Re: [Japi] RuntimeException on the latest japitools against BEA JRockit1.5

[Stefano Mazzocchi]

Stuart Ballard wrote:
> Well my proposal would detect when any public class is corrupt (as
> long as we could get enough data out of it to determine for sure that
> it's public).
> 
> And yes there ought to be a way to express that in the Japi results.
> 
> But I'm not sure we ought to go any further as far as trying to signal
> individual methods etc as corrupt. If a class file isn't valid I'd
> tend to consider that renders everything about it suspect.
> 
> But I don't know the class file format or what kinds of bugs we're
> likely to find well enough to know whether this is really true in
> practice.
> 
> Thoughts?

There are plenty of 'quality control' tools out there that do class
analysis and the like, so I think japi should just focus on API
footprint recognition.

So, any time an error in the class format can trigger an error in the
footprint recognition, japi should deal with it and mention it in the
reports. If not, not.

not sure this helps in practice though ;-)

-- 
Stefano.