Benjamin Saunders

This (and/or inline structs) comes up a lot in game development; hand-inlining and renumbering transform data (big piles of floats) rapidly becomes tedious, though it does get the job done.

Various fixed-size quantities of floats is an extremely common need in game development; I've taken to copypasting and renumbering blocks of them everywhere, which is tedious (especially to serialize/deserialize).

In my experience, this pattern is extremely convenient even when constrained by inventory. We should implement this.

Is the corruption persistent, or does it recover in a frame or two, or after the next time the glyph cache is touched?

I wonder if there's a race between multiple uploads that occur in rapid succession, maybe even the same frame if there's no automatic aggregation.

Have you considered using dynamically-sized matrices? What are some typical dimensions for your application?

I can't specifically help with the generics here, but bear in mind that copying 800 byte structures all over the stack is not necessarily going to be the optimal strategy....

Can you provide a self-contained example of code that seems to require this? I believe we did evaluate the use of extension structs in output parameters and found that it...

> For example [...] That's a code fragment, not a self-contained example. I don't know what `insert` is, and there's no easy way to test changes. Can you provide a...

Thanks for the report! At a glance, I think the worst-case scenario is the peer closing the connection with a transport error if we send an MTU probe that exceeds...