[FEATURE REQUEST] Primitives for Module support

[pante]

[FEATURE REQUEST] Primitives for Module support

[CC007]

My main concern is how we’re going to detect modules for @inline and @classpath. Or is it even necessary to do so in those cases?

If those pose a problem, maybe modules could be added more explicitly

I thought about it and it shouldn't matter. Thinking about it, not specifying a module when compiling @Inline and @Classpath will cause it to be treated as pre-JDK9 code which is fine.

The only concern is if they're compiling a module-info with that. However I believe that to be rare enough of an edge case that I won't bother with it.

Well... I was debating using that, since I will eventually need to test my @NotNullByDefault and @NullableByDefault annotations for modules. It would be hard to test both of those using @Introspect and I don't even know how I would even target those using that annotation, so I think that I'm going to need @Inline for those tests.

[CC007]

What happens if the current module isn't named (aka a project without module-info.java), but it does rely on classes inside a named module? How does that behave and do any changes need to be made to support that?

[Pante]

What happens if the current module isn't named (aka a project without module-info.java), but it does rely on classes inside a named module? How does that behave and do any changes need to be made to support that?

From my testing, the current set-up already handles non-module code calling module code. The classes will be available on the class-path. The only case that’s not currently supported is module code to module code.

[Pante]

My main concern is how we’re going to detect modules for @inline and @classpath. Or is it even necessary to do so in those cases?

[CC007]

My main concern is how we’re going to detect modules for @inline and @classpath. Or is it even necessary to do so in those cases?

If those pose a problem, maybe modules could be added more explicitly

[Pante]

My main concern is how we’re going to detect modules for @inline and @classpath. Or is it even necessary to do so in those cases?

If those pose a problem, maybe modules could be added more explicitly

I thought about it and it shouldn't matter. Thinking about it, not specifying a module when compiling @Inline and @Classpath will cause it to be treated as pre-JDK9 code which is fine.

The only concern is if they're compiling a module-info with that. However I believe that to be rare enough of an edge case that I won't bother with it.

[Pante]

I think the issue in general is that there's no good way to retrieve the class information from an inline source since it's not yet compiled. I'm honestly not sure how to handle such a use-case.