
PyUnicode_1BYTE_DATA(from) + from_start,
PyUnicode_1BYTE_DATA(from) + from_start + how_many,
PyUnicode_4BYTE_DATA(to) + to_start
);
}
else if (from_kind == PyUnicode_2BYTE_KIND
&& to_kind == PyUnicode_4BYTE_KIND)
{
_PyUnicode_CONVERT_BYTES(
Py_UCS2, Py_UCS4,
PyUnicode_2BYTE_DATA(from) + from_start,
PyUnicode_2BYTE_DATA(from) + from_start + how_many,
PyUnicode_4BYTE_DATA(to) + to_start
);
}
else {
assert (PyUnicode_MAX_CHAR_VALUE(from) > PyUnicode_MAX_CHAR_VALUE(to));
if (!check_maxchar) {
if (from_kind == PyUnicode_2BYTE_KIND
&& to_kind == PyUnicode_1BYTE_KIND)
{
_PyUnicode_CONVERT_BYTES(
Py_UCS2, Py_UCS1,
PyUnicode_2BYTE_DATA(from) + from_start,
PyUnicode_2BYTE_DATA(from) + from_start + how_many,
PyUnicode_1BYTE_DATA(to) + to_start
);
}
else if (from_kind == PyUnicode_4BYTE_KIND
&& to_kind == PyUnicode_1BYTE_KIND)
{
_PyUnicode_CONVERT_BYTES(
Py_UCS4, Py_UCS1,
PyUnicode_4BYTE_DATA(from) + from_start,
PyUnicode_4BYTE_DATA(from) + from_start + how_many,
PyUnicode_1BYTE_DATA(to) + to_start
);
}
else if (from_kind == PyUnicode_4BYTE_KIND
&& to_kind == PyUnicode_2BYTE_KIND)
{
_PyUnicode_CONVERT_BYTES(
Py_UCS4, Py_UCS2,
PyUnicode_4BYTE_DATA(from) + from_start,
PyUnicode_4BYTE_DATA(from) + from_start + how_many,
PyUnicode_2BYTE_DATA(to) + to_start
);
}
else {
Py_UNREACHABLE();
}
}
else {
const Py_UCS4 to_maxchar = PyUnicode_MAX_CHAR_VALUE(to);
Py_UCS4 ch;
Py_ssize_t i;
for (i=0; i < how_many; i++) {
ch = PyUnicode_READ(from_kind, from_data, from_start + i);
if (ch > to_maxchar)
return -1;
PyUnicode_WRITE(to_kind, to_data, to_start + i, ch);
}
}
}
return 0;
}
void
_PyUnicode_FastCopyCharacters(
PyObject *to, Py_ssize_t to_start,
PyObject *from, Py_ssize_t from_start, Py_ssize_t how_many)
{
(void)_copy_characters(to, to_start, from, from_start, how_many, 0);
}
Py_ssize_t
PyUnicode_CopyCharacters(PyObject *to, Py_ssize_t to_start,
PyObject *from, Py_ssize_t from_start,
Py_ssize_t how_many)
{
int err;
if (!PyUnicode_Check(from) || !PyUnicode_Check(to)) {
PyErr_BadInternalCall();
return -1;
}
if ((size_t)from_start > (size_t)PyUnicode_GET_LENGTH(from)) {
PyErr_SetString(PyExc_IndexError, "string index out of range");
return -1;
}
if ((size_t)to_start > (size_t)PyUnicode_GET_LENGTH(to)) {
PyErr_SetString(PyExc_IndexError, "string index out of range");
return -1;
}
if (how_many < 0) {
PyErr_SetString(PyExc_SystemError, "how_many cannot be negative");
return -1;
}
how_many = Py_MIN(PyUnicode_GET_LENGTH(from)-from_start, how_many);
if (to_start + how_many > PyUnicode_GET_LENGTH(to)) {
PyErr_Format(PyExc_SystemError,
"Cannot write %zi characters at %zi "
"in a string of %zi characters",
how_many, to_start, PyUnicode_GET_LENGTH(to));
return -1;
}
if (how_many == 0)
return 0;
if (unicode_check_modifiable(to))
return -1;
err = _copy_characters(to, to_start, from, from_start, how_many, 1);
if (err) {
PyErr_Format(PyExc_SystemError,
"Cannot copy %s characters "
"into a string of %s characters",
unicode_kind_name(from),
unicode_kind_name(to));
return -1;
}
return how_many;
}
/* Find the maximum code point and count the number of surrogate pairs so a
correct string length can be computed before converting a string to UCS4.
This function counts single surrogates as a character and not as a pair.
Return 0 on success, or -1 on error. */
static int
find_maxchar_surrogates(const wchar_t *begin, const wchar_t *end,
Py_UCS4 *maxchar, Py_ssize_t *num_surrogates)
{
const wchar_t *iter;
Py_UCS4 ch;
assert(num_surrogates != NULL && maxchar != NULL);
*num_surrogates = 0;
*maxchar = 0;
for (iter = begin; iter < end; ) {
#if SIZEOF_WCHAR_T == 2
if (Py_UNICODE_IS_HIGH_SURROGATE(iter[0])
&& (iter+1) < end
&& Py_UNICODE_IS_LOW_SURROGATE(iter[1]))
{
ch = Py_UNICODE_JOIN_SURROGATES(iter[0], iter[1]);
++(*num_surrogates);
iter += 2;
}
else
#endif
{
ch = *iter;
iter++;
}
if (ch > *maxchar) {
*maxchar = ch;
if (*maxchar > MAX_UNICODE) {
PyErr_Format(PyExc_ValueError,
"character U+%x is not in range [U+0000; U+%x]",
ch, MAX_UNICODE);
return -1;
}
}
}
return 0;
}
static void
unicode_dealloc(PyObject *unicode)
{
#ifdef Py_DEBUG
if (!unicode_is_finalizing() && unicode_is_singleton(unicode)) {
_Py_FatalRefcountError("deallocating an Unicode singleton");
}
#endif
/* This should never get called, but we also don't want to SEGV if
* we accidentally decref an immortal string out of existence. Since
* the string is an immortal object, just re-set the reference count.
*/
if (PyUnicode_CHECK_INTERNED(unicode)) {
_Py_SetImmortal(unicode);
return;
}
if (_PyUnicode_HAS_UTF8_MEMORY(unicode)) {
PyObject_Free(_PyUnicode_UTF8(unicode));
}
if (!PyUnicode_IS_COMPACT(unicode) && _PyUnicode_DATA_ANY(unicode)) {
PyObject_Free(_PyUnicode_DATA_ANY(unicode));
}
Py_TYPE(unicode)->tp_free(unicode);
}
#ifdef Py_DEBUG
static int
unicode_is_singleton(PyObject *unicode)
{
if (unicode == &_Py_STR(empty)) {
return 1;
}
PyASCIIObject *ascii = _PyASCIIObject_CAST(unicode);
if (ascii->length == 1) {
Py_UCS4 ch = PyUnicode_READ_CHAR(unicode, 0);
if (ch < 256 && LATIN1(ch) == unicode) {
return 1;
}
}
return 0;
}
#endif
static int
unicode_modifiable(PyObject *unicode)
{
assert(_PyUnicode_CHECK(unicode));
if (Py_REFCNT(unicode) != 1)
return 0;
if (_PyUnicode_HASH(unicode) != -1)
return 0;
if (PyUnicode_CHECK_INTERNED(unicode))
return 0;
if (!PyUnicode_CheckExact(unicode))
return 0;
#ifdef Py_DEBUG
/* singleton refcount is greater than 1 */
assert(!unicode_is_singleton(unicode));
#endif
return 1;
}
static int
unicode_resize(PyObject **p_unicode, Py_ssize_t length)
{
PyObject *unicode;
Py_ssize_t old_length;
assert(p_unicode != NULL);
unicode = *p_unicode;
assert(unicode != NULL);
assert(PyUnicode_Check(unicode));
assert(0 <= length);
old_length = PyUnicode_GET_LENGTH(unicode);
if (old_length == length)
return 0;
if (length == 0) {
PyObject *empty = unicode_new_empty();
Py_SETREF(*p_unicode, empty);
return 0;
}
if (!unicode_modifiable(unicode)) {
PyObject *copy = resize_copy(unicode, length);
if (copy == NULL)
return -1;
Py_SETREF(*p_unicode, copy);
return 0;
}
if (PyUnicode_IS_COMPACT(unicode)) {
PyObject *new_unicode = resize_compact(unicode, length);
if (new_unicode == NULL)
return -1;
*p_unicode = new_unicode;
return 0;
}
return resize_inplace(unicode, length);
}
int
PyUnicode_Resize(PyObject **p_unicode, Py_ssize_t length)
{
PyObject *unicode;
if (p_unicode == NULL) {
PyErr_BadInternalCall();
return -1;
}
unicode = *p_unicode;
if (unicode == NULL || !PyUnicode_Check(unicode) || length < 0)
{
PyErr_BadInternalCall();
return -1;
}
return unicode_resize(p_unicode, length);
}
/* Copy an ASCII or latin1 char* string into a Python Unicode string.
WARNING: The function doesn't copy the terminating null character and
doesn't check the maximum character (may write a latin1 character in an
ASCII string). */
static void
unicode_write_cstr(PyObject *unicode, Py_ssize_t index,
const char *str, Py_ssize_t len)
{
int kind = PyUnicode_KIND(unicode);
const void *data = PyUnicode_DATA(unicode);
const char *end = str + len;
assert(index + len <= PyUnicode_GET_LENGTH(unicode));
switch (kind) {
case PyUnicode_1BYTE_KIND: {
#ifdef Py_DEBUG
if (PyUnicode_IS_ASCII(unicode)) {
Py_UCS4 maxchar = ucs1lib_find_max_char(
(const Py_UCS1*)str,
(const Py_UCS1*)str + len);
assert(maxchar < 128);
}
#endif
memcpy((char *) data + index, str, len);
break;
}
case PyUnicode_2BYTE_KIND: {
Py_UCS2 *start = (Py_UCS2 *)data + index;
Py_UCS2 *ucs2 = start;
for (; str < end; ++ucs2, ++str)
*ucs2 = (Py_UCS2)*str;
assert((ucs2 - start) <= PyUnicode_GET_LENGTH(unicode));
break;
}
case PyUnicode_4BYTE_KIND: {
Py_UCS4 *start = (Py_UCS4 *)data + index;
Py_UCS4 *ucs4 = start;
for (; str < end; ++ucs4, ++str)
*ucs4 = (Py_UCS4)*str;
assert((ucs4 - start) <= PyUnicode_GET_LENGTH(unicode));
break;
}
default:
Py_UNREACHABLE();
}
}
static PyObject*
get_latin1_char(Py_UCS1 ch)
{
return Py_NewRef(LATIN1(ch));
}
static PyObject*
unicode_char(Py_UCS4 ch)
{
PyObject *unicode;
assert(ch <= MAX_UNICODE);
if (ch < 256) {
return get_latin1_char(ch);
}
unicode = PyUnicode_New(1, ch);
if (unicode == NULL)
return NULL;
assert(PyUnicode_KIND(unicode) != PyUnicode_1BYTE_KIND);
if (PyUnicode_KIND(unicode) == PyUnicode_2BYTE_KIND) {
PyUnicode_2BYTE_DATA(unicode)[0] = (Py_UCS2)ch;
} else {
assert(PyUnicode_KIND(unicode) == PyUnicode_4BYTE_KIND);
PyUnicode_4BYTE_DATA(unicode)[0] = ch;
}
assert(_PyUnicode_CheckConsistency(unicode, 1));
return unicode;
}
PyObject *
PyUnicode_FromWideChar(const wchar_t *u, Py_ssize_t size)
{
PyObject *unicode;
Py_UCS4 maxchar = 0;
Py_ssize_t num_surrogates;
if (u == NULL && size != 0) {
PyErr_BadInternalCall();
return NULL;
}
if (size == -1) {
size = wcslen(u);
}
/* If the Unicode data is known at construction time, we can apply
some optimizations which share commonly used objects. */
/* Optimization for empty strings */
if (size == 0)
_Py_RETURN_UNICODE_EMPTY();
#ifdef HAVE_NON_UNICODE_WCHAR_T_REPRESENTATION
/* Oracle Solaris uses non-Unicode internal wchar_t form for
non-Unicode locales and hence needs conversion to UCS-4 first. */
if (_Py_LocaleUsesNonUnicodeWchar()) {
wchar_t* converted = _Py_DecodeNonUnicodeWchar(u, size);
if (!converted) {
return NULL;
}
PyObject *unicode = _PyUnicode_FromUCS4(converted, size);
PyMem_Free(converted);
return unicode;
}
#endif
/* Single character Unicode objects in the Latin-1 range are
shared when using this constructor */
if (size == 1 && (Py_UCS4)*u < 256)
return get_latin1_char((unsigned char)*u);
/* If not empty and not single character, copy the Unicode data
into the new object */
if (find_maxchar_surrogates(u, u + size,
&maxchar, &num_surrogates) == -1)
return NULL;
unicode = PyUnicode_New(size - num_surrogates, maxchar);
if (!unicode)
return NULL;
switch (PyUnicode_KIND(unicode)) {
case PyUnicode_1BYTE_KIND:
_PyUnicode_CONVERT_BYTES(Py_UNICODE, unsigned char,
u, u + size, PyUnicode_1BYTE_DATA(unicode));
break;
case PyUnicode_2BYTE_KIND:
#if Py_UNICODE_SIZE == 2
memcpy(PyUnicode_2BYTE_DATA(unicode), u, size * 2);
#else
_PyUnicode_CONVERT_BYTES(Py_UNICODE, Py_UCS2,
u, u + size, PyUnicode_2BYTE_DATA(unicode));
#endif
break;
case PyUnicode_4BYTE_KIND:
#if SIZEOF_WCHAR_T == 2
/* This is the only case which has to process surrogates, thus
a simple copy loop is not enough and we need a function. */
unicode_convert_wchar_to_ucs4(u, u + size, unicode);
#else
assert(num_surrogates == 0);
memcpy(PyUnicode_4BYTE_DATA(unicode), u, size * 4);
#endif
break;
default:
Py_UNREACHABLE();
}
return unicode_result(unicode);
}
PyObject *
PyUnicode_FromStringAndSize(const char *u, Py_ssize_t size)
{
if (size < 0) {
PyErr_SetString(PyExc_SystemError,
"Negative size passed to PyUnicode_FromStringAndSize");
return NULL;
}
if (u != NULL) {
return PyUnicode_DecodeUTF8Stateful(u, size, NULL, NULL);
}
if (size > 0) {
PyErr_SetString(PyExc_SystemError,
"NULL string with positive size with NULL passed to PyUnicode_FromStringAndSize");
return NULL;
}
return unicode_new_empty();
}
PyObject *
PyUnicode_FromString(const char *u)
{
size_t size = strlen(u);
if (size > PY_SSIZE_T_MAX) {
PyErr_SetString(PyExc_OverflowError, "input too long");
return NULL;
}
return PyUnicode_DecodeUTF8Stateful(u, (Py_ssize_t)size, NULL, NULL);
}
PyObject *
_PyUnicode_FromId(_Py_Identifier *id)
{
PyInterpreterState *interp = _PyInterpreterState_GET();
struct _Py_unicode_ids *ids = &interp->unicode.ids;
Py_ssize_t index = _Py_atomic_size_get(&id->index);
if (index < 0) {
struct _Py_unicode_runtime_ids *rt_ids = &interp->runtime->unicode_state.ids;
PyThread_acquire_lock(rt_ids->lock, WAIT_LOCK);
// Check again to detect concurrent access. Another thread can have
// initialized the index while this thread waited for the lock.
index = _Py_atomic_size_get(&id->index);
if (index < 0) {
assert(rt_ids->next_index < PY_SSIZE_T_MAX);
index = rt_ids->next_index;
rt_ids->next_index++;
_Py_atomic_size_set(&id->index, index);
}
PyThread_release_lock(rt_ids->lock);
}
assert(index >= 0);
PyObject *obj;
if (index < ids->size) {
obj = ids->array[index];
if (obj) {
// Return a borrowed reference
return obj;
}
}
obj = PyUnicode_DecodeUTF8Stateful(id->string, strlen(id->string),
NULL, NULL);
if (!obj) {
return NULL;
}
PyUnicode_InternInPlace(&obj);
if (index >= ids->size) {
// Overallocate to reduce the number of realloc
Py_ssize_t new_size = Py_MAX(index * 2, 16);
Py_ssize_t item_size = sizeof(ids->array[0]);
PyObject **new_array = PyMem_Realloc(ids->array, new_size * item_size);
if (new_array == NULL) {
PyErr_NoMemory();
return NULL;
}
memset(&new_array[ids->size], 0, (new_size - ids->size) * item_size);
ids->array = new_array;
ids->size = new_size;
}
// The array stores a strong reference
ids->array[index] = obj;
// Return a borrowed reference
return obj;
}
static void
unicode_clear_identifiers(struct _Py_unicode_state *state)
{
struct _Py_unicode_ids *ids = &state->ids;
for (Py_ssize_t i=0; i < ids->size; i++) {
Py_XDECREF(ids->array[i]);
}
ids->size = 0;
PyMem_Free(ids->array);
ids->array = NULL;
// Don't reset _PyRuntime next_index: _Py_Identifier.id remains valid
// after Py_Finalize().
}
/* Internal function, doesn't check maximum character */
PyObject*
_PyUnicode_FromASCII(const char *buffer, Py_ssize_t size)
{
const unsigned char *s = (const unsigned char *)buffer;
PyObject *unicode;
if (size == 1) {
#ifdef Py_DEBUG
assert((unsigned char)s[0] < 128);
#endif
return get_latin1_char(s[0]);
}
unicode = PyUnicode_New(size, 127);
if (!unicode)
return NULL;
memcpy(PyUnicode_1BYTE_DATA(unicode), s, size);
assert(_PyUnicode_CheckConsistency(unicode, 1));
return unicode;
}
static Py_UCS4
kind_maxchar_limit(int kind)
{
switch (kind) {
case PyUnicode_1BYTE_KIND:
return 0x80;
case PyUnicode_2BYTE_KIND:
return 0x100;
case PyUnicode_4BYTE_KIND:
return 0x10000;
default:
Py_UNREACHABLE();
}
}
static PyObject*
_PyUnicode_FromUCS1(const Py_UCS1* u, Py_ssize_t size)
{
PyObject *res;
unsigned char max_char;
if (size == 0) {
_Py_RETURN_UNICODE_EMPTY();
}
assert(size > 0);
if (size == 1) {
return get_latin1_char(u[0]);
}
max_char = ucs1lib_find_max_char(u, u + size);
res = PyUnicode_New(size, max_char);
if (!res)
return NULL;
memcpy(PyUnicode_1BYTE_DATA(res), u, size);
assert(_PyUnicode_CheckConsistency(res, 1));
return res;
}
static PyObject*
_PyUnicode_FromUCS2(const Py_UCS2 *u, Py_ssize_t size)
{
PyObject *res;
Py_UCS2 max_char;
if (size == 0)
_Py_RETURN_UNICODE_EMPTY();
assert(size > 0);
if (size == 1)
return unicode_char(u[0]);
max_char = ucs2lib_find_max_char(u, u + size);
res = PyUnicode_New(size, max_char);
if (!res)
return NULL;
if (max_char >= 256)
memcpy(PyUnicode_2BYTE_DATA(res), u, sizeof(Py_UCS2)*size);
else {
_PyUnicode_CONVERT_BYTES(
Py_UCS2, Py_UCS1, u, u + size, PyUnicode_1BYTE_DATA(res));
}
assert(_PyUnicode_CheckConsistency(res, 1));
return res;
}
static PyObject*
_PyUnicode_FromUCS4(const Py_UCS4 *u, Py_ssize_t size)
{
PyObject *res;
Py_UCS4 max_char;
if (size == 0)
_Py_RETURN_UNICODE_EMPTY();
assert(size > 0);
if (size == 1)
return unicode_char(u[0]);
max_char = ucs4lib_find_max_char(u, u + size);
res = PyUnicode_New(size, max_char);
if (!res)
return NULL;
if (max_char < 256)
_PyUnicode_CONVERT_BYTES(Py_UCS4, Py_UCS1, u, u + size,
PyUnicode_1BYTE_DATA(res));
else if (max_char < 0x10000)
_PyUnicode_CONVERT_BYTES(Py_UCS4, Py_UCS2, u, u + size,
PyUnicode_2BYTE_DATA(res));
else
memcpy(PyUnicode_4BYTE_DATA(res), u, sizeof(Py_UCS4)*size);
assert(_PyUnicode_CheckConsistency(res, 1));
return res;
}
PyObject*
PyUnicode_FromKindAndData(int kind, const void *buffer, Py_ssize_t size)
{
if (size < 0) {
PyErr_SetString(PyExc_ValueError, "size must be positive");
return NULL;
}
switch (kind) {
case PyUnicode_1BYTE_KIND:
return _PyUnicode_FromUCS1(buffer, size);
case PyUnicode_2BYTE_KIND:
return _PyUnicode_FromUCS2(buffer, size);
case PyUnicode_4BYTE_KIND:
return _PyUnicode_FromUCS4(buffer, size);
default:
PyErr_SetString(PyExc_SystemError, "invalid kind");
return NULL;
}
}
Py_UCS4
_PyUnicode_FindMaxChar(PyObject *unicode, Py_ssize_t start, Py_ssize_t end)
{
int kind;
const void *startptr, *endptr;
assert(0 <= start);
assert(end <= PyUnicode_GET_LENGTH(unicode));
assert(start <= end);
if (start == 0 && end == PyUnicode_GET_LENGTH(unicode))
return PyUnicode_MAX_CHAR_VALUE(unicode);
if (start == end)
return 127;
if (PyUnicode_IS_ASCII(unicode))
return 127;
kind = PyUnicode_KIND(unicode);
startptr = PyUnicode_DATA(unicode);
endptr = (char *)startptr + end * kind;
startptr = (char *)startptr + start * kind;
switch(kind) {
case PyUnicode_1BYTE_KIND:
return ucs1lib_find_max_char(startptr, endptr);
case PyUnicode_2BYTE_KIND:
return ucs2lib_find_max_char(startptr, endptr);
case PyUnicode_4BYTE_KIND:
return ucs4lib_find_max_char(startptr, endptr);
default:
Py_UNREACHABLE();
}
}
/* Ensure that a string uses the most efficient storage, if it is not the
case: create a new string with of the right kind. Write NULL into *p_unicode
on error. */
static void
unicode_adjust_maxchar(PyObject **p_unicode)
{
PyObject *unicode, *copy;
Py_UCS4 max_char;
Py_ssize_t len;
int kind;
assert(p_unicode != NULL);
unicode = *p_unicode;
if (PyUnicode_IS_ASCII(unicode))
return;
len = PyUnicode_GET_LENGTH(unicode);
kind = PyUnicode_KIND(unicode);
if (kind == PyUnicode_1BYTE_KIND) {
const Py_UCS1 *u = PyUnicode_1BYTE_DATA(unicode);
max_char = ucs1lib_find_max_char(u, u + len);
if (max_char >= 128)
return;
}
else if (kind == PyUnicode_2BYTE_KIND) {
const Py_UCS2 *u = PyUnicode_2BYTE_DATA(unicode);
max_char = ucs2lib_find_max_char(u, u + len);
if (max_char >= 256)
return;
}
else if (kind == PyUnicode_4BYTE_KIND) {
const Py_UCS4 *u = PyUnicode_4BYTE_DATA(unicode);
max_char = ucs4lib_find_max_char(u, u + len);
if (max_char >= 0x10000)
return;
}
else
Py_UNREACHABLE();
copy = PyUnicode_New(len, max_char);
if (copy != NULL)
_PyUnicode_FastCopyCharacters(copy, 0, unicode, 0, len);
Py_DECREF(unicode);
*p_unicode = copy;
}
PyObject*
_PyUnicode_Copy(PyObject *unicode)
{
Py_ssize_t length;
PyObject *copy;
if (!PyUnicode_Check(unicode)) {
PyErr_BadInternalCall();
return NULL;
}
length = PyUnicode_GET_LENGTH(unicode);
copy = PyUnicode_New(length, PyUnicode_MAX_CHAR_VALUE(unicode));
if (!copy)
return NULL;
assert(PyUnicode_KIND(copy) == PyUnicode_KIND(unicode));
memcpy(PyUnicode_DATA(copy), PyUnicode_DATA(unicode),
length * PyUnicode_KIND(unicode));
assert(_PyUnicode_CheckConsistency(copy, 1));
return copy;
}
/* Widen Unicode objects to larger buffers. Don't write terminating null
character. Return NULL on error. */
static void*
unicode_askind(int skind, void const *data, Py_ssize_t len, int kind)
{
void *result;
assert(skind < kind);
switch (kind) {
case PyUnicode_2BYTE_KIND:
result = PyMem_New(Py_UCS2, len);
if (!result)
return PyErr_NoMemory();
assert(skind == PyUnicode_1BYTE_KIND);
_PyUnicode_CONVERT_BYTES(
Py_UCS1, Py_UCS2,
(const Py_UCS1 *)data,
((const Py_UCS1 *)data) + len,
result);
return result;
case PyUnicode_4BYTE_KIND:
result = PyMem_New(Py_UCS4, len);
if (!result)
return PyErr_NoMemory();
if (skind == PyUnicode_2BYTE_KIND) {
_PyUnicode_CONVERT_BYTES(
Py_UCS2, Py_UCS4,
(const Py_UCS2 *)data,
((const Py_UCS2 *)data) + len,
result);
}
else {
assert(skind == PyUnicode_1BYTE_KIND);
_PyUnicode_CONVERT_BYTES(
Py_UCS1, Py_UCS4,
(const Py_UCS1 *)data,
((const Py_UCS1 *)data) + len,
result);
}
return result;
default:
Py_UNREACHABLE();
return NULL;
}
}
static Py_UCS4*
as_ucs4(PyObject *string, Py_UCS4 *target, Py_ssize_t targetsize,
int copy_null)
{
int kind;
const void *data;
Py_ssize_t len, targetlen;
kind = PyUnicode_KIND(string);
data = PyUnicode_DATA(string);
len = PyUnicode_GET_LENGTH(string);
targetlen = len;
if (copy_null)
targetlen++;
if (!target) {
target = PyMem_New(Py_UCS4, targetlen);
if (!target) {
PyErr_NoMemory();
return NULL;
}
}
else {
if (targetsize < targetlen) {
PyErr_Format(PyExc_SystemError,
"string is longer than the buffer");
if (copy_null && 0 < targetsize)
target[0] = 0;
return NULL;
}
}
if (kind == PyUnicode_1BYTE_KIND) {
const Py_UCS1 *start = (const Py_UCS1 *) data;
_PyUnicode_CONVERT_BYTES(Py_UCS1, Py_UCS4, start, start + len, target);
}
else if (kind == PyUnicode_2BYTE_KIND) {
const Py_UCS2 *start = (const Py_UCS2 *) data;
_PyUnicode_CONVERT_BYTES(Py_UCS2, Py_UCS4, start, start + len, target);
}
else if (kind == PyUnicode_4BYTE_KIND) {
memcpy(target, data, len * sizeof(Py_UCS4));
}
else {
Py_UNREACHABLE();
}
if (copy_null)
target[len] = 0;
return target;
}
Py_UCS4*
PyUnicode_AsUCS4(PyObject *string, Py_UCS4 *target, Py_ssize_t targetsize,
int copy_null)
{
if (target == NULL || targetsize < 0) {
PyErr_BadInternalCall();
return NULL;
}
return as_ucs4(string, target, targetsize, copy_null);
}
Py_UCS4*
PyUnicode_AsUCS4Copy(PyObject *string)
{
return as_ucs4(string, NULL, 0, 1);
}
/* maximum number of characters required for output of %jo or %jd or %p.
We need at most ceil(log8(256)*sizeof(intmax_t)) digits,
plus 1 for the sign, plus 2 for the 0x prefix (for %p),
plus 1 for the terminal NUL. */
#define MAX_INTMAX_CHARS (5 + (sizeof(intmax_t)*8-1) / 3)
static int
unicode_fromformat_write_str(_PyUnicodeWriter *writer, PyObject *str,
Py_ssize_t width, Py_ssize_t precision, int flags)
{
Py_ssize_t length, fill, arglen;
Py_UCS4 maxchar;
length = PyUnicode_GET_LENGTH(str);
if ((precision == -1 || precision >= length)
&& width <= length)
return _PyUnicodeWriter_WriteStr(writer, str);
if (precision != -1)
length = Py_MIN(precision, length);
arglen = Py_MAX(length, width);
if (PyUnicode_MAX_CHAR_VALUE(str) > writer->maxchar)
maxchar = _PyUnicode_FindMaxChar(str, 0, length);
else
maxchar = writer->maxchar;
if (_PyUnicodeWriter_Prepare(writer, arglen, maxchar) == -1)
return -1;
fill = Py_MAX(width - length, 0);
if (fill && !(flags & F_LJUST)) {
if (PyUnicode_Fill(writer->buffer, writer->pos, fill, ' ') == -1)
return -1;
writer->pos += fill;
}
_PyUnicode_FastCopyCharacters(writer->buffer, writer->pos,
str, 0, length);
writer->pos += length;
if (fill && (flags & F_LJUST)) {
if (PyUnicode_Fill(writer->buffer, writer->pos, fill, ' ') == -1)
return -1;
writer->pos += fill;
}
return 0;
}
static int
unicode_fromformat_write_cstr(_PyUnicodeWriter *writer, const char *str,
Py_ssize_t width, Py_ssize_t precision, int flags)
{
/* UTF-8 */
Py_ssize_t length;
PyObject *unicode;
int res;
if (precision == -1) {
length = strlen(str);
}
else {
length = 0;
while (length < precision && str[length]) {
length++;
}
}
unicode = PyUnicode_DecodeUTF8Stateful(str, length, "replace", NULL);
if (unicode == NULL)
return -1;
res = unicode_fromformat_write_str(writer, unicode, width, -1, flags);
Py_DECREF(unicode);
return res;
}
static int
unicode_fromformat_write_wcstr(_PyUnicodeWriter *writer, const wchar_t *str,
Py_ssize_t width, Py_ssize_t precision, int flags)
{
/* UTF-8 */
Py_ssize_t length;
PyObject *unicode;
int res;
if (precision == -1) {
length = wcslen(str);
}
else {
length = 0;
while (length < precision && str[length]) {
length++;
}
}
unicode = PyUnicode_FromWideChar(str, length);
if (unicode == NULL)
return -1;
res = unicode_fromformat_write_str(writer, unicode, width, -1, flags);
Py_DECREF(unicode);
return res;
}
#define F_LONG 1
#define F_LONGLONG 2
#define F_SIZE 3
#define F_PTRDIFF 4
#define F_INTMAX 5
static const char * const formats[] = {"%d", "%ld", "%lld", "%zd", "%td", "%jd"};
static const char * const formats_o[] = {"%o", "%lo", "%llo", "%zo", "%to", "%jo"};
static const char * const formats_u[] = {"%u", "%lu", "%llu", "%zu", "%tu", "%ju"};
static const char * const formats_x[] = {"%x", "%lx", "%llx", "%zx", "%tx", "%jx"};
static const char * const formats_X[] = {"%X", "%lX", "%llX", "%zX", "%tX", "%jX"};
static const char*
unicode_fromformat_arg(_PyUnicodeWriter *writer,
const char *f, va_list *vargs)
{
const char *p;
Py_ssize_t len;
int flags = 0;
Py_ssize_t width;
Py_ssize_t precision;
p = f;
f++;
if (*f == '%') {
if (_PyUnicodeWriter_WriteCharInline(writer, '%') < 0)
return NULL;
f++;
return f;
}
/* Parse flags. Example: "%-i" => flags=F_LJUST. */
/* Flags '+', ' ' and '#' are not particularly useful.
* They are not worth the implementation and maintenance costs.
* In addition, '#' should add "0" for "o" conversions for compatibility
* with printf, but it would confuse Python users. */
while (1) {
switch (*f++) {
case '-': flags |= F_LJUST; continue;
case '0': flags |= F_ZERO; continue;
}
f--;
break;
}
/* parse the width.precision part, e.g. "%2.5s" => width=2, precision=5 */
width = -1;
if (*f == '*') {
width = va_arg(*vargs, int);
if (width < 0) {
flags |= F_LJUST;
width = -width;
}
f++;
}
else if (Py_ISDIGIT((unsigned)*f)) {
width = *f - '0';
f++;
while (Py_ISDIGIT((unsigned)*f)) {
if (width > (PY_SSIZE_T_MAX - ((int)*f - '0')) / 10) {
PyErr_SetString(PyExc_ValueError,
"width too big");
return NULL;
}
width = (width * 10) + (*f - '0');
f++;
}
}
precision = -1;
if (*f == '.') {
f++;
if (*f == '*') {
precision = va_arg(*vargs, int);
if (precision < 0) {
precision = -2;
}
f++;
}
else if (Py_ISDIGIT((unsigned)*f)) {
precision = (*f - '0');
f++;
while (Py_ISDIGIT((unsigned)*f)) {
if (precision > (PY_SSIZE_T_MAX - ((int)*f - '0')) / 10) {
PyErr_SetString(PyExc_ValueError,
"precision too big");
return NULL;
}
precision = (precision * 10) + (*f - '0');
f++;
}
}
}
int sizemod = 0;
if (*f == 'l') {
if (f[1] == 'l') {
sizemod = F_LONGLONG;
f += 2;
}
else {
sizemod = F_LONG;
++f;
}
}
else if (*f == 'z') {
sizemod = F_SIZE;
++f;
}
else if (*f == 't') {
sizemod = F_PTRDIFF;
++f;
}
else if (*f == 'j') {
sizemod = F_INTMAX;
++f;
}
if (f[0] != '\0' && f[1] == '\0')
writer->overallocate = 0;
switch (*f) {
case 'd': case 'i': case 'o': case 'u': case 'x': case 'X':
break;
case 'c': case 'p':
if (sizemod || width >= 0 || precision >= 0) goto invalid_format;
break;
case 's':
case 'V':
if (sizemod && sizemod != F_LONG) goto invalid_format;
break;
default:
if (sizemod) goto invalid_format;
break;
}
switch (*f) {
case 'c':
{
int ordinal = va_arg(*vargs, int);
if (ordinal < 0 || ordinal > MAX_UNICODE) {
PyErr_SetString(PyExc_OverflowError,
"character argument not in range(0x110000)");
return NULL;
}
if (_PyUnicodeWriter_WriteCharInline(writer, ordinal) < 0)
return NULL;
break;
}
case 'd': case 'i':
case 'o': case 'u': case 'x': case 'X':
{
/* used by sprintf */
char buffer[MAX_INTMAX_CHARS];
const char *fmt = NULL;
switch (*f) {
case 'o': fmt = formats_o[sizemod]; break;
case 'u': fmt = formats_u[sizemod]; break;
case 'x': fmt = formats_x[sizemod]; break;
case 'X': fmt = formats_X[sizemod]; break;
default: fmt = formats[sizemod]; break;
}
int issigned = (*f == 'd' || *f == 'i');
switch (sizemod) {
case F_LONG:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, long)) :
sprintf(buffer, fmt, va_arg(*vargs, unsigned long));
break;
case F_LONGLONG:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, long long)) :
sprintf(buffer, fmt, va_arg(*vargs, unsigned long long));
break;
case F_SIZE:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, Py_ssize_t)) :
sprintf(buffer, fmt, va_arg(*vargs, size_t));
break;
case F_PTRDIFF:
len = sprintf(buffer, fmt, va_arg(*vargs, ptrdiff_t));
break;
case F_INTMAX:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, intmax_t)) :
sprintf(buffer, fmt, va_arg(*vargs, uintmax_t));
break;
default:
len = issigned ?
sprintf(buffer, fmt, va_arg(*vargs, int)) :
sprintf(buffer, fmt, va_arg(*vargs, unsigned int));
break;
}
assert(len >= 0);
int sign = (buffer[0] == '-');
len -= sign;
precision = Py_MAX(precision, len);
width = Py_MAX(width, precision + sign);
if ((flags & F_ZERO) && !(flags & F_LJUST)) {
precision = width - sign;
}
Py_ssize_t spacepad = Py_MAX(width - precision - sign, 0);
Py_ssize_t zeropad = Py_MAX(precision - len, 0);
if (_PyUnicodeWriter_Prepare(writer, width, 127) == -1)
return NULL;
if (spacepad && !(flags & F_LJUST)) {
if (PyUnicode_Fill(writer->buffer, writer->pos, spacepad, ' ') == -1)
return NULL;
writer->pos += spacepad;
}
if (sign) {
if (_PyUnicodeWriter_WriteChar(writer, '-') == -1)
return NULL;
}
if (zeropad) {
if (PyUnicode_Fill(writer->buffer, writer->pos, zeropad, '0') == -1)
return NULL;
writer->pos += zeropad;
}
if (_PyUnicodeWriter_WriteASCIIString(writer, &buffer[sign], len) < 0)
return NULL;
if (spacepad && (flags & F_LJUST)) {
if (PyUnicode_Fill(writer->buffer, writer->pos, spacepad, ' ') == -1)
return NULL;
writer->pos += spacepad;
}
break;
}
case 'p':
{
char number[MAX_INTMAX_CHARS];
len = sprintf(number, "%p", va_arg(*vargs, void*));
assert(len >= 0);
/* %p is ill-defined: ensure leading 0x. */
if (number[1] == 'X')
number[1] = 'x';
else if (number[1] != 'x') {
memmove(number + 2, number,
strlen(number) + 1);
number[0] = '0';
number[1] = 'x';
len += 2;
}
if (_PyUnicodeWriter_WriteASCIIString(writer, number, len) < 0)
return NULL;
break;
}
case 's':
{
if (sizemod) {
const wchar_t *s = va_arg(*vargs, const wchar_t*);
if (unicode_fromformat_write_wcstr(writer, s, width, precision, flags) < 0)
return NULL;
}
else {
/* UTF-8 */
const char *s = va_arg(*vargs, const char*);
if (unicode_fromformat_write_cstr(writer, s, width, precision, flags) < 0)
return NULL;
}
break;
}
case 'U':
{
PyObject *obj = va_arg(*vargs, PyObject *);
assert(obj && _PyUnicode_CHECK(obj));
if (unicode_fromformat_write_str(writer, obj, width, precision, flags) == -1)
return NULL;
break;
}
case 'V':
{
PyObject *obj = va_arg(*vargs, PyObject *);
const char *str;
const wchar_t *wstr;
if (sizemod) {
wstr = va_arg(*vargs, const wchar_t*);
}
else {
str = va_arg(*vargs, const char *);
}
if (obj) {
assert(_PyUnicode_CHECK(obj));
if (unicode_fromformat_write_str(writer, obj, width, precision, flags) == -1)
return NULL;
}
else if (sizemod) {
assert(wstr != NULL);
if (unicode_fromformat_write_wcstr(writer, wstr, width, precision, flags) < 0)
return NULL;
}
else {
assert(str != NULL);
if (unicode_fromformat_write_cstr(writer, str, width, precision, flags) < 0)
return NULL;
}
break;
}
case 'S':
{
PyObject *obj = va_arg(*vargs, PyObject *);
PyObject *str;
assert(obj);
str = PyObject_Str(obj);
if (!str)
return NULL;
if (unicode_fromformat_write_str(writer, str, width, precision, flags) == -1) {
Py_DECREF(str);
return NULL;
}
Py_DECREF(str);
break;
}
case 'R':
{
PyObject *obj = va_arg(*vargs, PyObject *);
PyObject *repr;
assert(obj);
repr = PyObject_Repr(obj);
if (!repr)
return NULL;
if (unicode_fromformat_write_str(writer, repr, width, precision, flags) == -1) {
Py_DECREF(repr);
return NULL;
}
Py_DECREF(repr);
break;
}
case 'A':
{
PyObject *obj = va_arg(*vargs, PyObject *);
PyObject *ascii;
assert(obj);
ascii = PyObject_ASCII(obj);
if (!ascii)
return NULL;
if (unicode_fromformat_write_str(writer, ascii, width, precision, flags) == -1) {
Py_DECREF(ascii);
return NULL;
}
Py_DECREF(ascii);
break;
}
default:
invalid_format:
PyErr_Format(PyExc_SystemError, "invalid format string: %s", p);
return NULL;
}
f++;
return f;
}
PyObject *
PyUnicode_FromFormatV(const char *format, va_list vargs)
{
va_list vargs2;
const char *f;
_PyUnicodeWriter writer;
_PyUnicodeWriter_Init(&writer);
writer.min_length = strlen(format) + 100;
writer.overallocate = 1;
// Copy varags to be able to pass a reference to a subfunction.
va_copy(vargs2, vargs);
for (f = format; *f; ) {
if (*f == '%') {
f = unicode_fromformat_arg(&writer, f, &vargs2);
if (f == NULL)
goto fail;
}
else {
const char *p;
Py_ssize_t len;
p = f;
do
{
if ((unsigned char)*p > 127) {
PyErr_Format(PyExc_ValueError,
"PyUnicode_FromFormatV() expects an ASCII-encoded format "
"string, got a non-ASCII byte: 0x%02x",
(unsigned char)*p);
goto fail;
}
p++;
}
while (*p != '\0' && *p != '%');
len = p - f;
if (*p == '\0')
writer.overallocate = 0;
if (_PyUnicodeWriter_WriteASCIIString(&writer, f, len) < 0)
goto fail;
f = p;
}
}
va_end(vargs2);
return _PyUnicodeWriter_Finish(&writer);
fail:
va_end(vargs2);
_PyUnicodeWriter_Dealloc(&writer);
return NULL;
}
PyObject *
PyUnicode_FromFormat(const char *format, ...)
{
PyObject* ret;
va_list vargs;
va_start(vargs, format);
ret = PyUnicode_FromFormatV(format, vargs);
va_end(vargs);
return ret;
}
static Py_ssize_t
unicode_get_widechar_size(PyObject *unicode)
{
Py_ssize_t res;
assert(unicode != NULL);
assert(_PyUnicode_CHECK(unicode));
res = _PyUnicode_LENGTH(unicode);
#if SIZEOF_WCHAR_T == 2
if (PyUnicode_KIND(unicode) == PyUnicode_4BYTE_KIND) {
const Py_UCS4 *s = PyUnicode_4BYTE_DATA(unicode);
const Py_UCS4 *end = s + res;
for (; s < end; ++s) {
if (*s > 0xFFFF) {
++res;
}
}
}
#endif
return res;
}
static void
unicode_copy_as_widechar(PyObject *unicode, wchar_t *w, Py_ssize_t size)
{
assert(unicode != NULL);
assert(_PyUnicode_CHECK(unicode));
if (PyUnicode_KIND(unicode) == sizeof(wchar_t)) {
memcpy(w, PyUnicode_DATA(unicode), size * sizeof(wchar_t));
return;
}
if (PyUnicode_KIND(unicode) == PyUnicode_1BYTE_KIND) {
const Py_UCS1 *s = PyUnicode_1BYTE_DATA(unicode);
for (; size--; ++s, ++w) {
*w = *s;
}
}
else {
#if SIZEOF_WCHAR_T == 4
assert(PyUnicode_KIND(unicode) == PyUnicode_2BYTE_KIND);
const Py_UCS2 *s = PyUnicode_2BYTE_DATA(unicode);
for (; size--; ++s, ++w) {
*