gfgtdf

I actually don't know whether the syntax has changes during the transition to lua, if so we'd need some info there(not mpa_gerneration=cave with invoke the lua cave generator)

> Anything left on this? Yeah the main thing left is the decision on whether automatic recognition by file name (`map_generation=cave` automaticially looks for a file names `/map_gernators/cave.lua`) or a...

well you could say the same about map_file, images etc... But the main reason is to be compatible with the previous syntax which is also map_generation=cave, map_generation=lua

Imo it makes sense to do such stuff rather sooner than later because otherwise people who might want to start using this in 1.18 will have to change their code...

i don't really know, but i also don't see how that's relevant at all.

> Is there a reliable way to search for where it's being used? Either way though, while I suppose that deprecating something doesn't technically break the feature freeze, adding a...

probably related to #10516 (specific to sdl 2.32)

i made a quick test with the simple coide taken from `mask_surface_scalar_remainder` and it seems like clang does generate simd instructions for us automatically but gcc and msvc do not.

Adding these as user-visible variable givea the impression that these are reccomended or balanced settings, and the artifact_invest_options option is imply not interesting enough to be a player-visible setting. So...

imo a custom struct with fields that have names would be better than tuples. (i know that this previously returned pairs which is just as bad as tuples, but since...