A type validation library for writing a zero-cost, declarative, understandable, generic code in zig.
Note: minimum zig version 0.12.0 is required to use zig-validate.
For many projects, the desired method of creating generic types is to do either of the following:
- Use a concrete type that results in code bloat due to repetitiveness.
- Use a concrete type that uses composition and hand-rolled vtables to emulate interfaces from other languages and incurring a runtime cost.
- Use an
anytypeand have the caller of the function hope that sufficient documentation was written or comb through the codebase to see what methods are required. - Hand-roll type validation for a specific type every time a generic type is required.
- Use a switch to hand-roll static dispatch in zig due to lack of function overloading.
If only there was a way to have traits or interfaces in zig without having to waste time in repetitively writing error-prone imperative code for type validation...
Zig-Validate provides a non-obtrusive solution to all of those problems with a simple-to-use interface and a rich error reporting solution for relaying to the user issues with non-conforming generic types.
The following is a taste of the error-reporting capabilities of Zig-Validate:
Yes, static and dynamic dispatch implementations are available in zig-validate with two different flavors of static dispatch!
For your statically-dispatched needs there is the more vanilla validate.static, and the more powerful and ergonomic validate.staticFnOverride that supports function overloading!
If you require a more dynamic approach, there exists validate.dynamic to easily generate vtables!
For more information on type validation using
zig-validate, go to the static dispatch and/or dynamic dispatch page(s).
What if I have an existing application/library that would require too much refactoring to use the output type?
If you have a codebase that would require a large amount of refactoring to use the output type, then you can simply not use the output type. Due to the unobtrusive nature of zig-validate, you can simply do _ = validate.static(Target, Validator) and still use the type validation power of the library. The only downside of discarding the output type is a generic trait/interface that is used to interact with generic Target types.
That's pretty much all that there is needed to know in order to use zig-validate! Now, stop writing boilerplate and bear witness to the power of zig-validate.
Still unsure on how to use zig-validate? Then Read zig-validate's documentation!