A place for some experiments indexing Java sources
The basic idea is to make a complete index of resolved sources and binaries. This means reference information can be reused in particular for immutable assets, for example a particular version of a jar file from maven or a particular JDK class library version.
The impetus behind these experiments is twofold:
-
Allow for sharing of resources in the case of language services at web scale (think online-IDE's). For example, you could cut down on startup of a new cloud workspace if you didn't have to index the JDK classes upon first startup.
-
Reuse the Eclipse parser infrastructure without the burden of the Eclipse project model. The days of the "my way or high way" IDE are over: project object models are determined by build systems these days, not IDE's (maven, gradle, etc.)