Abhilash Raj [Fri, 29 May 2020 00:04:59 +0000 (17:04 -0700)]
bpo-39040: Fix parsing of email mime headers with whitespace between encoded-words. (gh-17620)
* bpo-39040: Fix parsing of email headers with encoded-words inside a quoted string.
It is fairly common to find malformed mime headers (especially content-disposition
headers) where the parameter values, instead of being encoded to RFC
standards, are "encoded" by doing RFC 2047 "encoded word" encoding, and
then enclosing the whole thing in quotes. The processing of these malformed
headers was incorrectly leaving the spaces between encoded words in the decoded
text (whitespace between adjacent encoded words is supposed to be stripped on
decoding). This changeset fixes the encoded word processing inside quoted strings
(bare-quoted-string) to do correct RFC 2047 decoding by stripping that
whitespace.
Michał Górny [Thu, 28 May 2020 16:33:04 +0000 (18:33 +0200)]
bpo-1294959: Try to clarify the meaning of platlibdir (GH-20332)
Try to make the meaning of platlibdir clear. The previous wording could
be misinterpreted to suggest that it will be used to find all shared
libraries on the system, and not just Python extensions. Furthermore,
it was unclear whether it affects third-party (site-packages) extensions
or not. The new wording tries to make its dual purpose clear,
and provide the additional example of extensions in site-packages.
Victor Stinner [Thu, 28 May 2020 15:23:39 +0000 (17:23 +0200)]
bpo-25920: Remove socket.getaddrinfo() lock on macOS (GH-20177)
On macOS, socket.getaddrinfo() no longer uses an internal lock to
prevent race conditions when calling getaddrinfo(). getaddrinfo is
thread-safe is macOS 10.5, whereas Python 3.9 requires macOS 10.6 or
newer.
The lock was also used on FreeBSD older than 5.3, OpenBSD older than
201311 and NetBSD older than 4.
Skip new "racing" socket tests which fail randomly until someone fix
them, to ease analysis of buildbot failures (skip tests which are
known to be broken/unstable).
Serhiy Storchaka [Thu, 28 May 2020 07:33:45 +0000 (10:33 +0300)]
bpo-40792: Make the result of PyNumber_Index() always having exact type int. (GH-20443)
Previously, the result could have been an instance of a subclass of int.
Also revert bpo-26202 and make attributes start, stop and step of the range
object having exact type int.
Add private function _PyNumber_Index() which preserves the old behavior
of PyNumber_Index() for performance to use it in the conversion functions
like PyLong_AsLong().
* Use a more universal explanation of string interpolation rather than specifically referencing sprintf(), which depends on the reader having a C background.
check_impl_detail() of test.support now uses sys.implementation.name,
instead of platform.python_implementation().lower(). This change
prepares test.support to import the platform module lazily.
Victor Stinner [Wed, 27 May 2020 22:38:12 +0000 (00:38 +0200)]
bpo-40795: ctypes calls unraisablehook with an exception (GH-20452)
If ctypes fails to convert the result of a callback or if a ctypes
callback function raises an exception, sys.unraisablehook is now
called with an exception set. Previously, the error was logged into
stderr by PyErr_Print().
Pablo Galindo [Wed, 27 May 2020 20:48:12 +0000 (21:48 +0100)]
Fix compiler warnings in _zoneinfo.c (GH-20342)
```
D:\a\cpython\cpython\Modules\_zoneinfo.c(903,52): warning C4267: '=': conversion from 'size_t' to 'unsigned int', possible loss of data [D:\a\cpython\cpython\PCbuild\_zoneinfo.vcxproj]
D:\a\cpython\cpython\Modules\_zoneinfo.c(904,44): warning C4267: '=': conversion from 'size_t' to 'unsigned int', possible loss of data [D:\a\cpython\cpython\PCbuild\_zoneinfo.vcxproj]
D:\a\cpython\cpython\Modules\_zoneinfo.c(1772,31): warning C4244: '=': conversion from 'ssize_t' to 'uint8_t', possible loss of data [D:\a\cpython\cpython\PCbuild\_zoneinfo.vcxproj]
```
Christian Heimes [Wed, 27 May 2020 19:50:06 +0000 (21:50 +0200)]
bpo-40791: Use CRYPTO_memcmp() for compare_digest (#20456)
hashlib.compare_digest uses OpenSSL's CRYPTO_memcmp() function
when OpenSSL is available.
Note: The _operator module is a builtin module. I don't want to add
libcrypto dependency to libpython. Therefore I duplicated the wrapper
function and added a copy to _hashopenssl.c.
Sean Gillespie [Wed, 27 May 2020 15:22:07 +0000 (08:22 -0700)]
bpo-13097: ctypes: limit callback to 1024 arguments (GH-19914)
ctypes now raises an ArgumentError when a callback
is invoked with more than 1024 arguments.
The ctypes module allocates arguments on the stack in
ctypes_callproc() using alloca(), which is problematic
when large numbers of arguments are passed. Instead
of a stack overflow, this commit raises an ArgumentError
if more than 1024 parameters are passed.
Victor Stinner [Wed, 27 May 2020 12:55:10 +0000 (14:55 +0200)]
bpo-39573: Convert Py_REFCNT and Py_SIZE to functions (GH-20429)
Convert Py_REFCNT() and Py_SIZE() macros to static inline functions.
They cannot be used as l-value anymore: use Py_SET_REFCNT() and
Py_SET_SIZE() to set an object reference count and size.
Replace &Py_SIZE(self) with &((PyVarObject*)self)->ob_size
in arraymodule.c.
This change is backward incompatible on purpose, to prepare the C API
for an opaque PyObject structure.
Oleg Höfling [Wed, 27 May 2020 10:07:15 +0000 (12:07 +0200)]
Fix the link to ncurses patch download in macos installer build script (GH-20421)
Reason: the link `ftp://invisible-island.net/ncurses//5.9/ncurses-5.9-20120616-patch.sh.bz2` is dead, which prevents `Mac/BuildScript/build-installer.py` from completing. Looks like the host of the FTP server was changed to `ftp.invisible-island.net`, thus this proposal.
Guido van Rossum [Tue, 26 May 2020 17:58:44 +0000 (10:58 -0700)]
Add soft keywords (GH-20370)
These are like keywords but they only work in context; they are not reserved except when there is an exact match.
This would enable things like match statements without reserving `match` (which would be bad for the `re.match()` function and probably lots of other places).
It is cleaner and friendlier to default the 'extra' argument to None
instead of either forcing the subclasses of LoggerAdapter to pass a None
value directly or to override the constructor.
This change is backward compatible because existing calls to
`LoggerAdapter.__init__` are already passing a value for the second
argument.
bpo-38964: Print correct filename on a SyntaxError in an fstring (GH-20399)
When a `SyntaxError` in the expression part of a fstring is found,
the filename attribute of the `SyntaxError` is always `<fstring>`.
With this commit, it gets changed to always have the name of the file
the fstring resides in.
Co-authored-by: Pablo Galindo <Pablogsal@gmail.com>
bpo-40688: Use the correct parser in the peg_generator scripts (GH-20235)
The scripts in `Tools/peg_generator/scripts` mostly assume that
`ast.parse` and `compile` use the old parser, since this was the
state of things, while we were developing them. They need to be
updated to always use the correct parser. `_peg_parser` is being
extended to support both parsing and compiling with both parsers.
:mod:`hashlib` no longer falls back to builtin hash implementations when
OpenSSL provides a hash digest and the algorithm is blocked by security
policy.
Signed-off-by: Christian Heimes <christian@python.org>
Chris Jerdonek [Fri, 22 May 2020 20:33:27 +0000 (13:33 -0700)]
bpo-40696: Fix a hang that can arise after gen.throw() (GH-20287)
This updates _PyErr_ChainStackItem() to use _PyErr_SetObject()
instead of _PyErr_ChainExceptions(). This prevents a hang in
certain circumstances because _PyErr_SetObject() performs checks
to prevent cycles in the exception context chain while
_PyErr_ChainExceptions() doesn't.
Huon Wilson [Fri, 22 May 2020 14:18:51 +0000 (00:18 +1000)]
bpo-40630: Add tracemalloc.reset_peak (GH-20102)
The reset_peak function sets the peak memory size to the current size,
representing a resetting of that metric. This allows for recording the
peak of specific sections of code, ignoring other code that may have
had a higher peak (since the most recent `tracemalloc.start()` or
tracemalloc.clear_traces()` call).
bpo-40334: Produce better error messages for non-parenthesized genexps (GH-20153)
The error message, generated for a non-parenthesized generator expression
in function calls, was still the generic `invalid syntax`, when the generator expression wasn't appearing as the first argument in the call. With this patch, even on input like `f(a, b, c for c in d, e)`, the correct error message gets produced.
Batuhan Taskaya [Thu, 21 May 2020 19:57:52 +0000 (22:57 +0300)]
bpo-40334: Correctly generate C parser when assigned var is None (GH-20296)
When there are 2 negative lookaheads in the same rule, let's say `!"(" blabla "," !")"`, there will the 2 `FunctionCall`'s where assigned value is None. Currently when the `add_var` is called
the first one will be ignored but when the second lookahead's var is sent to dedupe it
will be returned as `None_1` and this won't be ignored by the declaration generator in the `visit_Alt`. This patch adds an explicit check to `add_var` to distinguish whether if there is a variable or not.