Merge branch 'main' into increase-trashcan-overhead

markshannon · web-flow · commit c91af34f22ea · 2025-02-25T19:31:12.000Z
diff --git a/Doc/library/string.rst b/Doc/library/string.rst
@@ -319,14 +319,19 @@ non-empty format specification typically modifies the result.
 The general form of a *standard format specifier* is:
 
 .. productionlist:: format-spec
-   format_spec: [[`fill`]`align`][`sign`]["z"]["#"]["0"][`width`][`grouping_option`]["." `precision`][`type`]
+   format_spec: [`options`][`width_and_precision`][`type`]
+   options: [[`fill`]`align`][`sign`]["z"]["#"]["0"]
    fill: <any character>
    align: "<" | ">" | "=" | "^"
    sign: "+" | "-" | " "
+   width_and_precision: [`width_with_grouping`][`precision_with_grouping`]
+   width_with_grouping: [`width`][`grouping_option`]
+   precision_with_grouping: "." [`precision`]`grouping_option`
    width: `~python-grammar:digit`+
    grouping_option: "_" | ","
    precision: `~python-grammar:digit`+
-   type: "b" | "c" | "d" | "e" | "E" | "f" | "F" | "g" | "G" | "n" | "o" | "s" | "x" | "X" | "%"
+   type: "b" | "c" | "d" | "e" | "E" | "f" | "F" | "g"
+       : | "G" | "n" | "o" | "s" | "x" | "X" | "%"
 
 If a valid *align* value is specified, it can be preceded by a *fill*
 character that can be any character and defaults to a space if omitted.
@@ -458,6 +463,13 @@ indicates the maximum field size - in other words, how many characters will be
 used from the field content.  The *precision* is not allowed for integer
 presentation types.
 
+The ``'_'`` or ``','`` option after *precision* means the use of an underscore
+or a comma for a thousands separator of the fractional part for floating-point
+presentation types.
+
+.. versionchanged:: 3.14
+   Support thousands separators for the fractional part.
+
 Finally, the *type* determines how the data should be presented.
 
 The available string presentation types are:
@@ -704,10 +716,18 @@ Replacing ``%x`` and ``%o`` and converting the value to different bases::
    >>> "int: {0:d};  hex: {0:#x};  oct: {0:#o};  bin: {0:#b}".format(42)
    'int: 42;  hex: 0x2a;  oct: 0o52;  bin: 0b101010'
 
-Using the comma as a thousands separator::
+Using the comma or the underscore as a thousands separator::
 
    >>> '{:,}'.format(1234567890)
    '1,234,567,890'
+   >>> '{:_}'.format(1234567890)
+   '1_234_567_890'
+   >>> '{:_}'.format(123456789.123456789)
+   '123_456_789.12345679'
+   >>> '{:._}'.format(123456789.123456789)
+   '123456789.123_456_79'
+   >>> '{:_._}'.format(123456789.123456789)
+   '123_456_789.123_456_79'
 
 Expressing a percentage::
 
diff --git a/Doc/whatsnew/3.14.rst b/Doc/whatsnew/3.14.rst
@@ -336,6 +336,11 @@ Other language changes
   making it a :term:`generic type`.
   (Contributed by Brian Schubert in :gh:`126012`.)
 
+* Support underscore and comma as thousands separators in the fractional part
+  for floating-point presentation types of the new-style string formatting
+  (with :func:`format` or :ref:`f-strings`).
+  (Contrubuted by Sergey B Kirpichev in :gh:`87790`.)
+
 * ``\B`` in :mod:`regular expression <re>` now matches empty input string.
   Now it is always the opposite of ``\b``.
   (Contributed by Serhiy Storchaka in :gh:`124130`.)
diff --git a/Include/cpython/object.h b/Include/cpython/object.h
@@ -5,6 +5,7 @@
 PyAPI_FUNC(void) _Py_NewReference(PyObject *op);
 PyAPI_FUNC(void) _Py_NewReferenceNoTotal(PyObject *op);
 PyAPI_FUNC(void) _Py_ResurrectReference(PyObject *op);
+PyAPI_FUNC(void) _Py_ForgetReference(PyObject *op);
 
 #ifdef Py_REF_DEBUG
 /* These are useful as debugging aids when chasing down refleaks. */
diff --git a/Include/internal/pycore_object.h b/Include/internal/pycore_object.h
@@ -730,6 +730,9 @@ _PyObject_ResurrectStart(PyObject *op)
 #else
     Py_SET_REFCNT(op, 1);
 #endif
+#ifdef Py_TRACE_REFS
+    _Py_ResurrectReference(op);
+#endif
 }
 
 // Undoes an object resurrection by decrementing the refcount without calling
@@ -743,13 +746,22 @@ _PyObject_ResurrectEnd(PyObject *op)
 #endif
 #ifndef Py_GIL_DISABLED
     Py_SET_REFCNT(op, Py_REFCNT(op) - 1);
-    return Py_REFCNT(op) != 0;
+    if (Py_REFCNT(op) == 0) {
+# ifdef Py_TRACE_REFS
+        _Py_ForgetReference(op);
+# endif
+        return 0;
+    }
+    return 1;
 #else
     uint32_t local = _Py_atomic_load_uint32_relaxed(&op->ob_ref_local);
     Py_ssize_t shared = _Py_atomic_load_ssize_acquire(&op->ob_ref_shared);
     if (_Py_IsOwnedByCurrentThread(op) && local == 1 && shared == 0) {
         // Fast-path: object has a single refcount and is owned by this thread
         _Py_atomic_store_uint32_relaxed(&op->ob_ref_local, 0);
+# ifdef Py_TRACE_REFS
+        _Py_ForgetReference(op);
+# endif
         return 0;
     }
     // Slow-path: object has a shared refcount or is not owned by this thread
diff --git a/Include/internal/pycore_unicodeobject.h b/Include/internal/pycore_unicodeobject.h
@@ -246,7 +246,8 @@ extern Py_ssize_t _PyUnicode_InsertThousandsGrouping(
     Py_ssize_t min_width,
     const char *grouping,
     PyObject *thousands_sep,
-    Py_UCS4 *maxchar);
+    Py_UCS4 *maxchar,
+    int forward);
 
 /* --- Misc functions ----------------------------------------------------- */
 
diff --git a/Include/pythonrun.h b/Include/pythonrun.h
@@ -28,7 +28,7 @@ PyAPI_DATA(int) (*PyOS_InputHook)(void);
 #if defined(_Py_ADDRESS_SANITIZER) || defined(_Py_THREAD_SANITIZER)
 #  define PYOS_STACK_MARGIN 4096
 #elif defined(Py_DEBUG) && defined(WIN32)
-#  define PYOS_STACK_MARGIN 3072
+#  define PYOS_STACK_MARGIN 4096
 #elif defined(__wasi__)
    /* Web assembly has two stacks, so this isn't really a size */
 #  define PYOS_STACK_MARGIN 500
diff --git a/Lib/test/libregrtest/tsan.py b/Lib/test/libregrtest/tsan.py
@@ -6,6 +6,7 @@
     'test_capi.test_mem',
     'test_capi.test_pyatomic',
     'test_code',
+    'test_concurrent_futures',
     'test_enum',
     'test_functools',
     'test_httpservers',
diff --git a/Lib/test/test_concurrent_futures/util.py b/Lib/test/test_concurrent_futures/util.py
@@ -74,6 +74,7 @@ class ThreadPoolMixin(ExecutorMixin):
     executor_type = futures.ThreadPoolExecutor
 
 
+@support.skip_if_sanitizer("gh-129824: data races in InterpreterPool tests", thread=True)
 class InterpreterPoolMixin(ExecutorMixin):
     executor_type = futures.InterpreterPoolExecutor
 
diff --git a/Lib/test/test_float.py b/Lib/test/test_float.py
@@ -754,6 +754,28 @@ def test_format(self):
         self.assertEqual(format(INF, 'f'), 'inf')
         self.assertEqual(format(INF, 'F'), 'INF')
 
+        # thousands separators
+        x = 123_456.123_456
+        self.assertEqual(format(x, '_f'), '123_456.123456')
+        self.assertEqual(format(x, ',f'), '123,456.123456')
+        self.assertEqual(format(x, '._f'), '123456.123_456')
+        self.assertEqual(format(x, '.,f'), '123456.123,456')
+        self.assertEqual(format(x, '_._f'), '123_456.123_456')
+        self.assertEqual(format(x, ',.,f'), '123,456.123,456')
+        self.assertEqual(format(x, '.10_f'), '123456.123_456_000_0')
+        self.assertEqual(format(x, '.10,f'), '123456.123,456,000,0')
+        self.assertEqual(format(x, '>21._f'), '       123456.123_456')
+        self.assertEqual(format(x, '<21._f'), '123456.123_456       ')
+        self.assertEqual(format(x, '+.11_e'), '+1.234_561_234_56e+05')
+        self.assertEqual(format(x, '+.11,e'), '+1.234,561,234,56e+05')
+
+        self.assertRaises(ValueError, format, x, '._6f')
+        self.assertRaises(ValueError, format, x, '.,_f')
+        self.assertRaises(ValueError, format, x, '.6,_f')
+        self.assertRaises(ValueError, format, x, '.6_,f')
+        self.assertRaises(ValueError, format, x, '.6_n')
+        self.assertRaises(ValueError, format, x, '.6,n')
+
     @support.requires_IEEE_754
     def test_format_testfile(self):
         with open(format_testfile, encoding="utf-8") as testfile:
diff --git a/Lib/test/test_format.py b/Lib/test/test_format.py
@@ -515,11 +515,15 @@ def test_with_a_commas_and_an_underscore_in_format_specifier(self):
         error_msg = re.escape("Cannot specify both ',' and '_'.")
         with self.assertRaisesRegex(ValueError, error_msg):
             '{:,_}'.format(1)
+        with self.assertRaisesRegex(ValueError, error_msg):
+            '{:.,_f}'.format(1.1)
 
     def test_with_an_underscore_and_a_comma_in_format_specifier(self):
         error_msg = re.escape("Cannot specify both ',' and '_'.")
         with self.assertRaisesRegex(ValueError, error_msg):
             '{:_,}'.format(1)
+        with self.assertRaisesRegex(ValueError, error_msg):
+            '{:._,f}'.format(1.1)
 
     def test_better_error_message_format(self):
         # https://bugs.python.org/issue20524
diff --git a/Lib/test/test_sys_setprofile.py b/Lib/test/test_sys_setprofile.py
@@ -493,6 +493,24 @@ class A:
         # The last c_call is the call to sys.setprofile
         self.assertEqual(events, ['c_call', 'c_return', 'c_call'])
 
+        class B:
+            f = classmethod(max)
+        events = []
+        sys.setprofile(lambda frame, event, args: events.append(event))
+        # Not important, we only want to trigger INSTRUMENTED_CALL_KW
+        B().f(1, key=lambda x: 0)
+        sys.setprofile(None)
+        # The last c_call is the call to sys.setprofile
+        self.assertEqual(
+            events,
+            ['c_call',
+             'call', 'return',
+             'call', 'return',
+             'c_return',
+             'c_call'
+            ]
+        )
+
 
 if __name__ == "__main__":
     unittest.main()
diff --git a/Misc/NEWS.d/next/Core_and_Builtins/2024-10-11-10-41-05.gh-issue-87790.mlfEGl.rst b/Misc/NEWS.d/next/Core_and_Builtins/2024-10-11-10-41-05.gh-issue-87790.mlfEGl.rst
@@ -0,0 +1,3 @@
+Support underscore and comma as thousands separators in the fractional part for
+floating-point presentation types of the new-style string formatting (with
+:func:`format` or :ref:`f-strings`).  Patch by Sergey B Kirpichev.
diff --git a/Misc/NEWS.d/next/Core_and_Builtins/2025-02-23-20-48-31.gh-issue-122029.iW8GvA.rst b/Misc/NEWS.d/next/Core_and_Builtins/2025-02-23-20-48-31.gh-issue-122029.iW8GvA.rst
@@ -0,0 +1 @@
+``INSTRUMENTED_CALL_KW`` will expand the method before monitoring to reflect the actual behavior more accurately.
diff --git a/Objects/object.c b/Objects/object.c
@@ -496,10 +496,22 @@ _PyObject_ResurrectEndSlow(PyObject *op)
         // merge the refcount. This isn't necessary in all cases, but it
         // simplifies the implementation.
         Py_ssize_t refcount = _Py_ExplicitMergeRefcount(op, -1);
-        return refcount != 0;
+        if (refcount == 0) {
+#ifdef Py_TRACE_REFS
+            _Py_ForgetReference(op);
+#endif
+            return 0;
+        }
+        return 1;
     }
     int is_dead = _Py_DecRefSharedIsDead(op, NULL, 0);
-    return !is_dead;
+    if (is_dead) {
+#ifdef Py_TRACE_REFS
+        _Py_ForgetReference(op);
+#endif
+        return 0;
+    }
+    return 1;
 }
 
 
@@ -589,20 +601,24 @@ PyObject_CallFinalizerFromDealloc(PyObject *self)
                               Py_REFCNT(self) > 0,
                               "refcount is too small");
 
+    _PyObject_ASSERT(self,
+                    (!_PyType_IS_GC(Py_TYPE(self))
+                    || _PyObject_GC_IS_TRACKED(self)));
+
     /* Undo the temporary resurrection; can't use DECREF here, it would
      * cause a recursive call. */
-    if (!_PyObject_ResurrectEnd(self)) {
-        return 0;         /* this is the normal path out */
+    if (_PyObject_ResurrectEnd(self)) {
+        /* tp_finalize resurrected it!
+           gh-130202: Note that the object may still be dead in the free
+           threaded build in some circumstances, so it's not safe to access
+           `self` after this point. For example, the last reference to the
+           resurrected `self` may be held by another thread, which can
+           concurrently deallocate it. */
+        return -1;
     }
 
-    /* tp_finalize resurrected it!  Make it look like the original Py_DECREF
-     * never happened. */
-    _Py_ResurrectReference(self);
-
-    _PyObject_ASSERT(self,
-                     (!_PyType_IS_GC(Py_TYPE(self))
-                      || _PyObject_GC_IS_TRACKED(self)));
-    return -1;
+    /* this is the normal path out, the caller continues with deallocation. */
+    return 0;
 }
 
 int
@@ -2601,11 +2617,10 @@ _Py_ResurrectReference(PyObject *op)
 #endif
 }
 
-
-#ifdef Py_TRACE_REFS
 void
 _Py_ForgetReference(PyObject *op)
 {
+#ifdef Py_TRACE_REFS
     if (Py_REFCNT(op) < 0) {
         _PyObject_ASSERT_FAILED_MSG(op, "negative refcnt");
     }
@@ -2621,8 +2636,11 @@ _Py_ForgetReference(PyObject *op)
 #endif
 
     _PyRefchain_Remove(interp, op);
+#endif
 }
 
+
+#ifdef Py_TRACE_REFS
 static int
 _Py_PrintReference(_Py_hashtable_t *ht,
                    const void *key, const void *value,
diff --git a/Objects/stringlib/localeutil.h b/Objects/stringlib/localeutil.h
@@ -47,7 +47,7 @@ InsertThousandsGrouping_fill(_PyUnicodeWriter *writer, Py_ssize_t *buffer_pos,
                              PyObject *digits, Py_ssize_t *digits_pos,
                              Py_ssize_t n_chars, Py_ssize_t n_zeros,
                              PyObject *thousands_sep, Py_ssize_t thousands_sep_len,
-                             Py_UCS4 *maxchar)
+                             Py_UCS4 *maxchar, int forward)
 {
     if (!writer) {
         /* if maxchar > 127, maxchar is already set */
@@ -59,24 +59,39 @@ InsertThousandsGrouping_fill(_PyUnicodeWriter *writer, Py_ssize_t *buffer_pos,
     }
 
     if (thousands_sep) {
-        *buffer_pos -= thousands_sep_len;
-
+        if (!forward) {
+            *buffer_pos -= thousands_sep_len;
+        }
         /* Copy the thousands_sep chars into the buffer. */
         _PyUnicode_FastCopyCharacters(writer->buffer, *buffer_pos,
                                       thousands_sep, 0,
                                       thousands_sep_len);
+        if (forward) {
+            *buffer_pos += thousands_sep_len;
+        }
     }
 
-    *buffer_pos -= n_chars;
-    *digits_pos -= n_chars;
+    if (!forward) {
+        *buffer_pos -= n_chars;
+        *digits_pos -= n_chars;
+    }
     _PyUnicode_FastCopyCharacters(writer->buffer, *buffer_pos,
                                   digits, *digits_pos,
                                   n_chars);
+    if (forward) {
+        *buffer_pos += n_chars;
+        *digits_pos += n_chars;
+    }
 
     if (n_zeros) {
-        *buffer_pos -= n_zeros;
+        if (!forward) {
+            *buffer_pos -= n_zeros;
+        }
         int kind = PyUnicode_KIND(writer->buffer);
         void *data = PyUnicode_DATA(writer->buffer);
         unicode_fill(kind, data, '0', *buffer_pos, n_zeros);
+        if (forward) {
+            *buffer_pos += n_zeros;
+        }
     }
 }
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
diff --git a/Python/bytecodes.c b/Python/bytecodes.c
diff --git a/Python/formatter_unicode.c b/Python/formatter_unicode.c
diff --git a/Python/generated_cases.c.h b/Python/generated_cases.c.h
