Closed LemonInTheDark closed 4 months ago
I had to bump the multiplier up significantly, I am like 90% sure this is starting to edge on floating point rounding errors. Why do we multiply layer by 1000? is it to handle tiny baby values? could the final layer be a float 64 or would that break something? (Would it be worth removing applied effect layers (TOPDOWN_LAYER, EFFECT_LAYER, BACKGROUND_LAYER) as well as doing this to reduce the final value?)
Please advise
NVM I'm doing it
Done
@LemonInTheDark I've fixed the pipeline. Please update the PR to the latest main to proceed with CI
Description
The existing multiplier is lower then layer values encouraged by byond when combined with layering defines, so it breaks easy. I'm removing effect layers which should make the value at least a little more sane (though at a max of 4999 there is still risk of intermixing, should we be using float64 or someodd?)
I was not able to get the build process to work, so this is unfortunately untested. Apologies. Should close #292
Type of change