gh-110289: C API: Add PyUnicode_EqualToUTF8() and PyUnicode_EqualToUTF8AndSize() functions (GH-110297)

Author: serhiy-storchaka

Doc/c-api/unicode.rst

@@ -1396,6 +1396,28 @@ They all return ``NULL`` or ``-1`` if an exception occurs.
    :c:func:`PyErr_Occurred` to check for errors.
 
 
+.. c:function:: int PyUnicode_EqualToUTF8AndSize(PyObject *unicode, const char *string, Py_ssize_t size)
+
+   Compare a Unicode object with a char buffer which is interpreted as
+   being UTF-8 or ASCII encoded and return true (``1``) if they are equal,
+   or false (``0``) otherwise.
+   If the Unicode object contains surrogate characters or
+   the C string is not valid UTF-8, false (``0``) is returned.
+
+   This function does not raise exceptions.
+
+   .. versionadded:: 3.13
+
+
+.. c:function:: int PyUnicode_EqualToUTF8(PyObject *unicode, const char *string)
+
+   Similar to :c:func:`PyUnicode_EqualToUTF8AndSize`, but compute *string*
+   length using :c:func:`!strlen`.
+   If the Unicode object contains null characters, false (``0``) is returned.
+
+   .. versionadded:: 3.13
+
+
 .. c:function:: int PyUnicode_CompareWithASCIIString(PyObject *uni, const char *string)
 
    Compare a Unicode object, *uni*, with *string* and return ``-1``, ``0``, ``1`` for less

Doc/data/stable_abi.dat

@@ -1024,6 +1024,12 @@ New Features
   functions on Python 3.11 and 3.12.
   (Contributed by Victor Stinner in :gh:`107073`.)
 
+* Add :c:func:`PyUnicode_EqualToUTF8AndSize` and :c:func:`PyUnicode_EqualToUTF8`
+  functions: compare Unicode object with a :c:expr:`const char*` UTF-8 encoded
+  string and return true (``1``) if they are equal, or false (``0``) otherwise.
+  These functions do not raise exceptions.
+  (Contributed by Serhiy Storchaka in :gh:`110289`.)
+
 * Add :c:func:`PyThreadState_GetUnchecked()` function: similar to
   :c:func:`PyThreadState_Get()`, but don't kill the process with a fatal error
   if it is NULL. The caller is responsible to check if the result is NULL.

Doc/whatsnew/3.13.rst

@@ -957,6 +957,15 @@ PyAPI_FUNC(int) PyUnicode_CompareWithASCIIString(
     const char *right           /* ASCII-encoded string */
     );
 
+#if !defined(Py_LIMITED_API) || Py_LIMITED_API+0 >= 0x030D0000
+/* Compare a Unicode object with UTF-8 encoded C string.
+   Return 1 if they are equal, or 0 otherwise.
+   This function does not raise exceptions. */
+
+PyAPI_FUNC(int) PyUnicode_EqualToUTF8(PyObject *, const char *);
+PyAPI_FUNC(int) PyUnicode_EqualToUTF8AndSize(PyObject *, const char *, Py_ssize_t);
+#endif
+
 /* Rich compare two strings and return one of the following:
 
    - NULL in case an exception was raised

Include/unicodeobject.h

@@ -1297,6 +1297,118 @@ def test_comparewithasciistring(self):
         # CRASHES comparewithasciistring([], b'abc')
         # CRASHES comparewithasciistring(NULL, b'abc')
 
+    @support.cpython_only
+    @unittest.skipIf(_testcapi is None, 'need _testcapi module')
+    def test_equaltoutf8(self):
+        # Test PyUnicode_EqualToUTF8()
+        from _testcapi import unicode_equaltoutf8 as equaltoutf8
+        from _testcapi import unicode_asutf8andsize as asutf8andsize
+
+        strings = [
+            'abc', '\xa1\xa2\xa3', '\u4f60\u597d\u4e16',
+            '\U0001f600\U0001f601\U0001f602',
+            '\U0010ffff',
+        ]
+        for s in strings:
+            # Call PyUnicode_AsUTF8AndSize() which creates the UTF-8
+            # encoded string cached in the Unicode object.
+            asutf8andsize(s, 0)
+            b = s.encode()
+            self.assertEqual(equaltoutf8(s, b), 1)  # Use the UTF-8 cache.
+            s2 = b.decode()  # New Unicode object without the UTF-8 cache.
+            self.assertEqual(equaltoutf8(s2, b), 1)
+            self.assertEqual(equaltoutf8(s + 'x', b + b'x'), 1)
+            self.assertEqual(equaltoutf8(s + 'x', b + b'y'), 0)
+            self.assertEqual(equaltoutf8(s, b + b'\0'), 1)
+            self.assertEqual(equaltoutf8(s2, b + b'\0'), 1)
+            self.assertEqual(equaltoutf8(s + '\0', b + b'\0'), 0)
+            self.assertEqual(equaltoutf8(s + '\0', b), 0)
+            self.assertEqual(equaltoutf8(s2, b + b'x'), 0)
+            self.assertEqual(equaltoutf8(s2, b[:-1]), 0)
+            self.assertEqual(equaltoutf8(s2, b[:-1] + b'x'), 0)
+
+        self.assertEqual(equaltoutf8('', b''), 1)
+        self.assertEqual(equaltoutf8('', b'\0'), 1)
+
+        # embedded null chars/bytes
+        self.assertEqual(equaltoutf8('abc', b'abc\0def\0'), 1)
+        self.assertEqual(equaltoutf8('a\0bc', b'abc'), 0)
+        self.assertEqual(equaltoutf8('abc', b'a\0bc'), 0)
+
+        # Surrogate characters are always treated as not equal
+        self.assertEqual(equaltoutf8('\udcfe',
+                            '\udcfe'.encode("utf8", "surrogateescape")), 0)
+        self.assertEqual(equaltoutf8('\udcfe',
+                            '\udcfe'.encode("utf8", "surrogatepass")), 0)
+        self.assertEqual(equaltoutf8('\ud801',
+                            '\ud801'.encode("utf8", "surrogatepass")), 0)
+
+    @support.cpython_only
+    @unittest.skipIf(_testcapi is None, 'need _testcapi module')
+    def test_equaltoutf8andsize(self):
+        # Test PyUnicode_EqualToUTF8AndSize()
+        from _testcapi import unicode_equaltoutf8andsize as equaltoutf8andsize
+        from _testcapi import unicode_asutf8andsize as asutf8andsize
+
+        strings = [
+            'abc', '\xa1\xa2\xa3', '\u4f60\u597d\u4e16',
+            '\U0001f600\U0001f601\U0001f602',
+            '\U0010ffff',
+        ]
+        for s in strings:
+            # Call PyUnicode_AsUTF8AndSize() which creates the UTF-8
+            # encoded string cached in the Unicode object.
+            asutf8andsize(s, 0)
+            b = s.encode()
+            self.assertEqual(equaltoutf8andsize(s, b), 1)  # Use the UTF-8 cache.
+            s2 = b.decode()  # New Unicode object without the UTF-8 cache.
+            self.assertEqual(equaltoutf8andsize(s2, b), 1)
+            self.assertEqual(equaltoutf8andsize(s + 'x', b + b'x'), 1)
+            self.assertEqual(equaltoutf8andsize(s + 'x', b + b'y'), 0)
+            self.assertEqual(equaltoutf8andsize(s, b + b'\0'), 0)
+            self.assertEqual(equaltoutf8andsize(s2, b + b'\0'), 0)
+            self.assertEqual(equaltoutf8andsize(s + '\0', b + b'\0'), 1)
+            self.assertEqual(equaltoutf8andsize(s + '\0', b), 0)
+            self.assertEqual(equaltoutf8andsize(s2, b + b'x'), 0)
+            self.assertEqual(equaltoutf8andsize(s2, b[:-1]), 0)
+            self.assertEqual(equaltoutf8andsize(s2, b[:-1] + b'x'), 0)
+            # Not null-terminated,
+            self.assertEqual(equaltoutf8andsize(s, b + b'x', len(b)), 1)
+            self.assertEqual(equaltoutf8andsize(s2, b + b'x', len(b)), 1)
+            self.assertEqual(equaltoutf8andsize(s + '\0', b + b'\0x', len(b) + 1), 1)
+            self.assertEqual(equaltoutf8andsize(s2, b, len(b) - 1), 0)
+
+        self.assertEqual(equaltoutf8andsize('', b''), 1)
+        self.assertEqual(equaltoutf8andsize('', b'\0'), 0)
+        self.assertEqual(equaltoutf8andsize('', b'x', 0), 1)
+
+        # embedded null chars/bytes
+        self.assertEqual(equaltoutf8andsize('abc\0def', b'abc\0def'), 1)
+        self.assertEqual(equaltoutf8andsize('abc\0def\0', b'abc\0def\0'), 1)
+
+        # Surrogate characters are always treated as not equal
+        self.assertEqual(equaltoutf8andsize('\udcfe',
+                            '\udcfe'.encode("utf8", "surrogateescape")), 0)
+        self.assertEqual(equaltoutf8andsize('\udcfe',
+                            '\udcfe'.encode("utf8", "surrogatepass")), 0)
+        self.assertEqual(equaltoutf8andsize('\ud801',
+                            '\ud801'.encode("utf8", "surrogatepass")), 0)
+
+        def check_not_equal_encoding(text, encoding):
+            self.assertEqual(equaltoutf8andsize(text, text.encode(encoding)), 0)
+            self.assertNotEqual(text.encode(encoding), text.encode("utf8"))
+
+        # Strings encoded to other encodings are not equal to expected UTF8-encoding string
+        check_not_equal_encoding('Stéphane', 'latin1')
+        check_not_equal_encoding('Stéphane', 'utf-16-le')  # embedded null characters
+        check_not_equal_encoding('北京市', 'gbk')
+
+        # CRASHES equaltoutf8andsize('abc', b'abc', -1)
+        # CRASHES equaltoutf8andsize(b'abc', b'abc')
+        # CRASHES equaltoutf8andsize([], b'abc')
+        # CRASHES equaltoutf8andsize(NULL, b'abc')
+        # CRASHES equaltoutf8andsize('abc', NULL)
+
     @support.cpython_only
     @unittest.skipIf(_testcapi is None, 'need _testcapi module')
     def test_richcompare(self):

Lib/test/test_capi/test_unicode.py

@@ -0,0 +1 @@
+Add :c:func:`PyUnicode_EqualToUTF8AndSize` and :c:func:`PyUnicode_EqualToUTF8` functions.

Lib/test/test_stable_abi_ctypes.py

@@ -2462,3 +2462,7 @@
     added = '3.13'
 [function.Py_IsFinalizing]
     added = '3.13'
+[function.PyUnicode_EqualToUTF8]
+    added = '3.13'
+[function.PyUnicode_EqualToUTF8AndSize]
+    added = '3.13'

Misc/NEWS.d/next/C API/2023-10-03-19-01-20.gh-issue-110289.YBIHEz.rst

@@ -1429,6 +1429,48 @@ unicode_comparewithasciistring(PyObject *self, PyObject *args)
     return PyLong_FromLong(result);
 }
 
+/* Test PyUnicode_EqualToUTF8() */
+static PyObject *
+unicode_equaltoutf8(PyObject *self, PyObject *args)
+{
+    PyObject *left;
+    const char *right = NULL;
+    Py_ssize_t right_len;
+    int result;
+
+    if (!PyArg_ParseTuple(args, "Oz#", &left, &right, &right_len)) {
+        return NULL;
+    }
+
+    NULLABLE(left);
+    result = PyUnicode_EqualToUTF8(left, right);
+    assert(!PyErr_Occurred());
+    return PyLong_FromLong(result);
+}
+
+/* Test PyUnicode_EqualToUTF8AndSize() */
+static PyObject *
+unicode_equaltoutf8andsize(PyObject *self, PyObject *args)
+{
+    PyObject *left;
+    const char *right = NULL;
+    Py_ssize_t right_len;
+    Py_ssize_t size = -100;
+    int result;
+
+    if (!PyArg_ParseTuple(args, "Oz#|n", &left, &right, &right_len, &size)) {
+        return NULL;
+    }
+
+    NULLABLE(left);
+    if (size == -100) {
+        size = right_len;
+    }
+    result = PyUnicode_EqualToUTF8AndSize(left, right, size);
+    assert(!PyErr_Occurred());
+    return PyLong_FromLong(result);
+}
+
 /* Test PyUnicode_RichCompare() */
 static PyObject *
 unicode_richcompare(PyObject *self, PyObject *args)
@@ -2044,6 +2086,8 @@ static PyMethodDef TestMethods[] = {
     {"unicode_replace",          unicode_replace,                METH_VARARGS},
     {"unicode_compare",          unicode_compare,                METH_VARARGS},
     {"unicode_comparewithasciistring",unicode_comparewithasciistring,METH_VARARGS},
+    {"unicode_equaltoutf8",      unicode_equaltoutf8,            METH_VARARGS},
+    {"unicode_equaltoutf8andsize",unicode_equaltoutf8andsize,    METH_VARARGS},
     {"unicode_richcompare",      unicode_richcompare,            METH_VARARGS},
     {"unicode_format",           unicode_format,                 METH_VARARGS},
     {"unicode_contains",         unicode_contains,               METH_VARARGS},

Misc/stable_abi.toml

@@ -10673,6 +10673,82 @@ PyUnicode_CompareWithASCIIString(PyObject* uni, const char* str)
     }
 }
 
+int
+PyUnicode_EqualToUTF8(PyObject *unicode, const char *str)
+{
+    return PyUnicode_EqualToUTF8AndSize(unicode, str, strlen(str));
+}
+
+int
+PyUnicode_EqualToUTF8AndSize(PyObject *unicode, const char *str, Py_ssize_t size)
+{
+    assert(_PyUnicode_CHECK(unicode));
+    assert(str);
+
+    if (PyUnicode_IS_ASCII(unicode)) {
+        Py_ssize_t len = PyUnicode_GET_LENGTH(unicode);
+        return size == len &&
+            memcmp(PyUnicode_1BYTE_DATA(unicode), str, len) == 0;
+    }
+    if (PyUnicode_UTF8(unicode) != NULL) {
+        Py_ssize_t len = PyUnicode_UTF8_LENGTH(unicode);
+        return size == len &&
+            memcmp(PyUnicode_UTF8(unicode), str, len) == 0;
+    }
+
+    Py_ssize_t len = PyUnicode_GET_LENGTH(unicode);
+    if ((size_t)len >= (size_t)size || (size_t)len < (size_t)size / 4) {
+        return 0;
+    }
+    const unsigned char *s = (const unsigned char *)str;
+    const unsigned char *ends = s + (size_t)size;
+    int kind = PyUnicode_KIND(unicode);
+    const void *data = PyUnicode_DATA(unicode);
+    /* Compare Unicode string and UTF-8 string */
+    for (Py_ssize_t i = 0; i < len; i++) {
+        Py_UCS4 ch = PyUnicode_READ(kind, data, i);
+        if (ch < 0x80) {
+            if (ends == s || s[0] != ch) {
+                return 0;
+            }
+            s += 1;
+        }
+        else if (ch < 0x800) {
+            if ((ends - s) < 2 ||
+                s[0] != (0xc0 | (ch >> 6)) ||
+                s[1] != (0x80 | (ch & 0x3f)))
+            {
+                return 0;
+            }
+            s += 2;
+        }
+        else if (ch < 0x10000) {
+            if (Py_UNICODE_IS_SURROGATE(ch) ||
+                (ends - s) < 3 ||
+                s[0] != (0xe0 | (ch >> 12)) ||
+                s[1] != (0x80 | ((ch >> 6) & 0x3f)) ||
+                s[2] != (0x80 | (ch & 0x3f)))
+            {
+                return 0;
+            }
+            s += 3;
+        }
+        else {
+            assert(ch <= MAX_UNICODE);
+            if ((ends - s) < 4 ||
+                s[0] != (0xf0 | (ch >> 18)) ||
+                s[1] != (0x80 | ((ch >> 12) & 0x3f)) ||
+                s[2] != (0x80 | ((ch >> 6) & 0x3f)) ||
+                s[3] != (0x80 | (ch & 0x3f)))
+            {
+                return 0;
+            }
+            s += 4;
+        }
+    }
+    return s == ends;
+}
+
 int
 _PyUnicode_EqualToASCIIString(PyObject *unicode, const char *str)
 {