The big thing I would say I actually learned and would intentionally apply to other languages is SIMD programming. Otherwise, I'd say it gave me a much clearer mental model of memory management that helps me understand other languages much more fundamentally. Along with getting my hands directly on custom allocators for the first time, a question that took me time to figure out but gave me a lot of clarity in answering was "why can't you do closures in Zig?" Programming in Zig feels very Go-like, and not having closures was actually one of the biggest hiccups for me. I don't think this really changed how I write in other languages, but definitely how I think about other languages.
Snap! I also played around with closures a tonne in Zig. Definitely possible but not... ergonomic. Haven't ended up using them much.
And agree with allocators; in C I always considered using custom allocators but never really needed to. Having them just available in the zig std means I actually use them. The testing allocator is particularly useful IMO.
Never used Go but if it's Zig-like I might give it a shot! Thanks!
I'll make a list of the things that both languages have in common that make them feel similar to me:
- structs and functions are the main means of composition
- the pattern of: allocate resource, immediately defer deallocating the resource
- errors are values, handled very similarly (multiple return values vs error unions)
- built in json <-> struct support
- especially with the 0.16.0 Io changes in Zig, the concurrency story (std.Io.async[0] is equivalent to the go keyword[1], std.Io.Queue[2] is equivalent to channels[3], std.Io.select[4] is equivalent to the select keyword[5])
- batteries included but not sprawling stdlib
- git based dependencies
- built in testing
