0

We are looking into adopting a more CI friendly approach to managing our on premise jar dependency file server or doing away with it entirely for a better alternative. Right now it is just a Linux server with some form of file structure for example:

http://server_name/vault/apache.org/axis/version_number/axis.jar
http://server_name/vault/apache.org/axis/version_number/jaxrpc.jar
http://server_name/vault/apache.org/axis/version_number/axis.jar

Was wondering what type of standard adopted enterprise CI practices for managing these dependencies when you are doing automated builds.

  • Would storing them in AWS S3 be a good idea?
  • Are there any applications that help manage the jar dependency library? (Looking for open source solutions preferably). For example something similar to where you simply add a jar you need, and pull it down using an API when doing a build.
  • Assuming that apache ant is used for building out the project. And if that doesn't play well in the CI realm, what would be a good alternative?
alexfvolk
  • 1,810
  • 4
  • 20
  • 40

1 Answers1

1

I would recommend using Apache ivy to manage your build's dependencies. Ivy can be configured to pull your dependencies from Maven repositories which is pretty much the defacto standard now for storing Java binaries.

Maven Central is the largest repository of Java binaries. To host your private binaries there are various ready made options: Nexus, Artifactory or Archiva

Community
  • 1
  • 1
Mark O'Connor
  • 76,015
  • 10
  • 139
  • 185