MenTaLguY wrote:
On Sat, 2008-08-16 at 15:58 +0200, Ola Bini wrote:
So, I'm thinking about the new change that makes Java Exceptions be
outside the regular hierarchy of exceptions. This is quite nice in many
circumstances, but it's also kinda bad for many of the circumstances
where you want to catch everything and handle it correctly. As a typical
example: when spec:ji fails on a Java exception, the whole test run
immediately dies, because there is no Java code that can actually catch
the exceptions from Java without extra code.
Is this really what we want?
I don't think it is. This kind of thing is going to create bad problems
for parts of Omnibus as well.
I would clarify one thing though: the propagating Java exceptions will
still cause ensured sections to execute, since we use finally. So if
it's ensured sections you're worried about, they should be ok.
But either way I agree we should be able to treat propagating Java
exceptions as Exception subtypes, either actually or virtually, so it
should be a moot point.
And someone should file this as a bug so we don't forget.
- Charlie
---------------------------------------------------------------------
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
