test_ssl.test_pha_required_nocert() now uses
support.catch_threading_exception() to ignore the expected SSLError
in ConnectionHandler of ThreadedEchoServer (it is only raised
sometimes on Windows).
(cherry picked from commit 73ea54620a6f91c3f2e53880373dd47813691a21)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
bpo-18374: fix wrong col_offset of some ast.BinOp instances (GH-14607)
Nested BinOp instances (e.g. a+b+c) had a wrong col_offset for the
second BinOp (e.g. 2 instead of 0 in the example). Fix it by using the
correct st node to copy the line and col_offset from in ast.c.
(cherry picked from commit 110a47c4f42cf4db88edc1876899fff8f05190fb)
Co-authored-by: Carl Friedrich Bolz-Tereick <cfbolz@gmx.de>
bpo-37421: multiprocessing tests now stop ForkServer (GH-14601)
multiprocessing tests now stop the ForkServer instance if it's
running: close the "alive" file descriptor to ask the server to stop
and then remove its UNIX address.
(cherry picked from commit 8fbeb14312b4c1320d31ad86e69749515879d1c3)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
The distutils bdist_wininst command is now deprecated, use
bdist_wheel (wheel packages) instead.
(cherry picked from commit 1da4462765b084dfa8d869b6cb5855e8f6014a11)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
Fix multiprocessing.util.get_temp_dir() finalizer: clear also the
'tempdir' configuration of the current process, so next call to
get_temp_dir() will create a new temporary directory, rather than
reusing the removed temporary directory.
(cherry picked from commit 9d40554e0da09a44a8547f3f3a2b9dedfeaf7928)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
test_distutils.test_build_ext() is now able to remove the temporary
directory on Windows: don't import the newly built C extension ("xx")
in the current process, but test it in a separated process.
(cherry picked from commit 74c9dd57771f4f061ee83b069c8e7b37de41246b)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
test_winconsoleio doesn't leak a temporary file anymore: use
tempfile.TemporaryFile() to remove it when the test completes.
(cherry picked from commit b71d8d67959f3b5efbdfe00066589ac0d8f98aad)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
ssl.match_hostname() no longer accepts IPv4 addresses with additional text
after the address and only quad-dotted notation without trailing
whitespaces. Some inet_aton() implementations ignore whitespace and all data
after whitespace, e.g. '127.0.0.1 whatever'.
Short notations like '127.1' for '127.0.0.1' were already filtered out.
The bug was initially found by Dominik Czarnota and reported by Paul Kehrer.
multiprocessing tests now call explicitly _run_finalizers() to remove
immediately temporary directories created by
multiprocessing.util.get_temp_dir().
(cherry picked from commit 039fb49c185570ab7b02f13fbdc51c859cfd831e)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
Python initialization now ensures that sys stream encoding
names are always normalized by codecs.lookup(encoding).name.
Simplify test_c_locale_coercion: it doesn't have to normalize
encoding names anymore.
(cherry picked from commit 61bf97e91620e020939d57a36918ab22579920ff)
Under some conditions the earlier fix for bpo-18075, "Infinite recursion
tests triggering a segfault on Mac OS X", now causes failures on macOS
when attempting to change stack limit with resource.setrlimit
resource.RLIMIT_STACK, like regrtest does when running the test suite.
The reverted change had specified a non-default stack size when linking
the python executable on macOS. As of macOS 10.14.4, the previous
code causes a hard failure when running tests, although similar
failures had been seen under some conditions under some earlier
systems. Reverting the change to the interpreter stack size at link
time helped for release builds but caused some tests to fail when
built --with-pydebug. Try the opposite approach: continue to build
the interpreter with an increased stack size on macOS and remove
the failing setrlimit call in regrtest initialization. This will
definitely avoid the resource.RLIMIT_STACK error and should have
no, or fewer, side effects.
(cherry picked from commit 5bbbc733e6cc0804f19b071944af8d4719e26ae6)
Stop using "static PyConfig", PyConfig must now always use
dynamically allocated strings: use PyConfig_SetString(),
PyConfig_SetArgv() and PyConfig_Clear().
(cherry picked from commit 67310023f299b5a2fad71fca449b46d280036690)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
bpo-37467: Fix PyErr_Display() for bytes filename (GH-14504)
Fix sys.excepthook() and PyErr_Display() if a filename is a bytes
string. For example, for a SyntaxError exception where the filename
attribute is a bytes string.
Cleanup also test_sys:
* Sort imports.
* Rename numruns global var to INTERN_NUMRUNS.
* Add DisplayHookTest and ExceptHookTest test case classes.
* Don't save/restore sys.stdout and sys.displayhook using
setUp()/tearDown(): do it in each test method.
* Test error case (call hook with no argument) after the success case.
(cherry picked from commit f9b7457bd7f438263e0d2dd1f70589ad56a2585e)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
bpo-10945: Drop support for bdist_wininst on non-Windows systems (GH-14506)
bdist_wininst depends on MBCS codec, unavailable on non-Windows,
and bdist_wininst have not worked since at least Python 3.2, possibly
never on Python 3.
Here we document that bdist_wininst is only supported on Windows,
and we mark it unsupported otherwise to skip tests.
Distributors of Python 3 can now safely drop the bdist_wininst .exe files
without the need to skip bdist_wininst related tests.
(cherry picked from commit 72cd653c4ed7a4f8f8fb06ac364b08a97085a2b5)
[3.8] bpo-37428: Don't set PHA verify flag on client side (GH-14494)
SSLContext.post_handshake_auth = True no longer sets
SSL_VERIFY_POST_HANDSHAKE verify flag for client connections. Although the
option is documented as ignored for clients, OpenSSL implicitly enables cert
chain validation when the flag is set.
bpo-30754: Document textwrap.dedent blank line behavior. (GH-14469)
* Added documentation for textwrap.dedent behavior.
* Remove an obsolete note about pre-2.5 behavior from the docstring.
(cherry picked from commit eb97b9211e7c99841d6cae8c63893b3525d5a401)
Steve Dower [Sat, 29 Jun 2019 21:28:59 +0000 (14:28 -0700)]
bpo-37369: Fix initialization of sys members when launched via an app container (GH-14467)
sys._base_executable is now always defined on all platforms, and can be overridden through configuration.
Also adds test.support.PythonSymlink to encapsulate platform-specific logic for symlinking sys.executable
* Fix typo in supports_file2file_sendfile(); ensure that dst is
removed
* Fix test_copytree_custom_copy_function(): remove dst tree.
Use support.rmtree() rather than shutil.rmtree() to remove
temporary directories: support tries harder.
(cherry picked from commit 4c26abd14f1b7242998eb2f7756aa375e0fe714f)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
bpo-29412: Fix indexError when parsing a header value ending unexpectedly (GH-14387) (GH-14411)
* patched string index out of range error in get_word function of _header_value_parser.py and created tests in test__header_value_parser.py for CFWS.
* Raise HeaderParseError instead of continuing when parsing a word.
(cherry picked from commit 7213df7bbfd85378c6e42e1ac63144d5974bdcf6)
Fix test_wsgiref.testEnviron() to no longer depend on the environment
variables (don't fail if "X" variable is set).
testEnviron() now overrides os.environ to get a deterministic
environment. Test full TestHandler.environ content: not only a few
selected variables.
(cherry picked from commit 5150d327924959639215ed0a78feffc0d88258da)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
bpo-37412: os.getcwdb() now uses UTF-8 on Windows (GH-14396)
The os.getcwdb() function now uses the UTF-8 encoding on Windows,
rather than the ANSI code page: see PEP 529 for the rationale. The
function is no longer deprecated on Windows.
os.getcwd() and os.getcwdb() now detect integer overflow on memory
allocations. On Unix, these functions properly report MemoryError on
memory allocation failure.
(cherry picked from commit 689830ee6243126798a6c519c05aa11ba73db7cd)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
When certain malformed messages have content-type set to 'mulitpart/*' but
still have a single part body, iter_attachments can raise AttributeError. This
patch fixes it by returning a None value instead when the body is single part.
(cherry picked from commit 02257012f6d3821d816cb6a7e8461a88a05b9a08)
Use os.getgroups() rather than grp.getgrall() to get groups.
Rename also the test to test_chown_gid().
(cherry picked from commit d7c87d982d4ec4ba201bcee14324ae5e0e90581f)
Co-authored-by: Victor Stinner <vstinner@redhat.com>
bpo-24214: Fixed the UTF-8 and UTF-16 incremental decoders. (GH-14304)
* The UTF-8 incremental decoders fails now fast if encounter
a sequence that can't be handled by the error handler.
* The UTF-16 incremental decoders with the surrogatepass error
handler decodes now a lone low surrogate with final=False.
(cherry picked from commit 894263ba80af4b7733c2df95b527e96953922656)
Victor Stinner [Mon, 24 Jun 2019 10:51:47 +0000 (12:51 +0200)]
bpo-37359: Add --cleanup option to python3 -m test (GH-14332) (GH-14334)
* regrtest: Add --cleanup option to remove "test_python_*" directories
of previous failed test jobs.
* Add "make cleantest" to run "python3 -m test --cleanup".
asyncio: Fix docs for default event loop (GH-14308)
When the Windows default event loop changed, `asyncio-policy.rst` was updated but `asyncio-eventloop.rst` was missed.
(cherry picked from commit 9ffca670ed4df42eb256e7144619a9d3a830e7c0)
Rephrase and clarify that "the entire Python program exits when only daemon threads are left". This matches the documentation at https://docs.python.org/3/library/threading.htmlGH-thread-objects.
(cherry picked from commit bb110cc2ed81447fb48805f31146cf31323a8fc3)
projects / python / log