impl TryFrom between vector types (#104)

[joeftiger]

As conversions between vector types were not possible yet, I made some macros and made those possible.

As f32 to i32/u32 may be problematic due to either one of those:

• f32 is NaN, or
• f32 is infinity or negative infinity, or
• f32 is greater or less than the maximum/minimum of the target type I implemented TryFrom to give a descriptive error in such cases.

Implementing TryFrom is not possible for types out of the crate, so I used a module-private TryFromExt to circumvent it.

In case of converting from float vectors to int vectors (like Vec2 -> IVec2), any float precision will be lost, resulting in a lossy conversion that is practically "flooring" the float value to the next integer.

I wrote some simple tests for Vec2 conversions, these could be extended but they showcase the correctness of the above mentioned possible float issues.

I am open for feedback and possible changes to this pull request.

[fu5ha]

Released a new version with these and previous changes just fyi :)

[joeftiger]

Released a new version with these and previous changes just fyi :)

nice! I can delete the git path in my Cargo.toml now :D