A bit of a long shot but does compiling the class by hand (outside gradle) make any difference?
On Mon, Apr 23, 2018 at 8:05 PM, Schalk Cronjé <ysb...@gmail.com> wrote: > Hi all, > > I have build an hierarchy that effectively looks like: > > Abstract Base #1 (Java) > ^ > | > Abstract Base #2 (Groovy, CompileStatic) > ^ > | > Abstract Base #3 (Groovy, CompileStatic) > ^ > | > Abstract Base #4 (Groovy, CompileStatic) <- Implements methodA from > Abstract Base #1 > ^ ^ ^ > | | | > ExampleG (Groovy) ExampleJ (Java) ExampleK (Kotlin) > > Finally there are three examples - one each in Groovy, Java and otlin > example. They extend #4 with concrete classes. The Groovy and Kotlin > examples compile fine, but the Java complains that methodA is not > implemented. The real signature of methodA is actually > > SeekableByteChannel newByteChannel(Path path, Set<? extends OpenOption> > options, FileAttribute<?>[] attrs) throws IOException > > This could be down to what the Groovy compiler (2.4.11) has done with the > generics, but it has me me confused. I am even more surprised that only the > Java compiler complains and not the Kotlin compiler. In IntelliJ this is > not flagged up as an issue either. It only occurs when compiling the > project with Gradle. > > I know my example is slightly vague, but it removes all of the noise. The > actual code of the offending class is here -> > https://gitlab.com/ysb33rOrg/java-nio2-providers/blob/ > development/commons-compress-provider-core/src/test/java/ > org/ysb33r/nio/provider/commons_core/examples/java/ > GzFileSystemProvider.java > > Now I'm hoping if anyone has seen issues like this before or have an idea > where I can start diagnosing it. > > -- > Schalk W. Cronjé > Twitter / Ello / Toeter : @ysb33r > >
