Add PyUnicode_EqualToUTF8() function (python#78)

vstinner · web-flow · commit 99ab0d3767b2 · 2023-10-12T12:00:37.000+02:00
Add PyUnicode_EqualToUTF8() and PyUnicode_EqualToUTF8AndSize()
functions.
diff --git a/docs/api.rst b/docs/api.rst
@@ -113,6 +113,14 @@ Python 3.13
 
    Available on Python 3.5.2 and newer.
 
+.. c:function:: int PyUnicode_EqualToUTF8(PyObject *unicode, const char *str)
+
+   See `PyUnicode_EqualToUTF8() documentation <https://docs.python.org/dev/c-api/unicode.html#c.PyUnicode_EqualToUTF8>`__.
+
+.. c:function:: int PyUnicode_EqualToUTF8AndSize(PyObject *unicode, const char *str, Py_ssize_t size)
+
+   See `PyUnicode_EqualToUTF8AndSize() documentation <https://docs.python.org/dev/c-api/unicode.html#c.PyUnicode_EqualToUTF8AndSize>`__.
+
 
 Python 3.12
 -----------
diff --git a/docs/changelog.rst b/docs/changelog.rst
@@ -1,6 +1,11 @@
 Changelog
 =========
 
+* 2023-10-04: Add functions:
+
+  * ``PyUnicode_EqualToUTF8()``
+  * ``PyUnicode_EqualToUTF8AndSize()``
+
 * 2023-10-03: Add functions:
 
   * ``PyObject_VisitManagedDict()``
diff --git a/pythoncapi_compat.h b/pythoncapi_compat.h
@@ -939,6 +939,79 @@ PyThreadState_GetUnchecked(void)
 }
 #endif
 
+// gh-110289 added PyUnicode_EqualToUTF8() and PyUnicode_EqualToUTF8AndSize()
+// to Python 3.13.0a1
+#if PY_VERSION_HEX < 0x030D00A1
+static inline int
+PyUnicode_EqualToUTF8AndSize(PyObject *unicode, const char *str, Py_ssize_t str_len)
+{
+    Py_ssize_t len;
+    const void *utf8;
+    PyObject *exc_type, *exc_value, *exc_tb;
+    int res;
+
+    // API cannot report errors so save/restore the exception
+    PyErr_Fetch(&exc_type, &exc_value, &exc_tb);
+
+    // Python 3.3.0a1 added PyUnicode_AsUTF8AndSize()
+#if PY_VERSION_HEX >= 0x030300A1
+    if (PyUnicode_IS_ASCII(unicode)) {
+        utf8 = PyUnicode_DATA(unicode);
+        len = PyUnicode_GET_LENGTH(unicode);
+    }
+    else {
+        utf8 = PyUnicode_AsUTF8AndSize(unicode, &len);
+        if (utf8 == NULL) {
+            // Memory allocation failure. The API cannot report error,
+            // so ignore the exception and return 0.
+            res = 0;
+            goto done;
+        }
+    }
+
+    if (len != str_len) {
+        res = 0;
+        goto done;
+    }
+    res = (memcmp(utf8, str, (size_t)len) == 0);
+#else
+    PyObject *bytes = PyUnicode_AsUTF8String(unicode);
+    if (bytes == NULL) {
+        // Memory allocation failure. The API cannot report error,
+        // so ignore the exception and return 0.
+        res = 0;
+        goto done;
+    }
+
+#if PY_VERSION_HEX >= 0x03000000
+    len = PyBytes_GET_SIZE(bytes);
+    utf8 = PyBytes_AS_STRING(bytes);
+#else
+    len = PyString_GET_SIZE(bytes);
+    utf8 = PyString_AS_STRING(bytes);
+#endif
+    if (len != str_len) {
+        Py_DECREF(bytes);
+        res = 0;
+        goto done;
+    }
+
+    res = (memcmp(utf8, str, (size_t)len) == 0);
+    Py_DECREF(bytes);
+#endif
+
+done:
+    PyErr_Restore(exc_type, exc_value, exc_tb);
+    return res;
+}
+
+static inline int
+PyUnicode_EqualToUTF8(PyObject *unicode, const char *str)
+{
+    return PyUnicode_EqualToUTF8AndSize(unicode, str, (Py_ssize_t)strlen(str));
+}
+#endif
+
 
 #ifdef __cplusplus
 }
diff --git a/tests/test_pythoncapi_compat_cext.c b/tests/test_pythoncapi_compat_cext.c
@@ -1361,6 +1361,40 @@ test_managed_dict(PyObject *Py_UNUSED(module), PyObject *Py_UNUSED(args))
 #endif  // PY_VERSION_HEX >= 0x030B00A3
 
 
+static PyObject *
+test_unicode(PyObject *Py_UNUSED(module), PyObject *Py_UNUSED(args))
+{
+    PyObject *abc = PyUnicode_FromString("abc");
+    if (abc == NULL) {
+        return NULL;
+    }
+
+    PyObject *abc0def = PyUnicode_FromStringAndSize("abc\0def", 7);
+    if (abc == NULL) {
+        return NULL;
+    }
+
+    // PyUnicode_EqualToUTF8() and PyUnicode_EqualToUTF8AndSize() can be called
+    // with an exception raised and they must not clear the current exception.
+    PyErr_NoMemory();
+
+    assert(PyUnicode_EqualToUTF8AndSize(abc, "abc", 3) == 1);
+    assert(PyUnicode_EqualToUTF8AndSize(abc, "Python", 6) == 0);
+    assert(PyUnicode_EqualToUTF8AndSize(abc0def, "abc\0def", 7) == 1);
+
+    assert(PyUnicode_EqualToUTF8(abc, "abc") == 1);
+    assert(PyUnicode_EqualToUTF8(abc, "Python") == 0);
+    assert(PyUnicode_EqualToUTF8(abc0def, "abc\0def") == 0);
+
+    assert(PyErr_ExceptionMatches(PyExc_MemoryError));
+    PyErr_Clear();
+
+    Py_DECREF(abc);
+    Py_DECREF(abc0def);
+    Py_RETURN_NONE;
+}
+
+
 static struct PyMethodDef methods[] = {
     {"test_object", test_object, METH_NOARGS, _Py_NULL},
     {"test_py_is", test_py_is, METH_NOARGS, _Py_NULL},
@@ -1390,6 +1424,7 @@ static struct PyMethodDef methods[] = {
 #ifdef TEST_MANAGED_DICT
     {"test_managed_dict", test_managed_dict, METH_NOARGS, _Py_NULL},
 #endif
+    {"test_unicode", test_unicode, METH_NOARGS, _Py_NULL},
     {_Py_NULL, _Py_NULL, 0, _Py_NULL}
 };
 
