| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
Sorry, something went wrong.
|
Thanks @brandtbucher for the PR, and @markshannon for merging it 🌮🎉.. I'm working now to backport this PR to: 3.11. |
Sorry, something went wrong.
|
GH-98002 is a backport of this pull request to the 3.11 branch. |
Sorry, something went wrong.
… one `PyFrameObject` (pythonGH-97996) (cherry picked from commit 21a2d9f) Co-authored-by: Brandt Bucher <brandtbucher@microsoft.com>
* main: pythonGH-97002: Prevent `_PyInterpreterFrame`s from backing more than one `PyFrameObject` (pythonGH-97996) pythongh-97973: Return all necessary information from the tokenizer (pythonGH-97984) fixes pythongh-96078: os.sched_yield release the GIL while calling sched_yield(2). (pythongh-97965) pythongh-65961: Do not rely solely on `__cached__` (pythonGH-97990) pythongh-97850: Remove the open issues section from the import reference (python#97935) Docs: pin sphinx-lint (pythonGH-97992) pythongh-94590: add signatures to operator itemgetter, attrgetter, methodcaller (python#94591) Add Pynche's move to the What's new in 3.11 (python#97974) pythongh-97781: Apply changes from importlib_metadata 5. (pythonGH-97785) pythongh-86482: Document assignment expression need for ()s (python#23291) pythongh-97943: PyFunction_GetAnnotations should return a borrowed reference. (python#97949)
… one `PyFrameObject` (pythonGH-97996)
| // Just pretend that we have an owned, cleared frame so frame_dealloc | ||
| // doesn't make the situation worse: | ||
| f->f_frame = (_PyInterpreterFrame *)f->_f_frame_data; | ||
| f->f_frame->owner = FRAME_CLEARED; |
There was a problem hiding this comment.
FRAME_CLEARED is a value of enum _framestate. Shouldn’t ->owner be a value of enum _frameowner instead?
Sorry, something went wrong.
| Back | FazBrowse Home | New Git URL |
There is a nasty situation where a GC collection during PyFrameObject allocation could create multiple PyFrameObjects for the same _PyInterpreterFrame, which breaks our ownership rules and creates all sorts of opportunities for accessing corrupted memory.
If this happens, we just keep the PyFrameObject that has already been exposed to user code, and throw away the duplicate.