On 11/09/2010 07:07 PM, Rex Hoffman wrote:
I'm leaning to using the (javadoc) annotations
The trouble with Javadoc pseudo-annotations is that they are unavailable in bytecode and thus invisible to a tool like sigtest. I don't know about Clirr. I suspect this particular set of Javadoc tags was designed with Eclipse editor hints in mind, which probably have access to HTML Javadoc of APIs if not the original source project, and possibly predated JSR 175.
Furthermore, they will (ironically enough) not appear in the Javadoc HTML output by default, unlike a true @Documented annotation. The burden is on the library developer to remember to pass a complex argument such as -tag noimplement:t:"Do Not Implement" to the javadoc tool for each such annotation in use.
Also consider other tools which might want to analyze this information - such as FindBugs, or a hypothetical annotation processor that verified that a @NoImplement interface was not in fact implemented except by authorized providers.
--------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
