Sverker Eriksson

I recommend - Run released version. OTP-25.0.1 just released today. Maybe these are bugs already fixed. - Try run debug compiled emulator to hopefully get earlier crashes with nicer core...

I've looked at the segfault above with pending[i] == NULL. I don't see how that can happen without some wild memory corruption bug. Another way to detect memory corruption bugs...

I'm closing this as it looks similar to #6247 fixed by #6258 in OTP 25.1.

This is not prioritized for us right now. PR is welcome. Help Wanted label set.

Fix #6080 merged to maint for OTP 25.1.

@max-au What do you think about #6345?

I see two problems with the API: 1. It's not functional. It's possible to create cyclic terms (by mistake). 2. It kind of forces the implementation to allocate the list...

Proposal: ``` ERL_NIF_TERM enif_generate_list(ErlNifEnv*, ErlNifGenerateListFn* callback, size_t length_hint); typedef ERL_NIF_TERM ErlNifGenerateListFn(ErlNifEnv*, ERL_NIF_TERM* end_tail, void* arg); ``` To terminate the list, the callback would do `*end_tail = enif_make_list(env,0);` or set something...

Thanks for the contribution. I pushed a memory leak fix in `copy_struct_in_order_x` to your `copy` branch. Question: Did you contemplate looping over `copy_struct` for each element instead of implementing `copy_struct_in_order_x`?

> Thank you for catching the memory leak, I'm sorry for missing it. I just ran `ets_SUITE` with `debug` compiled emulator. ``` > cd $ERL_TOP > (cd erts/emulator && make...