Noticeable client lag when breaking/assembling large machines
erogenousbeef-zz opened this issue ยท 1 comments
Because the reactors' graphics are determined by metadata, assembling/breaking the reactor sends a giant wad of updates to the client whenever the reactor is assembled or broken.
With very large reactors (11x11 or bigger), this results in a noticeable hiccup/lag while the client processes all the updates.
To fix this, migrate the Reactor Part block and class to not use metadata to communicate state; instead, send a "reactor assembled" and "reactor disassembled" packet which triggers client-only visual updates based on the block's connection state.
This is a relatively large change and will require some significant rework, but should result in a faster, cleaner system.