Taha Aziz Ben Ali

I believe it won't be possible to share objects between traits without compromising on the "one trait multiple implementations" setup that we've got. I think that if we truly desire...

> One way to share some state would be to go the other way: reduce the number of different readouts to only one and every implementation can define their own...

> I also don't know how (and if) this would work with the proposed feature flags. I almost managed to get this to work, until I hit a roadblock I'm...

I'll try and get some benchmarks, but I don't want to get your hopes up, the only benefit of this is less allocations (which is still good, to be honest)....

> Even though it would certainly be better to do the allocation only once, I agree that there is not much performance to be gained here. I think the second...

I've already reverted the changes I previously pushed, the complexity (of my half-solution) that it introduced is just not worth it, hopefully we can pick this up some other day.

> I'm not very familiar how features work in Rust It's my first time making use of features (that's a lie, it's my second, first feature was to only expose...

Thank you for offering to help :heart: I certainly would love to see more and more readouts get implemented for Windows, as its current state is pretty unfortunate, to say...

Your request contradicts what the RFC proposes. If we were to implement shared objects (or a connection for Windows), it MUST be done at the root level of the platform...

Before you start hacking away at libmacchina, I highly suggest that you take a look at #110 and to tell us your thoughts. If the RFC succeeds to convince our...