DS

Many thanks for the tests! The heap corruption was due to a UaF. `addParticleSpawner` used to leak spawners when an id is reused, and with unique ptrs the leak became...

Comments addressed. I've run the same tests as Krock, with sanitizers. They seem to pass now. (At-player-pos version of the tween params test: `//lua local p = core.get_player_by_name("singleplayer"):get_pos() minetest.add_particlespawner({amount=50,time=3,texture="heart.png",pos_tween={p:add({x=0,y=1,z=0}),p:add({x=3,y=3,z=3}),style="pulse",reps=5}})`)

I was planning to rebase #13834 after this is merged. (I didn't want to promise anything before looking at how it works. I've done that now, rebasing should be doable...

~~Can't reproduce either, using bash and gnome-terminal.~~ Edit: Oops, haven't read reproduction steps correctly. You could try piping the stdout and stderr into a file and upload it here. And...

> @AFCMS only with `--server` or in general? >Run minetest server => With `--server`. Can reproduce.

It would of course be *possible*, but it's non-trivial. The client would have to track where the node tiles are copied to, and update block meshes and whatnot. Nodedef would...

No core-dev has explicitly expressed support for this. Marked as low priority. Don't expect large PRs for this to be easily merged. https://irc.minetest.net/minetest-dev/2023-04-30#i_6080402

I'd prefer a simple locking API, e.g.: ```lua local lock = minetest.lock_world_independent_mod_data(modname) -- read and write stuff lock:free() ``` Or preferable something more flexible: ```lua local lock = minetest.lock_lock_file(path .....

Added unittests and ConstViews. In case you're getting uaf warnings from gcc12 for SharedBuffer, I'm pretty sure that's false positives. A small web-search also reveals that there are also other...

> Did you consider using e.g. [span-lite](https://github.com/martinmoene/span-lite) instead of writing own `View` classes? I didn't. After all, the View class is very simple. Also, doing it like this worked out...