bpo-38291: Fix a spurious warning when using help(object) (#27039)
help(object) via pydoc.TextDoc.docclass(object) iterates over the
subclasses of object, which includes typing.io and typing.re if typing
is imported. It tries to access cls.__module__ for each of those
sub-classes. This change suppresses warnings when accessing
cls.__module__.
Ma Lin [Mon, 5 Jul 2021 01:10:44 +0000 (09:10 +0800)]
bpo-41486: zlib uses an UINT32_MAX sliding window for the output buffer (GH-26143)
* zlib uses an UINT32_MAX sliding window for the output buffer
These funtions have an initial output buffer size parameter:
- zlib.decompress(data, /, wbits=MAX_WBITS, bufsize=DEF_BUF_SIZE)
- zlib.Decompress.flush([length])
If the initial size > UINT32_MAX, use an UINT32_MAX sliding window, instead of clamping to UINT32_MAX.
Speed up when (the initial size == the actual size).
This fixes a memory consumption and copying performance regression in earlier 3.10 beta releases if someone used an output buffer larger than 4GiB with zlib.decompress.
bpo-43950: Print columns in tracebacks (PEP 657) (GH-26958)
The traceback.c and traceback.py mechanisms now utilize the newly added code.co_positions and PyCode_Addr2Location
to print carets on the specific expressions involved in a traceback.
This changes the documentation for `EnvBuilder.ensure_directories(env_dir)` to match the actual behavior of that API call.
In particular, `ensure_directories()` is not affected by the state of the `upgrade` attribute, and will not cause an error to have existing directories whether or not the `clear` attribute is set.
This documentation change I believe should be valid to all python versions back to 3.6.
bpo-43950: optimize column table assembling with pre-sizing object (GH-26997)
The new resizing system works like this;
```
$ cat t.py
a + a + a + b + c + a + a + a + b + c + a + a + a + b + c + a + a + a + b + c
[repeated 99 more times]
$ ./python t.py
RESIZE: prev len = 32, new len = 66
FINAL SIZE: 56
-----------------------------------------------------
RESIZE: prev len = 32, new len = 66
RESIZE: prev len = 66, new len = 134
RESIZE: prev len = 134, new len = 270
RESIZE: prev len = 270, new len = 542
RESIZE: prev len = 542, new len = 1086
RESIZE: prev len = 1086, new len = 2174
RESIZE: prev len = 2174, new len = 4350
RESIZE: prev len = 4350, new len = 8702
FINAL SIZE: 8004
```
So now we do considerably lower number of `_PyBytes_Resize` calls.
This PR is part of PEP 657 and augments the compiler to emit ending
line numbers as well as starting and ending columns from the AST
into compiled code objects. This allows bytecodes to be correlated
to the exact source code ranges that generated them.
This information is made available through the following public APIs:
* The `co_positions` method on code objects.
* The C API function `PyCode_Addr2Location`.
Remove the @asyncio.coroutine decorator
enabling legacy generator-based coroutines to be compatible with async/await
code; remove asyncio.coroutines.CoroWrapper used for wrapping
legacy coroutine objects in the debug mode.
The decorator has been deprecated
since Python 3.8 and the removal was initially scheduled for Python 3.10.
Victor Stinner [Thu, 1 Jul 2021 00:30:46 +0000 (02:30 +0200)]
bpo-44531: Add _PyType_AllocNoTrack() function (GH-26947)
Add an internal _PyType_AllocNoTrack() function to allocate an object
without tracking it in the GC.
Modify dict_new() to use _PyType_AllocNoTrack(): dict subclasses are
now only tracked once all PyDictObject members are initialized.
Calling _PyObject_GC_UNTRACK() is no longer needed for the dict type.
Similar change in tuple_subtype_new() for tuple subclasses.
Replace tuple_gc_track() with _PyObject_GC_TRACK().
Serhiy Storchaka [Tue, 29 Jun 2021 08:28:15 +0000 (11:28 +0300)]
bpo-44471: Change error type for bad objects in ExitStack.enter_context() (GH-26820)
A TypeError is now raised instead of an AttributeError in
ExitStack.enter_context() and AsyncExitStack.enter_async_context()
for objects which do not support the context manager or
asynchronous context manager protocols correspondingly.
Serhiy Storchaka [Tue, 29 Jun 2021 08:27:04 +0000 (11:27 +0300)]
bpo-12022: Change error type for bad objects in "with" and "async with" (GH-26809)
A TypeError is now raised instead of an AttributeError in
"with" and "async with" statements for objects which do not
support the context manager or asynchronous context manager
protocols correspondingly.
Batuhan Taskaya [Sun, 27 Jun 2021 14:58:32 +0000 (17:58 +0300)]
bpo-40528: fix is_simple(sum)s behavior for attributes (GH-26918)
This is something I noticed while (now discontinued) experimenting
with the idea of annotating operators with location information. Unfortunately
without this addition, adding any `attributes` to stuff like `unaryop`
doesn't change anything since the code assumes they are singletons and
caches all instances. This patch fixes this assumption with including
the attributes as well as constructor fields.
Barry Warsaw [Thu, 24 Jun 2021 19:37:26 +0000 (12:37 -0700)]
bpo-44498: Issue a deprecation warning on asynchat, asyncore and smtpd import (#26882)
* Issue a deprecation warning on smtpd import
* Also issue DeprecationWarnings for asynchat and asyncore
* Fix some tests
* test___all__ requires the word 'module' or 'package' in the deprecation
warning text, so add those to smtpd, asynchat, and asyncore.
* In test_support, use pprint now instead of asyncore as the landmark.
Petr Viktorin [Thu, 24 Jun 2021 12:57:28 +0000 (14:57 +0200)]
bpo-40939: Remove documentation for `PyParser_*` & add porting notes (GH-26855)
I tried to be relatively thorough and give lots of links.
One reason is that this wasn't deprecated very long; also it seems people running into this tend to not be familiar with similar APIs.
Co-authored-by: Ken Jin <28750310+Fidget-Spinner@users.noreply.github.com>
Victor Stinner [Wed, 23 Jun 2021 13:51:47 +0000 (15:51 +0200)]
bpo-39947: Remove old private trashcan C API functions (GH-26869)
Remove 4 C API private trashcan functions which were only kept for
the backward compatibility of the stable ABI with Python 3.8 and
older, since the trashcan API was not usable with the limited C API
on Python 3.8 and older. The trashcan API was excluded from the
limited C API in Python 3.9.
The trashcan C API was never usable with the limited C API, since old
trashcan macros accessed directly PyThreadState members like
"_tstate->trash_delete_nesting", whereas the PyThreadState structure
is opaque in the limited C API.
Exclude also the PyTrash_UNWIND_LEVEL constant from the C API.
Victor Stinner [Wed, 23 Jun 2021 13:40:27 +0000 (15:40 +0200)]
bpo-43770: Cleanup PyModuleDef_Init() (GH-26879)
PyModuleDef_Init() no longer tries to make PyModule_Type type: it's
already done by _PyTypes_Init() at Python startup. Replace
PyType_Ready() call with an assertion.
Py_RunMain() now resets PyImport_Inittab to its initial value at
exit. It must be possible to call PyImport_AppendInittab() or
PyImport_ExtendInittab() at each Python initialization.
Batuhan Taskaya [Tue, 22 Jun 2021 16:29:42 +0000 (19:29 +0300)]
bpo-40528: Implement a metadata system for ASDL Generator (GH-20193)
ASDL Generator was lack of proper annotation related to generated
module. This patch implements a MetadataVisitor that produces a
metadata object to pass to other visitors that are visiting that
same module. For the inital patch, it dynamically retrieves int
sequences (like cmpop), that was previously hardcoded. It offers
an interface that is easy to extend.
Fix asyncio test_popen() of test_windows_utils by using a longer
timeout. Use military grade battle-tested test.support.SHORT_TIMEOUT
timeout rather than a hardcoded timeout of 10 seconds: it's 30
seconds by default, but it is made longer on slow buildbots.
WaitForMultipleObjects() timeout argument is in milliseconds.
_thread.start_new_thread() no longer calls PyThread_exit_thread()
explicitly at the thread exit, the call was redundant.
On Linux with the glibc, pthread_cancel() loads dynamically the
libgcc_s.so.1 library. dlopen() can fail if there is no more
available file descriptor to open the file. In this case, the process
aborts with the error message:
"libgcc_s.so.1 must be installed for pthread_cancel to work"
pthread_cancel() unwinds back to the thread's wrapping function that
calls the thread entry point.
The unwind function is dynamically loaded from the libgcc_s library
since it is tightly coupled to the C compiler (GCC). The unwinder
depends on DWARF, the compiler generates DWARF, so the unwinder
belongs to the compiler.
Thanks Florian Weimer and Carlos O'Donell for their help on
investigating this issue.