annotations

[Alex Buckley]

Jesse Glick wrote:
> A third possibility is to do a bit more parsing inside the annotation 
> values:
> 
> @Requires({"com.foo.bar >= 4.0", "com.foo.baz >= 5.0"})

My initial mail of examples already had that, i.e.

@Requires({ "com.foo.bar @ 4.0+" })
@RequiresOptional({ "com.foo.baz @ 5.0+", "com.quux.nab @ 6.0+" })
module com.foo.app @ 1.0;

import org.osgi.annotations.*;
@RequireBundles({ "com.foo.bar @ [4.0,5.0)" })
@ImportPackages({ "com.foo.baz @ [5.0,6.0)", "com.quux.nab @ 6.0" })
module com.foo.app @ 1.0;

> An annotation processor can reject syntactically malformed values with a 
> descriptive message, so this need not be any laxer. (Put the processor 
> in the same JAR as the annotation itself and register it in 
> META-INF/services/javax.annotation.processing.Processor so that it is 
> run automatically by javac.)

Now you hit Evan's chicken-and-egg problem. javac has to find that JAR, 
yet the modules on which the module depends are not yet read. Is the 
answer to put the JAR on the bootclasspath? Maybe, but that's likely to 
be subsumed by bootmodulepath. You're essentially asking for annotation 
processors to enjoy different visibility rules than annotated classes, 
and it's not clear how to do this.

Alex