simon-downes / spf-contracts
Defines the interfaces and traits for the services used throughout SPF
Requires
- php: ~7.2
This package is auto-updated.
Last update: 2025-02-07 01:32:49 UTC
README
Introduction
This package defines the interfaces and traits for the core services used
throughout SPF. e.g. spf\contracts\database\DatabaseConnection
defines the contract for accessing a database, while the
spf\contracts\support\collections\Dictionary
defines methods for operating on
a collection of key/value pairs.
Each directory corresponds to a SPF package that provides implementations of
the interfaces defined in that directory. For example, the simon-downes/spf
package contains implementions of spf\contracts\log\Logger
.
Naming Conventions
Interfaces in this package are named in one of two ways:
- As a singular noun (e.g.
Dictionary
,DatabaseConnection
,Filter
) - As a noun or verb suffixed with "-able" (e.g.
Arrayable
,Runnable
)
The implementations of those interfaces within SPF packages are prefixed with Base
.
Dictionary
➡BaseDictionary
DatabaseConnection
➡BaseConnection
ConnectionManager
➡BaseConnectionManager
Profiler
➡BaseProfiler
Background
This package is based on gamernetwork/yolk-contracts. For more information on the history and relationship between Yolk and SPF please see the README of simon-downes/spf-core.
Why Contracts?
A separate contracts package provides two primary benefits:
- A quick reference point for all available contracts.
- A single, decoupled package that promotes loose-coupling by separating interfaces from implementations.
Quick Reference
Having all of the framework's services neatly defined within interfaces provides succinct documentation to available features.
Rather than tracking down which methods are available within a large, complicated class, you can refer to a simpler, cleaner interface.
Loose Coupling
SPF packages provide default and usually basic implementations of each contract; sometimes these might not be sufficient or quite right for a particular project.
By specifying contracts, code is no longer coupled to these specific implementations and you may easily write alternatives that better suit your project.
Likewise, you can switch between implementations without modifying any consuming code. The classic example being able to switch caching backends (e.g. Memory/Redis/Memcache) without modifying any code that uses the cache.