directly call into CPython's Unicode writers to format float/int values,...

directly call into CPython's Unicode writers to format float/int values, rather than going all the way through PyObject_Format() and its method lookups

Cython/Compiler/ExprNodes.py

@@ -3114,12 +3114,15 @@ class FormattedValueNode(ExprNode):
             fn = self.find_conversion_func(conversion_char)
             assert fn is not None, "invalid conversion character found: '%s'" % conversion_char
             value_result = '%s(%s)' % (fn, value_result)
-            code.globalstate.use_utility_code(UtilityCode.load_cached("PyObjectFormatAndDecref", "StringTools.c"))
+            code.globalstate.use_utility_code(
+                UtilityCode.load_cached("PyObjectFormatAndDecref", "StringTools.c"))
             format_func += 'AndDecref'
-        elif not self.format_spec:
-            code.globalstate.use_utility_code(UtilityCode.load_cached("PyObjectFormatSimple", "StringTools.c"))
+        elif self.format_spec:
+            code.globalstate.use_utility_code(
+                UtilityCode.load_cached("PyObjectFormat", "StringTools.c"))
         else:
-            format_func = 'PyObject_Format'
+            code.globalstate.use_utility_code(
+                UtilityCode.load_cached("PyObjectFormatSimple", "StringTools.c"))
 
         code.putln("%s = %s(%s, %s); %s" % (
             self.result(),

Cython/Utility/ModuleSetupCode.c

@@ -45,6 +45,8 @@
   #define CYTHON_USE_PYLIST_INTERNALS 0
   #undef CYTHON_USE_UNICODE_INTERNALS
   #define CYTHON_USE_UNICODE_INTERNALS 0
+  #undef CYTHON_USE_UNICODE_WRITER
+  #define CYTHON_USE_UNICODE_WRITER 0
   #undef CYTHON_USE_PYLONG_INTERNALS
   #define CYTHON_USE_PYLONG_INTERNALS 0
   #undef CYTHON_AVOID_BORROWED_REFS
@@ -71,6 +73,8 @@
   #ifndef CYTHON_USE_UNICODE_INTERNALS
     #define CYTHON_USE_UNICODE_INTERNALS 1
   #endif
+  #undef CYTHON_USE_UNICODE_WRITER
+  #define CYTHON_USE_UNICODE_WRITER 0
   #undef CYTHON_USE_PYLONG_INTERNALS
   #define CYTHON_USE_PYLONG_INTERNALS 0
   #ifndef CYTHON_AVOID_BORROWED_REFS
@@ -111,6 +115,12 @@
   #ifndef CYTHON_USE_UNICODE_INTERNALS
     #define CYTHON_USE_UNICODE_INTERNALS 1
   #endif
+  #if PY_VERSION_HEX < 0x030300F0
+    #undef CYTHON_USE_UNICODE_WRITER
+    #define CYTHON_USE_UNICODE_WRITER 0
+  #elif !defined(CYTHON_USE_UNICODE_WRITER)
+    #define CYTHON_USE_UNICODE_WRITER 1
+  #endif
   #ifndef CYTHON_AVOID_BORROWED_REFS
     #define CYTHON_AVOID_BORROWED_REFS 0
   #endif

Cython/Utility/StringTools.c

@@ -974,6 +974,48 @@ static CYTHON_INLINE int __Pyx_PyByteArray_Append(PyObject* bytearray, int value
 }
 
 
+//////////////////// PyObjectFormat.proto ////////////////////
+
+#if CYTHON_USE_UNICODE_WRITER
+static PyObject* __Pyx_PyObject_Format(PyObject* s, PyObject* f);
+#else
+#define __Pyx_PyObject_Format(s, f) PyObject_Format(s, f)
+#endif
+
+//////////////////// PyObjectFormat ////////////////////
+
+#if CYTHON_USE_UNICODE_WRITER
+static PyObject* __Pyx_PyObject_Format(PyObject* obj, PyObject* format_spec) {
+    int ret;
+    _PyUnicodeWriter writer;
+
+    if (likely(PyFloat_CheckExact(obj))) {
+        // copied from CPython 3.5 "float__format__()" in floatobject.c
+        _PyUnicodeWriter_Init(&writer);
+        ret = _PyFloat_FormatAdvancedWriter(
+            &writer,
+            obj,
+            format_spec, 0, PyUnicode_GET_LENGTH(format_spec));
+    } else if (likely(PyLong_CheckExact(obj))) {
+        // copied from CPython 3.5 "long__format__()" in longobject.c
+        _PyUnicodeWriter_Init(&writer);
+        ret = _PyLong_FormatAdvancedWriter(
+            &writer,
+            obj,
+            format_spec, 0, PyUnicode_GET_LENGTH(format_spec));
+    } else {
+        return PyObject_Format(obj, format_spec);
+    }
+
+    if (unlikely(ret == -1)) {
+        _PyUnicodeWriter_Dealloc(&writer);
+        return NULL;
+    }
+    return _PyUnicodeWriter_Finish(&writer);
+}
+#endif
+
+
 //////////////////// PyObjectFormatSimple.proto ////////////////////
 
 #if CYTHON_COMPILING_IN_PYPY

tests/run/fstring.pyx

@@ -254,9 +254,9 @@ def format_strings(str s, unicode u):
     return a, b, c, d, e, f, g
 
 
-def format_str(str s1, str s2):
+def format_pystr(str s1, str s2):
     """
-    >>> a, b, c, d = format_str('abc', 'xyz')
+    >>> a, b, c, d = format_pystr('abc', 'xyz')
     >>> print(a)
     abcxyz
     >>> print(b)
@@ -275,3 +275,49 @@ def format_str(str s1, str s2):
     d = f"s{s1}u{s2}"
     assert isinstance(d, unicode), type(d)
     return a, b, c, d
+
+
+def raw_fstring(value):
+    """
+    >>> print(raw_fstring('abc'))
+    abc\\x61
+    """
+    return fr'{value}\x61'
+
+
+def format_repr(value):
+    """
+    >>> a, b = format_repr('abc')
+    >>> print('x{value!r}x'.format(value='abc'))
+    x'abc'x
+    >>> print('x{value!r:6}x'.format(value='abc'))
+    x'abc' x
+    >>> print(a)
+    x'abc'x
+    >>> print(b)
+    x'abc' x
+    """
+    a = f'x{value!r}x'
+    assert isinstance(a, unicode), type(a)
+    b = f'x{value!r:6}x'
+    assert isinstance(b, unicode), type(b)
+    return a, b
+
+
+def format_str(value):
+    """
+    >>> a, b = format_str('abc')
+    >>> print('x{value!s}x'.format(value='abc'))
+    xabcx
+    >>> print('x{value!s:6}x'.format(value='abc'))
+    xabc   x
+    >>> print(a)
+    xabcx
+    >>> print(b)
+    xabc   x
+    """
+    a = f'x{value!s}x'
+    assert isinstance(a, unicode), type(a)
+    b = f'x{value!s:6}x'
+    assert isinstance(b, unicode), type(b)
+    return a, b