Andrii

> As well as reducing binary size, with a small runtime cost. 1. It doesn't reduce the final binary size AT ALL, the types are still generated and all the...

I am just gona argue the first one, I can't argue on hypotheticals untill I have a POC, so wait for that. > Moving away from dynamic dispatch would require...

> This should probably be moved over to a discussion, none of this is relevant to this pr as it currently stands Titus liked it the first time :smile:

I am not entirely against abstractions in this case, but as you said it yourself > Abstractions can indeed create more work. It's just that in this case, with static...

> > > > > > > > > Titus liked it the first time 😄 > > Please look at the original pull request which implemented floats with static...

@lj3954 Just to make it clear, we are arguing about stuff outside this PR. But do you approve this PR in its current form?

I just wanted to mention it, the binary size is so large not because we did something wrong, but because we are using GIANT crates like vt100, and if the...

> > Of course. I'm familiar with the binary size of Rust programs, especially when using library crates. A default release build of Hello World is 400KB. Static linking is...

@lj3954 When I say duplicated, I don't mean "Needs to be re-written", I mean the generated code will not have duplicates because of the different types involved, as there are...

@lj3954 And I told I wouldn't argue about hypotheticals, and I couldn't keep my word, but I really can't. I have this really cool idea on how to organize all...