Victor Stinner [Wed, 29 Nov 2017 16:20:38 +0000 (17:20 +0100)]
bpo-32030: Rework memory allocators (#4625)
* 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
Barry Warsaw [Tue, 28 Nov 2017 22:26:04 +0000 (17:26 -0500)]
bpo-32107 - Improve MAC address calculation and fix test_uuid.py (#4600)
``uuid.getnode()`` now preferentially returns universally administered MAC addresses if available, over locally administered MAC addresses. This makes a better guarantee for global uniqueness of UUIDs returned from ``uuid.uuid1()``. If only locally administered MAC addresses are available, the first such one found is returned.
Also improve internal code style by being explicit about ``return None`` rather than falling off the end of the function.
Victor Stinner [Tue, 28 Nov 2017 21:29:32 +0000 (22:29 +0100)]
bpo-32159: Remove tools for CVS and Subversion (#4615)
CPython migrated from CVS to Subversion, to Mercurial, and then to
Git. CVS and Subversion are not more used to develop CPython.
* platform module: drop support for sys.subversion. The
sys.subversion attribute has been removed in Python 3.3.
* Remove Misc/svnmap.txt
* Remove Tools/scripts/svneol.py
* Remove Tools/scripts/treesync.py
Victor Stinner [Tue, 28 Nov 2017 14:30:32 +0000 (15:30 +0100)]
bpo-32155: Bugfixes found by flake8 F841 warnings (#4608)
* distutils.config: Use the PyPIRCCommand.realm attribute if set
* turtledemo: wait until macOS osascript command completes to not
create a zombie process
* Tools/scripts/treesync.py: declare 'default_answer' and
'create_files' as globals to modify them with the command line
arguments. Previously, -y, -n, -f and -a options had no effect.
flake8 warning: "F841 local variable 'p' is assigned to but never
used".
Victor Stinner [Tue, 28 Nov 2017 14:19:56 +0000 (15:19 +0100)]
bpo-32154: Remove asyncio.selectors (#4605)
* Remove asyncio.selectors and asyncio._overlapped symbols from the
namespace of the asyncio module
* Replace "from asyncio import selectors" with "import selectors"
* Replace "from asyncio import _overlapped" with "import _overlapped"
asyncio.selectors was added to support Python 3.3, which doesn't have
selectors in its standard library, and Python 3.4 in the same code
base. Same rationale for asyncio._overlapped. Python 3.3 reached its
end of life, and asyncio is no more maintained as a third party
module on PyPI.
Victor Stinner [Tue, 28 Nov 2017 13:43:52 +0000 (14:43 +0100)]
asyncio: Remove asyncio/compat.py (#4606)
The asyncio/compat.py file was written to support Python < 3.5 and
Python < 3.5.2. But Python 3.5 doesn't accept bugfixes anymore, only
security fixes. There is no more need to backport bugfixes to Python
3.5, and so no need to have a single code base for Python 3.5, 3.6
and 3.7.
Say hello (again) to "async" and "await", who became real keywords in
Python 3.7 ;-)
Nick Coghlan [Mon, 27 Nov 2017 22:11:51 +0000 (08:11 +1000)]
bpo-32136: Separate embedding tests from C API tests (GH-4567)
Some parts of the C API are only relevant to larger
applications embedding CPython as a runtime engine.
The helpers to test those APIs are already separated
out into Programs/_testembed.c, this update moves
the associated test cases out into their own dedicated
test file.
Barry Warsaw [Mon, 27 Nov 2017 19:40:10 +0000 (14:40 -0500)]
bpo-32107 - Better merge of #4494 (#4576)
Improve UUID1 MAC address calculation and related tests.
There are two bits in the MAC address that are relevant to UUID1. The first is the locally administered vs. universally administered bit (second least significant of the first octet). Physical network interfaces such as ethernet ports and wireless adapters will always be universally administered, but some interfaces --such as the interface that MacBook Pros communicate with their Touch Bars-- are locally administered. The former are guaranteed to be globally unique, while the latter are demonstrably *not* globally unique and are in fact the same on every MBP with a Touch Bar. With this bit is set, the MAC is locally administered; with it unset it is universally administered.
The other bit is the multicast bit (least significant bit of the first octet). When no other MAC address can be found, RFC 4122 mandates that a random 48-bit number be generated. This randomly generated number *must* have the multicast bit set.
The improvements in uuid.py include:
* Preferentially return a universally administered MAC address, falling back to a locally administered address if none of the former can be found.
* Improve several coding style issues, such as adding explicit returns of None, using a more readable bitmask pattern, and assuming that the ultimate fallback, random MAC generation will not fail (and propagating any exception there instead of swallowing them).
Improvements in test_uuid.py include:
* Always testing the calculated MAC for universal administration, unless explicitly disabled (i.e. for the random case), or implicitly disabled due to running in the Travis environment. Travis test machines have *no* universally administered MAC address at the time of this writing.
Nick Coghlan [Sun, 26 Nov 2017 03:04:46 +0000 (13:04 +1000)]
bpo-11063: Handle uuid.h being in default include path (GH-4565)
find_file() returns an empty list if it finds the requested
header on the standard include path, so header existence
checks need to be explicitly against "is not None".
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]
}
^
The NNTP server currently has troubles with SSL, whereas we don't
have the control on this server. This test blocks all CIs, so disable
it until a fix can be found.
* 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
Py_GetPath() and Py_Main() now call
_PyMainInterpreterConfig_ReadEnv() to share the same code to get
environment variables.
Changes:
* Add _PyMainInterpreterConfig_ReadEnv()
* Add _PyMainInterpreterConfig_Clear()
* Add _PyMem_RawWcsdup()
* _PyMainInterpreterConfig: rename pythonhome to home
* Rename _Py_ReadMainInterpreterConfig() to
_PyMainInterpreterConfig_Read()
* Use _Py_INIT_USER_ERR(), instead of _Py_INIT_ERR(), for decoding
errors: the user is able to fix the issue, it's not a bug in
Python. Same change was made in _Py_INIT_NO_MEMORY().
* Remove _Py_GetPythonHomeWithConfig()
Will White [Fri, 24 Nov 2017 17:28:12 +0000 (17:28 +0000)]
Improve the String tutorial docs (GH-4541)
The paragraph that contains example of string literal concatenation was placed
after the section about concatenation using the '+' sign.
Moved the paragraph to the appropriate section.
The test.support.skip_unless_bind_unix_socket() decorator is used to skip
asyncio tests that fail because the platform lacks a functional bind()
function for unix domain sockets (as it is the case for non root users on the
recent Android versions that run now SELinux in enforcing mode).
* Move structures back from Include/internal/mem.h to
Objects/obmalloc.c
* Remove _PyObject_Initialize() and _PyMem_Initialize()
* Remove Include/internal/pymalloc.h
* Add test_capi.test_pre_initialization_api():
Make sure that it's possible to call Py_DecodeLocale(), and then call
Py_SetProgramName() with the decoded string, before Py_Initialize().
PyMem_RawMalloc() and Py_DecodeLocale() can be called again before
_PyRuntimeState_Init().
Co-Authored-By: Eric Snow <ericsnowcurrently@gmail.com>
Victor Stinner [Thu, 23 Nov 2017 16:13:44 +0000 (17:13 +0100)]
bpo-27535: Cleanup create_filter() (#4516)
create_filter() now expects the action as a _Py_Identifier which
avoids string comparison, and more important, to avoid handling the
"unknown action" annoying case.
Victor Stinner [Thu, 23 Nov 2017 16:03:20 +0000 (17:03 +0100)]
bpo-32030: Rewrite calculate_path() (#4521)
* calculate_path() rewritten in Modules/getpath.c and PC/getpathp.c
* Move global variables into a new PyPathConfig structure.
* calculate_path():
* Split the huge calculate_path() function into subfunctions.
* Add PyCalculatePath structure to pass data between subfunctions.
* Document PyCalculatePath fields.
* Move cleanup code into a new calculate_free() subfunction
* calculate_init() now handles Py_DecodeLocale() failures properly
* calculate_path() is now atomic: only replace PyPathConfig
(path_config) at once on success.
* _Py_GetPythonHomeWithConfig() now returns an error on failure
* Add _Py_INIT_NO_MEMORY() helper: report a memory allocation failure
* Coding style fixes (PEP 7)
* Py_Main() now reads the PYTHONHOME environment variable
* Add _Py_GetPythonHomeWithConfig() private function
* Add _PyWarnings_InitWithConfig()
* init_filters() doesn't get the current core configuration from the
current interpreter or Python thread anymore. Pass explicitly the
configuration to _PyWarnings_InitWithConfig().
* _Py_InitializeCore() now fails on _PyWarnings_InitWithConfig()
failure.
* Pass configuration as constant
Jesse-Bakker [Thu, 23 Nov 2017 00:23:28 +0000 (01:23 +0100)]
bpo-10049: Add a "no-op" (null) context manager to contextlib (GH-4464)
Adds a simpler and faster alternative to ExitStack for handling
single optional context managers without having to change the
lexical structure of your code.
* Py_Main() initializes _PyCoreConfig.module_search_path_env from
the PYTHONPATH environment variable.
* PyInterpreterState_New() now initializes core_config and config
fields
* Compute sys.path a little bit ealier in
_Py_InitializeMainInterpreter() and new_interpreter()
* Add _Py_GetPathWithConfig() private function.
Victor Stinner [Wed, 22 Nov 2017 22:51:42 +0000 (23:51 +0100)]
bpo-27535: Optimize warnings.warn() (#4508)
* Optimize warnings.filterwarnings(). Replace re.compile('') with
None to avoid the cost of calling a regex.match() method, whereas
it always matchs.
* Optimize get_warnings_attr(): replace PyObject_GetAttrString() with
_PyObject_GetAttrId().
Cleanup also create_filter():
* Use _Py_IDENTIFIER() to allow to cleanup strings at Python
finalization
* Replace Py_FatalError() with a regular exceptions
Victor Stinner [Tue, 21 Nov 2017 23:34:02 +0000 (15:34 -0800)]
bpo-31324: Optimize support._match_test() (#4421)
* Rename support._match_test() to support.match_test(): make it
public
* Remove support.match_tests global variable. It is replaced with a
new support.set_match_tests() function, so match_test() doesn't
have to check each time if patterns were modified.
* Rewrite match_test(): use different code paths depending on the
kind of patterns for best performances.
Victor Stinner [Tue, 21 Nov 2017 01:32:40 +0000 (17:32 -0800)]
bpo-32089: Fix warnings filters in dev mode (#4482)
The developer mode (-X dev) now creates all default warnings filters
to order filters in the correct order to always show ResourceWarning
and make BytesWarning depend on the -b option.
Write a functional test to make sure that ResourceWarning is logged
twice at the same location in the developer mode.
Moses Koledoye [Sat, 18 Nov 2017 22:49:15 +0000 (23:49 +0100)]
[Doc] Update opcode for var-positional arguments (#4446)
`BUILD_MAP_UNPACK_WITH_CALL` was duplicated as the opcode for both var-positional and var-keyword arguments. The opcode for the former was updated as `BUILD_TUPLE_UNPACK_WITH_CALL`.
xdegaye [Sat, 18 Nov 2017 17:17:16 +0000 (18:17 +0100)]
bpo-29185: Fix `test_distutils` failures on Android (GH-4438)
* Run gzip with separate command line options (Android understands '-f9' as the name of a file).
* Creation of a hard link is controled by SELinux on Android.