* Revert "bpo-43693: Compute deref offsets in compiler (gh-25152)"
This reverts commit b2bf2bc1ec.
* Revert "bpo-43693: Add new internal code objects fields: co_fastlocalnames and co_fastlocalkinds. (gh-26388)"
This reverts commit 2c1e2583fd.
These two commits are breaking the refleak buildbots.
test_disallow_instantiation and test_readonly_types try to test all the available
digests, however under FIPS mode, while the algorithms are available, trying to use
them will fail with a ValueError.
Merges locals and cells into a single array.
Saves a pointer in the interpreter and means that we don't need the LOAD_CLOSURE opcode any more
https://bugs.python.org/issue43693
When compiling an AST object with a direct / indirect reference
cycles, on the conversion phase because of exceeding amount of
calls, a segfault was raised. This patch adds recursion guards to
places for preventing user inputs to not to crash AST but instead
raise a RecursionError.
The sqlite3 module now fully implements the GC protocol, so there's no
need for this workaround anymore.
- Add and use managed resource helper for connections using TESTFN
- Sort test imports
- Split open-tests into their own test case
Automerge-Triggered-By: GH:vstinner
A number of places in the code base (notably ceval.c and frameobject.c) rely on mapping variable names to indices in the frame "locals plus" array (AKA fast locals), and thus opargs. Currently the compiler indirectly encodes that information on the code object as the tuples co_varnames, co_cellvars, and co_freevars. At runtime the dependent code must calculate the proper mapping from those, which isn't ideal and impacts performance-sensitive sections. This is something we can easily address in the compiler instead.
This change addresses the situation by replacing internal use of co_varnames, etc. with a single combined tuple of names in locals-plus order, along with a minimal array mapping each to its kind (local vs. cell vs. free). These two new PyCodeObject fields, co_fastlocalnames and co_fastllocalkinds, are not exposed to Python code for now, but co_varnames, etc. are still available with the same values as before (though computed lazily).
Aside from the (mild) performance impact, there are a number of other benefits:
* there's now a clear, direct relationship between locals-plus and variables
* code that relies on the locals-plus-to-name mapping is simpler
* marshaled code objects are smaller and serialize/de-serialize faster
Also note that we can take this approach further by expanding the possible values in co_fastlocalkinds to include specific argument types (e.g. positional-only, kwargs). Doing so would allow further speed-ups in _PyEval_MakeFrameVector(), which is where args get unpacked into the locals-plus array. It would also allow us to shrink marshaled code objects even further.
https://bugs.python.org/issue43693
Obviously, the former was a little misleading: not only rationals
may be considered as "infinite-precision, real numbers" - but, for
example, any real finite extension of the rationals.
On macOS, tk defers to system setting for button background when in normal state.
Give turtledemo button text a color that works on either light or dark background.
* Move up the comment about fields using in hashing/comparision.
* Group the fields more clearly.
* Add co_ncellvars and co_nfreevars.
* Raise ValueError if nlocals != len(varnames), rather than aborting.
The list lookups become a big burden for very long lists.
This patch changes the "happy flow" path of 2 lookups into 1 lookup.
Automerge-Triggered-By: GH:vsajip
Fix a regression in type() when a metaclass raises an exception. The
C function type_new() must properly report the exception when a
metaclass constructor raises an exception and the winner class is not
the metaclass.
A previous commit broke a check in sysconfig when building cpython itself.
This caused builds of the standard library modules to search a wrong
location (the installed location rather than the source directory) for
header files with the net effect that a ``make install``
incorrectly caused all extension modules to be rebuilt again and
with incorrect include file paths.
When building Python, we need two distinct "include" directories:
- source .h files
- install target for .h files
Note that this doesn't matter except when building Python from source.
Historically:
- source .h files were in the distutils scheme under 'include'
- the install directory was in the distutils.command.install scheme
under 'headers'
GH-24549 merged these; sysconfig is now the single source of truth and
distutils is derived from it.
This commit introduces a "secret" scheme path, 'headers', which contains
the install target. It is only present when building Python.
The distutils code uses it if present, and falls back to 'include'.
Co-authored-by: Ned Deily <nad@python.org>
The invalid assignment rules are very delicate since the parser can
easily raise an invalid assignment when a keyword argument is provided.
As they are very deep into the grammar tree, is very difficult to
specify in which contexts these rules can be used and in which don't.
For that, we need to use a different version of the rule that doesn't do
error checking in those situations where we don't want the rule to raise
(keyword arguments and generator expressions).
We also need to check if we are in left-recursive rule, as those can try
to eagerly advance the parser even if the parse will fail at the end of
the expression. Failing to do this allows the parser to start parsing a
call as a tuple and incorrectly identify a keyword argument as an
invalid assignment, before it realizes that it was not a tuple after all.
Fix a crash at Python exit when a deallocator function removes the
last strong reference to a heap type.
Don't read type memory after calling basedealloc() since
basedealloc() can deallocate the type and free its memory.
_PyMem_IsPtrFreed() argument is now constant.
* Remove 'zombie' frames. We won't need them once we are allocating fixed-size frames.
* Add co_nlocalplus field to code object to avoid recomputing size of locals + frees + cells.
* Move locals, cells and freevars out of frame object into separate memory buffer.
* Use per-threadstate allocated memory chunks for local variables.
* Move globals and builtins from frame object to per-thread stack.
* Move (slow) locals frame object to per-thread stack.
* Move internal frame functions to internal header.
Moreover, Py_FrozenMain() relies on Py_InitializeFromConfig() to
handle the PYTHONUNBUFFERED environment variable and configure C
stdio streams like stdout (make the stream unbuffered).
`shutil.unpack_archive()` tries to read the whole file into memory, making no use of any kind of smaller buffer. Process crashes for really large files: I.e. archive: ~1.7G, unpacked: ~10G. Before the crash it can easily take away all available RAM on smaller systems. Had to pull the code form `zipfile.Zipfile.extractall()` to fix this
Automerge-Triggered-By: GH:gpshead
For IPv4 mapped IPv6 addresses, defer privacy check to the mapped IPv4 address. Solves bug where public mapped IPv4 addresses are considered private by the IPv6 check.
Automerge-Triggered-By: GH:gpshead
When a Thread is not joined after it has stopped, its lock may remain in the _shutdown_locks set until interpreter shutdown. If many threads are created this way, the _shutdown_locks set could therefore grow endlessly. To avoid such a situation, purge expired locks each time a new one is added or removed.
Added two new attributes to ``_GenericAlias``:
* ``_typevar_types``, a single type or tuple of types indicating what types are treated as a ``TypeVar``. Used for ``isinstance`` checks.
* ``_paramspec_tvars ``, a boolean flag which guards special behavior for dealing with ``ParamSpec``. Setting it to ``True`` means this class deals with ``ParamSpec``.
Automerge-Triggered-By: GH:gvanrossum
Switch to lru_cache in urllib.parse.
urllib.parse now uses functool.lru_cache for its internal URL splitting and
quoting caches instead of rolling its own like its the 90s.
The undocumented internal Quoted class API is now deprecated
as it had no reason to be public and no existing OSS users were found.
The clear_cache() API remains undocumented but gets an explicit test as it
is used in a few projects' (twisted, gevent) tests as well as our own regrtest.
Added in bpo-41730 (GH-22186), the test apparently causes refleaks. The
test isn't worth hunting them down, so it's simply reverted.
This partially reverts commit 4a2d98a1e9.
Pablo Galindo
Raymond Hettinger
Erlend Egeberg Aasland
stratakis
Mark Shannon
Victor Stinner
Batuhan Taskaya
Eric Snow
andrei kulakov
Gregory P. Smith
Tal Einat
Mariusz Felisiak
Serhiy Storchaka
Jason R. Coombs
Dong-hee Na
Sergey B Kirpichev
Terry Jan Reedy
Andreas Jansson
Sean Grady
Jürgen Gmach
Yonatan Goldschmidt
Zack Kneupper
Petr Viktorin
Filipe Laíns
Irit Katriel
Anthony Sottile
E-Paine
Zackery Spytz
Barney Gale
native-api
Catherine Devlin
Ian Norton
uniocto
Igor Bolshakov
Pete Wicken
Sergey B Kirpichev
flizzywine
Ashwin Ramaswami
Batuhan Taskaya
Ken Jin
Rory Yorke
Antoine Pitrou
Miguel Brito
Yurii Karabas
Karl Dubost
Sebastian Rittau
Giampaolo Rodola
Zachary Ware