Wenzel Jakob [Fri, 27 May 2022 08:27:39 +0000 (10:27 +0200)]
gh-60074: add new stable API function PyType_FromMetaclass (GH-93012)
Added a new stable API function ``PyType_FromMetaclass``, which mirrors
the behavior of ``PyType_FromModuleAndSpec`` except that it takes an
additional metaclass argument. This is, e.g., useful for language
binding tools that need to store additional information in the type
object.
Avoid _PyCodec_Lookup() and PyCodec_LookupError() for most common
built-in encodings and error handlers to avoid creating a temporary
Unicode string object, whereas these encodings and error handlers are
known to be valid.
Christian Heimes [Wed, 25 May 2022 13:57:26 +0000 (15:57 +0200)]
gh-90473: Misc test fixes for WASI (GH-93218)
* ``sys.executable`` is not set
* WASI does not support subprocess
* ``pwd`` module is not available
* WASI checks ``open`` syscall flags more strict, needs r, w, rw flag.
* ``umask`` is not available
* ``/dev/null`` may not be accessible
Victor Stinner [Wed, 25 May 2022 12:21:36 +0000 (14:21 +0200)]
gh-93202: Always use %zd printf formatter (#93201)
Python now always use the ``%zu`` and ``%zd`` printf formats to
format a size_t or Py_ssize_t number. Building Python 3.12 requires a
C11 compiler, so these printf formats are now always supported.
* PyObject_Print() and _PyObject_Dump() now use the printf %zd format
to display an object reference count.
* Update PY_FORMAT_SIZE_T comment.
* Remove outdated notes about the %zd format in PyBytes_FromFormat()
and PyUnicode_FromFormat() documentations.
* configure no longer checks for the %zd format and no longer defines
PY_FORMAT_SIZE_T macro in pyconfig.h.
* pymacconfig.h no longer undefines PY_FORMAT_SIZE_T: macOS 10.4 is
no longer supported. Python 3.12 now requires macOS 10.6 (Snow
Leopard) or newer.
Victor Stinner [Wed, 25 May 2022 09:45:33 +0000 (11:45 +0200)]
gh-91924: Fix __lltrace__ for non-UTF-8 stdout encoding (#93199)
Fix __lltrace__ debug feature if the stdout encoding is not UTF-8.
If the stdout encoding is not UTF-8, the first call to
lltrace_resume_frame() indirectly sets lltrace to 0 when calling
unicode_check_encoding_errors() which calls
encodings.search_function().
Tobin Yehle [Wed, 25 May 2022 01:16:20 +0000 (19:16 -0600)]
gh-93035: [Enum] Fix IntFlag crash when no single-bit members (GH-93076)
`EnumType` attempts to create a custom docstring for each enum/flag, but that was failing with pathological flags that had no members (only multi-bit aliases).
The Py_DecodeLocale() and Py_EncodeLocale() now use
_PyRuntime.preconfig, rather than Py_UTF8Mode and
Py_LegacyWindowsFSEncodingFlag global configuration varibles, to
decide if the UTF-8 encoding is used or not.
As documented, these functions must not be called before Python is
preinitialized. The new PyConfig API should now be used, rather than
using deprecated functions like Py_SetPath() or PySys_SetArgv().
goldsteinn [Tue, 24 May 2022 01:45:31 +0000 (20:45 -0500)]
gh-93033: Use wmemchr in stringlib (GH-93034)
Generally comparable perf for the "good" case where memchr doesn't
return any collisions (false matches on lower byte) but clearly faster
with collisions.
Update documentation of PyUnicode_DecodeFSDefault(),
PyUnicode_DecodeFSDefaultAndSize() and PyUnicode_EncodeFSDefault():
they now use the filesystem encoding and error handler of PyConfig,
Py_FileSystemDefaultEncoding and Py_FileSystemDefaultEncodeErrors
variables are no longer used.
Christian Heimes [Mon, 23 May 2022 08:39:57 +0000 (10:39 +0200)]
gh-90473: WASI: skip gethostname tests (GH-93092)
- WASI's ``gethostname()`` is a stub that always fails with OSError
``ENOTSUP``
- skip mailcap ``test`` if subprocess is not available
- WASI process_time clock does not work.
Nicolas Haller [Mon, 23 May 2022 04:58:07 +0000 (00:58 -0400)]
gh-89158: Add some REPL secondary prompt markers (#93073)
This fixes an issue on tutorial/classes.rst section 9.4 where the example "class Warehouse"
was truncated when pressing the >>> button to hide the prompts and output.
CAM Gerlach [Fri, 20 May 2022 21:47:51 +0000 (15:47 -0600)]
gh-92611: Clarify planned removal version in PEP 594-deprecated modules (GH-92793)
As discussed in #92611 and #92564 and as a followup to PR #92612 , this 3.11+ only PR uses the proper `deprecated-removed` role for the modules deprecated by PEP 593 (PEP-594) to clearly indicate to users that a removal version is planned and what it is, so they can prepare accordingly or voice any unanticipated impacts.
Related to #92792 ; if we decide to backport that PR, the upgrade to using `deprecated-removed` on those functions can be moved to this one.
Michael Osipov [Thu, 19 May 2022 15:42:21 +0000 (17:42 +0200)]
gh-78630: Drop invalid HP aCC compiler switch -fPIC on HP-UX (#8847)
At compile time, '+z' is already properly used with HP aCC, and shared
libraries are correctly linked with '+b'. The '-fPIC' switch can safely be
dropped.
Martin Fischer [Thu, 19 May 2022 15:34:58 +0000 (17:34 +0200)]
bpo-46879: Fix incorrect sphinx object names in doc (GH-31615)
API members documented in sphinx have an object name, which allow the
documentation to be linked from other projects. Sphinx calculates the
object name by prefixing the current module name to the directive
argument, e.g:
.. module:: foo
.. function:: bar.baz
becomes foo.bar.baz. Since these anchors aren't displayed in the
documentation, some mistakes have crept in, namely the Python stdlib
documentation currently contains the objects:
This commit fixes this by making use of the :module: option which
without an argument makes sphinx take the directive argument as is
for the object name (avoiding the prefixing of the current module
name that led to these broken object names).