Re:Zig's new plan for asynchronous programs
by debugnik on Wednesday December 03, 2025 @12:17PM
(#46137145)
Right, and there's also no need for await syntax at all, they can be then-ed, ContinueWith-ed or whatever the language calls them, but people keep bringing syntax into a semantics battle, so I had to mention it.
There's no need for the do-syntax at all. The (IO a) is no different to other generic types that can be fmap-ed, pointfree-composed with other expressions, and joined/folded when required. The only difference is the fact that they represent actions that affect the real world, so that ordering of things suddenly becomes important.