Modules/_testcapi/unicode.c

#include <stddef.h>               // ptrdiff_t

#include "parts.h"
#include "util.h"

/* Test PyUnicode_New() */
static PyObject *
unicode_new(PyObject *self, PyObject *args)
{
    Py_ssize_t size;
    unsigned int maxchar;
    PyObject *result;

    if (!PyArg_ParseTuple(args, "nI", &size, &maxchar)) {
        return NULL;
    }

    result = PyUnicode_New(size, (Py_UCS4)maxchar);
    if (!result) {
        return NULL;
    }
    if (size > 0 && maxchar <= 0x10ffff &&
        PyUnicode_Fill(result, 0, size, (Py_UCS4)maxchar) < 0)
    {
        Py_DECREF(result);
        return NULL;
    }
    return result;
}


static PyObject *
unicode_copy(PyObject *unicode)
{
    PyObject *copy;

    if (!unicode) {
        return NULL;
    }
    if (!PyUnicode_Check(unicode)) {
        Py_INCREF(unicode);
        return unicode;
    }

    copy = PyUnicode_New(PyUnicode_GET_LENGTH(unicode),
                         PyUnicode_MAX_CHAR_VALUE(unicode));
    if (!copy) {
        return NULL;
    }
    if (PyUnicode_CopyCharacters(copy, 0, unicode,
                                 0, PyUnicode_GET_LENGTH(unicode)) < 0)
    {
        Py_DECREF(copy);
        return NULL;
    }
    return copy;
}


/* Test PyUnicode_Fill() */
static PyObject *
unicode_fill(PyObject *self, PyObject *args, PyObject *kwargs)
{
    static char *kwlist[] = {"to", "start", "length", "fill_char",
                             "incref", NULL};
    PyObject *to, *to_copy;
    Py_ssize_t start, length, filled;
    unsigned int fill_char;
    int incref = 0;

    if (!PyArg_ParseTupleAndKeywords(args, kwargs,
                                     "OnnI|p", kwlist,
                                     &to, &start, &length,
                                     &fill_char, &incref))
    {
        return NULL;
    }

    NULLABLE(to);
    if (!(to_copy = unicode_copy(to)) && to) {
        return NULL;
    }

    if (incref) {
        Py_INCREF(to_copy);
    }
    filled = PyUnicode_Fill(to_copy, start, length, (Py_UCS4)fill_char);
    if (incref) {
        Py_DECREF(to_copy);
    }

    if (filled == -1 && PyErr_Occurred()) {
        Py_DECREF(to_copy);
        return NULL;
    }
    return Py_BuildValue("(Nn)", to_copy, filled);
}


/* Test PyUnicode_FromKindAndData() */
static PyObject *
unicode_fromkindanddata(PyObject *self, PyObject *args)
{
    int kind;
    void *buffer;
    Py_ssize_t bsize;
    Py_ssize_t size = -100;

    if (!PyArg_ParseTuple(args, "iz#|n", &kind, &buffer, &bsize, &size)) {
        return NULL;
    }

    if (size == -100) {
        size = bsize;
    }
    if (kind && size % kind) {
        PyErr_SetString(PyExc_AssertionError,
                        "invalid size in unicode_fromkindanddata()");
        return NULL;
    }
    return PyUnicode_FromKindAndData(kind, buffer, kind ? size / kind : 0);
}


// Test PyUnicode_AsUCS4().
// Part of the limited C API, but the test needs PyUnicode_FromKindAndData().
static PyObject *
unicode_asucs4(PyObject *self, PyObject *args)
{
    PyObject *unicode, *result;
    Py_UCS4 *buffer;
    int copy_null;
    Py_ssize_t str_len, buf_len;

    if (!PyArg_ParseTuple(args, "Onp:unicode_asucs4", &unicode, &str_len, &copy_null)) {
        return NULL;
    }

    NULLABLE(unicode);
    buf_len = str_len + 1;
    buffer = PyMem_NEW(Py_UCS4, buf_len);
    if (buffer == NULL) {
        return PyErr_NoMemory();
    }
    memset(buffer, 0, sizeof(Py_UCS4)*buf_len);
    buffer[str_len] = 0xffffU;

    if (!PyUnicode_AsUCS4(unicode, buffer, buf_len, copy_null)) {
        PyMem_Free(buffer);
        return NULL;
    }

    result = PyUnicode_FromKindAndData(PyUnicode_4BYTE_KIND, buffer, buf_len);
    PyMem_Free(buffer);
    return result;
}


// Test PyUnicode_AsUCS4Copy().
// Part of the limited C API, but the test needs PyUnicode_FromKindAndData().
static PyObject *
unicode_asucs4copy(PyObject *self, PyObject *args)
{
    PyObject *unicode;
    Py_UCS4 *buffer;
    PyObject *result;

    if (!PyArg_ParseTuple(args, "O", &unicode)) {
        return NULL;
    }

    NULLABLE(unicode);
    buffer = PyUnicode_AsUCS4Copy(unicode);
    if (buffer == NULL) {
        return NULL;
    }
    result = PyUnicode_FromKindAndData(PyUnicode_4BYTE_KIND,
                                       buffer,
                                       PyUnicode_GET_LENGTH(unicode) + 1);
    PyMem_FREE(buffer);
    return result;
}


/* Test PyUnicode_AsUTF8() */
static PyObject *
unicode_asutf8(PyObject *self, PyObject *args)
{
    PyObject *unicode;
    Py_ssize_t buflen;
    const char *s;

    if (!PyArg_ParseTuple(args, "On", &unicode, &buflen))
        return NULL;

    NULLABLE(unicode);
    s = PyUnicode_AsUTF8(unicode);
    if (s == NULL)
        return NULL;

    return PyBytes_FromStringAndSize(s, buflen);
}


/* Test PyUnicode_CopyCharacters() */
static PyObject *
unicode_copycharacters(PyObject *self, PyObject *args, PyObject *kwargs)
{
    static char *kwlist[] = {"to", "to_start", "from", "from_start",
                             "howmany", "incref", NULL};
    PyObject *from, *to, *to_copy;
    Py_ssize_t from_start, to_start, how_many, copied;
    int incref = 0;

    if (!PyArg_ParseTupleAndKeywords(args, kwargs,
                                     "UnOnn|p", kwlist,
                                     &to, &to_start, &from, &from_start,
                                     &how_many, &incref)) {
        return NULL;
    }

    NULLABLE(from);
    if (!(to_copy = PyUnicode_New(PyUnicode_GET_LENGTH(to),
                                  PyUnicode_MAX_CHAR_VALUE(to)))) {
        return NULL;
    }
    if (PyUnicode_Fill(to_copy, 0, PyUnicode_GET_LENGTH(to_copy), 0U) < 0) {
        Py_DECREF(to_copy);
        return NULL;
    }

    if (incref) {
        Py_INCREF(to_copy);
    }
    copied = PyUnicode_CopyCharacters(to_copy, to_start, from,
                                      from_start, how_many);
    if (incref) {
        Py_DECREF(to_copy);
    }

    if (copied == -1 && PyErr_Occurred()) {
        Py_DECREF(to_copy);
        return NULL;
    }

    return Py_BuildValue("(Nn)", to_copy, copied);
}


// Test PyUnstable_Unicode_GET_CACHED_HASH()
static PyObject*
unicode_GET_CACHED_HASH(PyObject *self, PyObject *arg)
{
    return PyLong_FromSsize_t(PyUnstable_Unicode_GET_CACHED_HASH(arg));
}


// Test the deprecated _Py_Identifier C API:
// - _Py_IDENTIFIER()
// - _Py_static_string()
// - _Py_static_string_init()
// - _PyObject_CallMethodId()
// - _PyObject_GetAttrId()
// - _PyUnicode_FromId()
//
// _testembed also has tests on _PyUnicode_FromId().
static PyObject*
test_py_identifier(PyObject *self, PyObject *Py_UNUSED(args))
{
// Ignore deprecation warnings
_Py_COMP_DIAG_PUSH
_Py_COMP_DIAG_IGNORE_DEPR_DECLS

    _Py_IDENTIFIER(hello);
    PyObject *str = _PyUnicode_FromId(&PyId_hello);  // borrowed ref
    if (str == NULL) {
        return NULL;
    }
    assert(PyUnicode_EqualToUTF8(str, "hello") == 1);

    // Calling twice return the same object
    PyObject *str2 = _PyUnicode_FromId(&PyId_hello);  // borrowed ref
    assert(str2 == str);

    PyObject *number = Py_GetConstant(Py_CONSTANT_ONE);  // immortal
    _Py_static_string(to_bytes_id, "to_bytes");
    PyObject *res = _PyObject_CallMethodId(number, &to_bytes_id, NULL);
    if (res == NULL) {
        return NULL;
    }
    Py_DECREF(res);

    static _Py_Identifier real_id = _Py_static_string_init("real");
    res = _PyObject_GetAttrId(number, &real_id);
    if (res == NULL) {
        return NULL;
    }
    assert(res == number);
    Py_DECREF(res);

    Py_RETURN_NONE;

_Py_COMP_DIAG_POP
}


// Write into an immutable str object to test _PyStaticObjects_CheckAll()
static PyObject *
corrupt_unicode(PyObject *Py_UNUSED(module), PyObject *args)
{
    PyObject *obj, *override;
    if (!PyArg_ParseTuple(args, "OO", &obj, &override)) {
        return NULL;
    }
    assert(PyUnicode_KIND(obj) == PyUnicode_1BYTE_KIND);
    assert(PyUnicode_KIND(override) == PyUnicode_1BYTE_KIND);

    Py_UCS1 *dst = PyUnicode_1BYTE_DATA(obj);
    Py_UCS1 *src = PyUnicode_1BYTE_DATA(override);
    Py_ssize_t size = PyUnicode_GET_LENGTH(override);
    memcpy(dst, src, size);
    Py_RETURN_NONE;
}


/* Test PyUnicode_READ_CHAR() */
static PyObject *
unicode_read_char(PyObject *self, PyObject *args)
{
    PyObject *unicode;
    Py_ssize_t index;
    if (!PyArg_ParseTuple(args, "On", &unicode, &index)) {
        return NULL;
    }
    NULLABLE(unicode);

    Py_UCS4 result = PyUnicode_READ_CHAR(unicode, index);
    return PyLong_FromUnsignedLong(result);
}


/* Test PyUnicode_READ() macro */
static PyObject *
unicode_read(PyObject *self, PyObject *args)
{
    PyObject *unicode;
    Py_ssize_t index;
    if (!PyArg_ParseTuple(args, "On", &unicode, &index)) {
        return NULL;
    }
    NULLABLE(unicode);

    int kind = PyUnicode_KIND(unicode);
    const void *data = PyUnicode_DATA(unicode);
    Py_UCS4 result = PyUnicode_READ(kind, data, index);
    return PyLong_FromUnsignedLong(result);
}


/* Test PyUnicode_WRITE() macro */
static PyObject *
unicode_write(PyObject *self, PyObject *args)
{
    PyObject *unicode;
    Py_ssize_t index;
    unsigned int character;
    if (!PyArg_ParseTuple(args, "OnI", &unicode, &index, &character)) {
        return NULL;
    }
    NULLABLE(unicode);

    PyObject *copy = unicode_copy(unicode);
    if (copy == NULL) {
        return NULL;
    }

    int kind = PyUnicode_KIND(copy);
    const void *data = PyUnicode_DATA(copy);
    PyUnicode_WRITE(kind, data, index, character);
    // Same return value than _testlimitedcapi unicode_writechar():
    // always use 0 as the function result
    return Py_BuildValue("(Ni)", copy, 0);
}


/* Test PyUnicode_KIND() macro */
static PyObject *
unicode_kind(PyObject *self, PyObject *unicode)
{
    NULLABLE(unicode);

    int kind = PyUnicode_KIND(unicode);
    return PyLong_FromLong(kind);
}


/* Test PyUnicode_MAX_CHAR_VALUE() macro */
static PyObject *
unicode_max_char_value(PyObject *self, PyObject *unicode)
{
    NULLABLE(unicode);

    Py_UCS4 maxchar = PyUnicode_MAX_CHAR_VALUE(unicode);
    return PyLong_FromUnsignedLong(maxchar);
}


/* Test PyUnicode_GET_LENGTH() macro */
static PyObject *
unicode_getlength_macro(PyObject *self, PyObject *arg)
{
    NULLABLE(arg);
    return PyLong_FromSsize_t(PyUnicode_GET_LENGTH(arg));
}


/* Test PyUnicode_Equal() macro */
static PyObject *
unicode_equal(PyObject *self, PyObject *args)
{
    PyObject *str1, *str2;
    if (!PyArg_ParseTuple(args, "OnI", &str1, &str2)) {
        return NULL;
    }
    NULLABLE(str1);
    NULLABLE(str2);

    RETURN_INT(PyUnicode_Equal(str1, str2));
}


/* Test PyUnicode_CHECK_INTERNED() macro */
static PyObject *
unicode_check_interned(PyObject *self, PyObject *arg)
{
    NULLABLE(arg);
    RETURN_UINT(PyUnicode_CHECK_INTERNED(arg));
}


/* Test PyUnicode_IS_ASCII() macro */
static PyObject *
unicode_is_ascii(PyObject *self, PyObject *arg)
{
    NULLABLE(arg);
    RETURN_UINT(PyUnicode_IS_ASCII(arg));
}


/* Test PyUnicode_IS_COMPACT() macro */
static PyObject *
unicode_is_compact(PyObject *self, PyObject *arg)
{
    NULLABLE(arg);
    RETURN_UINT(PyUnicode_IS_COMPACT(arg));
}


/* Test PyUnicode_IS_COMPACT_ASCII() macro */
static PyObject *
unicode_is_compact_ascii(PyObject *self, PyObject *arg)
{
    NULLABLE(arg);
    RETURN_INT(PyUnicode_IS_COMPACT_ASCII(arg));
}


// --- PyUnicodeWriter type -------------------------------------------------

typedef struct {
    PyObject_HEAD
    PyUnicodeWriter *writer;
} WriterObject;


static PyObject *
writer_new(PyTypeObject *type, PyObject *args, PyObject *kwargs)
{
    WriterObject *self = (WriterObject *)type->tp_alloc(type, 0);
    if (!self) {
        return NULL;
    }
    self->writer = NULL;
    return (PyObject*)self;
}


static int
writer_init(PyObject *self_raw, PyObject *args, PyObject *kwargs)
{
    WriterObject *self = (WriterObject *)self_raw;

    Py_ssize_t size;
    if (!PyArg_ParseTuple(args, "n", &size)) {
        return -1;
    }

    if (self->writer) {
        PyUnicodeWriter_Discard(self->writer);
    }

    self->writer = PyUnicodeWriter_Create(size);
    if (self->writer == NULL) {
        return -1;
    }
    return 0;
}


static void
writer_dealloc(PyObject *self_raw)
{
    WriterObject *self = (WriterObject *)self_raw;
    PyTypeObject *tp = Py_TYPE(self);
    if (self->writer) {
        PyUnicodeWriter_Discard(self->writer);
    }
    tp->tp_free(self);
    Py_DECREF(tp);
}


static inline int
writer_check(WriterObject *self)
{
    if (self->writer == NULL) {
        PyErr_SetString(PyExc_ValueError, "operation on finished writer");
        return -1;
    }
    return 0;
}


static PyObject*
writer_write_char(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    unsigned int ch;
    if (!PyArg_ParseTuple(args, "I", &ch)) {
        return NULL;
    }

    if (PyUnicodeWriter_WriteChar(self->writer, (Py_UCS4)ch) < 0) {
        return NULL;
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_write_utf8(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    const char *str;
    Py_ssize_t bsize, size;
    if (!PyArg_ParseTuple(args, "z#n", &str, &bsize, &size)) {
        return NULL;
    }

    if (PyUnicodeWriter_WriteUTF8(self->writer, str, size) < 0) {
        return NULL;
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_write_ascii(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    const char *str;
    Py_ssize_t bsize, size;
    if (!PyArg_ParseTuple(args, "z#n", &str, &bsize, &size)) {
        return NULL;
    }

    if (PyUnicodeWriter_WriteASCII(self->writer, str, size) < 0) {
        return NULL;
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_write_widechar(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    const char *s;
    Py_ssize_t bsize;
    Py_ssize_t size = -100;

    if (!PyArg_ParseTuple(args, "z#|n", &s, &bsize, &size)) {
        return NULL;
    }
    if (size == -100) {
        if (bsize % SIZEOF_WCHAR_T) {
            PyErr_SetString(PyExc_AssertionError,
                            "invalid size in writer.write_widechar()");
            return NULL;
        }
        size = bsize / SIZEOF_WCHAR_T;
    }

    int res = PyUnicodeWriter_WriteWideChar(self->writer, (const wchar_t *)s, size);
    if (res < 0) {
        return NULL;
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_write_ucs4(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    const char *s;
    Py_ssize_t bsize;
    Py_ssize_t size = -100;

    if (!PyArg_ParseTuple(args, "z#|n", &s, &bsize, &size)) {
        return NULL;
    }
    if (size == -100) {
        if (bsize % sizeof(Py_UCS4)) {
            PyErr_SetString(PyExc_AssertionError,
                            "invalid size in writer.write_ucs4()");
            return NULL;
        }
        size = bsize / sizeof(Py_UCS4);
    }

    int res = PyUnicodeWriter_WriteUCS4(self->writer, (const Py_UCS4 *)s, size);
    if (res < 0) {
        return NULL;
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_write_str(PyObject *self_raw, PyObject *obj)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    NULLABLE(obj);
    if (PyUnicodeWriter_WriteStr(self->writer, obj) < 0) {
        return NULL;
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_write_repr(PyObject *self_raw, PyObject *obj)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    NULLABLE(obj);
    if (PyUnicodeWriter_WriteRepr(self->writer, obj) < 0) {
        return NULL;
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_write_substring(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    PyObject *str;
    Py_ssize_t start, end;
    if (!PyArg_ParseTuple(args, "Onn", &str, &start, &end)) {
        return NULL;
    }
    NULLABLE(str);

    if (PyUnicodeWriter_WriteSubstring(self->writer, str, start, end) < 0) {
        return NULL;
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_write_latin1(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    const char *str;
    Py_ssize_t bsize, size;
    if (!PyArg_ParseTuple(args, "z#n", &str, &bsize, &size)) {
        return NULL;
    }

    _PyUnicodeWriter *writer = (_PyUnicodeWriter*)self->writer;
_Py_COMP_DIAG_PUSH
_Py_COMP_DIAG_IGNORE_DEPR_DECLS
    if (_PyUnicodeWriter_WriteLatin1String(writer, str, size) < 0) {
        return NULL;
    }
_Py_COMP_DIAG_POP

    Py_RETURN_NONE;
}


static PyObject*
writer_decodeutf8stateful(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    const char *str;
    Py_ssize_t bsize, len;
    const char *errors;
    int use_consumed = 0;
    if (!PyArg_ParseTuple(args, "z#nz#|p", &str, &bsize, &len, &errors, &bsize, &use_consumed)) {
        return NULL;
    }

    Py_ssize_t consumed = 12345;
    Py_ssize_t *pconsumed = use_consumed ? &consumed : NULL;
    if (PyUnicodeWriter_DecodeUTF8Stateful(self->writer, str, len,
                                           errors, pconsumed) < 0) {
        if (use_consumed) {
            assert(consumed == 0);
        }
        return NULL;
    }

    if (use_consumed) {
        return PyLong_FromSsize_t(consumed);
    }
    Py_RETURN_NONE;
}


static PyObject*
writer_get_pointer(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    return PyLong_FromVoidPtr(self->writer);
}


static PyObject*
writer_get_buffer(PyObject *self_raw, PyObject *args)
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    _PyUnicodeWriter *writer = (_PyUnicodeWriter*)self->writer;
    PyObject *allocated;
    Py_UCS4 maxchar;
    if (writer->buffer) {
        allocated = PyLong_FromSsize_t(PyUnicode_GET_LENGTH(writer->buffer));
        maxchar = PyUnicode_MAX_CHAR_VALUE(writer->buffer);
    }
    else {
        allocated = Py_None;
        maxchar = writer->min_char;
    }
    return Py_BuildValue("(NkN)",
                         allocated, (unsigned long)maxchar,
                         PyBool_FromLong(writer->readonly));
}


static PyObject*
writer_finish(PyObject *self_raw, PyObject *Py_UNUSED(args))
{
    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    PyObject *str = PyUnicodeWriter_Finish(self->writer);
    self->writer = NULL;
    return str;
}


static PyObject*
writer_finish_with_size(PyObject *self_raw, PyObject *args)
{
    PyAPI_FUNC(PyObject*) _PyUnicodeWriter_FinishWithSize(
        _PyUnicodeWriter *writer,
        Py_ssize_t size);

    WriterObject *self = (WriterObject *)self_raw;
    if (writer_check(self) < 0) {
        return NULL;
    }

    Py_ssize_t size;
    if (!PyArg_ParseTuple(args, "n", &size)) {
        return NULL;
    }

    _PyUnicodeWriter *writer = (_PyUnicodeWriter*)self->writer;
    PyObject *str = _PyUnicodeWriter_FinishWithSize(writer, size);
    PyUnicodeWriter_Discard(self->writer);
    self->writer = NULL;
    return str;
}


static PyMethodDef writer_methods[] = {
    {"write_char", _PyCFunction_CAST(writer_write_char), METH_VARARGS},
    {"write_utf8", _PyCFunction_CAST(writer_write_utf8), METH_VARARGS},
    {"write_ascii", _PyCFunction_CAST(writer_write_ascii), METH_VARARGS},
    {"write_widechar", _PyCFunction_CAST(writer_write_widechar), METH_VARARGS},
    {"write_ucs4", _PyCFunction_CAST(writer_write_ucs4), METH_VARARGS},
    {"write_str", _PyCFunction_CAST(writer_write_str), METH_O},
    {"write_repr", _PyCFunction_CAST(writer_write_repr), METH_O},
    {"write_substring", _PyCFunction_CAST(writer_write_substring), METH_VARARGS},
    {"write_latin1", _PyCFunction_CAST(writer_write_latin1), METH_VARARGS},
    {"decodeutf8stateful", _PyCFunction_CAST(writer_decodeutf8stateful), METH_VARARGS},
    {"get_pointer", _PyCFunction_CAST(writer_get_pointer), METH_VARARGS},
    {"get_buffer", _PyCFunction_CAST(writer_get_buffer), METH_VARARGS},
    {"finish", _PyCFunction_CAST(writer_finish), METH_NOARGS},
    {"finish_with_size", _PyCFunction_CAST(writer_finish_with_size), METH_VARARGS},
    {NULL,              NULL}           /* sentinel */
};

static PyType_Slot Writer_Type_slots[] = {
    {Py_tp_new, writer_new},
    {Py_tp_init, writer_init},
    {Py_tp_dealloc, writer_dealloc},
    {Py_tp_methods, writer_methods},
    {0, 0},  /* sentinel */
};

static PyType_Spec Writer_spec = {
    .name = "_testcapi.PyUnicodeWriter",
    .basicsize = sizeof(WriterObject),
    .flags = Py_TPFLAGS_DEFAULT,
    .slots = Writer_Type_slots,
};


static PyMethodDef TestMethods[] = {
    {"unicode_new",              unicode_new,                    METH_VARARGS},
    {"unicode_fill",             _PyCFunction_CAST(unicode_fill), METH_VARARGS | METH_KEYWORDS},
    {"unicode_fromkindanddata",  unicode_fromkindanddata,        METH_VARARGS},
    {"unicode_asucs4",           unicode_asucs4,                 METH_VARARGS},
    {"unicode_asucs4copy",       unicode_asucs4copy,             METH_VARARGS},
    {"unicode_asutf8",           unicode_asutf8,                 METH_VARARGS},
    {"unicode_copycharacters",   _PyCFunction_CAST(unicode_copycharacters), METH_VARARGS | METH_KEYWORDS},
    {"unicode_GET_CACHED_HASH",  unicode_GET_CACHED_HASH,        METH_O},
    {"test_py_identifier",       test_py_identifier,             METH_NOARGS},
    {"corrupt_unicode",          corrupt_unicode,                METH_VARARGS},
    {"unicode_read_char",        unicode_read_char,              METH_VARARGS},
    {"unicode_read",             unicode_read,                   METH_VARARGS},
    {"unicode_write",            unicode_write,                  METH_VARARGS},
    {"unicode_kind",             unicode_kind,                   METH_O},
    {"unicode_max_char_value",   unicode_max_char_value,         METH_O},
    {"unicode_getlength_macro",  unicode_getlength_macro,        METH_O},
    {"unicode_equal",            unicode_equal,                  METH_VARARGS},
    {"unicode_check_interned",   unicode_check_interned,         METH_O},
    {"unicode_is_ascii",         unicode_is_ascii,               METH_O},
    {"unicode_is_compact",       unicode_is_compact,             METH_O},
    {"unicode_is_compact_ascii", unicode_is_compact_ascii,       METH_O},
    {NULL},
};

int
_PyTestCapi_Init_Unicode(PyObject *m) {
    if (PyModule_AddFunctions(m, TestMethods) < 0) {
        return -1;
    }

    PyTypeObject *writer_type = (PyTypeObject *)PyType_FromSpec(&Writer_spec);
    if (writer_type == NULL) {
        return -1;
    }
    if (PyModule_AddType(m, writer_type) < 0) {
        Py_DECREF(writer_type);
        return -1;
    }
    Py_DECREF(writer_type);

    return 0;
}