Created attachment 355651
Fault on OOM in the cycle collector, rev. 1
We weren't null-checking the result of PL_DHASH_ADD, obviously. It's documented that in OOM conditions PL_DHashTableOperate can return a null entry. I don't think this bug is a blocker, but we should try to get it landed, along with bug 423473 so that this "fix" doesn't cause even more leaking.
Created attachment 355651
Fault on OOM in the cycle collector, rev. 1
We weren't null-checking the result of PL_DHASH_ADD, obviously. It's documented that in OOM conditions PL_DHashTableOp erate can return a null entry. I don't think this bug is a blocker, but we should try to get it landed, along with bug 423473 so that this "fix" doesn't cause even more leaking.