gh-110052: Fix faulthandler for freed tstate (#110069)
faulthandler now detected freed interp and freed tstate, and no
longer dereference them.
Backport to 3.11: add pycore_pymem.h include to traceback.c.
(cherry picked from commit 2e37a38bcbfbe1357436e030538290e7d00b668d)
gh-108987: Fix _thread.start_new_thread() race condition (#109135)
Fix _thread.start_new_thread() race condition. If a thread is created
during Python finalization, the newly spawned thread now exits
immediately instead of trying to access freed memory and lead to a
crash.
thread_run() calls PyEval_AcquireThread() which checks if the thread
must exit. The problem was that tstate was dereferenced earlier in
_PyThreadState_Bind() which leads to a crash most of the time.
Move _PyThreadState_CheckConsistency() from thread_run() to
_PyThreadState_Bind().
(cherry picked from commit 517cd82ea7d01b344804413ef05610934a43a241)
gh-104690: thread_run() checks for tstate dangling pointer (#109056)
thread_run() of _threadmodule.c now calls
_PyThreadState_CheckConsistency() to check if tstate is a dangling
pointer when Python is built in debug mode.
Rename ceval_gil.c is_tstate_valid() to
_PyThreadState_CheckConsistency() to reuse it in _threadmodule.c.
(cherry picked from commit f63d37877ad166041489a968233b57540f8456e8)
Functions like PyErr_SetFromErrno() and SetFromWindowsErr() should be
called immediately after using the C API which sets errno or the Windows
error code.
(cherry picked from commit 2b15536fa94d07e9e286826c23507402313ec7f4)
Such C API functions as PyErr_SetString(), PyErr_Format(),
PyErr_SetFromErrnoWithFilename() and many others no longer crash or
ignore errors if it failed to format the error message or decode the
filename. Instead, they keep a corresponding error.
(cherry picked from commit 633ea217a85f6b6ba5bdbc73094254d5811b3485)
* EOFError no longer overrides other errors such as MemoryError or OSError at
the start of the object.
* Raise more relevant error when the NULL object occurs as a code object
component.
* Minimize an overhead of calling PyErr_Occurred().
(cherry picked from commit 8bf6904b229583033035d91a3800da5604dcaad4)
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
[3.11] gh-106033: Get rid of new occurrences of PyDict_GetItem and PyObject_HasAttr (GH-106034)
These functions are broken by design because they discard any exceptions raised
inside, including MemoryError and KeyboardInterrupt. They should not be
used in new code..
(cherry picked from commit 1d33d5378058671bfabb6f4d4b5bfd4726973ff9)
* PyUnicode_CompareWithASCIIString() only works if the second argument
is ASCII string.
* Refleak in get_suggestions_for_name_error.
* Add some missing error checks.
(cherry picked from commit c8c162ef5294cddb7ac75fe93ab918e5661c68ee)
GH-105840: Fix assertion failures when specializing calls with too many __defaults__ (GH-105847)
(cherry picked from commit 2beab5bdef5fa2a00a59371e6137f769586b7404)
Co-authored-by: Brandt Bucher <brandtbucher@microsoft.com>
In sys_add_xoption(), 'value' may be uninitialized for some error paths.
(cherry picked from commit a8d69fe92c65d636fc454cfb1825c357eb2e6325)
Co-authored-by: Nikita Sobolev <mail@sobolevn.me>
(cherry picked from commit 41cddc2e93a285b81fa30ac542b088bd9d0112e9)
In _PySys_AddXOptionWithError() and sys_add_xoption(),
bail on first error to prevent exceptions from possibly being
overwritten.
Bail on first error to prevent exceptions from possibly being
overwritten.
(cherry picked from commit 567d6ae8e77579173510fc948ac06b2ababf3d40)
Co-authored-by: Erlend E. Aasland <erlend.aasland@protonmail.com>
gh-105184: document that marshal functions can fail and need to be checked with PyErr_Occurred (GH-105185)
(cherry picked from commit ee26ca13a129da8cf549409d0a1b2e892ff2b4ec)
Co-authored-by: Irit Katriel <1055913+iritkatriel@users.noreply.github.com>
gh-104018: remove unused format "z" handling in string formatfloat() (GH-104107)
This is a cleanup overlooked in PR GH-104033.
(cherry picked from commit 69621d1b09c996e43a1e13d2fa4c317d3dd4d738)
Co-authored-by: John Belmonte <john@neggie.net>
The error-handling code in new_interpreter() has been broken for a while. We hadn't noticed because those code mostly doesn't fail. (I noticed while working on gh-101660.) The problem is that we try to clear/delete the newly-created thread/interpreter using itself, which just failed. The solution is to switch back to the calling thread state first.
(cherry picked from commit d1b883b52a99427d234c20e4a92ddfa6a1da8880)
Co-authored-by: Eric Snow <ericsnowcurrently@gmail.com>
https: //github.com/python/cpython/issues/98608
Add thrashcan macros to the deallocator of the filter objects to protect against deeply nested destruction of chains of nested filters.
(cherry picked from commit 66aa78cbe604a7c5731f074b869f92174a8e3b64)
Co-authored-by: Marta Gómez Macías <mgmacias@google.com>
[3.11] GH-102126: fix deadlock at shutdown when clearing thread states (GH-102222)
(cherry picked from commit 5f11478ce7fda826d399530af4c5ca96c592f144)
