bpo-42972: Fully implement GC protocol for re types #26368
Conversation
| @@ -1417,6 +1436,7 @@ _sre_compile_impl(PyObject *module, PyObject *pattern, int flags, | |||
| break; | |||
| } | |||
| } | |||
| PyObject_GC_Track(self); | |||
There was a problem hiding this comment.
@pablogsal I'm unsure if this it the correct place to start tracking the newly created pattern object. My reasoning was that AFAICS, it seems to be fully initialised here, unless I've missed something.
There was a problem hiding this comment.
IMO it's the right place.
It must be called before the first DECREF(self), otherwise PyObject_GC_UnTrack() crash in dealloc.
Usually, it's better to only track an object once it's fully initialized, but here it's more convient to do it here to be able to simply call DECREF() on error. The bare minimum is that traverse must be crash when an object started to be tracked. I added a test in debug mode for that: see traverse(visit_validate) in PyObject_GC_Track().
I'm not sure that calling PyObject_ClearWeakRefs() in a clear function is safe. What do you think?
| @@ -1417,6 +1436,7 @@ _sre_compile_impl(PyObject *module, PyObject *pattern, int flags, | |||
| break; | |||
| } | |||
| } | |||
| PyObject_GC_Track(self); | |||
There was a problem hiding this comment.
IMO it's the right place.
It must be called before the first DECREF(self), otherwise PyObject_GC_UnTrack() crash in dealloc.
Usually, it's better to only track an object once it's fully initialized, but here it's more convient to do it here to be able to simply call DECREF() on error. The bare minimum is that traverse must be crash when an object started to be tracked. I added a test in debug mode for that: see traverse(visit_validate) in PyObject_GC_Track().
… refs in tp_clear
|
Thanks @erlend-aasland for the PR, and @vstinner for merging it |
|
GH-26411 is a backport of this pull request to the 3.10 branch. |
(cherry picked from commit fba42d1) Co-authored-by: Erlend Egeberg Aasland <erlend.aasland@innova.no>
|
Thanks @erlend-aasland for the PR, and @vstinner for merging it |
(cherry picked from commit fba42d1) Co-authored-by: Erlend Egeberg Aasland <erlend.aasland@innova.no>
|
GH-26414 is a backport of this pull request to the 3.10 branch. |
https://bugs.python.org/issue42972
The text was updated successfully, but these errors were encountered: