For the next major version, PlantUML will be developped using Haxe language.

Moving to Haxe means that we will have a native implementation of PlantUML in all following languages:

• Javascript
• Java (JVM)
• Python 3
• PHP7
• C#
• C++

This way, PlantUML could run almost everywhere.

PlantUML 2 is really a rewrite of the existing Java code.

Haxe provides everything (and more!) that we need for PlantUML (oriented-object, regular expression, nice standard library), however re-implementing all that PlantUML is currently doing will takes time (that is, at least 2 years).

This means that we are going to slow down on adding new features in current PlantUML 1. We will only focus on bug fixes.

Current version of PlantUML 2 provides only a tiny part of what PlantUML 1 is capable of:

• only Mindmap diagrams
• only SVG output
• no preprocessor
• no standard library

Next step is now to add Sequence Diagram.

Globally, PlantUML 2 will be able to parse the same files than PlantUML 1.

Some features will not be re-implemented, but the most important will:

• MindMap diagram (done)
• Sequence diagram
• Usecase diagram
• Activity diagram
• Class diagram
• Object diagram
• Component diagram
• Deployment diagram
• State diagram
• Timing diagram
• JSON data
• YAML data
• Network diagram (nwdiag)
• Wireframe graphical interface or UI mockups (salt)
• Archimate diagram
• Specification and Description Language (SDL)
• Gantt diagram
• Work Breakdown Structure diagram (WBS)
• Entity Relationship diagram (IE/ER)

We are unsure about the following one:

• Mathematic with AsciiMath or JLaTeXMath notation
• Activity diagram legacy

Features that will be excluded:

• Ditaa

We will port to Haxe the Smetana project which means that PlantUML 2 will not need any Graphviz/dot executable to run.

About the output, we currently generate only SVG, but we plan to support:

• SVG (done)
• PNG
• LaTeX
• EPS
• TXT (Ascii Art)
• XMI
• SCXML

As said, we will be generally compatible with PlantUML 1 syntax. However, we take this new release as an opportunity to improve some inconsistent choices that have been made in the past.

• In class diagram, package do not define any namespace, which may be confusing. We have added namespace keywords to mitigate this. However, in Plantuml 2, package and namespace will be synonym, which may break some existing diagrams.

Let us know if you know any other inconsistencies that you would like to see resolved. :-)

You first have to install Haxe 4.

The two following libraries are used:

haxelib install utest
haxelib install haxe-strings

If you have some HTTPS certificate issues (because of proxy for example), you can use:

haxelib install utest -R https://lib.haxe.org/
haxelib install haxe-strings -R https://lib.haxe.org/

For building, you have to launch the haxe command with some .hxml build file.

As IDE, we recommand the use of VSCode.

To build the unit tests, you have to launch the following command:

haxe unit-test.hxml 

Then you can run the unit tests in Java:

java -jar unit-test/MainTest.jar

Or in python3:

python3 unit-test.py

You can also open unit-test.html file in your favorite browser to run unitary tests in Javascript.

A basic Command Line Interface (CLI) is included for Java and Python.

To build them, just run:

haxe CLI.hxml 

Then you can run the python CLI:

$ python3 plantuml-cli.py 
usage: <PlantUML> foo.puml foo.svg

or the Java one:

$ java -jar plantuml-cli/MainCLI.jar 
usage: <PlantUML> foo.puml foo.svg

This CLI is right now really minimalist.

To build the PlantUML library, you must run the following command:

haxe lib.hxml 

This will build the following files:

• libplantuml.js
• libplantuml.py
• libplantuml/Libplantuml.jar

You can use libplantuml.js this way:

<script type="text/javascript" src="libplantuml.js"></script>
<script type="text/javascript">
	function update() {
		var plantuml = new PlantumlV2();
		var tt = document.getElementById('tt').value;
		plantuml.addLines(tt);
		document.getElementById('result').innerHTML = plantuml.getSvg();
	}
</script>

You can use libplantuml.py this way:

#!/usr/bin/python
import libplantuml

if __name__ == "__main__":
    plantuml = libplantuml.PlantumlV2()
    plantuml.addLineSingle("@startmindmap")
    plantuml.addLineSingle("* World")
    plantuml.addLineSingle("** America")
    plantuml.addLineSingle("** Europe")
    plantuml.addLineSingle("@endmindmap")
    svg = plantuml.getSvg()
    print(svg)

The very same example running in Java:

import com.plantuml.api.v2.Plantuml;

public class Main {

	public static void main(String[] args) {
		
		final Plantuml plantuml = new Plantuml();
		plantuml.addLineSingle("@startmindmap");
		plantuml.addLineSingle("* World");
		plantuml.addLineSingle("** America");
		plantuml.addLineSingle("** Europe");
		plantuml.addLineSingle("@endmindmap");
		
		final String svg = plantuml.getSvg();
		System.out.println(svg);

	}

}