When an unawaited coroutine is collected very late in shutdown --
like, during the final GC at the end of PyImport_Cleanup -- then it
was triggering an interpreter abort, because we'd try to look up the
"warnings" module and not only was it missing (we were prepared for
that), but the entire module system was missing (which we were not
prepared for).
I've tried to fix this at the source, by making the utility function
get_warnings_attr robust against this in general. Note that it already
has the convention that it can return NULL without setting an error,
which is how it signals that the attribute it was asked to fetch is
missing, and that all callers already check for NULL returns.
There's a similar check for being late in shutdown at the top of
warn_explicit, which might be unnecessary after this fix, but I'm not
sure so I'm going to leave it.
* Document `from __future__ import annotations`
* Provide plumbing and tests for `from __future__ import annotations`
* Implement unparsing the AST back to string form
This is required for PEP 563 and as such only implements a part of the
unparsing process that covers expressions.
The refleak in question wasn't really important, as context vars
are usually created at the toplevel and live as long as the interpreter
lives, so the context var name isn't ever GCed anyways.
PyUnicode_DecodeLocaleAndSize(), PyUnicode_DecodeLocale() and
PyUnicode_EncodeLocale() now use always use the UTF-8 encoding on
Android, instead of the current locale encoding.
On Android API 19, mbstowcs() and wcstombs() are broken and cannot be
used.
The previous version was correct in terms of behaviour, but
checking the return value of PyErr_WarnFormat allows to
avoid calling PyErr_Occurred and silences the coverity alarm.
* Add coro.cr_origin and sys.set_coroutine_origin_tracking_depth
* Use coroutine origin information in the unawaited coroutine warning
* Stop using set_coroutine_wrapper in asyncio debug mode
* In BaseEventLoop.set_debug, enable debugging in the correct thread
* Use wider types (int => Py_ssize_t) to avoid integer overflows.
* Fix gc.get_freeze_count(): use Py_ssize_t type rather than int, since gc_list_size() returns a Py_ssize_t.
AttributeError was raised always when attribute is not found.
This commit skip raising AttributeError when `tp_getattro` is `PyObject_GenericGetAttr`.
It makes hasattr() and getattr() about 4x faster when attribute is not found.
* Add _Py_GetLocaleconvNumeric() function: decode decimal_point and
thousands_sep fields of localeconv() from the LC_NUMERIC encoding,
rather than decoding from the LC_CTYPE encoding.
* Modify locale.localeconv() and "n" formatter of str.format() (for
int, float and complex to use _Py_GetLocaleconvNumeric()
internally.
Modify locale.localeconv(), time.tzname, os.strerror() and other
functions to ignore the UTF-8 Mode: always use the current locale
encoding.
Changes:
* Add _Py_DecodeLocaleEx() and _Py_EncodeLocaleEx(). On decoding or
encoding error, they return the position of the error and an error
message which are used to raise Unicode errors in
PyUnicode_DecodeLocale() and PyUnicode_EncodeLocale().
* Replace _Py_DecodeCurrentLocale() with _Py_DecodeLocaleEx().
* PyUnicode_DecodeLocale() now uses _Py_DecodeLocaleEx() for all
cases, especially for the strict error handler.
* Add _Py_DecodeUTF8Ex(): return more information on decoding error
and supports the strict error handler.
* Rename _Py_EncodeUTF8_surrogateescape() to _Py_EncodeUTF8Ex().
* Replace _Py_EncodeCurrentLocale() with _Py_EncodeLocaleEx().
* Ignore the UTF-8 mode to encode/decode localeconv(), strerror()
and time zone name.
* Remove PyUnicode_DecodeLocale(), PyUnicode_DecodeLocaleAndSize()
and PyUnicode_EncodeLocale() now ignore the UTF-8 mode: always use
the "current" locale.
* Remove _PyUnicode_DecodeCurrentLocale(),
_PyUnicode_DecodeCurrentLocaleAndSize() and
_PyUnicode_EncodeCurrentLocale().
Add new fuctions ignoring the UTF-8 mode:
* _Py_DecodeCurrentLocale()
* _Py_EncodeCurrentLocale()
* _PyUnicode_DecodeCurrentLocaleAndSize()
* _PyUnicode_EncodeCurrentLocale()
Modify the readline module to use these functions.
Re-enable test_readline.test_nonascii().
- primary change is to add a new default filter entry for
'default::DeprecationWarning:__main__'
- secondary change is an internal one to cope with plain
strings in the warning module's internal filter list
(this avoids the need to create a compiled regex object
early on during interpreter startup)
- assorted documentation updates, including many more
examples of configuring the warnings settings
- additional tests to ensure that both the pure Python and
the C accelerated warnings modules have the expected
default configuration
Third party projects may wish to hide their own internal machinery in
order to present more comprehensible tracebacks to end users
(e.g. Jinja2 and Trio both do this).
Previously such projects have had to rely on ctypes to do so:
fe3dadacdf/jinja2/debug.py (L345)1e86b1aee8/trio/_core/_multierror.py (L296)
This provides a Python level API for creating and modifying real
Traceback objects, allowing tracebacks to be edited at runtime.
Patch by Nathaniel Smith.
This module has never been enabled by default, never worked correctly
on x86-64, and caused ABI problems that caused C extension
compatibility. See bpo-29137 for details/discussion.
Add a new _Py_FindEnvConfigValue() function: code shared between
Windows and Unix implementations of _PyPathConfig_Calculate() to read
the pyenv.cfg file.
_Py_FindEnvConfigValue() now uses _Py_DecodeUTF8_surrogateescape()
instead of using a Python Unicode string, the Python API must not be
used early during Python initialization. Same change in Unix
search_for_exec_prefix(): use _Py_DecodeUTF8_surrogateescape().
Cleanup also encode_current_locale(): PyMem_RawFree/PyMem_Free can be
called with NULL.
Fix also "NUL byte" => "NULL byte" typo.
Replace Py_EncodeLocale() with _Py_EncodeLocaleRaw() in:
* _Py_wfopen()
* _Py_wreadlink()
* _Py_wrealpath()
* _Py_wstat()
* pymain_open_filename()
These functions are called early during Python intialization, only
the RAW memory allocator must be used.
Py_EncodeLocale() now uses _Py_EncodeUTF8_surrogateescape(), instead
of using temporary unicode and bytes objects. So Py_EncodeLocale()
doesn't use the Python C API anymore.
* _Py_InitializeCore() doesn't call _PyMem_SetupAllocators() anymore
if the PYTHONMALLOC environment variable is not set.
* pymain_cmdline() now sets the allocator to the default, instead of
setting the allocator in subfunctions.
* Py_SetStandardStreamEncoding() now calls
_PyMem_SetDefaultAllocator() to get a known allocator, to be able
to release the memory with the same allocator.
* Add _PyCoreConfig.install_signal_handlers
* Remove _PyMain.config: _PyMainInterpreterConfig usage is now
restricted to pymain_init_python_main().
* Rename _PyMain.core_config to _PyMain.config
* _PyMainInterpreterConfig_Read() now creates the xoptions dictionary
from the core config
* Fix _PyMainInterpreterConfig_Read(): don't replace xoptions and
argv if they are already set.
Merge _PyCoreConfig_ReadEnv() into _PyCoreConfig_Read(), and
_Py_CommandLineDetails usage is now restricted to pymain_cmdline().
Changes:
* _PyCoreConfig: Add nxoption, xoptions, nwarnoption and warnoptions
* Add _PyCoreConfig.program: argv[0] or ""
* Move filename, command, module and xoptions from
_Py_CommandLineDetails to _PyMain. xoptions _Py_OptList becomes
(int, wchar_t**) list.
* Add pymain_cmdline() function
* Rename copy_argv() to copy_wstrlist(). Rename clear_argv() to
clear_wstrlist(). Remove _Py_OptList structure: use (int,
wchar_t**) list instead.
* Rename pymain_set_flag_from_env() to pymain_get_env_flag()
* Rename pymain_set_flags_from_env() to pymain_get_env_flags()
* _PyMainInterpreterConfig_Read() now creates the warnoptions from
_PyCoreConfig.warnoptions
* Inline pymain_add_warning_dev_mode() and
pymain_add_warning_bytes_flag() into config_init_warnoptions()
* Inline pymain_get_program_name() into _PyCoreConfig_Read()
* _Py_CommandLineDetails: Replace warning_options with nwarnoption
and warnoptions. Replace env_warning_options with nenv_warnoption
and env_warnoptions.
* pymain_warnings_envvar() now has a single implementation for
Windows and Unix: use config_get_env_var_dup() to also get the
variable as wchar_t* on Unix.
* Reorganize pymain_main() to make the code more flat
* Clear configurations before pymain_update_sys_path()
* Mark Py_FatalError() and _Py_FatalInitError() with _Py_NO_RETURN
* Replace _PyMain.run_code variable with a new RUN_CODE() macro
* Move _PyMain.cf into a local variable in pymain_run_python()
Fix compiler warnings in Py_FinalizeEx(): only define variables if
they are needed, add #ifdef.
Other cleanup changes:
* _PyWarnings_InitWithConfig() is no more needed: call
_PyWarnings_Init() instead.
* Inline pymain_init_main_interpreter() in its caller. This
subfunction is no more justifed.
Constant folding was moved to AST optimizer.
But compiler may emit LOAD_CONSTs + BUILD_TUPLE.
For example, default arguments can be constant tuple
if all arguments are constant.
This commit makes peephole's tuple folding simple.
It doesn't support nested tuples because nested
tuples are folded by AST optimizer already.
As a result of 92a3c6f493, the compiler complains:
Python/import.c:2311:21: warning: comparison of integers of different signs: 'long' and 'unsigned long' [-Wsign-compare]
if ((i + n + 1) <= PY_SSIZE_T_MAX / sizeof(struct _inittab)) {
~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This overflow is extremely unlikely to happen, but let's avoid undefined
behavior anyway.
bpo-29240, bpo-32030: If the encoding change (C locale coerced or
UTF-8 Mode changed), Py_Main() now reads again the configuration with
the new encoding.
Changes:
* Add _Py_UnixMain() called by main().
* Rename pymain_free_pymain() to pymain_clear_pymain(), it can now be
called multipled times.
* Rename pymain_parse_cmdline_envvars() to pymain_read_conf().
* Py_Main() now clears orig_argc and orig_argv at exit.
* Remove argv_copy2, Py_Main() doesn't modify argv anymore. There is
no need anymore to get two copies of the wchar_t** argv.
* _PyCoreConfig: add coerce_c_locale and coerce_c_locale_warn.
* Py_UTF8Mode is now initialized to -1.
* Locale coercion (PEP 538) now respects -I and -E options.
It no longer spends much time doing complex calculations and no
longer consumes much memory for creating large constants that will
be dropped later.
This fixes also bpo-21074.
Each interpreter now has its core_config and main_config copy:
* Add _PyCoreConfig_Copy() and _PyMainInterpreterConfig_Copy()
* Move _PyCoreConfig_Read(), _PyCoreConfig_Clear() and
_PyMainInterpreterConfig_Clear() from Python/pylifecycle.c to
Modules/main.c
* Fix _Py_InitializeEx_Private(): call _PyCoreConfig_ReadEnv() before
_Py_InitializeCore()
bpo-32329, bpo-32030:
* The -R option now turns on hash randomization when the
PYTHONHASHSEED environment variable is set to 0 Previously, the
option was ignored.
* sys.flags.hash_randomization is now properly set to 0 when hash
randomization is turned off by PYTHONHASHSEED=0.
* _PyCoreConfig_ReadEnv() now reads the PYTHONHASHSEED environment
variable. _Py_HashRandomization_Init() now only apply the
configuration, it doesn't read PYTHONHASHSEED anymore.
Conversions like `not a is b -> a is not b` are implemented
in AST optimizer in previous commit (7ea143a).
So this commit removes them from peephole optimizer.
_PyMainInterpreterConfig now contains Python objects, whereas
_PyCoreConfig contains wchar_t* strings.
Core config:
* Rename _PyMainInterpreterConfig_ReadEnv() to _PyCoreConfig_ReadEnv()
* Move 3 strings from _PyMainInterpreterConfig to _PyCoreConfig:
module_search_path_env, home, program_name.
* Add _PyCoreConfig_Clear()
* _PyPathConfig_Calculate() now takes core config rather than main
config
* _PyMainInterpreterConfig_Read() now requires also a core config
Main config:
* Add _PyMainInterpreterConfig.module_search_path: sys.path list
* Add _PyMainInterpreterConfig.argv: sys.argv list
* _PyMainInterpreterConfig_Read() now computes module_search_path
Changes:
* Split _PySys_SetArgvWithError() into subfunctions for Py_Main():
* Create the Python list object
* Set sys.argv to the list
* Compute argv0
* Prepend argv0 to sys.path
* Add _PyPathConfig_ComputeArgv0()
* Remove _PySys_SetArgvWithError()
* Py_Main() now splits the code to compute sys.argv/path0 and the
code to update the sys module: add pymain_compute_argv()
subfunction.
bpo-29240, bpo-32030:
* Rename pymain_set_argv() to pymain_set_sys_argv()
* pymain_set_sys_argv() now creates of copy of argv and modify the
copy, rather than modifying pymain->argv
* Call pymain_set_sys_argv() earlier: before pymain_run_python(), but
after pymain_get_importer().
* Add _PySys_SetArgvWithError() to handle errors
* Add -X utf8 command line option, PYTHONUTF8 environment variable
and a new sys.flags.utf8_mode flag.
* If the LC_CTYPE locale is "C" at startup: enable automatically the
UTF-8 mode.
* Add _winapi.GetACP(). encodings._alias_mbcs() now calls
_winapi.GetACP() to get the ANSI code page
* locale.getpreferredencoding() now returns 'UTF-8' in the UTF-8
mode. As a side effect, open() now uses the UTF-8 encoding by
default in this mode.
* Py_DecodeLocale() and Py_EncodeLocale() now use the UTF-8 encoding
in the UTF-8 Mode.
* Update subprocess._args_from_interpreter_flags() to handle -X utf8
* Skip some tests relying on the current locale if the UTF-8 mode is
enabled.
* Add test_utf8mode.py.
* _Py_DecodeUTF8_surrogateescape() gets a new optional parameter to
return also the length (number of wide characters).
* pymain_get_global_config() and pymain_set_global_config() now
always copy flag values, rather than only copying if the new value
is greater than the old value.
Fix the warning:
Python/import.c: warning: comparison between signed and unsigned integer expressions
if ((i + n + 1) <= PY_SSIZE_T_MAX / sizeof(struct _inittab)) {
Rather than supporting dev mode directly in the warnings module, this
instead adjusts the initialisation code to add an extra 'default'
entry to sys.warnoptions when dev mode is enabled.
This ensures that dev mode behaves *exactly* as if `-Wdefault` had
been passed on the command line, including in the way it interacts
with `sys.warnoptions`, and with other command line flags like `-bb`.
Fix also bpo-20361: have -b & -bb options take precedence over any
other warnings options.
Patch written by Nick Coghlan, with minor modifications of Victor Stinner.
Reference siphash takes the keys as a bytes, so it makes sense to byte swap
when reifying the keys as 64-bit integers. However, Python's siphash takes host
integers in to start with.
Python now supports checking bytecode cache up-to-dateness with a hash of the
source contents rather than volatile source metadata. See the PEP for details.
While a fairly straightforward idea, quite a lot of code had to be modified due
to the pervasiveness of pyc implementation details in the codebase. Changes in
this commit include:
- The core changes to importlib to understand how to read, validate, and
regenerate hash-based pycs.
- Support for generating hash-based pycs in py_compile and compileall.
- Modifications to our siphash implementation to support passing a custom
key. We then expose it to importlib through _imp.
- Updates to all places in the interpreter, standard library, and tests that
manually generate or parse pyc files to grok the new format.
- Support in the interpreter command line code for long options like
--check-hash-based-pycs.
- Tests and documentation for all of the above.
PyImport_ExtendInittab() now uses PyMem_RawRealloc() rather than
PyMem_Realloc(). PyImport_ExtendInittab() can be called before
Py_Initialize() whereas only the PyMem_Raw allocator is supposed to
be used before Py_Initialize().
Add _PyImport_Fini2() to release the memory allocated by
PyImport_ExtendInittab() at exit. PyImport_ExtendInittab() now forces
the usage of the default raw allocator, to be able to release memory
in _PyImport_Fini2().
Don't export these functions anymore to be C API, only to
Py_BUILD_CORE:
* _PyExc_Fini()
* _PyImport_Fini()
* _PyGC_DumpShutdownStats()
* _PyGC_Fini()
* _PyType_Fini()
* _Py_HashRandomization_Fini()
* Py_Main() now starts by reading Py_xxx configuration variables to
only work on its own private structure, and then later writes back
the configuration into these variables.
* Replace Py_GETENV() with pymain_get_env_var() which ignores empty
variables.
* Add _PyCoreConfig.dump_refs
* Add _PyCoreConfig.malloc_stats
* _PyObject_DebugMallocStats() is now responsible to check if debug
hooks are installed. The function returns 1 if stats were written,
or 0 if the hooks are disabled. Mark _PyMem_PymallocEnabled() as
static.
* Simplify _PyCoreConfig_INIT, _PyMainInterpreterConfig_INIT,
_PyPathConfig_INIT macros: no need to set fields to 0/NULL, it's
redundant (the C language sets them to 0/NULL for us).
* Fix typo: pymain_run_statup() => pymain_run_startup()
* Remove a few XXX/TODO
Add import_find_and_load() helper function. The addition of
the importtime option has made PyImport_ImportModuleLevelObject() large
and so using a helper seems worthwhile. It also makes it clearer that
abs_name is the only argument needed by _find_and_load().
Previously, CO_NOFREE was set in the compiler, which meant
it could end up being set incorrectly when code objects
were created directly. Setting it in the constructor based
on freevars and cellvars ensures it is always accurate,
regardless of how the code object is defined.
_PyPathConfig_Init() now also initialize home and program_name:
* Rename existing _PyPathConfig_Init() to _PyPathConfig_Calculate().
Add a new _PyPathConfig_Init() function in pathconfig.c which
handles the _Py_path_config variable and call
_PyPathConfig_Calculate().
* Add home and program_name fields to _PyPathConfig.home
* _PyPathConfig_Init() now initialize home and program_name
from main_config
* Py_SetProgramName(), Py_SetPythonHome() and Py_GetPythonHome() now
calls Py_FatalError() on failure, instead of silently ignoring
failures.
* config_init_home() now gets directly _Py_path_config.home to only
get the value set by Py_SetPythonHome(), or NULL if
Py_SetPythonHome() was not called.
* config_get_program_name() now gets directly
_Py_path_config.program_name to only get the value set by
Py_SetProgramName(), or NULL if Py_SetProgramName() was not called.
* pymain_init_python() doesn't call Py_SetProgramName() anymore,
_PyPathConfig_Init() now always sets the program name
* Call _PyMainInterpreterConfig_Read() in
pymain_parse_cmdline_envvars_impl() to control the memory allocator
* C API documentation: it's no more safe to call Py_GetProgramName()
before Py_Initialize().
* Factorize code from PC/getpathp.c and Modules/getpath.c to remove
duplicated code
* rename pathconfig_clear() to _PyPathConfig_Clear()
* Inline _PyPathConfig_Fini() in pymain_impl() and then remove it,
since it's a oneliner
Changes:
* _PyPathConfig_Fini() cannot be called in Py_FinalizeEx().
Py_Initialize() and Py_Finalize() can be called multiple times, but
it must not "forget" parameters set by Py_SetProgramName(),
Py_SetPath() or Py_SetPythonHome(), whereas _PyPathConfig_Fini()
clear all these parameters.
* config_get_program_name() and calculate_program_full_path() now
also decode paths using Py_DecodeLocale() to use the
surrogateescape error handler, rather than decoding using
mbstowcs() which is strict.
* Change _Py_CheckPython3() prototype: () => (void)
* Truncate a few lines which were too long
The current behaviour of yield expressions inside comprehensions and
generator expressions is essentially an accident of implementation - it
arises implicitly from the way the compiler handles yield expressions inside
nested functions and generators.
Since the current behaviour wasn't deliberately designed, and is inherently
confusing, we're deprecating it, with no current plans to reintroduce it.
Instead, our advice will be to use a named nested generator definition
for cases where this behaviour is desired.
When PyGILState_Ensure() is called in a non-Python thread before
PyEval_InitThreads(), only call PyEval_InitThreads() after calling
PyThreadState_New() to fix a crash.
Add an unit test in test_embed.
_Py_InitializeEx_Private() now calls
_PyMainInterpreterConfig_ReadEnv() to read environment variables
PYTHONHOME and PYTHONPATH, and set the program name.
* bpo-32101: Add sys.flags.dev_mode flag
Rename also the "Developer mode" to the "Development mode".
* bpo-32101: Add PYTHONDEVMODE environment variable
Mention it in the development chapiter.
* Fix _PyMem_SetupAllocators("debug"): always restore allocators to
the defaults, rather than only caling _PyMem_SetupDebugHooks().
* Add _PyMem_SetDefaultAllocator() helper to set the "default"
allocator.
* Add _PyMem_GetAllocatorsName(): get the name of the allocators
* main() now uses debug hooks on memory allocators if Py_DEBUG is
defined, rather than calling directly malloc()
* Document default memory allocators in C API documentation
* _Py_InitializeCore() now fails with a fatal user error if
PYTHONMALLOC value is an unknown memory allocator, instead of
failing with a fatal internal error.
* Add new tests on the PYTHONMALLOC environment variable
* Add support.with_pymalloc()
* Add the _testcapi.WITH_PYMALLOC constant and expose it as
support.with_pymalloc().
* sysconfig.get_config_var('WITH_PYMALLOC') doesn't work on Windows, so
replace it with support.with_pymalloc().
* pythoninfo: add _testcapi collector for pymem
The warnings module doesn't leak memory anymore in the hidden
warnings registry for the "ignore" action of warnings filters.
The warn_explicit() function doesn't add the warning key to the
registry anymore for the "ignore" action.
clang can't figure out that fatal_error is noreturn itself and emits warnings:
../cpython/Python/pylifecycle.c:2116:1: warning: function declared 'noreturn' should not return [-Winvalid-noreturn]
}
^
../cpython/Python/pylifecycle.c:2125:1: warning: function declared 'noreturn' should not return [-Winvalid-noreturn]
}
^
* Py_Main() now calls Py_SetProgramName() earlier to be able to get
the program name in _PyMainInterpreterConfig_ReadEnv().
* Rename prog to program_name
* Rename progpath to program_name
Benjamin Peterson
Yury Selivanov
Victor Stinner
Xiang Zhang
Nathaniel J. Smith
Guido van Rossum
Dmitry Alimov
Serhiy Storchaka
hui shang
INADA Naoki
Barry Warsaw
Nick Coghlan
Marcel Plch
Ivan Levkivskyi
Antoine Pitrou
Neil Schemenauer