MrTJP

Allow z-fighting, you say? Preposterous! Microblocks have mechanics to resolve such a thing. I'm sure that can be used..

Most elegant solution I found to this is a mechanism to allow a manual flush of the queued multipart write streams.

I don't think anything would be incorrect with being able to manually send the data during the tick. For example, before the tile is actually moved, if the data was...

Project Red will be updated to the newer FMP changes next version. Until then, it will crash.

As of the latest version, the inverted lamp microblocks are added automatically. They also glow when placed in-world. The standard configs can be used to add the non-inverted lamps as...

Not in production, at the moment. I will leave a comment in my code and link you to it when I push the commit.

Here is a link to a piece of code that can produce the problem. https://github.com/MrTJP/ProjectRed/blob/57cded5ca6f448727f635256e81bc20a1429b4ba/src/mrtjp/projectred/fabrication/tileicprinter.scala#L677

Actually while we are on the subject, you never actually unified the UV and vector transformations. (See the very ugly UVT class that was created for ProjectRed model generation)

Maybe its block update related? Does it happen if you were to freeze your circuit (i.e. no more signal changes that would trigger block updates)?

This is an FMP issue. I will look into it and push a fix.