bpo-33738: Fix macros which contradict PEP 384 (GH-7477)

During development of the limited API support for PySide,
we saw an error in a macro that accessed a type field.

This patch fixes the 7 errors in the Python headers.
Macros which were not written as capitals were implemented
as function.

To do the necessary analysis again, a script was included that
parses all headers and looks for "->tp_" in serctions which can
be reached with active limited API.

It is easily possible to call this script as a test.

Error listing:

../../Include/objimpl.h:243
#define PyObject_IS_GC(o) (PyType_IS_GC(Py_TYPE(o)) && \
    (Py_TYPE(o)->tp_is_gc == NULL || Py_TYPE(o)->tp_is_gc(o)))
Action: commented only

../../Include/objimpl.h:362
#define PyType_SUPPORTS_WEAKREFS(t) ((t)->tp_weaklistoffset > 0)
Action: commented only

../../Include/objimpl.h:364
#define PyObject_GET_WEAKREFS_LISTPTR(o) \
    ((PyObject **) (((char *) (o)) + Py_TYPE(o)->tp_weaklistoffset))
Action: commented only

../../Include/pyerrors.h:143
#define PyExceptionClass_Name(x) \
     ((char *)(((PyTypeObject*)(x))->tp_name))
Action: implemented function

../../Include/abstract.h:593
#define PyIter_Check(obj) \
    ((obj)->ob_type->tp_iternext != NULL && \
     (obj)->ob_type->tp_iternext != &_PyObject_NextNotImplemented)
Action: implemented function

../../Include/abstract.h:713
#define PyIndex_Check(obj)                              \
    ((obj)->ob_type->tp_as_number != NULL &&            \
     (obj)->ob_type->tp_as_number->nb_index != NULL)
Action: implemented function

../../Include/abstract.h:924
#define PySequence_ITEM(o, i)\
    ( Py_TYPE(o)->tp_as_sequence->sq_item(o, i) )
Action: commented only
This commit is contained in:
Christian Tismer 2018-06-09 20:32:25 +02:00 committed by Ned Deily
parent 3f45f5da8e
commit ea62ce7f4f
8 changed files with 198 additions and 0 deletions

View File

@ -590,9 +590,16 @@ PyAPI_FUNC(PyObject *) PyObject_Format(PyObject *obj,
returns itself. */
PyAPI_FUNC(PyObject *) PyObject_GetIter(PyObject *);
/* Returns 1 if the object 'obj' provides iterator protocols, and 0 otherwise.
This function always succeeds. */
#ifndef Py_LIMITED_API
#define PyIter_Check(obj) \
((obj)->ob_type->tp_iternext != NULL && \
(obj)->ob_type->tp_iternext != &_PyObject_NextNotImplemented)
#else
PyAPI_FUNC(int) PyIter_Check(PyObject*);
#endif
/* Takes an iterator object and calls its tp_iternext slot,
returning the next value.
@ -710,9 +717,15 @@ PyAPI_FUNC(PyObject *) PyNumber_Xor(PyObject *o1, PyObject *o2);
This is the equivalent of the Python expression: o1 | o2. */
PyAPI_FUNC(PyObject *) PyNumber_Or(PyObject *o1, PyObject *o2);
/* Returns 1 if obj is an index integer (has the nb_index slot of the
tp_as_number structure filled in), and 0 otherwise. */
#ifndef Py_LIMITED_API
#define PyIndex_Check(obj) \
((obj)->ob_type->tp_as_number != NULL && \
(obj)->ob_type->tp_as_number->nb_index != NULL)
#else
PyAPI_FUNC(int) PyIndex_Check(PyObject *);
#endif
/* Returns the object 'o' converted to a Python int, or NULL with an exception
raised on failure. */
@ -921,8 +934,10 @@ PyAPI_FUNC(PyObject *) PySequence_Fast(PyObject *o, const char* m);
/* Assume tp_as_sequence and sq_item exist and that 'i' does not
need to be corrected for a negative index. */
#ifndef Py_LIMITED_API
#define PySequence_ITEM(o, i)\
( Py_TYPE(o)->tp_as_sequence->sq_item(o, i) )
#endif
/* Return a pointer to the underlying item array for
an object retured by PySequence_Fast */

View File

@ -240,8 +240,10 @@ PyAPI_FUNC(Py_ssize_t) _PyGC_CollectIfEnabled(void);
#define PyType_IS_GC(t) PyType_HasFeature((t), Py_TPFLAGS_HAVE_GC)
/* Test if an object has a GC head */
#ifndef Py_LIMITED_API
#define PyObject_IS_GC(o) (PyType_IS_GC(Py_TYPE(o)) && \
(Py_TYPE(o)->tp_is_gc == NULL || Py_TYPE(o)->tp_is_gc(o)))
#endif
PyAPI_FUNC(PyVarObject *) _PyObject_GC_Resize(PyVarObject *, Py_ssize_t);
#define PyObject_GC_Resize(type, op, n) \
@ -359,10 +361,12 @@ PyAPI_FUNC(void) PyObject_GC_Del(void *);
/* Test if a type supports weak references */
#ifndef Py_LIMITED_API
#define PyType_SUPPORTS_WEAKREFS(t) ((t)->tp_weaklistoffset > 0)
#define PyObject_GET_WEAKREFS_LISTPTR(o) \
((PyObject **) (((char *) (o)) + Py_TYPE(o)->tp_weaklistoffset))
#endif
#ifdef __cplusplus
}

View File

@ -140,8 +140,12 @@ PyAPI_FUNC(void) _PyErr_ChainExceptions(PyObject *, PyObject *, PyObject *);
#define PyExceptionInstance_Check(x) \
PyType_FastSubclass((x)->ob_type, Py_TPFLAGS_BASE_EXC_SUBCLASS)
#ifndef Py_LIMITED_API
#define PyExceptionClass_Name(x) \
((char *)(((PyTypeObject*)(x))->tp_name))
#else
PyAPI_FUNC(char *) PyExceptionClass_Name(PyObject*);
#endif
#define PyExceptionInstance_Class(x) ((PyObject*)((x)->ob_type))

View File

@ -0,0 +1,3 @@
Seven macro incompatibilities with the Limited API were fixed, and the
macros PyIter_Check, PyIndex_Check and PyExceptionClass_Name were added as
functions. A script for automatic macro checks was added.

View File

@ -1244,6 +1244,14 @@ PyNumber_Absolute(PyObject *o)
return type_error("bad operand type for abs(): '%.200s'", o);
}
#undef PyIndex_Check
int
PyIndex_Check(PyObject *obj)
{
return obj->ob_type->tp_as_number != NULL &&
obj->ob_type->tp_as_number->nb_index != NULL;
}
/* Return a Python int from the object item.
Raise TypeError if the result is not an int
or if the object cannot be interpreted as an index.
@ -2535,6 +2543,13 @@ PyObject_GetIter(PyObject *o)
}
}
#undef PyIter_Check
int PyIter_Check(PyObject *obj)
{
return obj->ob_type->tp_iternext != NULL &&
obj->ob_type->tp_iternext != &_PyObject_NextNotImplemented;
}
/* Return next item.
* If an error occurs, return NULL. PyErr_Occurred() will be true.
* If the iteration terminates normally, return NULL and clear the

View File

@ -342,6 +342,12 @@ PyException_SetContext(PyObject *self, PyObject *context)
Py_XSETREF(((PyBaseExceptionObject *)self)->context, context);
}
#undef PyExceptionClass_Name
char *
PyExceptionClass_Name(PyObject *ob)
{
return ((PyTypeObject*)ob)->tp_name;
}
static struct PyMemberDef BaseException_members[] = {
{"__suppress_context__", T_BOOL,

View File

@ -248,6 +248,7 @@ EXPORTS
PyExc_Warning=python38.PyExc_Warning DATA
PyExc_WindowsError=python38.PyExc_WindowsError DATA
PyExc_ZeroDivisionError=python38.PyExc_ZeroDivisionError DATA
PyExceptionClass_Name=python38.PyExceptionClass_Name
PyException_GetCause=python38.PyException_GetCause
PyException_GetContext=python38.PyException_GetContext
PyException_GetTraceback=python38.PyException_GetTraceback
@ -294,9 +295,11 @@ EXPORTS
PyImport_ImportModuleLevelObject=python38.PyImport_ImportModuleLevelObject
PyImport_ImportModuleNoBlock=python38.PyImport_ImportModuleNoBlock
PyImport_ReloadModule=python38.PyImport_ReloadModule
PyIndex_Check=python38.PyIndex_Check
PyInterpreterState_Clear=python38.PyInterpreterState_Clear
PyInterpreterState_Delete=python38.PyInterpreterState_Delete
PyInterpreterState_New=python38.PyInterpreterState_New
PyIter_Check=python38.PyIter_Check
PyIter_Next=python38.PyIter_Next
PyListIter_Type=python38.PyListIter_Type DATA
PyListRevIter_Type=python38.PyListRevIter_Type DATA

View File

@ -0,0 +1,148 @@
"""
pep384_macrocheck.py
This programm tries to locate errors in the relevant Python header
files where macros access type fields when they are reachable from
the limided API.
The idea is to search macros with the string "->tp_" in it.
When the macro name does not begin with an underscore,
then we have found a dormant error.
Christian Tismer
2018-06-02
"""
import sys
import os
import re
DEBUG = False
def dprint(*args, **kw):
if DEBUG:
print(*args, **kw)
def parse_headerfiles(startpath):
"""
Scan all header files which are reachable fronm Python.h
"""
search = "Python.h"
name = os.path.join(startpath, search)
if not os.path.exists(name):
raise ValueError("file {} was not found in {}\n"
"Please give the path to Python's include directory."
.format(search, startpath))
errors = 0
with open(name) as python_h:
while True:
line = python_h.readline()
if not line:
break
found = re.match(r'^\s*#\s*include\s*"(\w+\.h)"', line)
if not found:
continue
include = found.group(1)
dprint("Scanning", include)
name = os.path.join(startpath, include)
if not os.path.exists(name):
name = os.path.join(startpath, "../PC", include)
errors += parse_file(name)
return errors
def ifdef_level_gen():
"""
Scan lines for #ifdef and track the level.
"""
level = 0
ifdef_pattern = r"^\s*#\s*if" # covers ifdef and ifndef as well
endif_pattern = r"^\s*#\s*endif"
while True:
line = yield level
if re.match(ifdef_pattern, line):
level += 1
elif re.match(endif_pattern, line):
level -= 1
def limited_gen():
"""
Scan lines for Py_LIMITED_API yes(1) no(-1) or nothing (0)
"""
limited = [0] # nothing
unlimited_pattern = r"^\s*#\s*ifndef\s+Py_LIMITED_API"
limited_pattern = "|".join([
r"^\s*#\s*ifdef\s+Py_LIMITED_API",
r"^\s*#\s*(el)?if\s+!\s*defined\s*\(\s*Py_LIMITED_API\s*\)\s*\|\|",
r"^\s*#\s*(el)?if\s+defined\s*\(\s*Py_LIMITED_API"
])
else_pattern = r"^\s*#\s*else"
ifdef_level = ifdef_level_gen()
status = next(ifdef_level)
wait_for = -1
while True:
line = yield limited[-1]
new_status = ifdef_level.send(line)
dir = new_status - status
status = new_status
if dir == 1:
if re.match(unlimited_pattern, line):
limited.append(-1)
wait_for = status - 1
elif re.match(limited_pattern, line):
limited.append(1)
wait_for = status - 1
elif dir == -1:
# this must have been an endif
if status == wait_for:
limited.pop()
wait_for = -1
else:
# it could be that we have an elif
if re.match(limited_pattern, line):
limited.append(1)
wait_for = status - 1
elif re.match(else_pattern, line):
limited.append(-limited.pop()) # negate top
def parse_file(fname):
errors = 0
with open(fname) as f:
lines = f.readlines()
type_pattern = r"^.*?->\s*tp_"
define_pattern = r"^\s*#\s*define\s+(\w+)"
limited = limited_gen()
status = next(limited)
for nr, line in enumerate(lines):
status = limited.send(line)
line = line.rstrip()
dprint(fname, nr, status, line)
if status != -1:
if re.match(define_pattern, line):
name = re.match(define_pattern, line).group(1)
if not name.startswith("_"):
# found a candidate, check it!
macro = line + "\n"
idx = nr
while line.endswith("\\"):
idx += 1
line = lines[idx].rstrip()
macro += line + "\n"
if re.match(type_pattern, macro, re.DOTALL):
# this type field can reach the limited API
report(fname, nr + 1, macro)
errors += 1
return errors
def report(fname, nr, macro):
f = sys.stderr
print(fname + ":" + str(nr), file=f)
print(macro, file=f)
if __name__ == "__main__":
p = sys.argv[1] if sys.argv[1:] else "../../Include"
errors = parse_headerfiles(p)
if errors:
# somehow it makes sense to raise a TypeError :-)
raise TypeError("These {} locations contradict the limited API."
.format(errors))