slipher

I guess there's nothing that distinguishes the constructors very well so maybe it would be easier to go ahead and define the subclasses even if all the functionality is in...

Disabling such functions would help to test all our code. I provided a platform-independent version of CountTrailingZeroes, but I don't think it's possible to actually use it on any supported...

> So this option is not to configure how compiler-provided functions behave. We should trust those compiler-provided functions. The two cases seem pretty similar to me. Yes, we trust the...

> engine/renderer/tr_types.h(142,15): warning C4324: > 'refBone_t': structure was padded due to alignment specifier I also saw this warning when I tried implementing `USE_RECOMMENDED_CPU_FEATURES` for MSVC so I could test with...

> I implemented a fix (see the branch `slipher/align-stuff`), but it is a pretty silly warning so I'd probably just disable it if it annoyed us again. Oops, apparently the...

New warning theory: - You have a struct A - A's alignment is greater than its natural aligment (considered by recursively ignoring alignment specifiers), due to a user-requested alignment (via...

> Can it be possible that `xmmintrin.h` just do `#pragma warning(disable : 4324)` ? It does not. It's not about whether SSE headers are included; it's whether we use `__m128`...

> Basically the idea is to not sleep the remaining time to fill the expected frametime, but to let enough time after sleeping to process events in a loop until...

> > Are you saying you are going to put it in a busy loop? > > In some way, yes. That's bad. For example if you are using a...

For engine only?