This makes a couple related changes to inspect.signature's behaviour
when parsing a signature from `__text_signature__`.
First, `inspect.signature` is documented as only raising ValueError or
TypeError. However, in some cases, we could raise RuntimeError. This PR
changes that, thereby fixing GH-83685.
(Note that the new ValueErrors in RewriteSymbolics are caught and then
reraised with a message)
Second, `inspect.signature` could randomly drop parameters that it
didn't understand (corresponding to `return None` in the `p` function).
This is the core issue in GH-85267. I think this is very surprising
behaviour and it seems better to fail outright.
Third, adding this new failure broke a couple tests. To fix them (and to
e.g. allow `inspect.signature(select.epoll.register)` as in GH-85267), I
add constant folding of a couple binary operations to RewriteSymbolics.
(There's some discussion of making signature expression evaluation
arbitrary powerful in GH-68155. I think that's out of scope. The
additional constant folding here is pretty straightforward, useful, and
not much of a slippery slope)
Fourth, while GH-85267 is incorrect about the cause of the issue, it turns
out if you had consecutive newlines in __text_signature__, you'd get
`tokenize.TokenError`.
Finally, the `if name is invalid:` code path was dead, since
`parse_name` never returned `invalid`..
(cherry picked from commit 79311cbfe7)
Co-authored-by: Shantanu <12621235+hauntsaninja@users.noreply.github.com>
test.test_ast_recursion_limit() now uses infinite_recursion() of
test.support to prevent crashes on debug builds.
Before this change, the test crashed on ARM64 Windows 3.x buildbot
worker which builds Python in debug mode.
(cherry picked from commit cd67c1bb30)
Co-authored-by: Victor Stinner <vstinner@python.org>
When checking for auto() instances, only top-level usage is supported,
which means either alone or as part of a regular tuple. Other
containers, such as lists, dicts, or namedtuples, will not have auto()
transformed into a value.
(cherry picked from commit ded02ca54d)
Co-authored-by: Ethan Furman <ethan@stoneleaf.us>
Also \ escape \s in the http.server BaseHTTPRequestHandler.log_message so
that it is technically possible to parse the line and reconstruct what the
original data was. Without this a \xHH is ambiguious as to if it is a hex
replacement we put in or the characters r"\x" came through in the original
request line.
(cherry picked from commit 7e29398407)
Co-authored-by: Gregory P. Smith <greg@krypto.org>
Replace control characters in http.server.BaseHTTPRequestHandler.log_message with an escaped \xHH sequence to avoid causing problems for the terminal the output is printed to.
(cherry picked from commit d8ab0a4dfa)
Co-authored-by: Gregory P. Smith <greg@krypto.org>
In case if only True/False be supported as boolean arguments in future,
we should continue to support 1/0 here.
(cherry picked from commit 922a6cf6c2)
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
Skip test_normalization() of test_unicodedata if it fails to download
NormalizationTest.txt file from pythontest.net.
(cherry picked from commit 2488c1e1b6)
Co-authored-by: Victor Stinner <vstinner@python.org>
On FreeBSD, skip test_script_as_dev_fd() of test_cmd_line_script if
fdescfs is not mounted (at /dev/fd).
(cherry picked from commit 038b151963)
Co-authored-by: Victor Stinner <vstinner@python.org>
Existing elements do produce different output on x86_64, but they
do not on x86. Let's make the data longer to ensure it differs.
(cherry picked from commit c68573b339)
Co-authored-by: Alexander Kanavin <alex.kanavin@gmail.com>
(cherry picked from commit 367f552129)
Co-authored-by: Christoph Anton Mitterer <calestyo@scientia.org>
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Co-authored-by: Eryk Sun <eryksun@gmail.com>
Co-authored-by: Jelle Zijlstra <jelle.zijlstra@gmail.com>
Fixes private checks for network objects. The previous method would incorrectly return True for a private check in cases such as "0.0.0.0/0".
(cherry picked from commit ed391090cc)
Co-authored-by: Pete Wicken <2273100+JamoBox@users.noreply.github.com>
* Fix substitution of TypeVarTuple and ParamSpec together in user generics.
* Fix substitution of ParamSpec followed by TypeVarTuple in generic aliases.
* Check the number of arguments in substitution in user generics containing a
TypeVarTuple and one or more TypeVar.
(cherry picked from commit 8f2fb7dfe7)
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
On macOS all file descriptors for a particular file in /dev/fd
share the same file offset, that is ``open("/dev/fd/9", "r")`` behaves
more like ``dup(9)`` than a regular open.
This causes problems when a user tries to run "/dev/fd/9" as a script
because zipimport changes the file offset to try to read a zipfile
directory. Therefore change zipimport to reset the file offset after
trying to read the zipfile directory.
(cherry picked from commit d08fb25769)
Co-authored-by: Ronald Oussoren <ronaldoussoren@mac.com>
The tests in question were added in 0eec6276fd by Serhiy. Apparently,
sqlite3 changed exceptions raised in those cases in the mean time but
the tests never ran because they require a high `-M` setting in the
test runner.
(cherry picked from commit 2781ec9b0e)
Co-authored-by: Łukasz Langa <lukasz@langa.pl>
[3.11] gh-99103: Normalize specialized traceback anchors against the current line (GH-99145)
Automerge-Triggered-By: GH:isidentical.
(cherry picked from commit 57be545959)
Co-authored-by: Batuhan Taskaya <isidentical@gmail.com>
Fix a reference bug in _imp.create_builtin() after the creation of
the first sub-interpreter for modules "builtins" and "sys".
(cherry picked from commit cb2ef8b2ac)
Co-authored-by: Victor Stinner <vstinner@python.org>
When build with shared enabled, we need to set `LD_LIBRARY_PATH`
for the non-installed python environment in
test_zippath_from_non_installed_posix so that the python binary
and find and link the libpython.so.
(cherry picked from commit 1df0752b4d)
Co-authored-by: Kai Zhang <kylerzhang11@gmail.com>
Before python3.11, when in a venv the zip path is calculated
from prefix on POSIX platforms. In python3.11 the behavior is
accidentally changed to calculating from default prefix. This
change will break venv created from a non-installed python
with a stdlib zip file. This commit restores the behavior back
to before python3.11.
(cherry picked from commit e3d4fed074)
Co-authored-by: Kai Zhang <kylerzhang11@gmail.com>
