内容简介:PublishedWe all know testing is important. And, I hope, it might even be familiar. You’re coding away and catch yourself thinking “I should probably test this”. And before you know it you’ve mashed on your keyboard and there appearsThese are unit tests whi
Published onJoshleeb's blog
We all know testing is important. And, I hope, it might even be familiar. You’re coding away and catch yourself thinking “I should probably test this”. And before you know it you’ve mashed on your keyboard and there appears
#[cfg(test)] mod tests { ... }
These are unit tests which are super easy to setup. No added dependencies, no added tooling
But what about integration tests?
Rust supports those too, which you probably already know. They’re only slightly more involved to setup so they’re still fairly straightforward.
But there are some caveats that I personally ran into.
For one, The Rust Book explains that
each file in the tests directory is a separate crate.
Which on its own doesn’t sound too bad. But it has the implication that having code shared between your integration tests isn’t intuitive. I think it’s because the code structure goes back to the Rust 2015 edition of defining modules, using mod.rs
.
And two, which is arguably the bigger issue, is that it’s not obvious how to run only all integration tests. Or only all unit tests for that matter, but we’ll get to that too.
A Better Setup
I went looking through docs and crates and eventually opened up the source for diesel-rs/diesel , which isn’t the first time it’s been a great reference.
Now it looks like diesel has a separate crate, internal to the workspace, called diesel_tests
which appears to contain the integration tests. But I don’t want to setup a new crate and workspace just to extend my test suite. Instead let’s take a look at the Cargo.toml
for that crate and borrow what we can borrow.
For my new crate, mycrate
, I’ve setup the Cargo.toml
as
[package] name = "mycrate" ... autotests = false [[test]] name = "integration" path = "tests/lib.rs"
Specifically, we’ve added two sections .
Autotests
The autotests = false
disables automatic test discovery. This sounds bad but isn’t needed because we’re manually telling Cargo where to find our tests with…
Test Target
This may look familiar if you’ve seen [[bin]]
or [[example]]
where we explicitly specify an additional target for Cargo to work with. This is no different. We are simply telling Cargo there’s a test target called “integration” which has an entry point at tests/lib.rs
.
Single Integration Test Crate
This config solves the first caveat. We no longer have to deal with a crate per integration test file, and so the way we handle code shared between the tests is the same way we would handle sharing code in the implementation.
And everything is still in the tests
directory, so no changes there either.
Running Tests
We’ve also made segmenting our tests a bit clearer.
# Run all tests cargo test # Run only unit tests cargo test --lib # Run only integration tests cargo test --test integration # Run only integration tests, single threaded # (you’ll probably want this one) cargo test --test integration -- --test-threads=1
以上就是本文的全部内容,希望本文的内容对大家的学习或者工作能带来一定的帮助,也希望大家多多支持 码农网
猜你喜欢:本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们。
亿级流量网站架构核心技术
张开涛 / 电子工业出版社 / 2017-4 / 99
《亿级流量网站架构核心技术》一书总结并梳理了亿级流量网站高可用和高并发原则,通过实例详细介绍了如何落地这些原则。本书分为四部分:概述、高可用原则、高并发原则、案例实战。从负载均衡、限流、降级、隔离、超时与重试、回滚机制、压测与预案、缓存、池化、异步化、扩容、队列等多方面详细介绍了亿级流量网站的架构核心技术,让读者看后能快速运用到实践项目中。 不管是软件开发人员,还是运维人员,通过阅读《亿级流......一起来看看 《亿级流量网站架构核心技术》 这本书的介绍吧!