Re: psycopg2 2.0.8 - segmentation fault

>
> > 1. All / most of the classes in psycopg2 code seem to have
> > Py_TPFLAGS_HAVE_GC, so they need to a PyObject_GC_UnTrack as their first
> > step: http://docs.python.org/c-api/gcsupport.html . But since psycopg2
> > classes haven't implemented tp_clear either, the actual double dealloc
> > probably won't happen, unless people inherit these classes. (Heap classes
> > get the subtype_clear, tp_clear implementation.)
>
> If the type only uses Py_CLEAR() or equivalent to free its resources,
> then it isn't necessary to untrack the object (although it probably
> won't hurt).
>
> This macro essentially does the following:
>
>    tmp = self->whatever;
>    self->whatever = NULL;
>    if (tmp != NULL) {
>        /* free tmp */
>    }
>
> Given that the code is protected by the GIL, there isn't a race in
> unsetting the object.


That is true. But there are times when the code voluntarily releases the GIL
from inside the destructor, such as the conn_close in my above 2 fixes.
Then, the garbage collector can end up running concurrent to the current
dealloc. In those cases, we need to untrack the object first.

Btw, I'm trying to decide whether I should manually patch psycopg for my
organization or if I should wait for a release of psycopg with both my
patches in. Can someone please make an estimate on when there will be a
release with both these patches applied (double_dealloc_fix.diff and
cursor_dealloc_fix.diff) ?

Thanks,
Gangadharan

_______________________________________________
Psycopg mailing list
Psycopg-IAPFreCvJWPBWskQ1e/[email protected]
http://lists.initd.org/mailman/listinfo/psycopg