The plugin uses Reflections to scan and index your project classes at build-time, allowing run-time querying without the indexing performance hit.
The idea is to use the plugin with Gradle to embed a pre-scanned metadata index in your jar,
then utilise the embedded index at ruintime using Reflections.collect(). See the Gradle and Java sections bellow
for an example.
The plugin is published in Gradle plugin portal,
so it is rather easy to use. This will add the pre-scanned
metadata index in your jar as META-INF/reflections/PROJECTNAME-reflections.xml
, with
PROJECTNAME substituted by your actual project name.
// import the reflections plugin
plugins {
id "io.github.manosbatsis.gradle.plugin.reflections" version "1.1"
}
// you probably need this...
apply plugin: 'java'
// reflections plugin needs the compiled
// project classes, so either chain tasks
// with dependsOn as bellow or execute tasks explicitly
// when using the command line
reflections{
dependsOn classes
}
jar{
dependsOn reflections
}
// Use jcenter for resolving your dependencies.
repositories {
jcenter()
}
// Add Reflections and dom4j dependencies
dependencies {
compile 'org.reflections:reflections:0.9.10'
compile 'dom4j:dom4j:1.6.1'
testCompile 'junit:junit:4.12'
}
To utilise the pre-scanned index simply create a Reflections instance as:
// Collect and merge pre-scanned Reflection xml resources
// and merge them into a Reflections instance
Reflections reflections = Reflections.collect();
// use the instance, e.g.
reflections.getSubTypesOf(Foobar.class);
You don't need this to use the plugin. This section is meant for developers that want to build the plugin code themselves.
This section assumes you have a JDK and a recent Gradle installed and available in your path.
Clone:
git clone https://github.com/manosbatsis/gradle-reflections-plugin.git
Build:
gradle build
The build has some unit
and integration
tests. The build creates the report at build/reports/tests/test/index.html
: