corda/gradle-plugins/api-scanner
Chris Rankin 2c84d07e8e CORDA-704: Implement @DoNotImplement annotation (#1903)
* Enhance the API Scanner plugin to monitor class annotations.
* Implement @DoNotImplement annotation, and apply it.
* Update API definition.
* Update API change detection to handle @DoNotImplement.
* Document the `@DoNotImplement` annotation.
2017-10-19 17:18:35 +01:00
..
src/main CORDA-704: Implement @DoNotImplement annotation (#1903) 2017-10-19 17:18:35 +01:00
build.gradle Gradle API Scanner plugin (#1669) 2017-09-29 16:55:26 +01:00
README.md Gradle API Scanner plugin (#1669) 2017-09-29 16:55:26 +01:00

API Scanner

Generates a text summary of Corda's public API that we can check for API-breaking changes.

$ gradlew generateApi

See here for Corda's public API strategy. We will need to apply this plugin to other modules in future Corda releases as those modules' APIs stabilise.

Basically, this plugin will document a module's public and protected classes/methods/fields, excluding those from our *.internal.* packgages, any synthetic methods, bridge methods, or methods identified as having Kotlin's internal scope. (Kotlin doesn't seem to have implemented internal scope for classes or fields yet as these are currently public inside the .class file.)

Usage

Include this line in the build.gradle file of every Corda module that exports public API:

apply plugin: 'net.corda.plugins.api-scanner'

This will create a Gradle task called scanApi which will analyse that module's Jar artifacts. More precisely, it will analyse all of the Jar artifacts that have not been assigned a Maven classifier, on the basis that these should be the module's main artifacts.

The scanApi task supports the following configuration options:

scanApi {
    // Make the classpath-scanning phase more verbose.
    verbose = {true|false}

    // Enable / disable the task within this module.
    enabled = {true|false}

    // Names of classes that should be excluded from the output.
    excludeClasses = [
        ...
    ]
}

All of the ScanApi tasks write their output files to their own $buildDir/api directory, where they are collated into a single output file by the GenerateApi task. The GenerateApi task is declared in the root project's build.gradle file:

task generateApi(type: net.corda.plugins.GenerateApi){
    baseName = "api-corda"
}

The final API file is written to $buildDir/api/$baseName-$project.version.txt

Sample Output

public interface net.corda.core.contracts.Attachment extends net.corda.core.contracts.NamedByHash
  public abstract void extractFile(String, java.io.OutputStream)
  @org.jetbrains.annotations.NotNull public abstract List getSigners()
  @org.jetbrains.annotations.NotNull public abstract java.io.InputStream open()
  @org.jetbrains.annotations.NotNull public abstract jar.JarInputStream openAsJAR()
##
public interface net.corda.core.contracts.AttachmentConstraint
  public abstract boolean isSatisfiedBy(net.corda.core.contracts.Attachment)
##
public final class net.corda.core.contracts.AttachmentResolutionException extends net.corda.core.flows.FlowException
  public <init>(net.corda.core.crypto.SecureHash)
  @org.jetbrains.annotations.NotNull public final net.corda.core.crypto.SecureHash getHash()
##

Notes

The GenerateApi task will collate the output of every ScanApi task found either in the same project, or in any of that project's subprojects. So it is theoretically possible also to collate the API output from subtrees of modules simply by defining a new GenerateApi task at the root of that subtree.

Plugin Installation

See here for full installation instructions.