PyJarModularizer

This is the pythoner brother of JarModularizer project. It's a lightweight command line python tool for make non-modular JAR files compatible with Java Platform Module System. It's a translate of JarModularizer code with many changes since python is very different from java.

Usage

The way PyJarModularizer work is simple and almost identical to how JarModularizer does it. It need a configuration JSON file, called modularization descriptor, with the description of how each JAR file, called artifact, will should become in a java module, source directory path which is where non-modular JARs files will be found; and JDK home directory.

Prerequisites

JDK 9 or later because is requires for compile the module descriptors.
Python 3.7.

Example

From python:

python jarmod.py DESCRIPTOR SOURCE --jdk-home /path/to/jdk

Using an native executable (go to dist):

jarmod DESCRIPTOR SOURCE --jdk-home /path/to/jdk

Command above will deposit the new modular JAR copies in SOURCE/mods directory with fallowing name pattern: original.jar-mod.jar (ej. for log4j-1.2.17.jar new file name will be log4j-1.2.17.jar-mod.jar).

Note: If --jdk-home param is missing, path to JDK home directory will be taken from JAVA_HOME environment variable.

Important!

Only that files which name (including .jar extension) match with an entry in modularization descriptor will be processed.

Making executable files (optional)

If you want to build an executable native file (like found in dist) for any platform (Windows, Linux and Mac OS X), PyInstaller could be used (or any other tool compatible with Python 3.7).

Installing PyInstaller

Most easy way is from PyPI:

pip install pyinstaller

Build executable

For bundle all the project in a single file, we must to use following command: (see exe-build-commands.txt)

Windows

pyinstaller --onefile --name jarmod-<version>-win<arch> --icon icon\jigsaw-64.ico jarmod.py

Linux and Mac OS X (Mac OS X theoretically)

pyinstaller --onefile --name jarmod-<version>-<osname><arch> jarmod.py

Go to PyInstaller offical documentation for more details and options.

Getting help

If --help param is using, tool's help will be displayed in the terminal.

Modularization descriptor format

As was mentioned above, the modularization descriptor is a JSON file. Below show it format.

Note: JSON format doesn't allow comments, so java-style inline comment texts are only for descriptive purposes.

[
    {
        "name": "log4j-1.2.17.jar",// artifact name
        "module": {// module entry
            "name": "log4j",// future module name
            "exportsPackages": [// (optional) list of artifact's packages to be exported by the future module. Default is all artifact non-empty packages
                "org.apache.log4j",
                "org.apache.log4j.net"
            ],
            "requiresModules": [// (optional) list of requires directive to be included in module-info.java. Dafault is none diretive
                "java.base",
                "java.desktop",
                "java.management",
                "java.naming",
                "java.sql",
                "java.xml"
            ]
        }
    },
    ...
]

Author

Eduardo Betanzos @ebetanzosm

License

This project is licensed under the MIT License - see the LICENSE file for details.

betanzos / py-jar-modularizer