gh-91912: Long object free list for medium value #91938
Conversation
penguin-wwy
requested review from
a team,
pablogsal and
markshannon
as code owners
There was a problem hiding this comment.
In the process of refactoring, what used to be PyObject_Free calls (matching the PyObject_Malloc calls used to allocate in the first place) became PyObject_Del calls (which pair with PyObject_New, which is not used here). I don't know whether the two happen to be compatible by coincidence, but even if they are, the calls should still pair with the other calls in the family; New with Del, Malloc with Free, while this patch now pairs Malloc with Del.
| else | ||
| #endif | ||
| { | ||
| Py_TYPE(op)->tp_free(op); |
There was a problem hiding this comment.
This feels like it should maybe be calling the type's tp_dealloc, not tp_free? Not going to swear to this, but it feels off to me (then again, calling their tp_dealloc could recurse)... Don't take this comment as gospel, I'm probably wrong here.
There was a problem hiding this comment.
The argument passed in should be long or a subclass of long, continued use of tp_dealloc has a high probability of leading to unlimited recursive calls.
|
A Python core developer has requested some changes be made to your pull request before we can consider merging it. If you could please address their requests along with any other requests in other reviews from core developers that would be appreciated. Once you have made the requested changes, please leave a comment on this pull request containing the phrase |
|
@iritkatriel Resolved conflicts with the main branch. |
|
Thanks for making the requested changes! @iritkatriel: please review the changes made to this pull request. |
#91912