configuration
Branch: master
Clone or download
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
configure
configure_derive Increment version of derive. Jan 24, 2018
.gitignore Move configure into own directory. Nov 25, 2017
.gitlab-ci.yml
README.md

README.md

configure - pull in configuration from the environment

"The 12 Factor App" has this very good advice about managing configuration:

Store config in the environment

An app’s config is everything that is likely to vary between deploys (staging, production, developer environments, etc). This includes:

  • Resource handles to the database, Memcached, and other backing services
  • Credentials to external services such as Amazon S3 or Twitter
  • Per-deploy values such as the canonical hostname for the deploy

For reasons laid out in the linked web page, it is poor practice to store this information inline in the source code. Setting aside even their reasons, it is especially inconvenient in compiled languages like Rust, as it requires recompiling the entire project between deployment environments, and it means performing an entire redeploy to change one of these values.

However, most libraries today take these kinds of configuration as an argument to their constructor, leaving application authors responsible for developing their own system for pulling those configurations from the environment. Configure is an attempt to create a standardized way for libraries to pull configuration from the environment, making it easier for end users to follow best practices regarding configuration.

This doesn't apply to all kinds of "configuration." For example, if it is performance critical that a configuration be applied at compile time (say, to drive monomorphization in order to get inlining benefits), using this library would not be appropriate. If the configuration could vary every time an operation is performed (say because its configuring a kind of pretty printer or formatter), using this library would probably not be appropriate either.

The Configure trait

Libraries adopting this model should put the appropriate configuration all into a struct:

struct Config {
    socket_addr: SocketAddr,
    tls_cert: Option<PathBuf>,
    // ... etc
}

This struct that needs to implement Configure. The easiest way to get Configure implemented correctly is to derive it. Deriving Configure requires the struct to also implement Deserialize, which can also be derived:

#[macro_use] extern crate configure;

extern crate serde;
#[macro_use] extern crate serde_derive;

#[derive(Deserialize, Configure)]
struct Config {
    socket_addr: SocketAddr,
    tls_cert: Option<PathBuf>,
    // ... etc
}

The Configure trait provides two functions:

  • Configure::generate, a constructor which generates the configuration from the environment.
  • Configure::regenerate, a method that updates the configuration by pulling from the environment again.

The generated implementation of Configure all pull the configuration from a configuration source, which is controlled by the end application.

Configuration sources

Ultimately, the end application controls the source from which configuration is generation. This control is applied through the CONFIGURATION static. The Configure impl will access the source using CONFIGURATION.get.

A default source is provided for you, so applications for which the default is satisfactory don't need to do anything. Applications which want to store configuration can override that source.

The default source

By default, configure provides a source of configuration that users can rely on. Users can use this source using the use_default_config! macro at the beginning of their main function. The default source is targeted at network services, and may not be appropriate to all other domains.

It works like this:

  1. By default, it looks up configuration values using environment variables. For example, if the library foo had a config struct with the field bar, that field would be controlled by the FOO_BAR environment variable.
  2. If no environment variable is set, it will fall back to looking in the Cargo.toml. If there is a Cargo.toml present, and it contains a [package.metadata.foo] table (where foo is the name of the library), the bar member of that table will control the bar field in foo's config struct.

In general, it is recommended that most environments use env vars to control configuration. The Cargo.toml fallback is intended for development environments, so that you can check these values into the Cargo.toml and have them be consistent across every developer's machine.

Custom configuration source

Users can override the default configuration for their app using the use_config_from! macro. This macro should only be invoked once, in the final binary. First, you will need to prepare a type that implements ConfigSource to be used as the source of configuration.

This allows users to control their configuration source without recompiling all of the library code that depends on it, as would occur if the configuration source were a type parameter.