This PR adds the ability to enable the GIL if it was disabled at
interpreter startup, and modifies the multi-phase module initialization
path to enable the GIL when loading a module, unless that module's spec
includes a slot indicating it can run safely without the GIL.
PEP 703 called the constant for the slot `Py_mod_gil_not_used`; I went
with `Py_MOD_GIL_NOT_USED` for consistency with gh-104148.
A warning will be issued up to once per interpreter for the first
GIL-using module that is loaded. If `-v` is given, a shorter message
will be printed to stderr every time a GIL-using module is loaded
(including the first one that issues a warning).
Remove the following private functions from the public C API:
* _Py_CheckFunctionResult()
* _PyObject_CallMethod()
* _PyObject_CallMethodId()
* _PyObject_CallMethodIdNoArgs()
* _PyObject_CallMethodIdObjArgs()
* _PyObject_CallMethodIdOneArg()
* _PyObject_MakeTpCall()
* _PyObject_VectorcallMethodId()
* _PyStack_AsDict()
Move these functions to the internal C API (pycore_call.h).
No longer export the following functions:
* _PyObject_Call()
* _PyObject_CallMethod()
* _PyObject_CallMethodId()
* _PyObject_CallMethodIdObjArgs()
* _PyObject_Call_Prepend()
* _PyObject_FastCallDictTstate()
* _PyStack_AsDict()
The following functions are still exported for stdlib shared
extensions:
* _Py_CheckFunctionResult()
* _PyObject_MakeTpCall()
Mark the following internal functions as extern:
* _PyStack_UnpackDict()
* _PyStack_UnpackDict_Free()
* _PyStack_UnpackDict_FreeNoDecRef()
Here we are doing no more than adding the value for Py_mod_multiple_interpreters and using it for stdlib modules. We will start checking for it in gh-104206 (once PyInterpreterState.ceval.own_gil is added in gh-104204).
We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules.
The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings).
https://bugs.python.org/issue46541#msg411799 explains the rationale for this change.
The core of the change is in:
* (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros
* Include/internal/pycore_runtime_init.h - added the static initializers for the global strings
* Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState
* Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers
I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config.
The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *.
The following are not changed (yet):
* stop using _Py_IDENTIFIER() in the stdlib modules
* (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API
* (maybe) intern the strings during runtime init
https://bugs.python.org/issue46541
Fix invalid function cast warnings with gcc 8
for method conventions different from METH_NOARGS, METH_O and
METH_VARARGS excluding Argument Clinic generated code.
Remove aliases from the C module. Always implement bisect() and insort()
aliases in bisect.py
Remove also the "# backward compatibility" command, there is no plan to
deprecate nor remove these aliases. When keys are equal, it makes sense to use
bisect.bisect() and bisect.insort().
svn+ssh://pythondev@svn.python.org/python/trunk
........
r64722 | georg.brandl | 2008-07-05 12:13:36 +0200 (Sat, 05 Jul 2008) | 4 lines
#2663: support an *ignore* argument to shutil.copytree(). Patch by Tarek Ziade.
This is a new feature, but Barry authorized adding it in the beta period.
........
r64729 | mark.dickinson | 2008-07-05 13:33:52 +0200 (Sat, 05 Jul 2008) | 5 lines
Issue 3188: accept float('infinity') as well as float('inf'). This
makes the float constructor behave in the same way as specified
by various other language standards, including C99, IEEE 754r,
and the IBM Decimal standard.
........
r64753 | gregory.p.smith | 2008-07-06 05:35:58 +0200 (Sun, 06 Jul 2008) | 4 lines
- Issue #2862: Make int and float freelist management consistent with other
freelists. Changes their CompactFreeList apis into ClearFreeList apis and
calls them via gc.collect().
........
r64845 | raymond.hettinger | 2008-07-10 16:03:19 +0200 (Thu, 10 Jul 2008) | 1 line
Issue 3301: Bisect functions behaved badly when lo was negative.
........
r64846 | raymond.hettinger | 2008-07-10 16:34:57 +0200 (Thu, 10 Jul 2008) | 1 line
Issue 3285: Fractions from_float() and from_decimal() accept Integral arguments.
........
r64849 | andrew.kuchling | 2008-07-10 16:43:31 +0200 (Thu, 10 Jul 2008) | 1 line
Wording changes
........
r64871 | raymond.hettinger | 2008-07-11 14:00:21 +0200 (Fri, 11 Jul 2008) | 1 line
Add cautionary note on the use of PySequence_Fast_ITEMS.
........
r64880 | amaury.forgeotdarc | 2008-07-11 23:28:25 +0200 (Fri, 11 Jul 2008) | 5 lines
#3317 in zipfile module, restore the previous names of global variables:
some applications relied on them.
Also remove duplicated lines.
........
r64881 | amaury.forgeotdarc | 2008-07-11 23:45:06 +0200 (Fri, 11 Jul 2008) | 3 lines
#3342: In tracebacks, printed source lines were not indented since r62555.
#3343: Py_DisplaySourceLine should be a private function. Rename it to _Py_DisplaySourceLine.
........
r64882 | josiah.carlson | 2008-07-12 00:17:14 +0200 (Sat, 12 Jul 2008) | 2 lines
Fix for the AttributeError in test_asynchat.
........
r64885 | josiah.carlson | 2008-07-12 01:26:59 +0200 (Sat, 12 Jul 2008) | 2 lines
Fixed test for asyncore.
........
r64888 | matthias.klose | 2008-07-12 09:51:48 +0200 (Sat, 12 Jul 2008) | 2 lines
- Fix bashisms in Tools/faqwiz/move-faqwiz.sh
........
r64897 | benjamin.peterson | 2008-07-12 22:16:19 +0200 (Sat, 12 Jul 2008) | 1 line
fix various doc typos #3320
........
r64900 | alexandre.vassalotti | 2008-07-13 00:06:53 +0200 (Sun, 13 Jul 2008) | 2 lines
Fixed typo.
........
r64901 | benjamin.peterson | 2008-07-13 01:41:19 +0200 (Sun, 13 Jul 2008) | 1 line
#1778443 robotparser fixes from Aristotelis Mikropoulos
........
r64915 | nick.coghlan | 2008-07-13 16:52:36 +0200 (Sun, 13 Jul 2008) | 1 line
Fix issue 3221 by emitting a RuntimeWarning instead of raising SystemError when the parent module can't be found during an absolute import (likely due to non-PEP 361 aware code which sets a module level __package__ attribute)
........
r64926 | martin.v.loewis | 2008-07-13 22:31:49 +0200 (Sun, 13 Jul 2008) | 2 lines
Add turtle into the module index.
........
r64927 | alexandre.vassalotti | 2008-07-13 22:42:44 +0200 (Sun, 13 Jul 2008) | 3 lines
Issue #3274: Use a less common identifier for the temporary variable
in Py_CLEAR().
........
r64928 | andrew.kuchling | 2008-07-13 23:43:25 +0200 (Sun, 13 Jul 2008) | 1 line
Re-word
........
r64929 | andrew.kuchling | 2008-07-13 23:43:52 +0200 (Sun, 13 Jul 2008) | 1 line
Add various items; move ctypes items into a subsection of their own
........
r64938 | andrew.kuchling | 2008-07-14 02:35:32 +0200 (Mon, 14 Jul 2008) | 1 line
Typo fixes
........
r64939 | andrew.kuchling | 2008-07-14 02:40:55 +0200 (Mon, 14 Jul 2008) | 1 line
Typo fix
........
r64940 | andrew.kuchling | 2008-07-14 03:18:16 +0200 (Mon, 14 Jul 2008) | 1 line
Typo fix
........
r64941 | andrew.kuchling | 2008-07-14 03:18:31 +0200 (Mon, 14 Jul 2008) | 1 line
Expand the multiprocessing section
........
r64944 | gregory.p.smith | 2008-07-14 08:06:48 +0200 (Mon, 14 Jul 2008) | 7 lines
Fix posix.fork1() / os.fork1() to only call PyOS_AfterFork() in the child
process rather than both parent and child.
Does anyone actually use fork1()? It appears to be a Solaris thing
but if Python is built with pthreads on Solaris, fork1() and fork()
should be the same.
........
r64961 | jesse.noller | 2008-07-15 15:47:33 +0200 (Tue, 15 Jul 2008) | 1 line
multiprocessing/connection.py patch to remove fqdn oddness for issue 3270
........
r64966 | nick.coghlan | 2008-07-15 17:40:22 +0200 (Tue, 15 Jul 2008) | 1 line
Add missing NEWS entry for r64962
........
r64973 | jesse.noller | 2008-07-15 20:29:18 +0200 (Tue, 15 Jul 2008) | 1 line
Revert 3270 patch: self._address is in pretty widespread use, need to revisit
........
In C++, it's an error to pass a string literal to a char* function
without a const_cast(). Rather than require every C++ extension
module to put a cast around string literals, fix the API to state the
const-ness.
I focused on parts of the API where people usually pass literals:
PyArg_ParseTuple() and friends, Py_BuildValue(), PyMethodDef, the type
slots, etc. Predictably, there were a large set of functions that
needed to be fixed as a result of these changes. The most pervasive
change was to make the keyword args list passed to
PyArg_ParseTupleAndKewords() to be a const char *kwlist[].
One cast was required as a result of the changes: A type object
mallocs the memory for its tp_doc slot and later frees it.
PyTypeObject says that tp_doc is const char *; but if the type was
created by type_new(), we know it is safe to cast to char *.
Brett Simmers
Inada Naoki
Victor Stinner
Eric Snow
Raymond Hettinger
Owain Davies
Dennis Sweeney
oda-gitso
Dong-hee Na
Serhiy Storchaka
hrchu
Dong-hee Na
Shantanu
Victor Stinner
Raymond Hettinger
Martin v. Löwis
Antoine Pitrou
Mark Dickinson
Benjamin Peterson
Georg Brandl
Christian Heimes
Jeremy Hylton
Michael W. Hudson