Create a Jython Script Task
Jython Script task contains a Jython script that is executed on the Release server. It is an automated task that completes when the script is successfully executed.
For detailed information about the contents of the task, see Task Drawer for Tasks.
In the Overview tab of the Jython Script task, you can:
- Add tags to your Jython Script task in the Task Tags field for filtering.
- Type or paste a Jython script into the Script field. Click Save to save the changes or click Revert to undo your changes.
You can toggle on the Ignore variable interpolation toggle button to ignore variables from being used in the script. See the Variable Interpolation section below for more information on variable interpolation.
Release supports Jython 2.7. Jython is the Java implementation of Python. This means that you have access to standard Python as well as the Java libraries included in Java 8.
The output of the Jython Script task is in markdown format. For more information, refer to Using Markdown in Release.
You can access and modify release variables in your scripts using the dictionary named releaseVariables
. This sample script shows how to access and modify a variable:
print(releaseVariables['xldeployPackage'])
releaseVariables['xldeployPackage'] = 'Release'
In a similar way, you can modify global variables using the dictionary named globalVariables
.
If you want to update global variables, the Run automated tasks as user variable must be set to a user that has the Edit global variables permission.
To set this variable:
- In the left navigation bar, click Releases.
- Click a release.
- In the Show dropdown, click properties.
- In the Run automated tasks as user field, set a user.
- Click Save.
In the release flow editor, Jython Script tasks have a gray border.
Security and Jython Script tasks
When a Jython Script task becomes active, the script is executed in a sandbox environment on the Release server. This means that the script has very restricted permissions. By default, access to the file system, network, and non API related classes is not allowed.
To remove these restrictions, add a script.policy
file to the XL_RELEASE_SERVER_HOME/conf
directory.
This is a standard Java Security Policy file that contains the permissions that a script should have.
To enable the use of additional Java packages or classes in the script, use the following Release specific RuntimePermission
:
permission com.xebialabs.xlrelease.script.security.RuntimePermission "accessClass.com.company.domain.*";
permission com.xebialabs.xlrelease.script.security.RuntimePermission "accessClass.com.company.utils.HelperClass";
To allow the use of a restricted built-in Jython module in the script, please remove the module from the list of restricted modules in the XL_RELEASE_SERVER_HOME/conf/xl-release.conf
file:
xl.security.scripting.sandbox.jython.restricted-modules = ["importlib", "imp", "tempfile", "linecache", "shutil", "os",
"getpass", "platform", "threading", "thread", "subprocess", "webbrowser", "cmd",
"pdb", "bdb", "gc", "user", "code", "codeop", "zipimport", "pkgutil", "modulefinder",
"runpy", "distutils", "compiler", "posix", "pwd", "grp", "posixfile", "commands"]
Note: Although we do not advise doing so, it is possible to disable the sandbox environment of the script task by updating the XL_RELEASE_SERVER_HOME/conf/xl-release.conf
file:
xl.security.scripting.sandbox.enabled = false
By default, all password properties for release
, phase
, and task
are encrypted in script task context. It is possible to get decrypted password properties by updating the XL_RELEASE_SERVER_HOME/conf/xl-release.conf
file:
xl.security.scripting.sandbox.decryptPasswords = true
This is a deprecated feature and it will be removed in future releases.
As a part of security improvement, release will now restrict users to pass encrypted values for any create or update operation.
From Release 9.7.7, you can fall back to previous behaviour by updating the script task in the XL_RELEASE_SERVER_HOME/conf/xl-release.conf
with the following config:
xl {
security {
accept-encrypted-secrets {
enabled = false
}
}
}
You must restart the Release server after creating or changing the XL_RELEASE_SERVER_HOME/conf/script.policy
or the XL_RELEASE_SERVER_HOME/conf/xl-release.conf
file.
Using Jython datetime in scripts
Sometimes it is desirable to write Java code while writing Jython scripts.
However if you try using Jython datetime
in the following way:
import datetime
currentDT = datetime.datetime.now() # here this will be converted to java.sql.Timestamp
print(str(currentDT))
the printed result will be incorrect.
You can overcome this problem by calling Jython datetime
directly like this:
import datetime
print(datetime.datetime.now())
print(datetime.datetime.now().day)
The Jython will sometimes convert Jython instances that are assigned to a variable to Java, and because you can't know what kind of Java instance it will be, it is better to use Java in this case when you are dealing with Jython datetime
.
from java.text import SimpleDateFormat
from java.util import Date
date = Date()
sdf = SimpleDateFormat("MM/dd/yyyy HH:mm:ss")
print(sdf.format(date))
Sample scripts
This sample script shows how to download a file from a web site and save it locally:
import httplib
url = 'www.xebialabs.com'
file = '/tmp/xebialabs.html'
xl = httplib.HTTPConnection(url)
xl.request('GET', '/')
response = xl.getresponse()
myFile = open(file, 'w')
myFile.write(response.read())
myFile.close()
print "Save %s to %s" % (url, file)
This example grants access to the network using Python httplib
and read/write access to the /tmp
directory on the Release server:
grant {
// Network access
permission java.lang.RuntimePermission "accessClassInPackage.sun.nio.ch";
permission java.net.SocketPermission "*", "connect, resolve";
// File access
permission java.io.FilePermission "/tmp/*", "read, write";
permission java.util.PropertyPermission "line.separator", "read";
};
Variable Interpolation
You can insert release, folder, or global variables directly into the script using the standard ${variable_name}
syntax. However, users should exercise caution when interpolating variables because of the potential for insecurities and code injections, and it is not a recommended practice. This capability will be removed in a future update.
Variable interpolation replaces the ${variable_name}
text in the script with the variable value "as is", without any escaping. This means that it will definitely fail if your variable value contains double quotes or new lines and if your Jython code looks like this:
x = "${myvariable}"
You can still work around this issue using triple quotes, but only in the case that your variable itself does not contain triple quotes, for example:
x = """${myvariable}"""
If you wish to use this feature, it is strongly recommended to avoid such characters as quotes, parentheses, and so on in the body of your variable. It is much preferred to use the releaseVariables['variable_name']
syntax as mentioned above.
Reserved variables
Script execution environment defines and uses set of variables, which you should consider as reserved variables and do not assign in your scripts.
DANGER: assigning to any of these variables may result in unpredictable behavior of your script.
- any variables starting with underscore, please avoid using variables starting with underscore
release
phase
task
globalVariables
folderVariables
releaseVariables
- API helper objects, functions and classes listed here