>What I like to do is to make the last arg take a struct such that you indirectly have named (and optional) arguments:
Rob Pike, and Dave Cheney both posted about this [1][2]. They summarised that using self-referential functions were a more optimal way for handling options to a function. This gives the benefit of allowing the options to be easily extensible by yourself, and any users that would be interacting with your API.
[1] - https://commandcenter.blogspot.com/2014/01/self-referential-...
[2] - https://dave.cheney.net/2014/10/17/functional-options-for-fr...
Re:One Program Written in Python, Go, and Rust
by nemo1618 on Tuesday July 02, 2019 @10:28AM
(#20335493)
Functional options were exciting when they were first discovered, and they're used in a few popular APIs (e.g. gRPC), but they haven't become as ubiquitous as you'd expect if they were truly superior. My guess is that a struct is the more direct and obvious approach for most people. Functional options also make it awkward to save and reuse a set of options later.
Re:One Program Written in Python, Go, and Rust
by jpittis on Tuesday July 02, 2019 @12:31PM
(#20336716)
And then you get the lovely benefit of not knowing whether someone explicitly passed in zero or didn't specify an option at all. There are a lot of ways around this: Use pointers, specify default arguments, etc. I does frustrate me though whenever the language creators bless a hack approach to get around deficiencies in the base language.
What I like to do is to make the last arg take a struct such that you indirectly have named (and optional) arguments:
> Never needing to pause for garbage collection could also be a factor [in Rust's greater speed compared to Go].Would be nice if the author had checked
to see if there was actually any garbage collection performed.