/src/cpython/Python/Python-tokenize.c

Source (jump to first uncovered line)
#include "Python.h"
#include "errcode.h"
#include "internal/pycore_critical_section.h"   // Py_BEGIN_CRITICAL_SECTION
#include "../Parser/lexer/state.h"
#include "../Parser/lexer/lexer.h"
#include "../Parser/tokenizer/tokenizer.h"
#include "../Parser/pegen.h"                    // _PyPegen_byte_offset_to_character_offset()

static struct PyModuleDef _tokenizemodule;

typedef struct {
    PyTypeObject *TokenizerIter;
} tokenize_state;

static tokenize_state *
get_tokenize_state(PyObject *module) {
    return (tokenize_state *)PyModule_GetState(module);
}

#define _tokenize_get_state_by_type(type) \
    get_tokenize_state(PyType_GetModuleByDef(type, &_tokenizemodule))

#include "pycore_runtime.h"
#include "clinic/Python-tokenize.c.h"

/*[clinic input]
module _tokenizer
class _tokenizer.tokenizeriter "tokenizeriterobject *" "_tokenize_get_state_by_type(type)->TokenizerIter"
[clinic start generated code]*/
/*[clinic end generated code: output=da39a3ee5e6b4b0d input=96d98ee2fef7a8bc]*/

typedef struct
{
    PyObject_HEAD struct tok_state *tok;
    int done;

    /* Needed to cache line for performance */
    PyObject *last_line;
    Py_ssize_t last_lineno;
    Py_ssize_t last_end_lineno;
    Py_ssize_t byte_col_offset_diff;
} tokenizeriterobject;

/*[clinic input]
@classmethod
_tokenizer.tokenizeriter.__new__ as tokenizeriter_new

    readline: object
    /
    *
    extra_tokens: bool
    encoding: str(c_default="NULL") = 'utf-8'
[clinic start generated code]*/

static PyObject *
tokenizeriter_new_impl(PyTypeObject *type, PyObject *readline,
                       int extra_tokens, const char *encoding)
/*[clinic end generated code: output=7501a1211683ce16 input=f7dddf8a613ae8bd]*/
{
    tokenizeriterobject *self = (tokenizeriterobject *)type->tp_alloc(type, 0);
    if (self == NULL) {
        return NULL;
    }
    PyObject *filename = PyUnicode_FromString("<string>");
    if (filename == NULL) {
        return NULL;
    }
    self->tok = _PyTokenizer_FromReadline(readline, encoding, 1, 1);
    if (self->tok == NULL) {
        Py_DECREF(filename);
        return NULL;
    }
    self->tok->filename = filename;
    if (extra_tokens) {
        self->tok->tok_extra_tokens = 1;
    }
    self->done = 0;

    self->last_line = NULL;
    self->byte_col_offset_diff = 0;
    self->last_lineno = 0;
    self->last_end_lineno = 0;

    return (PyObject *)self;
}

static int
_tokenizer_error(tokenizeriterobject *it)
{
    _Py_CRITICAL_SECTION_ASSERT_OBJECT_LOCKED(it);
    if (PyErr_Occurred()) {
        return -1;
    }

    const char *msg = NULL;
    PyObject* errtype = PyExc_SyntaxError;
    struct tok_state *tok = it->tok;
    switch (tok->done) {
        case E_TOKEN:
            msg = "invalid token";
            break;
        case E_EOF:
            PyErr_SetString(PyExc_SyntaxError, "unexpected EOF in multi-line statement");
            PyErr_SyntaxLocationObject(tok->filename, tok->lineno,
                                       tok->inp - tok->buf < 0 ? 0 : (int)(tok->inp - tok->buf));
            return -1;
        case E_DEDENT:
            msg = "unindent does not match any outer indentation level";
            errtype = PyExc_IndentationError;
            break;
        case E_INTR:
            if (!PyErr_Occurred()) {
                PyErr_SetNone(PyExc_KeyboardInterrupt);
            }
            return -1;
        case E_NOMEM:
            PyErr_NoMemory();
            return -1;
        case E_TABSPACE:
            errtype = PyExc_TabError;
            msg = "inconsistent use of tabs and spaces in indentation";
            break;
        case E_TOODEEP:
            errtype = PyExc_IndentationError;
            msg = "too many levels of indentation";
            break;
        case E_LINECONT: {
            msg = "unexpected character after line continuation character";
            break;
        }
        default:
            msg = "unknown tokenization error";
    }

    PyObject* errstr = NULL;
    PyObject* error_line = NULL;
    PyObject* tmp = NULL;
    PyObject* value = NULL;
    int result = 0;

    Py_ssize_t size = tok->inp - tok->buf;
    assert(tok->buf[size-1] == '\n');
    size -= 1; // Remove the newline character from the end of the line
    error_line = PyUnicode_DecodeUTF8(tok->buf, size, "replace");
    if (!error_line) {
        result = -1;
        goto exit;
    }

    Py_ssize_t offset = _PyPegen_byte_offset_to_character_offset(error_line, tok->inp - tok->buf);
    if (offset == -1) {
        result = -1;
        goto exit;
    }
    tmp = Py_BuildValue("(OnnOOO)", tok->filename, tok->lineno, offset, error_line, Py_None, Py_None);
    if (!tmp) {
        result = -1;
        goto exit;
    }

    errstr = PyUnicode_FromString(msg);
    if (!errstr) {
        result = -1;
        goto exit;
    }

    value = PyTuple_Pack(2, errstr, tmp);
    if (!value) {
        result = -1;
        goto exit;
    }

    PyErr_SetObject(errtype, value);

exit:
    Py_XDECREF(errstr);
    Py_XDECREF(error_line);
    Py_XDECREF(tmp);
    Py_XDECREF(value);
    return result;
}

static PyObject *
_get_current_line(tokenizeriterobject *it, const char *line_start, Py_ssize_t size,
                  int *line_changed)
{
    _Py_CRITICAL_SECTION_ASSERT_OBJECT_LOCKED(it);
    PyObject *line;
    if (it->tok->lineno != it->last_lineno) {
        // Line has changed since last token, so we fetch the new line and cache it
        // in the iter object.
        Py_XDECREF(it->last_line);
        line = PyUnicode_DecodeUTF8(line_start, size, "replace");
        it->last_line = line;
        it->byte_col_offset_diff = 0;
    }
    else {
        line = it->last_line;
        *line_changed = 0;
    }
    return line;
}

static void
_get_col_offsets(tokenizeriterobject *it, struct token token, const char *line_start,
                 PyObject *line, int line_changed, Py_ssize_t lineno, Py_ssize_t end_lineno,
                 Py_ssize_t *col_offset, Py_ssize_t *end_col_offset)
{
    _Py_CRITICAL_SECTION_ASSERT_OBJECT_LOCKED(it);
    Py_ssize_t byte_offset = -1;
    if (token.start != NULL && token.start >= line_start) {
        byte_offset = token.start - line_start;
        if (line_changed) {
            *col_offset = _PyPegen_byte_offset_to_character_offset_line(line, 0, byte_offset);
            it->byte_col_offset_diff = byte_offset - *col_offset;
        }
        else {
            *col_offset = byte_offset - it->byte_col_offset_diff;
        }
    }

    if (token.end != NULL && token.end >= it->tok->line_start) {
        Py_ssize_t end_byte_offset = token.end - it->tok->line_start;
        if (lineno == end_lineno) {
            // If the whole token is at the same line, we can just use the token.start
            // buffer for figuring out the new column offset, since using line is not
            // performant for very long lines.
            Py_ssize_t token_col_offset = _PyPegen_byte_offset_to_character_offset_line(line, byte_offset, end_byte_offset);
            *end_col_offset = *col_offset + token_col_offset;
            it->byte_col_offset_diff += token.end - token.start - token_col_offset;
        }
        else {
            *end_col_offset = _PyPegen_byte_offset_to_character_offset_raw(it->tok->line_start, end_byte_offset);
            it->byte_col_offset_diff += end_byte_offset - *end_col_offset;
        }
    }
    it->last_lineno = lineno;
    it->last_end_lineno = end_lineno;
}

static PyObject *
tokenizeriter_next(PyObject *op)
{
    tokenizeriterobject *it = (tokenizeriterobject*)op;
    PyObject* result = NULL;

    Py_BEGIN_CRITICAL_SECTION(it);

    struct token token;
    _PyToken_Init(&token);

    int type = _PyTokenizer_Get(it->tok, &token);
    if (type == ERRORTOKEN) {
        if(!PyErr_Occurred()) {
            _tokenizer_error(it);
            assert(PyErr_Occurred());
        }
        goto exit;
    }
    if (it->done || type == ERRORTOKEN) {
        PyErr_SetString(PyExc_StopIteration, "EOF");
        it->done = 1;
        goto exit;
    }
    PyObject *str = NULL;
    if (token.start == NULL || token.end == NULL) {
        str = Py_GetConstant(Py_CONSTANT_EMPTY_STR);
    }
    else {
        str = PyUnicode_FromStringAndSize(token.start, token.end - token.start);
    }
    if (str == NULL) {
        goto exit;
    }

    int is_trailing_token = 0;
    if (type == ENDMARKER || (type == DEDENT && it->tok->done == E_EOF)) {
        is_trailing_token = 1;
    }

    const char *line_start = ISSTRINGLIT(type) ? it->tok->multi_line_start : it->tok->line_start;
    PyObject* line = NULL;
    int line_changed = 1;
    if (it->tok->tok_extra_tokens && is_trailing_token) {
        line = Py_GetConstant(Py_CONSTANT_EMPTY_STR);
    } else {
        Py_ssize_t size = it->tok->inp - line_start;
        if (size >= 1 && it->tok->implicit_newline) {
            size -= 1;
        }

        line = _get_current_line(it, line_start, size, &line_changed);
    }
    if (line == NULL) {
        Py_DECREF(str);
        goto exit;
    }

    Py_ssize_t lineno = ISSTRINGLIT(type) ? it->tok->first_lineno : it->tok->lineno;
    Py_ssize_t end_lineno = it->tok->lineno;
    Py_ssize_t col_offset = -1;
    Py_ssize_t end_col_offset = -1;
    _get_col_offsets(it, token, line_start, line, line_changed,
                     lineno, end_lineno, &col_offset, &end_col_offset);

    if (it->tok->tok_extra_tokens) {
        if (is_trailing_token) {
            lineno = end_lineno = lineno + 1;
            col_offset = end_col_offset = 0;
        }
        // Necessary adjustments to match the original Python tokenize
        // implementation
        if (type > DEDENT && type < OP) {
            type = OP;
        }
        else if (type == NEWLINE) {
            Py_DECREF(str);
            if (!it->tok->implicit_newline) {
                if (it->tok->start[0] == '\r') {
                    str = PyUnicode_FromString("\r\n");
                } else {
                    str = PyUnicode_FromString("\n");
                }
            }
            end_col_offset++;
        }
        else if (type == NL) {
            if (it->tok->implicit_newline) {
                Py_DECREF(str);
                str = Py_GetConstant(Py_CONSTANT_EMPTY_STR);
            }
        }

        if (str == NULL) {
            Py_DECREF(line);
            goto exit;
        }
    }

    result = Py_BuildValue("(iN(nn)(nn)O)", type, str, lineno, col_offset, end_lineno, end_col_offset, line);
exit:
    _PyToken_Free(&token);
    if (type == ENDMARKER) {
        it->done = 1;
    }

    Py_END_CRITICAL_SECTION();
    return result;
}

static void
tokenizeriter_dealloc(PyObject *op)
{
    tokenizeriterobject *it = (tokenizeriterobject*)op;
    PyTypeObject *tp = Py_TYPE(it);
    Py_XDECREF(it->last_line);
    _PyTokenizer_Free(it->tok);
    tp->tp_free(it);
    Py_DECREF(tp);
}

static PyType_Slot tokenizeriter_slots[] = {
    {Py_tp_new, tokenizeriter_new},
    {Py_tp_dealloc, tokenizeriter_dealloc},
    {Py_tp_getattro, PyObject_GenericGetAttr},
    {Py_tp_iter, PyObject_SelfIter},
    {Py_tp_iternext, tokenizeriter_next},
    {0, NULL},
};

static PyType_Spec tokenizeriter_spec = {
    .name = "_tokenize.TokenizerIter",
    .basicsize = sizeof(tokenizeriterobject),
    .flags = (Py_TPFLAGS_DEFAULT | Py_TPFLAGS_IMMUTABLETYPE),
    .slots = tokenizeriter_slots,
};

static int
tokenizemodule_exec(PyObject *m)
{
    tokenize_state *state = get_tokenize_state(m);
    if (state == NULL) {
        return -1;
    }

    state->TokenizerIter = (PyTypeObject *)PyType_FromModuleAndSpec(m, &tokenizeriter_spec, NULL);
    if (state->TokenizerIter == NULL) {
        return -1;
    }
    if (PyModule_AddType(m, state->TokenizerIter) < 0) {
        return -1;
    }

    return 0;
}

static PyMethodDef tokenize_methods[] = {
    {NULL, NULL, 0, NULL} /* Sentinel */
};

static PyModuleDef_Slot tokenizemodule_slots[] = {
    {Py_mod_exec, tokenizemodule_exec},
    {Py_mod_multiple_interpreters, Py_MOD_PER_INTERPRETER_GIL_SUPPORTED},
    {Py_mod_gil, Py_MOD_GIL_NOT_USED},
    {0, NULL}
};

static int
tokenizemodule_traverse(PyObject *m, visitproc visit, void *arg)
{
    tokenize_state *state = get_tokenize_state(m);
    Py_VISIT(state->TokenizerIter);
    return 0;
}

static int
tokenizemodule_clear(PyObject *m)
{
    tokenize_state *state = get_tokenize_state(m);
    Py_CLEAR(state->TokenizerIter);
    return 0;
}

static void
tokenizemodule_free(void *m)
{
    tokenizemodule_clear((PyObject *)m);
}

static struct PyModuleDef _tokenizemodule = {
    PyModuleDef_HEAD_INIT,
    .m_name = "_tokenize",
    .m_size = sizeof(tokenize_state),
    .m_slots = tokenizemodule_slots,
    .m_methods = tokenize_methods,
    .m_traverse = tokenizemodule_traverse,
    .m_clear = tokenizemodule_clear,
    .m_free = tokenizemodule_free,
};

PyMODINIT_FUNC
PyInit__tokenize(void)
{
    return PyModuleDef_Init(&_tokenizemodule);
}

Coverage Report

Created: 2025-07-04 06:49

Line	Count	Source (jump to first uncovered line)
1		#include "Python.h"
2		#include "errcode.h"
3		#include "internal/pycore_critical_section.h" // Py_BEGIN_CRITICAL_SECTION
4		#include "../Parser/lexer/state.h"
5		#include "../Parser/lexer/lexer.h"
6		#include "../Parser/tokenizer/tokenizer.h"
7		#include "../Parser/pegen.h" // _PyPegen_byte_offset_to_character_offset()
8
9		static struct PyModuleDef _tokenizemodule;
10
11		typedef struct {
12		PyTypeObject *TokenizerIter;
13		} tokenize_state;
14
15		static tokenize_state *
16	0	get_tokenize_state(PyObject *module) {
17	0	return (tokenize_state *)PyModule_GetState(module);
18	0	}
19
20		#define _tokenize_get_state_by_type(type) \
21		get_tokenize_state(PyType_GetModuleByDef(type, &_tokenizemodule))
22
23		#include "pycore_runtime.h"
24		#include "clinic/Python-tokenize.c.h"
25
26		/*[clinic input]
27		module _tokenizer
28		class _tokenizer.tokenizeriter "tokenizeriterobject *" "_tokenize_get_state_by_type(type)->TokenizerIter"
29		[clinic start generated code]*/
30		/[clinic end generated code: output=da39a3ee5e6b4b0d input=96d98ee2fef7a8bc]/
31
32		typedef struct
33		{
34		PyObject_HEAD struct tok_state *tok;
35		int done;
36
37		/* Needed to cache line for performance */
38		PyObject *last_line;
39		Py_ssize_t last_lineno;
40		Py_ssize_t last_end_lineno;
41		Py_ssize_t byte_col_offset_diff;
42		} tokenizeriterobject;
43
44		/*[clinic input]
45		@classmethod
46		_tokenizer.tokenizeriter.__new__ as tokenizeriter_new
47
48		readline: object
49		/
50		*
51		extra_tokens: bool
52		encoding: str(c_default="NULL") = 'utf-8'
53		[clinic start generated code]*/
54
55		static PyObject *
56		tokenizeriter_new_impl(PyTypeObject type, PyObject readline,
57		int extra_tokens, const char *encoding)
58		/[clinic end generated code: output=7501a1211683ce16 input=f7dddf8a613ae8bd]/
59	0	{
60	0	tokenizeriterobject self = (tokenizeriterobject )type->tp_alloc(type, 0);
61	0	if (self == NULL) {
62	0	return NULL;
63	0	}
64	0	PyObject *filename = PyUnicode_FromString("<string>");
65	0	if (filename == NULL) {
66	0	return NULL;
67	0	}
68	0	self->tok = _PyTokenizer_FromReadline(readline, encoding, 1, 1);
69	0	if (self->tok == NULL) {
70	0	Py_DECREF(filename);
71	0	return NULL;
72	0	}
73	0	self->tok->filename = filename;
74	0	if (extra_tokens) {
75	0	self->tok->tok_extra_tokens = 1;
76	0	}
77	0	self->done = 0;
78
79	0	self->last_line = NULL;
80	0	self->byte_col_offset_diff = 0;
81	0	self->last_lineno = 0;
82	0	self->last_end_lineno = 0;
83
84	0	return (PyObject *)self;
85	0	}
86
87		static int
88		_tokenizer_error(tokenizeriterobject *it)
89	0	{
90	0	_Py_CRITICAL_SECTION_ASSERT_OBJECT_LOCKED(it);
91	0	if (PyErr_Occurred()) {
92	0	return -1;
93	0	}
94
95	0	const char *msg = NULL;
96	0	PyObject* errtype = PyExc_SyntaxError;
97	0	struct tok_state *tok = it->tok;
98	0	switch (tok->done) {
99	0	case E_TOKEN:
100	0	msg = "invalid token";
101	0	break;
102	0	case E_EOF:
103	0	PyErr_SetString(PyExc_SyntaxError, "unexpected EOF in multi-line statement");
104	0	PyErr_SyntaxLocationObject(tok->filename, tok->lineno,
105	0	tok->inp - tok->buf < 0 ? 0 : (int)(tok->inp - tok->buf));
106	0	return -1;
107	0	case E_DEDENT:
108	0	msg = "unindent does not match any outer indentation level";
109	0	errtype = PyExc_IndentationError;
110	0	break;
111	0	case E_INTR:
112	0	if (!PyErr_Occurred()) {
113	0	PyErr_SetNone(PyExc_KeyboardInterrupt);
114	0	}
115	0	return -1;
116	0	case E_NOMEM:
117	0	PyErr_NoMemory();
118	0	return -1;
119	0	case E_TABSPACE:
120	0	errtype = PyExc_TabError;
121	0	msg = "inconsistent use of tabs and spaces in indentation";
122	0	break;
123	0	case E_TOODEEP:
124	0	errtype = PyExc_IndentationError;
125	0	msg = "too many levels of indentation";
126	0	break;
127	0	case E_LINECONT: {
128	0	msg = "unexpected character after line continuation character";
129	0	break;
130	0	}
131	0	default:
132	0	msg = "unknown tokenization error";
133	0	}
134
135	0	PyObject* errstr = NULL;
136	0	PyObject* error_line = NULL;
137	0	PyObject* tmp = NULL;
138	0	PyObject* value = NULL;
139	0	int result = 0;
140
141	0	Py_ssize_t size = tok->inp - tok->buf;
142	0	assert(tok->buf[size-1] == '\n');
143	0	size -= 1; // Remove the newline character from the end of the line
144	0	error_line = PyUnicode_DecodeUTF8(tok->buf, size, "replace");
145	0	if (!error_line) {
146	0	result = -1;
147	0	goto exit;
148	0	}
149
150	0	Py_ssize_t offset = _PyPegen_byte_offset_to_character_offset(error_line, tok->inp - tok->buf);
151	0	if (offset == -1) {
152	0	result = -1;
153	0	goto exit;
154	0	}
155	0	tmp = Py_BuildValue("(OnnOOO)", tok->filename, tok->lineno, offset, error_line, Py_None, Py_None);
156	0	if (!tmp) {
157	0	result = -1;
158	0	goto exit;
159	0	}
160
161	0	errstr = PyUnicode_FromString(msg);
162	0	if (!errstr) {
163	0	result = -1;
164	0	goto exit;
165	0	}
166
167	0	value = PyTuple_Pack(2, errstr, tmp);
168	0	if (!value) {
169	0	result = -1;
170	0	goto exit;
171	0	}
172
173	0	PyErr_SetObject(errtype, value);
174
175	0	exit:
176	0	Py_XDECREF(errstr);
177	0	Py_XDECREF(error_line);
178	0	Py_XDECREF(tmp);
179	0	Py_XDECREF(value);
180	0	return result;
181	0	}
182
183		static PyObject *
184		_get_current_line(tokenizeriterobject it, const char line_start, Py_ssize_t size,
185		int *line_changed)
186	0	{
187	0	_Py_CRITICAL_SECTION_ASSERT_OBJECT_LOCKED(it);
188	0	PyObject *line;
189	0	if (it->tok->lineno != it->last_lineno) {
190		// Line has changed since last token, so we fetch the new line and cache it
191		// in the iter object.
192	0	Py_XDECREF(it->last_line);
193	0	line = PyUnicode_DecodeUTF8(line_start, size, "replace");
194	0	it->last_line = line;
195	0	it->byte_col_offset_diff = 0;
196	0	}
197	0	else {
198	0	line = it->last_line;
199	0	*line_changed = 0;
200	0	}
201	0	return line;
202	0	}
203
204		static void
205		_get_col_offsets(tokenizeriterobject it, struct token token, const char line_start,
206		PyObject *line, int line_changed, Py_ssize_t lineno, Py_ssize_t end_lineno,
207		Py_ssize_t col_offset, Py_ssize_t end_col_offset)
208	0	{
209	0	_Py_CRITICAL_SECTION_ASSERT_OBJECT_LOCKED(it);
210	0	Py_ssize_t byte_offset = -1;
211	0	if (token.start != NULL && token.start >= line_start) {
212	0	byte_offset = token.start - line_start;
213	0	if (line_changed) {
214	0	*col_offset = _PyPegen_byte_offset_to_character_offset_line(line, 0, byte_offset);
215	0	it->byte_col_offset_diff = byte_offset - *col_offset;
216	0	}
217	0	else {
218	0	*col_offset = byte_offset - it->byte_col_offset_diff;
219	0	}
220	0	}
221
222	0	if (token.end != NULL && token.end >= it->tok->line_start) {
223	0	Py_ssize_t end_byte_offset = token.end - it->tok->line_start;
224	0	if (lineno == end_lineno) {
225		// If the whole token is at the same line, we can just use the token.start
226		// buffer for figuring out the new column offset, since using line is not
227		// performant for very long lines.
228	0	Py_ssize_t token_col_offset = _PyPegen_byte_offset_to_character_offset_line(line, byte_offset, end_byte_offset);
229	0	end_col_offset = col_offset + token_col_offset;
230	0	it->byte_col_offset_diff += token.end - token.start - token_col_offset;
231	0	}
232	0	else {
233	0	*end_col_offset = _PyPegen_byte_offset_to_character_offset_raw(it->tok->line_start, end_byte_offset);
234	0	it->byte_col_offset_diff += end_byte_offset - *end_col_offset;
235	0	}
236	0	}
237	0	it->last_lineno = lineno;
238	0	it->last_end_lineno = end_lineno;
239	0	}
240
241		static PyObject *
242		tokenizeriter_next(PyObject *op)
243	0	{
244	0	tokenizeriterobject it = (tokenizeriterobject)op;
245	0	PyObject* result = NULL;
246
247	0	Py_BEGIN_CRITICAL_SECTION(it);
248
249	0	struct token token;
250	0	_PyToken_Init(&token);
251
252	0	int type = _PyTokenizer_Get(it->tok, &token);
253	0	if (type == ERRORTOKEN) {
254	0	if(!PyErr_Occurred()) {
255	0	_tokenizer_error(it);
256	0	assert(PyErr_Occurred());
257	0	}
258	0	goto exit;
259	0	}
260	0	if (it->done \|\| type == ERRORTOKEN) {
261	0	PyErr_SetString(PyExc_StopIteration, "EOF");
262	0	it->done = 1;
263	0	goto exit;
264	0	}
265	0	PyObject *str = NULL;
266	0	if (token.start == NULL \|\| token.end == NULL) {
267	0	str = Py_GetConstant(Py_CONSTANT_EMPTY_STR);
268	0	}
269	0	else {
270	0	str = PyUnicode_FromStringAndSize(token.start, token.end - token.start);
271	0	}
272	0	if (str == NULL) {
273	0	goto exit;
274	0	}
275
276	0	int is_trailing_token = 0;
277	0	if (type == ENDMARKER \|\| (type == DEDENT && it->tok->done == E_EOF)) {
278	0	is_trailing_token = 1;
279	0	}
280
281	0	const char *line_start = ISSTRINGLIT(type) ? it->tok->multi_line_start : it->tok->line_start;
282	0	PyObject* line = NULL;
283	0	int line_changed = 1;
284	0	if (it->tok->tok_extra_tokens && is_trailing_token) {
285	0	line = Py_GetConstant(Py_CONSTANT_EMPTY_STR);
286	0	} else {
287	0	Py_ssize_t size = it->tok->inp - line_start;
288	0	if (size >= 1 && it->tok->implicit_newline) {
289	0	size -= 1;
290	0	}
291
292	0	line = _get_current_line(it, line_start, size, &line_changed);
293	0	}
294	0	if (line == NULL) {
295	0	Py_DECREF(str);
296	0	goto exit;
297	0	}
298
299	0	Py_ssize_t lineno = ISSTRINGLIT(type) ? it->tok->first_lineno : it->tok->lineno;
300	0	Py_ssize_t end_lineno = it->tok->lineno;
301	0	Py_ssize_t col_offset = -1;
302	0	Py_ssize_t end_col_offset = -1;
303	0	_get_col_offsets(it, token, line_start, line, line_changed,
304	0	lineno, end_lineno, &col_offset, &end_col_offset);
305
306	0	if (it->tok->tok_extra_tokens) {
307	0	if (is_trailing_token) {
308	0	lineno = end_lineno = lineno + 1;
309	0	col_offset = end_col_offset = 0;
310	0	}
311		// Necessary adjustments to match the original Python tokenize
312		// implementation
313	0	if (type > DEDENT && type < OP) {
314	0	type = OP;
315	0	}
316	0	else if (type == NEWLINE) {
317	0	Py_DECREF(str);
318	0	if (!it->tok->implicit_newline) {
319	0	if (it->tok->start[0] == '\r') {
320	0	str = PyUnicode_FromString("\r\n");
321	0	} else {
322	0	str = PyUnicode_FromString("\n");
323	0	}
324	0	}
325	0	end_col_offset++;
326	0	}
327	0	else if (type == NL) {
328	0	if (it->tok->implicit_newline) {
329	0	Py_DECREF(str);
330	0	str = Py_GetConstant(Py_CONSTANT_EMPTY_STR);
331	0	}
332	0	}
333
334	0	if (str == NULL) {
335	0	Py_DECREF(line);
336	0	goto exit;
337	0	}
338	0	}
339
340	0	result = Py_BuildValue("(iN(nn)(nn)O)", type, str, lineno, col_offset, end_lineno, end_col_offset, line);
341	0	exit:
342	0	_PyToken_Free(&token);
343	0	if (type == ENDMARKER) {
344	0	it->done = 1;
345	0	}
346
347	0	Py_END_CRITICAL_SECTION();
348	0	return result;
349	0	}
350
351		static void
352		tokenizeriter_dealloc(PyObject *op)
353	0	{
354	0	tokenizeriterobject it = (tokenizeriterobject)op;
355	0	PyTypeObject *tp = Py_TYPE(it);
356	0	Py_XDECREF(it->last_line);
357	0	_PyTokenizer_Free(it->tok);
358	0	tp->tp_free(it);
359	0	Py_DECREF(tp);
360	0	}
361
362		static PyType_Slot tokenizeriter_slots[] = {
363		{Py_tp_new, tokenizeriter_new},
364		{Py_tp_dealloc, tokenizeriter_dealloc},
365		{Py_tp_getattro, PyObject_GenericGetAttr},
366		{Py_tp_iter, PyObject_SelfIter},
367		{Py_tp_iternext, tokenizeriter_next},
368		{0, NULL},
369		};
370
371		static PyType_Spec tokenizeriter_spec = {
372		.name = "_tokenize.TokenizerIter",
373		.basicsize = sizeof(tokenizeriterobject),
374		.flags = (Py_TPFLAGS_DEFAULT \| Py_TPFLAGS_IMMUTABLETYPE),
375		.slots = tokenizeriter_slots,
376		};
377
378		static int
379		tokenizemodule_exec(PyObject *m)
380	0	{
381	0	tokenize_state *state = get_tokenize_state(m);
382	0	if (state == NULL) {
383	0	return -1;
384	0	}
385
386	0	state->TokenizerIter = (PyTypeObject *)PyType_FromModuleAndSpec(m, &tokenizeriter_spec, NULL);
387	0	if (state->TokenizerIter == NULL) {
388	0	return -1;
389	0	}
390	0	if (PyModule_AddType(m, state->TokenizerIter) < 0) {
391	0	return -1;
392	0	}
393
394	0	return 0;
395	0	}
396
397		static PyMethodDef tokenize_methods[] = {
398		{NULL, NULL, 0, NULL} /* Sentinel */
399		};
400
401		static PyModuleDef_Slot tokenizemodule_slots[] = {
402		{Py_mod_exec, tokenizemodule_exec},
403		{Py_mod_multiple_interpreters, Py_MOD_PER_INTERPRETER_GIL_SUPPORTED},
404		{Py_mod_gil, Py_MOD_GIL_NOT_USED},
405		{0, NULL}
406		};
407
408		static int
409		tokenizemodule_traverse(PyObject m, visitproc visit, void arg)
410	0	{
411	0	tokenize_state *state = get_tokenize_state(m);
412	0	Py_VISIT(state->TokenizerIter);
413	0	return 0;
414	0	}
415
416		static int
417		tokenizemodule_clear(PyObject *m)
418	0	{
419	0	tokenize_state *state = get_tokenize_state(m);
420	0	Py_CLEAR(state->TokenizerIter);
421	0	return 0;
422	0	}
423
424		static void
425		tokenizemodule_free(void *m)
426	0	{
427	0	tokenizemodule_clear((PyObject *)m);
428	0	}
429
430		static struct PyModuleDef _tokenizemodule = {
431		PyModuleDef_HEAD_INIT,
432		.m_name = "_tokenize",
433		.m_size = sizeof(tokenize_state),
434		.m_slots = tokenizemodule_slots,
435		.m_methods = tokenize_methods,
436		.m_traverse = tokenizemodule_traverse,
437		.m_clear = tokenizemodule_clear,
438		.m_free = tokenizemodule_free,
439		};
440
441		PyMODINIT_FUNC
442		PyInit__tokenize(void)
443	0	{
444	0	return PyModuleDef_Init(&_tokenizemodule);
445	0	}