Pedro J. Estébanez

I don't think this is in the roadmap for now. In other words, `Object` is not meant to be thread-safe. User has to apply their own lock/sync measures. The current...

If no scripting language is using thread enter/exit callbacks (and there are no near future plans for it), I'd even remove them. However, I'm a bit unsure about the needs...

Should this be debug only or dev only?

I agree the current code is unsafe. Now, I have a couple of thoughts: - I don't get why keeping limiting caller threads to the main one, now the method...

> My guess is the author of [9df77d2](https://github.com/godotengine/godot/commit/9df77d276593ef7082e3971d1c180b8f74b0cb2e) didn't manage to make `FUNC3R()` call `ERR_FAIL_COND_V(main_thread != Thread::get_caller_id(), false);`. I'm not very fond of that hypothesis. First of all, the author...

New thoughts: - Should maybe these functions have the same constraint as `body_get_direct_state()`?, that is, make it only usable from physics process. That way, the fact that the command queue...

Could you please paste the output of the commands `lscpu` and `sudo lshw -C CPU`?

> [...] but I went with whatever requires the least amount of knowledge from the reader. The graph is only created once per RenderingDevice so there's not much in the...

> Yes, the usage of thread_local is very deliberate across this and my other PRs as it's an effective way to avoid using stack memory and reusing an already allocated...

I'm not sure my feedback (the directly actionable part) has been addressed yet. Also, this may be better rebased on top of #91416, which already includes the upgrade of the...