cpython

cpython copied to clipboard

- Switch PyUnicode_InternInPlace to _PyUnicode_InternMortal - Keep immortalizing in functions that take `char*`, and are almost exclusively used with strings known at compile-time: - PyUnicode_InternFromString - PyDict_SetItemString - PyObject_SetAttrString -...

encukou

Interned strings are immortal, despite what the documentation says

36

# Bug report ### Bug description: The [`sys.intern` documentation](https://docs.python.org/3.13/library/sys.html#sys.intern) explicitly says: > Interned strings are not [immortal](https://docs.python.org/3.13/glossary.html#term-immortal); you must keep a reference to the return value of [`intern()`](https://docs.python.org/3.13/library/sys.html#sys.intern) around to...

encukou

There are validations in compile.c that can move to earlier stages (AST validation or symtable construction). And the compiler is modifying the symbol table (setting ``ste_coroutine``), which it really shouldn't...

iritkatriel

gh-121160: Add a test for readline.set_history_length

2

I could not reproduce the reported issue, but, I didn't try very hard to test on various readline libraries. Regardless, `set_history_length` was missing a test. This PR should show if...

encukou

gh-119993 ignore `NotADirectoryError` in `Path.unlink()` if `missing_ok` is `True`

1

Add a short note on when Path.unlink() raises `NotADirectoryError` This can occur "in the wild", for example, after calling `shutil.move(src, dst)` where `dst` is a non-existent directory. Debugging can cost...

MusicalNinjaDad

gh-113993: For string interning, do not rely on (or assert) _Py_IsImmortal

1

Older stable ABI extensions are allowed to make immortal objects mortal. Instead of `_Py_IsImmortal`, use `_PyUnicode_STATE` (`interned` and `statically_allocated`), or skip asserts altogether. * Issue: gh-113993

encukou

gh-121160: Note that readline libraries using different history formats.

1

This is not something we can do too much about, without help from the underlying libraries. The file formats seem to be treated as implementation details, hidden behind [API calls](https://www.man7.org/linux/man-pages/man3/history.3.html)....

encukou

test_pyrepl: test_not_wiping_history_file() fails (randomly?) on multiple buildbots

3

Example with s390x Fedora 3.x: https://buildbot.python.org/all/#/builders/223/builds/6180 ``` FAIL: test_not_wiping_history_file (test.test_pyrepl.test_pyrepl.TestMain.test_not_wiping_history_file) ---------------------------------------------------------------------- Traceback (most recent call last): File "/home/dje/cpython-buildarea/3.x.edelsohn-fedora-z/build/Lib/test/test_pyrepl/test_pyrepl.py", line 916, in test_not_wiping_history_file self.assertNotEqual(pathlib.Path(hfile.name).stat().st_size, 0) ~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ AssertionError: 0 == 0 ```...

vstinner

Suffix in pathlib is not behaving like a file extension

7

# Bug report ### Bug description: According to the docs [here](https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.suffix) a suffix is defined as: > The file extension of the final component, if any But pathlib doesn't behave...

bbilly1

GH-120754: Add a strace helper and test set of syscalls for open().read()

7

General goal here is to be able to add tests that behavior isn't unintentionally changed around some common python code patterns, in particular Path("README.rst").read_text(). This started with the code that...

cmaloney