* Windows: Fix counter name in WindowsLoadTracker. Counter names are
localized: use the registry to get the counter name. Original
change written by Lorenz Mende.
* Regrtest.main() now ensures that the Windows load tracker is also
killed if an exception is raised
* TestWorkerProcess now ensures that worker processes are no longer
running before exiting: kill also worker processes when an
exception is raised.
* Enhance regrtest messages and warnings: include test name,
duration, add a worker identifier, etc.
* Rename MultiprocessRunner to TestWorkerProcess
* Use print_warning() to display warnings.
Co-Authored-By: Lorenz Mende <Lorenz.mende@gmail.com>
Important work originally done by @emilyemorehouse two years ago and nearly ready to go in.
This bug has affected many people and in some cases has been a dealbreaker to the adoption of the otherwise wonderful pathlib and PEP519. https://stackoverflow.com/questions/33625931/copy-file-with-pathlib-in-python.
This adds the outstanding test request from that PR @vstinner (https://github.com/python/cpython/pull/5393).
Test fails without the change, passes with it, along with every other test in test_shutil.
Some variants were experimented with to make the one line change and the most performant one was picked.
# Added Test for PathLike directory destination, the current fail case
```
Lib/test/test_shutil.py::TestMove::test_move_file_pathlike FAILED [100%]
============================================================== FAILURES ===============================================================
__________________________________________________ TestMove.test_move_file_pathlike ___________________________________________________
self = <test.test_shutil.TestMove testMethod=test_move_file_pathlike>
def test_move_file_pathlike(self):
# Move a file to another location on the same filesystem.
src = pathlib.Path(self.src_file)
> self._check_move_file(src, self.dst_dir, self.dst_file)
Lib/test/test_shutil.py:1563:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Lib/test/test_shutil.py:1545: in _check_move_file
shutil.move(src, dst)
/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/shutil.py:562: in move
real_dst = os.path.join(dst, _basename(src))
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
path = PosixPath('/var/folders/r2/psq74t5x3nbfzlph8bh2pvdw0000gn/T/tmp9ie0wh9_/foo')
def _basename(path):
# A basename() variant which first strips the trailing slash, if present.
# Thus we always get the last component of the path, even for directories.
sep = os.path.sep + (os.path.altsep or '')
> return os.path.basename(path.rstrip(sep))
E AttributeError: 'PosixPath' object has no attribute 'rstrip'
/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/shutil.py:526: AttributeError
============================================== 1 failed, 102 deselected in 0.30 seconds ===============================================
```
After change:
```
========================================================= test session starts =========================================================
platform darwin -- Python 3.7.4, pytest-5.0.1, py-1.8.0, pluggy-0.12.0 -- /Users/maxwellmckinnon/.venvs/TA3.7/bin/python3.7
cachedir: .pytest_cache
rootdir: /Users/maxwellmckinnon/dev/cpython
plugins: cov-2.7.1, mock-1.10.4
collected 103 items / 102 deselected / 1 selected
Lib/test/test_shutil.py::TestMove::test_move_file_pathlike PASSED [100%]
============================================== 1 passed, 102 deselected in 0.06 seconds ===============================================
```
Running all the tests in test_shutil.py
```
╰─ pytest Lib/test/test_shutil.py -v
========================================================= test session starts =========================================================
platform darwin -- Python 3.7.4, pytest-5.0.1, py-1.8.0, pluggy-0.12.0 -- /Users/maxwellmckinnon/.venvs/TA3.7/bin/python3.7
cachedir: .pytest_cache
rootdir: /Users/maxwellmckinnon/dev/cpython
plugins: cov-2.7.1, mock-1.10.4
collected 103 items
Lib/test/test_shutil.py::TestShutil::test_chown PASSED [ 0%]
Lib/test/test_shutil.py::TestShutil::test_copy PASSED [ 1%]
...
Lib/test/test_shutil.py::TermsizeTests::test_stty_match SKIPPED [ 99%]
Lib/test/test_shutil.py::PublicAPITests::test_module_all_attribute PASSED [100%]
================================================ 96 passed, 7 skipped in 1.25 seconds =================================================
```
# Performance Considerations
Is it considered poor form to get rid of _basename altogether and make use of pathlib in the move function? I'm not sure if the idea is for all these modules to strictly avoid circular dependencies. They are already using os.path which is just as much a citizen in 3.8 as pathlib right?
e.g.
`real_dst = os.path.join(dst, _basename(src))`
becomes
`real_dst = Path(dst) / Path(src).name`
I've looked around and familiarized myself, and I now think importing pathlib here is fine. My only remaining concern is that of performance.
Here's the performance difference for this step.
```
In [46]: %timeit real_dst = os.path.join("a/b/c", _basename('b/'))
2.71 µs ± 62.6 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
In [47]: %timeit real_dst = Path("a/b/c") / Path('b/').name
12.4 µs ± 65.3 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
```
Is 10us significant or insignificant compared to the least expensive operation this function will do? I don't know. Let's find out.
```
In [55]: %timeit os.rename('/tmp/a/a.txt', '/tmp/a/b.txt'); os.rename('/tmp/a/b.txt', '/tmp/a/a.txt')
124 µs ± 2.18 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)
```
62us to rename. 10us seems significant enough that we wouldn't want to favor the Path sugar suggestion. 16% speed decrease from adding the 10us.
What do people think? I was hoping to get to use pathlib.Path here, but I suspect for this low level move, it should be as fast as possible, and 16% is not worth one line of sugary code to me.
https://bugs.python.org/issue32689
Automerge-Triggered-By: @gvanrossum
Fix warnings options priority: PyConfig.warnoptions has the highest
priority, as stated in the PEP 587.
* Document options order in PyConfig.warnoptions documentation.
* Make PyWideStringList_INIT macro private: replace "Py" prefix
with "_Py".
* test_embed: add test_init_warnoptions().
Document that lnotab can contain invalid bytecode offsets (because of
terrible reasons that are difficult to fix). Make dis.findlinestarts()
ignore invalid offsets in lnotab. All other uses of lnotab in CPython
(various reimplementations of addr2line or line2addr in Python, C and gdb)
already ignore this, because they take an address to look for, instead.
Add tests for the result of dis.findlinestarts() on wacky constructs in
test_peepholer.py, because it's the easiest place to add them.
* bpo-38216: Allow bypassing input validation
* bpo-36274: Also allow the URL encoding to be overridden.
* bpo-38216, bpo-36274: Add tests demonstrating a hook for overriding validation, test demonstrating override encoding, and a test to capture expectation of the interface for the URL.
* Call with skip_host to avoid tripping on the host checking in the URL.
* Remove obsolete comment.
* Make _prepare_path_encoding its own attr.
This makes overriding just that simpler.
Also, don't use the := operator to make backporting easier.
* Add a news entry.
* _prepare_path_encoding -> _encode_prepared_path()
* Once again separate the path validation and request encoding, drastically simplifying the behavior. Drop the guarantee that all processing happens in _prepare_path.
The "Slot" helper (descriptor) is leaking references due to its caching mechanism. The change includes a partial fix to Slot, but also adds Variable.storage to replace the problematic use of Slot.
https://bugs.python.org/issue38187
test_hmac and test_hashlib test built-in hashing implementations and
OpenSSL-based hashing implementations. Add more checks to skip OpenSSL
implementations when a strict crypto policy is active.
Use EVP_DigestInit_ex() instead of EVP_DigestInit() to initialize the
EVP context. The EVP_DigestInit() function clears alls flags and breaks
usedforsecurity flag again.
Signed-off-by: Christian Heimes <christian@python.org>
https://bugs.python.org/issue38270
test_ssl now handles disabled TLS/SSL versions better. OpenSSL's crypto
policy and run-time settings are recognized and tests for disabled versions
are skipped.
Signed-off-by: Christian Heimes <christian@python.org>
https://bugs.python.org/issue38275
* Py_InitializeFromConfig() now writes PyConfig path configuration to
the global path configuration (_Py_path_config).
* Add test_embed.test_get_pathconfig().
* Fix typo in _PyWideStringList_Join().
* Group tests for specific functions and groups of related functions
into separate classes.
* Clean up creating and cleaning up temporary directories.
* Simplify and make more robust monkey patching of shutil.open.
This avoids the buildbot failure on Windows:
```
FileNotFoundError: [WinError 206] The filename or extension is too long: 'd:\\temp\\tmp5r3z438t\\long\\1\\2\\3\\4\\5\\6\\7\\8\\9\\10\\11\\12\\13\\14\\15\\16\\17\\18\\19\\20\\21\\22\\23\\24\\25\\26\\27\\28\\29\\30\\31\\32\\33\\34\\35\\36\\37\\38\\39\\40\\41\\42\\43\\44\\45\\46\\47\\48\\49\\50\\51\\52\\53\\54\\55\\56\\57\\58\\59\\60\\61\\62\\63\\64\\65\\66\\67\\68\\69\\70\\71\\72\\73\\74\\75\\76\\77\\78'
```
Creates a path that's long but avoids OS restrictions.
https://bugs.python.org/issue38112
* Raise the limit of maximum path depth to actual recursion limit
* Add posibilities to adjust a path compiled in .pyc file.
Now, you can:
- Strip a part of path from a beggining of path into compiled file
example "-s /test /test/build/real/test.py" → "build/real/test.py"
- Append some new path to a beggining of path into compiled file
example "-p /boo real/test.py" → "/boo/real/test.py"
You can also use both options in the same time. In that case,
striping is done before appending.
* Add a possibility to specify multiple optimization levels
Each optimization level then leads to separated compiled file.
Use `action='append'` instead of `nargs='+'` for the -o option.
Instead of `-o 0 1 2`, specify `-o 0 -o 1 -o 2`. It's more to type,
but much more explicit.
* Add a symlinks limitation feature
This feature allows us to limit byte-compilation of symbolic
links if they are pointing outside specified dir (build root
for example).
Explicitly preinitializes with a Python preconfiguration to avoid
Py_SetPath() implicit preinitialization with a compat
preconfiguration.
Fix also test_init_setpath() and test_init_setpythonhome() on macOS:
use self.test_exe as the executable (and base_executable), rather
than shutil.which('python3').
* Add test_embed.test_init_setpath_config(): test Py_SetPath()
with PyConfig.
* test_init_setpath() and test_init_setpythonhome() no longer call
Py_SetProgramName(), but use the default program name.
* _PyPathConfig: isolated, site_import and base_executable
fields are now only available on Windows.
* If executable is set explicitly in the configuration, ignore
calculated base_executable: _PyConfig_InitPathConfig() copies
executable to base_executable.
* Complete path config documentation.
* Updated _hashopenssl.c to be PEP 384 compliant
* Remove refleak test from test_hashlib. The updated type no longer accepts random arguments to __init__.
The private keys for test_ssl were encrypted with 3DES in traditional
PKCS#5 format. 3DES and the digest algorithm of PKCS#5 are blocked by
some strict crypto policies. Use PKCS#8 format with AES256 encryption
instead.
Signed-off-by: Christian Heimes <christian@python.org>
https://bugs.python.org/issue38271
Automerge-Triggered-By: @tiran
bpo-38234: On macOS and FreeBSD, the temporary directory can be
symbolic link. For example, /tmp can be a symbolic link to /var/tmp.
Call realpath() to resolve all symbolic links.
Make it easier to run and test Python on systems with restrict crypto policies:
* add requires_hashdigest to test.support to check if a hash digest algorithm is available and working
* avoid MD5 in test_hmac
* replace MD5 with SHA256 in test_tarfile
* mark network tests that require MD5 for MD5-based digest auth or CRAM-MD5
https://bugs.python.org/issue38270
* subprocess.Popen now longer uses posix_spawn() if uid, gid or gids are set.
* test_subprocess: add "nobody" and "nfsnobody" group names for test_group().
* test_subprocess: test_user() and test_group() are now also tested with close_fds=False.
* search_for_prefix() directly calls reduce() if found is greater
than 0.
* Add calculate_pybuilddir() subfunction.
* search_for_prefix(): add path string buffer for readability.
* Fix some error handling code paths: release resources on error.
* calculate_read_pyenv(): rename tmpbuffer to filename.
* test.pythoninfo now also logs windows.dll_path
Add test_init_pybuilddir() and test_init_pyvenv_cfg() to test_embed
to test pyvenv.cfg and pybuilddir.txt configuration files.
Fix sysconfig._generate_posix_vars(): pybuilddir.txt uses UTF-8
encoding, not ASCII.
Multiprocessing test test_mymanager() now also expects -SIGTERM, not
only exitcode 0.
bpo-30356: BaseManager._finalize_manager() sends SIGTERM to the
manager process if it takes longer than 1 second to stop, which
happens on slow buildbots.
Even when the helper is not started yet.
This behavior follows conventional generator one.
There is no reason for `async_generator_athrow` to handle `gen.throw()` differently.
https://bugs.python.org/issue38013
When using multiprocesss (-jN), the main process now uses a timeout
of 60 seconds instead of the double of the --timeout value. The
buildbot server stops a job which does not produce any output in 1200
seconds.
In ArgumentClinic, value "NULL" should now be used only for unrepresentable default values
(like in the optional third parameter of getattr). "None" should be used if None is accepted
as argument and passing None has the same effect as not passing the argument at all.
* Fix a crash in comparing with float (and maybe other crashes).
* They are now never equal to strings and non-integer numbers.
* Comparison with a large number no longer raises OverflowError.
* Arbitrary exceptions no longer silenced in constructors and comparisons.
* TypeError raised in the constructor contains now the name of the type.
* Accept only ChannelID and int-like objects in channel functions.
* Accept only InterpreterId, int-like objects and str in the InterpreterId constructor.
* Accept int-like objects, not just int in interpreter related functions.
https://bugs.python.org/issue34706
Specifically in the case of a class that does not override its
constructor signature inherited from object.
These are Buck Evan @bukzor's changes cherrypicked from GH-9344.
Handle time comparison for cookies with `expires` attribute when `CookieJar.make_cookies` is called.
Co-authored-by: Demian Brecht <demianbrecht@gmail.com>
https://bugs.python.org/issue12144
Automerge-Triggered-By: @asvetlov
The usedforsecurity keyword only argument added to the hash constructors is useful for FIPS builds and similar restrictive environment with non-technical requirements that legacy algorithms be forbidden by their implementations without being explicitly annotated as not being used for any security related purposes. Linux distros with FIPS support benefit from this being standard rather than making up their own way(s) to do it.
Contributed and Signed-off-by: Christian Heimes christian@python.org
* subprocess: Add user, group and extra_groups paremeters to subprocess.Popen
This adds a `user` parameter to the Popen constructor that will call
setreuid() in the child before calling exec(). This allows processes
running as root to safely drop privileges before running the subprocess
without having to use a preexec_fn.
This also adds a `group` parameter that will call setregid() in
the child process before calling exec().
Finally an `extra_groups` parameter was added that will call
setgroups() to set the supplimental groups.
* 1. add test case with wrong behavior
* 2. fix bug when max_length == -1
* 3. allow b"" as valid input data for decompress_buf()
* 4. when max_length >= 0, let needs_input mechanism works
* add more asserts to test case
Test that they do not keep too many file descriptors open for the host OS in a reasonable test scenario.
See [bpo-37935](https://bugs.python.org/issue37935).
Unicode has grown since Python first gained support for it,
when Unicode itself was still rather new.
This pair of test cases was added in commit 6a20ee7de back in 2000,
and they haven't needed to change much since then. But do change
them to look beyond the Basic Multilingual Plane (range(0x10000))
and cover all 17 planes of Unicode's final form.
This adds about 5 seconds to the test suite's runtime. Mark the
tests as CPU-using accordingly.
* bits method and test_bits
* Cleaned up assert string
* blurb
* added docstring
* Faster method, per Eric Smith
* redoing as __format__
* added ipv6 method
* test cases and cleanup
* updated news
* cleanup and NEWS.d
* cleaned up old NEWS
* removed cut and paste leftover
* one more cleanup
* moved to regexp, moved away from v4- and v6-specific versions of __format__
* More cleanup, added ipv6 test cases
* more cleanup
* more cleanup
* cleanup
* cleanup
* cleanup per review, part 1
* addressed review comments around help string and regexp matching
* wrapped v6 test strings. contiguous integers: break at 72char. with underscores: break so that it looks clean.
* 's' and '' tests for pv4 and ipv6
* whitespace cleanup
* Remove trailing whitespace
* Remove more trailing whitespace
* Remove an excess blank line
The socket module now has the socket.send_fds() and socket.recv.fds() functions.
Contributed by Joannah Nanjekye, Shinya Okano (original patch)
and Victor Stinner.
Co-Authored-By: Victor Stinner <vstinner@redhat.com>
Before, running deactivate from a bash shell configured to treat undefined variables as errors (`set -u`) would produce a warning:
```
$ python3 -m venv test
$ source test/bin/activate
(test) $ deactivate
-bash: $1: unbound variable
```
* bpo-20504 : in cgi.py, fix bug when a multipart/form-data request has no content-length header
* Add Misc/NEWS.d/next file.
* Add rst formatting for NEWS.d/next file
* Reaplce assert by self.assertEqual
* bpo-35168: Documentation about shlex.punctuation_chars now states that it should be set in __init__.py
* bpo-35168: Convert shlex.punctuation_chars to read-only property
* Add NEWS.d entry
Different libc implementations have different behavior when presented with trailing % in strftime strings. To make test_strftime_trailing_percent more portable, compare the output of datetime.strftime directly to that of time.strftime rather than hardcoding.
Relative imports use resolve_name to get the absolute target name,
which first seeks the current module's absolute package name from the globals:
If __package__ (and __spec__.parent) are missing then
import uses __name__, truncating the last segment if
the module is a submodule rather than a package __init__.py
(which it guesses from whether __path__ is defined).
The __name__ attempt should fail if there is no parent package (top level modules),
if __name__ is '__main__' (-m entry points), or both (scripts).
That is, if both __name__ has no subcomponents and the module does not seem
to be a package __init__ module then import should fail.
A root cause of bpo-37936 is that it's easy to write a .gitignore
rule that's intended to apply to a specific file (e.g., the
`pyconfig.h` generated by `./configure`) but actually applies to all
similarly-named files in the tree (e.g., `PC/pyconfig.h`.)
Specifically, any rule with no non-trailing slashes is applied in an
"unrooted" way, to files anywhere in the tree. This means that if we
write the rules in the most obvious-looking way, then
* for specific files we want to ignore that happen to be in
subdirectories (like `Modules/config.c`), the rule will work
as intended, staying "rooted" to the top of the tree; but
* when a specific file we want to ignore happens to be at the root of
the repo (like `platform`), then the obvious rule (`platform`) will
apply much more broadly than intended: if someone tries to add a
file or directory named `platform` somewhere else in the tree, it
will unexpectedly get ignored.
That's surprising behavior that can make the .gitignore file's
behavior feel finicky and unpredictable.
To avoid it, we can simply always give a rule "rooted" behavior when
that's what's intended, by systematically using leading slashes.
Further, to help make the pattern obvious when looking at the file and
minimize any need for thinking about the syntax when adding new rules:
separate the rules into one group for each type, with brief comments
identifying them.
For most of these rules it's clear whether they're meant to be rooted
or unrooted, but in a handful of cases I've only guessed. In that
case the safer default (the choice that won't hide information) is the
narrower, rooted meaning, with a leading slash. If for some of these
the unrooted meaning is desired after all, it'll be easy to move them
to the unrooted section at the top.
Fixes a possible hang when using a timeout on subprocess.run() while
capturing output. If the child process spawned its own children or otherwise
connected its stdout or stderr handles with another process, we could hang
after the timeout was reached and our child was killed when attempting to read
final output from the pipes.
The instance destructor for a type is responsible for preparing
an instance for deallocation by decrementing the reference counts
of its referents.
If an instance belongs to a heap type, the type object of an instance
has its reference count decremented while for static types, which
are permanently allocated, the type object is unaffected by the
instance destructor.
Previously, the default instance destructor searched the class
hierarchy for an inherited instance destructor and, if present,
would invoke it.
Then, if the instance type is a heap type, it would decrement the
reference count of that heap type. However, this could result in the
premature destruction of a type because the inherited instance
destructor should have already decremented the reference count
of the type object.
This change avoids the premature destruction of the type object
by suppressing the decrement of its reference count when an
inherited, non-default instance destructor has been invoked.
Finally, an assertion on the Py_SIZE of a type was deleted. Heap
types have a non zero size, making this into an incorrect assertion.
https://github.com/python/cpython/pull/15323
Add functions with various calling conventions to `_testcapi`, expose them as module-level functions, bound methods, class methods, and static methods, and test calling them and introspecting them through GDB.
https://bugs.python.org/issue37499
Co-authored-by: Jeroen Demeyer <J.Demeyer@UGent.be>
Automerge-Triggered-By: @pganssle
This PR deprecate explicit loop parameters in all public asyncio APIs
This issues is split to be easier to review.
Third step: locks.py
https://bugs.python.org/issue36373
* PEP-384 _struct
* More PEP-384 fixes for _struct
Summary: Add a couple of more fixes for `_struct` that were previously missed such as removing `tp_*` accessors and using `PyBytesWriter` instead of calling `PyBytes_FromStringAndSize` with `NULL`. Also added a test to confirm that `iter_unpack` type is still uninstantiable.
* 📜🤖 Added by blurb_it.
Having these in a separate file from the one that's named after the
module in the usual way makes it very easy to miss them when looking
for tests for these two functions.
(In fact when working recently on is_normalized, I'd been surprised to
see no tests for it here and concluded the function had evaded being
tested at all. I'd gone as far as to write up some tests myself
before I spotted this other file.)
Mostly this just means moving all the one file's code into the other,
and moving code from the module toplevel to inside the test class to
keep it tidily separate from the rest of the file's code.
There's one substantive change, which reduces by a bit the amount of
code to be moved: we drop the `x > sys.maxunicode` conditional and all
the `RangeError` logic behind it. Now if that condition ever occurs
it will cause an error at `chr(x)`, and a test failure. That's the
right result because, since PEP 393 in Python 3.3, there is no longer
such a thing as an "unsupported character".
Accumulate certificates in a set instead of doing a costly list contain
operation. A Windows cert store can easily contain over hundred
certificates. The old code would result in way over 5,000 comparison
operations
Signed-off-by: Christian Heimes <christian@python.org>
Since PEP 393 in Python 3.3, this value is always 0x10ffff, the
maximum codepoint in Unicode; there's no longer such a thing as a
UCS-2 build of Python, which couldn't properly represent some
characters.
There are a couple of spots left where we still condition on the value
of this constant. Take them out.
weakref.WeakValueDictionary defines a local remove() function used as
callback for weak references. This function was created with a
closure. Modify the implementation to avoid the closure.
* bpo-26185: Fix repr() on empty ZipInfo object
It was failing on AttributeError due to inexistant
but required attributes file_size and compress_size.
They are now initialized to 0 in ZipInfo.__init__().
* Remove useless hasattr() in ZipInfo._open_to_write()
* Completely remove file_size setting in _open_to_write().
The purpose of the `unicodedata.is_normalized` function is to answer
the question `str == unicodedata.normalized(form, str)` more
efficiently than writing just that, by using the "quick check"
optimization described in the Unicode standard in UAX #15.
However, it turns out the code doesn't implement the full algorithm
from the standard, and as a result we often miss the optimization and
end up having to compute the whole normalized string after all.
Implement the standard's algorithm. This greatly speeds up
`unicodedata.is_normalized` in many cases where our partial variant
of quick-check had been returning MAYBE and the standard algorithm
returns NO.
At a quick test on my desktop, the existing code takes about 4.4 ms/MB
(so 4.4 ns per byte) when the partial quick-check returns MAYBE and it
has to do the slow normalize-and-compare:
$ build.base/python -m timeit -s 'import unicodedata; s = "\uf900"*500000' \
-- 'unicodedata.is_normalized("NFD", s)'
50 loops, best of 5: 4.39 msec per loop
With this patch, it gets the answer instantly (58 ns) on the same 1 MB
string:
$ build.dev/python -m timeit -s 'import unicodedata; s = "\uf900"*500000' \
-- 'unicodedata.is_normalized("NFD", s)'
5000000 loops, best of 5: 58.2 nsec per loop
This restores a small optimization that the original version of this
code had for the `unicodedata.normalize` use case.
With this, that case is actually faster than in master!
$ build.base/python -m timeit -s 'import unicodedata; s = "\u0338"*500000' \
-- 'unicodedata.normalize("NFD", s)'
500 loops, best of 5: 561 usec per loop
$ build.dev/python -m timeit -s 'import unicodedata; s = "\u0338"*500000' \
-- 'unicodedata.normalize("NFD", s)'
500 loops, best of 5: 512 usec per loop
* Use the 'p' format unit instead of manually called PyObject_IsTrue().
* Pass boolean value instead 0/1 integers to functions that needs boolean.
* Convert some arguments to boolean only once.
Fixes a case in which email._header_value_parser.get_unstructured hangs the system for some invalid headers. This covers the cases in which the header contains either:
- a case without trailing whitespace
- an invalid encoded word
https://bugs.python.org/issue37764
This fix should also be backported to 3.7 and 3.8
https://bugs.python.org/issue37764
- drop TargetScopeError in favour of raising SyntaxError directly
as per the updated PEP 572
- comprehension iteration variables are explicitly local, but
named expression targets in comprehensions are nonlocal or
global. Raise SyntaxError as specified in PEP 572
- named expression targets in the outermost iterable of a
comprehension have an ambiguous target scope. Avoid resolving
that question now by raising SyntaxError. PEP 572
originally required this only for cases where the bound name
conflicts with the iteration variable in the comprehension,
but CPython can't easily restrict the exception to that case
(as it doesn't know the target variable names when visiting
the outermost iterator expression)
* fix Path._add_implied_dirs to include all implied directories
* fix Path._add_implied_dirs to include all implied directories
* Optimize code by using sets instead of lists
* 📜🤖 Added by blurb_it.
* fix Path._add_implied_dirs to include all implied directories
* Optimize code by using sets instead of lists
* 📜🤖 Added by blurb_it.
* Add tests to zipfile.Path.iterdir() fix
* Update test for zipfile.Path.iterdir()
* remove whitespace from test file
* Rewrite NEWS blurb to describe the user-facing impact and avoid implementation details.
* remove redundant [] within set comprehension
* Update to use unique_everseen to maintain order and other suggestions in review
* remove whitespace and add back add_dirs in tests
* Add new standalone function parents using posixpath to get parents of a directory
* removing whitespace (sorry)
* Remove import pathlib from zipfile.py
* Rewrite _parents as a slice on a generator of the ancestry of a path.
* Remove check for '.' and '/', now that parents no longer returns those.
* Separate calculation of implied dirs from adding those
* Re-use _implied_dirs in tests for generating zipfile with dir entries.
* Replace three fixtures (abcde, abcdef, abde) with one representative example alpharep.
* Simplify implementation of _implied_dirs by collapsing the generation of parent directories for each name.
PyConfig_Read() is now responsible to handle early calls to
PySys_AddXOption() and PySys_AddWarnOption().
Options added by PySys_AddXOption() are now handled the same way than
PyConfig.xoptions and command line -X options.
For example, PySys_AddXOption(L"faulthandler") enables faulthandler
as expected.
Replace test.support.run_unittest() with unittest.main()
to automatically discover test cases, instead of having
an maintaing manually a list which may be incomplete.
Remove also an unused variable.
Special characters in email address header display names are normally
put within double quotes. However, encoded words (=?charset?x?...?=) are
not allowed withing double quotes. When the header contains a word with
special characters and another word that must be encoded, the first one
must also be encoded.
In the next example, the display name in the From header is quoted and
therefore the comma is allowed; in the To header, the comma is not
within quotes and not encoded, which is not allowed and therefore
rejected by some mail servers.
From: "Foo Bar, France" <foo@example.com>
To: Foo Bar, =?utf-8?q?Espa=C3=B1a?= <foo@example.com>
https://bugs.python.org/issue37482
bpo-37834: Normalise handling of reparse points on Windows
* ntpath.realpath() and nt.stat() will traverse all supported reparse points (previously was mixed)
* nt.lstat() will let the OS traverse reparse points that are not name surrogates (previously would not traverse any reparse point)
* nt.[l]stat() will only set S_IFLNK for symlinks (previous behaviour)
* nt.readlink() will read destinations for symlinks and junction points only
bpo-1311: os.path.exists('nul') now returns True on Windows
* nt.stat('nul').st_mode is now S_IFCHR (previously was an error)
* Write a message when killing a worker process
* Put a timeout on the second popen.communicate() call
(after killing the process)
* Put a timeout on popen.wait() call
* Catch popen.kill() and popen.wait() exceptions
There are plenty of legitimate scripts in the tree that begin with a
`#!`, but also a few that seem to be marked executable by mistake.
Found them with this command -- it gets executable files known to Git,
filters to the ones that don't start with a `#!`, and then unmarks
them as executable:
$ git ls-files --stage \
| perl -lane 'print $F[3] if (!/^100644/)' \
| while read f; do
head -c2 "$f" | grep -qxF '#!' \
|| chmod a-x "$f"; \
done
Looking at the list by hand confirms that we didn't sweep up any
files that should have the executable bit after all. In particular
* The `.psd` files are images from Photoshop.
* The `.bat` files sure look like things that can be run.
But we have lots of other `.bat` files, and they don't have
this bit set, so it must not be needed for them.
Automerge-Triggered-By: @benjaminp
If this service had thoroughly vanished, we could just ignore the
test until someone gets around to either recreating such a service
or redesigning the test to somehow work locally. The
`support.transient_internet` mechanism catches the failure to
resolve the domain name, and skips the test.
But in fact the domain snakebite.net does still exist, as do its
nameservers -- and they can be quite slow to reply. As a result
this test can easily take 20-30s before it gets auto-skipped.
So, skip the test explicitly up front.
The fact that keyword names are strings is now part of the vectorcall and `METH_FASTCALL` protocols. The biggest concrete change is that `_PyStack_UnpackDict` now checks that and raises `TypeError` if not.
CC @markshannon @vstinner
https://bugs.python.org/issue37540
The documented definition was much broader than the real one:
there are tons of characters with general category "Other",
and we don't (and shouldn't) treat most of them as whitespace.
Rewrite the definition to agree with the comment on
_PyUnicode_IsWhitespace, and with the logic in makeunicodedata.py,
which is what generates that function and so ultimately governs.
Add suitable breadcrumbs so that a reader who wants to pin down
exactly what this definition means (what's a "bidirectional class"
of "B"?) can do so. The `unicodedata` module documentation is an
appropriate central place for our references to Unicode's own copious
documentation, so point there.
Also add to the isspace() test a thorough check that the
implementation agrees with the intended definition.
This file started life as a script, before conversion to a
`unittest` test file. Clear out some legacies of that conversion
that are a bit confusing about how it works.
Most notably, it's unlikely there's still a good reason to try
to recover from `unicodedata` failing to import -- as there was
when that logic was first added, when the module was very new.
So take that out entirely. Keep `self.db` working, though, to
avoid a noisy diff.
This should fix the IndexError trying to retrieve `DisplayName.display_name` and `DisplayName.value` when the `value` is basically an empty string.
https://bugs.python.org/issue32178
DeprecationWarning will continue to be emitted for invalid escape
sequences in string and bytes literals just as it did in 3.7.
SyntaxWarning may be emitted in the future. But per mailing list
discussion, we don't yet know when because we haven't settled on how to
do so in a non-disruptive manner.
(Applies 4c5b6bac24 to the master branch).
(This is https://github.com/python/cpython/pull/15142 for master/3.9)
https://bugs.python.org/issue32912
Automerge-Triggered-By: @gpshead
This fixes an inconsistency between the Python and C implementations of
the datetime module. The pure python version of the code was not
accepting offsets greater than 23:59 but less than 24:00. This is an
accidental legacy of the original implementation, which was put in place
before tzinfo allowed sub-minute time zone offsets.
GH-14878
Use warnings.filterwarnings() when importing the deprecated parser
module.
@pablogsal
https://bugs.python.org/issue37268
Automerge-Triggered-By: @pablogsal
There was a discrepancy between the Python and C implementations.
Add singletons ALWAYS_EQ, LARGEST and SMALLEST in test.support
to test mixed type comparison.
Imports now raise `TypeError` instead of `ValueError` for relative import failures. This makes things consistent between `builtins.__import__` and `importlib.__import__` as well as using a more natural import for the failure.
https://bugs.python.org/issue37444
Automerge-Triggered-By: @brettcannon
Previously pdb checked the $HOME environmental variable
to find the user .pdbrc. If $HOME is not set, the user
.pdbrc would not be found.
Change pdb to use `os.path.expanduser('~')` to determine
the user's home directory. Thus, if $HOME is not set (as
in tox or on Windows), os.path.expanduser('~') falls
back on other techniques for locating the user's home
directory.
This follows pip's implementation for loading .piprc.
Co-authored-by: Dan Lidral-Porter <dlp@aperiodic.org>
Support for RFCOMM, L2CAP, HCI, SCO is based on the BTPROTO_* macros
being defined. Winsock only supports RFCOMM, even though it has a
BTHPROTO_L2CAP macro. L2CAP support would build on windows, but not
necessarily work.
This also adds some basic unittests for constants (all of which existed
prior to this commit, just not on windows) and creating sockets.
pair: Nate Duarte <slacknate@gmail.com>
* bpo-37742: Return the root logger when logging.getLogger('root') is called.
* Added type check guard on logger name in logging.getLogger() and refined a test.
BPO -16970: Adding error message for invalid args
Applied the patch argparse-v2 patch issue 16970, ran patch check and the test suite, test_argparse with 0 errors
https://bugs.python.org/issue16970
This changeset increases the default size of the stack
for threads on macOS to the size of the stack
of the main thread and reenables the relevant
recursion test.
Expose the CAN_BCM SocketCAN constants used in the bcm_msg_head struct
flags (provided by <linux/can/bcm.h>) under the socket library.
This adds the following constants with a CAN_BCM prefix:
* SETTIMER
* STARTTIMER
* TX_COUNTEVT
* TX_ANNOUNCE
* TX_CP_CAN_ID
* RX_FILTER_ID
* RX_CHECK_DLC
* RX_NO_AUTOTIMER
* RX_ANNOUNCE_RESUME
* TX_RESET_MULTI_IDX
* RX_RTR_FRAME
* CAN_FD_FRAME
The CAN_FD_FRAME flag was introduced in the 4.8 kernel, while the other
ones were present since SocketCAN drivers were mainlined in 2.6.25. As
such, it is probably unnecessary to guard against these constants being
missing.
Mark some individual tests to skip when --pgo is used. The tests
marked increase the PGO task time significantly and likely don't
help improve optimization of the final executable.
Deprecate the parser module and add a deprecation warning triggered on import and a warning block in the documentation.
https://bugs.python.org/issue37268
Automerge-Triggered-By: @pablogsal
* bpo-37399: Correctly attach tail text to the last element/comment/pi, even when comments or pis are discarded.
Also fixes the insertion of PIs when "insert_pis=True" is configured for a TreeBuilder.
Reduce the number of unit tests run for the PGO generation task. This
speeds up the task by a factor of about 15x. Running the full unit test
suite is slow. This change may result in a slightly less optimized build
since not as many code branches will be executed. If you are willing to
wait for the much slower build, the old behavior can be restored using
'./configure [..] PROFILE_TASK="-m test --pgo-extended"'. We make no
guarantees as to which PGO task set produces a faster build. Users who
care should run their own relevant benchmarks as results can depend on
the environment, workload, and compiler tool chain.
A bug in MSVC UCRT version 17763.615 (which has been fixed in newer versions) is causing test failures in some strptime/strftime tests when the default code page is c65001. This change selectively skips the tests affected by this.
* bpo-37461: Fix infinite loop in parsing of specially crafted email headers.
Some crafted email header would cause the get_parameter method to run in an
infinite loop causing a DoS attack surface when parsing those headers. This
patch fixes that by making sure the DQUOTE character is handled to prevent
going into an infinite loop.
As far as I can tell, this infinite loop would be triggered if:
1. The value being folded contains a single word (no spaces) longer than
max_line_length
2. The max_line_length is shorter than the encoding's name + 9
characters.
bpo-36564: https://bugs.python.org/issue36564
bpo-15386, bpo-37473: test_import, regrtest and libregrtest no longer
import importlib as soon as possible, as the first import, "to test
bpo-15386".
It is tested by test_import.test_there_can_be_only_one().
Sort test_import imports.
The `allow_abbrev` option for ArgumentParser is documented and intended to disable support for unique prefixes of --options, which may sometimes be ambiguous due to deferred parsing.
However, the initial implementation also broke parsing of grouped short flags, such as `-ab` meaning `-a -b` (or `-a=b`). Checking the argument for a leading `--` before rejecting it fixes this.
This was prompted by pytest-dev/pytest#5469, so a backport to at least 3.8 would be great 😄
And this is my first PR to CPython, so please let me know if I've missed anything!
https://bugs.python.org/issue26967
Returns NotImplemented for timedelta and time in __eq__ for different types in Python implementation, which matches the C implementation.
This also adds tests to enforce that these objects will fall back to the right hand side's __eq__ and/or __ne__ implementation.
bpo-37579
Move `threaded_import_hangers`, a dependency of `test_threaded_import`, to the directory `test_importlib/`. Also update the import references for `threaded_import_hangers` in `test_threaded_import`.
https://bugs.python.org/issue19696
With the addition of shared memory into Python 3.8, we now have three tests failing on Solaris, namely `test_multiprocessing_fork`, `test_multiprocessing_forkserver` and `test_multiprocessing_spawn`. The reason seems to be incorrect name handling which results in two slashes being prepended.
https://bugs.python.org/issue37558
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).
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.
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.
Replacing the deprecated method "random.choose" to "random.choice" was technically not part of the original issue. However, it was discussed in the talk page and involved one of the files being moved. I assumed this was too minor to justify the creation of a separate issue.
Also, I added my name to the contributors list in Misc/ACKS. This will be my third PR to cpython, forgot to do it in the previous ones.
https://bugs.python.org/issue19696
test_concurrent_futures now cleans up multiprocessing to remove
immediately temporary directories created by
multiprocessing.util.get_temp_dir().
The test now uses setUpModule() and tearDownModule().
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.
Signed-off-by: Christian Heimes <christian@python.org>
https://bugs.python.org/issue37463
urllib.request tests now call urlcleanup() to remove temporary files
created by urlretrieve() tests and to clear the _opener global
variable set by urlopen() and functions calling indirectly urlopen().
regrtest now checks if urllib.request._url_tempfiles and
urllib.request._opener are changed by tests.
multiprocessing tests now call explicitly _run_finalizers() to remove
immediately temporary directories created by
multiprocessing.util.get_temp_dir().
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.
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.
Stop using "static PyConfig", PyConfig must now always use
dynamically allocated strings: use PyConfig_SetString(),
PyConfig_SetArgv() and PyConfig_Clear().
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.
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.
Signed-off-by: Christian Heimes <christian@python.org>
https://bugs.python.org/issue37428
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
As noted by @eryksun in [1] and [2], using _cleanup and _active(in
__del__) is not necessary on Windows, since:
> Unlike Unix, a process in Windows doesn't have to be waited on by
> its parent to avoid a zombie. Keeping the handle open will actually
> create a zombie until the next _cleanup() call, which may be never
> if Popen() isn't called again.
This patch simply defines `subprocess._active` as `None`, for which we already
have the proper logic in place in `subprocess.Popen.__del__`, that prevents it
from trying to append the process to the `_active`. This patch also defines
`subprocess._cleanup` as a noop for Windows.
[1] https://bugs.python.org/issue37380#msg346333
[2] https://bugs.python.org/issue36067#msg336262
Signed-off-by: Ruslan Kuprieiev <ruslan@iterative.ai>
On Windows, test.pythoninfo now checks if support for long paths is
enabled using ntdll.RtlAreLongPathsEnabled() function.
Co-Authored-By: Eryk Sun <eryksun@gmail.com>
* 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.
* 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.
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.
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.
In development mode and in debug build, encoding and errors arguments
are now checked on string encoding and decoding operations. Examples:
open(), str.encode() and bytes.decode().
By default, for best performances, the errors argument is only
checked at the first encoding/decoding error, and the encoding
argument is sometimes ignored for empty strings.
* bpo-33972: Fix EmailMessage.iter_attachments raising AttributeError.
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.
Python now gets the absolute path of the script filename specified on
the command line (ex: "python3 script.py"): the __file__ attribute of
the __main__ module, sys.argv[0] and sys.path[0] become an absolute
path, rather than a relative path.
* Add _Py_isabs() and _Py_abspath() functions.
* _PyConfig_Read() now tries to get the absolute path of
run_filename, but keeps the relative path if _Py_abspath() fails.
* Reimplement os._getfullpathname() using _Py_abspath().
* Use _Py_isabs() in getpath.c.
* 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.
Remove sys.getcheckinterval() and sys.setcheckinterval() functions.
They were deprecated since Python 3.2. Use sys.getswitchinterval()
and sys.setswitchinterval() instead.
Remove also check_interval field of the PyInterpreterState structure.
* regrtest: Add --cleanup option to remove "test_python_*" directories
of previous failed test jobs.
* Add "make cleantest" to run "python3 -m test --cleanup".
At the moment you can definitely use UDPLITE sockets on Linux systems, but it would be good if this support were formalized such that you can detect support at runtime easily.
At the moment, to make and use a UDPLITE socket requires something like the following code:
```
>>> import socket
>>> a = socket.socket(socket.AF_INET, socket.SOCK_DGRAM, 136)
>>> b = socket.socket(socket.AF_INET, socket.SOCK_DGRAM, 136)
>>> a.bind(('localhost', 44444))
>>> b.sendto(b'test'*256, ('localhost', 44444))
>>> b.setsockopt(136, 10, 16)
>>> b.sendto(b'test'*256, ('localhost', 44444))
>>> b.setsockopt(136, 10, 32)
>>> b.sendto(b'test'*256, ('localhost', 44444))
>>> b.setsockopt(136, 10, 64)
>>> b.sendto(b'test'*256, ('localhost', 44444))
```
If you look at this through Wireshark, you can see that the packets are different in that the checksums and checksum coverages change.
With the pull request that I am submitting momentarily, you could do the following code instead:
```
>>> import socket
>>> a = socket.socket(socket.AF_INET, socket.SOCK_DGRAM, socket.IPPROTO_UDPLITE)
>>> b = socket.socket(socket.AF_INET, socket.SOCK_DGRAM, socket.IPPROTO_UDPLITE)
>>> a.bind(('localhost', 44444))
>>> b.sendto(b'test'*256, ('localhost', 44444))
>>> b.set_send_checksum_coverage(16)
>>> b.sendto(b'test'*256, ('localhost', 44444))
>>> b.set_send_checksum_coverage(32)
>>> b.sendto(b'test'*256, ('localhost', 44444))
>>> b.set_send_checksum_coverage(64)
>>> b.sendto(b'test'*256, ('localhost', 44444))
```
One can also detect support for UDPLITE just by checking
```
>>> hasattr(socket, 'IPPROTO_UDPLITE')
```
https://bugs.python.org/issue37345
When the test is ran with `PYTHONWARNINGS=error` the environment variable is passed to the python interpreter used in `assert_python_ok` where `DeprecationWarning` from `@asyncio.coroutine` is converted into an error. Ignore the `DeprecationWarning` in `assert_python_ok`.
https://bugs.python.org/issue37323
test_gdb no longer fails if it gets an "unexpected" message on
stderr: it now ignores stderr. The purpose of test_gdb is to test
that python-gdb.py commands work as expected, not to test gdb.
For datetime.datetime.strptime(), the leading zero for some two-digit formats is optional.
This adds a footnote to the strftime/strptime documentation to reflect this fact, and adds some tests to ensure that it is true.
bpo-34903
aifc.openfp() alias to aifc.open(), sunau.openfp() alias to
sunau.open(), and wave.openfp() alias to wave.open() have been
removed. They were deprecated since Python 3.7.
Add --upgrade-deps to venv module
- This allows for pip + setuptools to be automatically upgraded to the latest version on PyPI
- Update documentation to represent this change
bpo-34556: Add --upgrade to venv module
bpo-35031, bpo-35998: Reintroduce workaround on Windows and FreeBSD
in test_start_tls_server_1() of test_asyncio: disable TLS v1.3 on the
client context.
uuid could try fallback methods that had no chance of working on a particular
platform, and this could cause spurious test failures, as well as degraded
performance as fallback options were tried and failed.
This fixes both the uuid module and its test's SkipUnless logic to use a
prefiltered list of techniques that may at least potentially work on that platform.
Patch by Michael Felt (aixtools).
In a subinterpreter, spawning a daemon thread now raises an
exception. Daemon threads were never supported in subinterpreters.
Previously, the subinterpreter finalization crashed with a Pyton
fatal error if a daemon thread was still running.
* Add _thread._is_main_interpreter()
* threading.Thread.start() now raises RuntimeError if the thread is a
daemon thread and the method is called from a subinterpreter.
* The _thread module now uses Argument Clinic for the new function.
* Use textwrap.dedent() in test_threading.SubinterpThreadingTests
Document reference cycle and resurrected objects issues in
sys.unraisablehook() and threading.excepthook() documentation.
Fix test.support.catch_unraisable_exception(): __exit__() no longer
ignores unraisable exceptions.
Fix test_io test_writer_close_error_on_close(): use a second
catch_unraisable_exception() to catch the BufferedWriter unraisable
exception.
Join the thread to prevent leaking a running thread and leaking a
reference.
Cleanup also the test:
* asyncioWindowsProactorEventLoopPolicy became the default policy,
there is no need to set it manually.
* Only start the thread once the loop is running.
* Use a shorter sleep in the thread (100 ms rather than 1 sec).
* Use close_loop(loop) rather than loop.close().
* Use longer variable names.
Fix a regression introduced by af8646c805 that was causing code of the form:
if True and False:
do_something()
to be optimized incorrectly, eliminating the block.
The peephole optimizer was not optimizing correctly bytecode after negative deltas were introduced. This is due to the fact that some special values (255) were being searched for in both instruction pointer delta and line number deltas.
The __exit__() method of test.support.catch_unraisable_exception
context manager now ignores unraisable exception raised when clearing
self.unraisable attribute.
Allow pure Python implementation of pickle to work
even when the C _pickle module is unavailable.
Fix test_pickle when _pickle is missing: declare PyPicklerHookTests
outside "if has_c_implementation:" block.
Fix a race condition at Python shutdown when waiting for threads.
Wait until the Python thread state of all non-daemon threads get
deleted (join all non-daemon threads), rather than just wait until
Python threads complete.
* Add threading._shutdown_locks: set of Thread._tstate_lock locks
of non-daemon threads used by _shutdown() to wait until all Python
thread states get deleted. See Thread._set_tstate_lock().
* Add also threading._shutdown_locks_lock to protect access to
threading._shutdown_locks.
* Add test_finalization_shutdown() test.
regrtest now uses sys.unraisablehook() to mark a test as "environment
altered" (ENV_CHANGED) if it emits an "unraisable exception".
Moreover, regrtest logs a warning in this case.
Use "python3 -m test --fail-env-changed" to catch unraisable
exceptions in tests.
Use catch_unraisable_exception() to ignore 'Exception ignored in:'
error when the internal BufferedWriter of the BufferedRWPair is
destroyed. The C implementation doesn't give access to the
internal BufferedWriter, so just ignore the warning instead.
Rename compile() feature_version parameter to _feature_version and
convert it to a keyword-only parameter.
Update also test_type_comments to pass feature_version as a tuple.
Currently, inspect.getfile(str) will report nonsense:
```pytb
>>> inspect.getfile(str)
TypeError: <module 'builtins' (built-in)> is a built-in class
```
This fixes that
https://bugs.python.org/issue37173
There is a possibility that someone (like me) accidentally will omit parentheses with `FileType` arguments after `FileType`, and parser will contain wrong file until someone will try to use it.
Example:
```python
parser = argparse.ArgumentParser()
parser.add_argument('-x', type=argparse.FileType)
```
https://bugs.python.org/issue37150
Replace two Python function calls with a single one to ensure that no
memory allocation is done between the invalid object is created and
when _PyObject_IsFreed() is called.
* bpo-36520: reset the encoded word offset when starting a new
line during an email header folding operation
* 📜🤖 Added by blurb_it.
* bpo-36520: add an additional test case, and provide descriptive
comments for the test_folding_with_utf8_encoding_* tests
* bpo-36520: fix whitespace issue
* bpo-36520: changes per reviewer request -- remove extraneous
backslashes; add whitespace between terminating quotes and
line-continuation backslashes; use "bpo-" instead of
"issue #" in comments
In test_fromkeys() the derived test class now supports all arguments in its
constructor so that the class to be tested can use its own constructor in its
fromkeys() implementation.
In test_mutatingiteration() the test fails as soon as iterating over a
dictionary with one entry and adding new entries in the loop iterates more
than once (to avoid endless loops in faulty implementations).
https://bugs.python.org/issue2661
* bpo-21315: Fix parsing of encoded words with missing leading ws.
Because of missing leading whitespace, encoded word would get parsed as
unstructured token. This patch fixes that by looking for encoded words when
splitting tokens with whitespace.
Missing trailing whitespace around encoded word now register a defect
instead.
Original patch suggestion by David R. Murray on bpo-21315.
* bpo-30835: email: Fix AttributeError when parsing invalid Content-Transfer-Encoding
Parsing an email containing a multipart Content-Type, along with a
Content-Transfer-Encoding containing an invalid (non-ASCII-decodable) byte
will fail. email.feedparser.FeedParser._parsegen() gets the header and
attempts to convert it to lowercase before comparing it with the accepted
encodings, but as the header contains an invalid byte, it's returned as a
Header object rather than a str.
Cast the Content-Transfer-Encoding header to a str to avoid this.
Found using the AFL fuzzer.
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Andrew Donnellan <andrew@donnellan.id.au>
* Add email and NEWS entry for the bugfix.
* bpo-35805: Add parser for Message-ID header.
This parser is based on the definition of Identification Fields from RFC 5322
Sec 3.6.4.
This should also prevent folding of Message-ID header using RFC 2047 encoded
words and hence fix bpo-35805.
* Prevent folding of non-ascii message-id headers.
* Add fold method to MsgID token to prevent folding.
Replace asyncio.set_event_loop() with TestCase.set_event_loop() of
test_asyncio.utils: this method calls TestCase.close_loop() which
waits until the executor completes, to avoid leaking dangling
threads.
Inherit from test_asyncio.utils.TestCase rather than
unittest.TestCase.
Modify test_coroutines, test_cprofile, test_generators, test_raise,
test_ssl and test_yield_from to use
support.catch_unraisable_exception() rather than
support.captured_stderr().
test_thread: remove test_save_exception_state_on_error() which is now
updated. test_unraisable_exception() checks that sys.unraisablehook()
is called to handle _thread.start_new_thread() exception.
test_cprofile now rely on unittest for test discovery: replace
support.run_unittest() with unittest.main().
When inheriting a heap subclass from a vectorcall class that sets
`.tp_call=PyVectorcall_Call` (as recommended in PEP 590), the subclass does
not inherit `_Py_TPFLAGS_HAVE_VECTORCALL`, and thus `PyVectorcall_Call` does
not work for it.
This attempts to solve the issue by:
* always inheriting `tp_vectorcall_offset` unless `tp_call` is overridden
in the subclass
* inheriting _Py_TPFLAGS_HAVE_VECTORCALL for static types, unless `tp_call`
is overridden
* making `PyVectorcall_Call` ignore `_Py_TPFLAGS_HAVE_VECTORCALL`
This means it'll be ever more important to only call `PyVectorcall_Call`
on classes that support vectorcall. In `PyVectorcall_Call`'s intended role
as `tp_call` filler, that's not a problem.
Fixed QueueListener in order to avoid random deadlocks.
Unable to add regression tests atm due to time constraints, will add it in a bit.
Regarding implementation, although it's nested, it does not cause performance issues whatsoever, and does not call task_done() in case of an exception (which is the right thing to do IMHO).
https://bugs.python.org/issue36813
Adds a new option in trace that allows tracing runnable modules. It is
exposed as `--module module_name` as `-m` is already in use for another
argument.
The ssl module now can dump key material to a keylog file and trace TLS
protocol messages with a tracing callback. The default and stdlib
contexts also support SSLKEYLOGFILE env var.
The msg_callback and related enums are private members. The feature
is designed for internal debugging and not for end users.
Signed-off-by: Christian Heimes <christian@python.org>
This is an old feature request that appears from time to time. After a year of experimenting with various introspection capabilities in `typing_inspect` on PyPI, I propose to add these two most commonly used functions: `get_origin()` and `get_args()`. These are essentially thin public wrappers around private APIs: `__origin__` and `__args__`.
As discussed in the issue and on the typing tracker, exposing some public helpers instead of `__origin__` and `__args__` directly will give us more flexibility if we will decide to update the internal representation, while still maintaining backwards compatibility.
The implementation is very simple an is essentially a copy from `typing_inspect` with one exception: `ClassVar` was special-cased in `typing_inspect`, but I think this special-casing doesn't really help and only makes things more complicated.
From 3.8 async functions used with mock.patch return an `AsyncMock`. `_accept_connection2` is an async function where create_task is also mocked. Don't mock `create_task` so that tasks are created out of coroutine returned by `AsyncMock` and the tasks are completed.
https://bugs.python.org/issue37015
As it changes the way functions are called, the PEP 590 implementation
skipped the functions that the GDB integration is looking for
(by name) to find function calls.
Looking for the new helper `cfunction_call_varargs` hopefully fixes the
tests, and thus buildbots.
The changed frame nuber in test_gdb is due to there being fewer
C calls when calling a built-in method.
* bpo-26836: Add os.memfd_create()
* Use the glibc wrapper for memfd_create()
Co-Authored-By: Christian Heimes <christian@python.org>
* Fix deletions caused by autoreconf.
* Use MFD_CLOEXEC as the default value for *flags*.
* Add memset_s to configure.ac.
* Revert memset_s changes.
* Apply the requested changes.
* Tweak the docs.
* bpo-22385: Support output separators in hex methods.
Also in binascii.hexlify aka b2a_hex.
The underlying implementation behind all hex generation in CPython uses the
same pystrhex.c implementation. This adds support to bytes, bytearray,
and memoryview objects.
The binascii module functions exist rather than being slated for deprecation
because they return bytes rather than requiring an intermediate step through a
str object.
This change was inspired by MicroPython which supports sep in its binascii
implementation (and does not yet support the .hex methods).
https://bugs.python.org/issue22385
Add explicit `asyncSetUp` and `asyncTearDown` methods.
The rest is the same as for #13228
`AsyncTestCase` create a loop instance for every test for the sake of test isolation.
Sometimes a loop shared between all tests can speed up tests execution time a lot but it requires control of closed resources after every test finish. Basically, it requires nested supervisors support that was discussed with @1st1 many times. Sorry, asyncio supervisors have no chance to land on Python 3.8.
The PR intentionally does not provide API for changing the used event loop or getting the test loop: use `asyncio.set_event_loop_policy()` and `asyncio.get_event_loop()` instead.
The PR adds four overridable methods to base `unittest.TestCase` class:
```
def _callSetUp(self):
self.setUp()
def _callTestMethod(self, method):
method()
def _callTearDown(self):
self.tearDown()
def _callCleanup(self, function, /, *args, **kwargs):
function(*args, **kwargs)
```
It allows using asyncio facilities with minimal influence on the unittest code.
The last but not least: the PR respects contextvars. The context variable installed by `asyncSetUp` is available on test, `tearDown` and a coroutine scheduled by `addCleanup`.
https://bugs.python.org/issue32972
_thread.start_new_thread() now logs uncaught exception raised by the
function using sys.unraisablehook(), rather than sys.excepthook(), so
the hook gets access to the function which raised the exception.
I tried to get rid of the `_ProtocolMeta`, but unfortunately it didn'y work. My idea to return a generic alias from `@runtime_checkable` made runtime protocols unpickleable. I am not sure what is worse (a custom metaclass or having some classes unpickleable), so I decided to stick with the status quo (since there were no complains so far). So essentially this is a copy of the implementation in `typing_extensions` with two modifications:
* Rename `@runtime` to `@runtime_checkable` (plus corresponding updates).
* Allow protocols that extend `collections.abc.Iterable` etc.
It has been documented as deprecated and to be removed in 3.8;
From a comment on another thread – which I can't find ; leave get_coro_wrapper() for now, but always return `None`.
https://bugs.python.org/issue36933
Add a new threading.excepthook() function which handles uncaught
Thread.run() exception. It can be overridden to control how uncaught
exceptions are handled.
threading.ExceptHookArgs is not documented on purpose: it should not
be used directly.
* threading.excepthook() and threading.ExceptHookArgs.
* Add _PyErr_Display(): similar to PyErr_Display(), but accept a
'file' parameter.
* Add _thread._excepthook(): C implementation of the exception hook
calling _PyErr_Display().
* Add _thread._ExceptHookArgs: structseq type.
* Add threading._invoke_excepthook_wrapper() which handles the gory
details to ensure that everything remains alive during Python
shutdown.
* Add unit tests.
When using the "=" debug functionality of f-strings, use another Constant node (or a merged constant node) instead of adding expr_text to the FormattedValue node.
This will address the common mistake many asyncio users make:
an "except Exception" clause breaking Tasks cancellation.
In addition to this change, we stop inheriting asyncio.TimeoutError
and asyncio.InvalidStateError from their concurrent.futures.*
counterparts. There's no point for these exceptions to share the
inheritance chain.
In 3.9 we'll focus on implementing supervisors and cancel scopes,
which should allow better handling of all exceptions, including
SystemExit and KeyboardInterrupt
* sys.unraisablehook: add 'err_msg' field to UnraisableHookArgs.
* Use _PyErr_WriteUnraisableMsg() in _ctypes _DictRemover_call()
and gc delete_garbage().
The implementation is straightforward, it just mimics `ClassVar` (since the latter is also a name/access qualifier, not really a type). Also it is essentially copied from `typing_extensions`.
Previously, it was hard to tell whether a function should be awaited. It was also incorrect (per PEP 484) to put this in the type hint for coroutine functions. Added this info to the output of builtins.help and pydoc.
https://bugs.python.org/issue36045
* _PyPreConfig_InitCompatConfig() sets utf8_mode to 0.
* Change Py_UTF8Mode default value to 0.
* Fix _PyPreConfig_Copy(): copy also _config_init attrbibute.
* _PyPreConfig_AsDict() exports _config_init
* Fix _PyPreConfig_GetGlobalConfig(): use Py_UTF8Mode if it's greater
than 0, even if utf8_mode >= 0.
* Add unit tests on environment variables using Python API.
In development (-X dev) mode and in a debug build, IOBase finalizer
of the _pyio module now logs the exception if the close() method
fails. The exception is ignored silently by default in release build.
test_io: test_error_through_destructor() now uses
support.catch_unraisable_exception() rather than capturing stderr.
PyErr_WriteUnraisable() now creates a traceback object if there is no
current traceback. Moreover, call PyErr_NormalizeException() and
PyException_SetTraceback() to normalize the exception value. Ignore
silently any error.
* _PyPreConfig_GetGlobalConfig() and _PyCoreConfig_GetGlobalConfig()
now do nothing if the configuration was not initialized with
_PyPreConfig_InitCompatConfig() and _PyCoreConfig_InitCompatConfig()
* Remove utf8_mode=-2 special case: use utf8_mode=-1 instead.
* Fix _PyPreConfig_InitPythonConfig():
* isolated = 0 instead of -1
* use_environment = 1 instead of -1
* Rename _PyConfig_INIT to _PyConfig_INIT_COMPAT
* Rename _PyPreConfig_Init() to _PyPreConfig_InitCompatConfig()
* Rename _PyCoreConfig_Init() to _PyCoreConfig_InitCompatConfig()
* PyInterpreterState_New() now uses _PyCoreConfig_InitPythonConfig()
as default configuration, but it's very quickly overriden anyway.
* _freeze_importlib.c uses _PyCoreConfig_SetString() to set
program_name.
* Cleanup preconfig_init_utf8_mode(): cmdline is always non-NULL.
* Copy test_exceptions.test_unraisable() to
test_sys.UnraisableHookTest().
* Use catch_unraisable_exception() in test_coroutines,
test_exceptions, test_generators.
* Fixes issue 24882
* Add news file entry for change.
* Change test_concurrent_futures.ThreadPoolShutdownTest
Adjust the shutdown test so that, after submitting three jobs
to the executor, the test checks for less than three threads,
instead of looking for exactly three threads.
If idle threads are being recycled properly, then we should have
less than three threads.
* Switched idle count to semaphor, Updated tests
As suggested by reviewer tomMoral, swapped lock-protected counter
with a semaphore to track the number of unused threads.
Adjusted test_threads_terminate to wait for completiton of the
previous future before submitting a new one (and checking the
number of threads used).
Also added a new test to confirm the thread pool can be saturated.
* Updates tests as requested by pitrou.
* Correct minor whitespace error.
* Make test_saturation faster
Wrap the callback call within the `add_done_callback` function within concurrent.futures, in order to behave in an identical manner to callbacks added to a running future are triggered once it has completed.
This disallows things like `# type: ignoreé`, which seems wrong.
Also switch to using Py_ISALNUM for the alnum check, for consistency
with other code (and maybe correctness re: locale issues?).
https://bugs.python.org/issue36878
CVE-2019-9948: Avoid file reading as disallowing the unnecessary URL
scheme in URLopener().open() and URLopener().retrieve()
of urllib.request.
Co-Authored-By: SH <push0ebp@gmail.com>
GH-13238 made extra text after a # type: ignore accepted by the parser.
This finishes the job and actually plumbs the extra text through the
parser and makes it available in the AST.
In order to support typing checks calling hex(), oct() and bin() on user-defined classes, a SupportIndex protocol is required. The ability to check these at runtime would be good to add for completeness sake. This is pretty much just a copy of SupportsInt with the names tweaked.
It adds a missing testcase for bpo-34125. This is testing code which is
affected by PEP 590, so missing this test might accidentally break
CPython if we screw up with implementing PEP 590.
Add new sys.unraisablehook() function which can be overridden to
control how "unraisable exceptions" are handled. It is called when an
exception has occurred but there is no way for Python to handle it.
For example, when a destructor raises an exception or during garbage
collection (gc.collect()).
Changes:
* Add an internal UnraisableHookArgs type used to pass arguments to
sys.unraisablehook.
* Add _PyErr_WriteUnraisableDefaultHook().
* The default hook now ignores exception on writing the traceback.
* test_sys now uses unittest.main() to automatically discover tests:
remove test_main().
* Add _PyErr_Init().
* Fix PyErr_WriteUnraisable(): hold a strong reference to sys.stderr
while using it
Victor Stinner
Maxwell A McKinnon
Yury Selivanov
Yury Selivanov
Giampaolo Rodola
Andrew Svetlov
T. Wouters
Jason R. Coombs
Dong-hee Na
Eric Snow
Christian Heimes
Michael Felt
Petr Viktorin
Serhiy Storchaka
Lumír 'Frenzy' Balhar
PatrikKopkan
Victor Stinner
Vinay Sajip
Ammar Askar
Eddie Elizondo
t k
Pablo Galindo
Zackery Spytz
plokmijnuhby
Benjamin Peterson
Gregory P. Smith
Ashwin Ramaswami
Xtreak
Rémi Lapeyre
Patrick McLean
animalize
Berker Peksag
Hai Shi
Divij Rajkumar
bggardner
Greg Price
ewosborne
Joannah Nanjekye
Daniel Andrade
Daniel Abrahamsson
Pavel Koneski
Jakub Kulík
hui shang
Géry Ogam
Pierre Quentel
Alex
Ben Lewis
Steve Dower
Eddie Elizondo
Gordon P. Hemsley
Jeroen Demeyer
Emmanuel Arias
hliu0
Dino Viehland
Vinay Sharma
Sergey Fedoseev
Anthony Sottile
Zeth
Mickaël Schoentgen
Daniel Hahler
Raymond Hettinger
HongWeipeng
Jon Janzen
Min ho Kim
Thomas A Caswell
Justin Blanchard
Ronald Oussoren
Daniel Fortunov
Inada Naoki
Forest Gregg
Flavian Hautbois
Nick Coghlan
shireenrao
Stefan Behnel
Ivan Levkivskyi
bsiem
Eric V. Smith
Abhilash Raj
Ngalim Siregar
Paul Monson
aiudirog
Timothy Hopper
Greg Bowser
tmblweed
karl ding
Neil Schemenauer
Markus Mohrhard
Walter Dörwald
jpic
Paul Ganssle
Zac Hatfield-Dodds
Kyle Stanley
Carl Friedrich Bolz-Tereick
thatneat
Ned Deily
Kyle Stanley
Ruslan Kuprieiev
Brandt Bucher
Carl Bordum Hansen
Pierre Glaser
David K. Hess
Gabe Appleton
Jörn Heissler
Mike Gleen
Cooper Lees
Joost Lek
Philipp A
zygocephalus
websurfer5
Augusto Hack
Mark Dickinson
Evan
Bar Harel
Yash Aggarwal
Mario Corchero
Rob Day
Ammar Askar
xdegaye
Alex Grönholm
Bo Bayles
依云
Matthias Bussonnier
sbstp
Antoine Pitrou
Chih-Hsuan Yen
Dan Rose
Matěj Cepl
Steve Dower
Batuhan Taşkaya
Sean
Sam Martin
Michael J. Sullivan
Victor Stinner
Jake Tesler
Jelle Zijlstra
Paul Dagnelie
Marcel Plch