Just reading up on this and have the basic question: what is the motivation for passing through the exceptions?
>From what I can see it is simply that the exceptions become visible to the >developer, which can be done by properly logging them (in the adapterfactories >etc.). It was mentioned that this decision (whether to log or not) depends on >the application = user of the adaptTo call. But I haven't seen an example of >that and are unsure that really is the case. The example Stefan gave [1] is just about removing the boilerplate of the null check + throwing a runtime exception, which could be handled using a static utility method (adaptOrThrow, but outside the adaptable interface). Also, this seems to be heavily depending on the style of the application code, whether exceptions are used a lot or null handling (using optionals etc.) is prefered and done consistently. Hence I am not sure if that requires a major change in the adaptable interface contract. [1] https://issues.apache.org/jira/browse/SLING-3714?focusedCommentId=14048040&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14048040 Cheers, Alex
