baseName is the name of the Eclipse Xtext runtime project. Other projects then add additional sub-names to this name. As far as I can tell, it's not used for anything else. We could probably actually simply omit it and make the Xtext tool generate a generic baseName...
Our code currently makes too much of baseName and uses languageName too little. This is wrong.
While this is getting fixed, there is a reasonable workaround for activities: given that the baseName really doesn't matter, as long as it's a Java qualified package name, one can simply add an artificial .foo to the end and make sure that the rest is the prefix of the languageName except the final segment.
The Xtext tool still makes wrong assumptions about how Xtext actually generates project structures. Specifically:
https://github.com/mdenet/platformtools/blob/6b1cb41c1d8a62ad51d379cd1288e126c9d4bea5/services/com.mde-network.ep.toolfunctions.xtext/src/main/java/com/mdenetnetwork/ep/toolfunctions/xtext/XtextTool.java#L73-L75
assumes that the
baseNameincludes the name of the language (a Java classname) and, consequently, removes the last item of this qualified name. This is wrong and leads to deriving the wrong package names. Looking at https://github.com/eclipse/xtext/blob/main/org.eclipse.xtext.xtext.wizard/src/org/eclipse/xtext/xtext/wizard/RuntimeProjectDescriptor.xtend,baseNameandlanguageNameare used differently:baseNameis the name of the Eclipse Xtext runtime project. Other projects then add additional sub-names to this name. As far as I can tell, it's not used for anything else. We could probably actually simply omit it and make the Xtext tool generate a generic baseName...languageNameis the fully qualified name of the.xtextfile (pretending for the moment that it was a class). In other words, its last element is a Java class name and everything before that is the so-calledbasePackage(defined in https://github.com/eclipse/xtext/blob/a189bea4cf97327ef221fa70d115ba8119e1b275/org.eclipse.xtext.xtext.wizard/src/org/eclipse/xtext/xtext/wizard/LanguageDescriptor.java#L48-L57). This is what's used to generate all the extra source files and to place the.xtextfile inside the project structure. It's the only thing that matters...Our code currently makes too much of
baseNameand useslanguageNametoo little. This is wrong.While this is getting fixed, there is a reasonable workaround for activities: given that the
baseNamereally doesn't matter, as long as it's a Java qualified package name, one can simply add an artificial.footo the end and make sure that the rest is the prefix of thelanguageNameexcept the final segment.