On the server I can use 5-char Unicode codepoints, eg <U+1F310>, see this example.
When I install ver.1.2017.15 locally and run the normal command
java -jar c:\prog\plantuml\plantuml.jar -charset UTF-8 %*
The codepoints eg <U+1F310> are rendered fine, but the embedded unicode chars are not.
If I run this command (intended to enable PDF rendition using a bunch of extra jars, see https://forum.plantuml.net/1026), such codepoints are not converted to the corresponding char:
java -cp c:\prog\plantuml\* net.sourceforge.plantuml.Run -charset UTF-8 %*