How to set an environment variable programmatically in Jenkins/Hudson?

Question

I have two scripts in the pre-build step in a Jenkins job, the first one a perl script, the second a system groovy script using the groovy plugin. I need information from the first perl script in my second groovy script. I think the best way would be to set some environment variable, and was wondering how that can be realized.

Or any other better way.

Thanks for your time.

Solution 1

The way to propagate environment variables among build steps is via EnvInject Plugin.

Here are some previous answers that show how to do it:

• How to set environment variables in Jenkins?
• Jenkins : Report results of intermediate [windows batch] build steps in email body

In your case, however, it may be simpler just to write to a file in one build step and read that file in another. To make sure you do not accidentally read from a previous version of the file you can incorporate BUILD_ID in the file name.

Solution 2

Using EnvInject Plugin from job configuration you should use Inject environment variables to the build process / Evaluated Groovy script.

Depending on the setup you may execute Groovy or shell command and save it in map containing environment variables:

Example

By either getting command result with execute method:

return [DATE: 'date'.execute().text]

or with Groovy equivalent if one exists:

return [DATE: new Date()]

Comments

[Nux]

Workspace is always separate for each build so there is no need to add a build_id. If you will add build_id you will end up with a lot of files eventually. It's better to simply start by removing old file if it exists (or simply overwrite it).

[Luis Elizondo]

It really depends on your use case. If you want to preserve the content of the file for a later usage you need to add the build-id in the file name. If you will only use the latest build info, then you can always replace it.