Zig and Rust
matklad.github.io
Zig and Rust
1–10 of 247 posts
Re: Zig and Rust
#2Probably just a small miss, but in the first section the author refers to stack space limitations possibly hit when calling malloc. That should probably refer to heap space. Or if it is stack space they meant then maybe not malloc but rather stack allocations / function calls.
Re: Zig and Rust
#3Probably just a small miss, but in the first section the author refers to stack space limitations possibly hit when calling malloc. That should probably refer to heap space. Or if it is stack space they meant then maybe not malloc but rather stack allocations / function calls.
Besides that little inaccuracy I very much agree with most things in the article and think it brings up good points. Specifically the part about allocator handling is a huge thing for me. Having the allocator be a global thing has always been a bit of a cheat imo and it solves so many memory issues just to remove that part. And opens up some optimization possibilities ad well.
I do agree with some of zigs Cons that you bring up as well. Currently I have really been into Linear types and feel like that would be the perfect "simple" type-based memory safety zig could incorporate without making it too c++ complicated. But at the same time I don't know if it would work with its comptime and maybe it too would compromise on the simpleness of the language. I don't know.
Re: Zig and Rust
#4Probably just a small miss, but in the first section the author refers to stack space limitations possibly hit when calling malloc. That should probably refer to heap space. Or if it is stack space they meant then maybe not malloc but rather stack allocations / function calls.
I think he's referring to the fact that as a function, calls to malloc result in a new stack frame.
Re: Zig and Rust
#5The rust-analyzer guy now writes Zig full time?
Re: Zig and Rust
#6This sort of convinced me to stay away from Zig
Re: Zig and Rust
#7This sort of convinced me to stay away from Zig
can you elaborate why
Re: Zig and Rust
#8> Collections are not parametrized by an allocator, like in C++ or (future) Rust.
What does he mean by this?
Re: Zig and Rust
#9The rust-analyzer guy now writes Zig full time?
> I now find myself writing Zig full-time
Seemed pretty clear to me. :)
Re: Zig and Rust
#10> Collections are not parametrized by an allocator, like in C++ or (future) Rust. What does he mean by this?
My understanding is, instead of the allocator being a generic parameter of the type, it's a value you pass to the constructor.