In message <[EMAIL PROTECTED]>, Hen
ri Yandell writes:
>Basically. Except it's:
>Fred fred = (Fred)doSomething();  in some cases.

Shouldn't you actually do something with the result as a precaution
to ensure the assignment doesn't get optimized out (either by the JIT
or the compiler)?  (Or make sure tests are compiled and run
with no optimizations and no jit with a compiler and jvm with
well-understood behavior.)

In any case, all metrics reports have to be taken with a grain of salt
and interpreted by a human rather than blindly trusted as indications
of errors.  So any general out of the box configuration is not going
to be suitable for all subprojects.  One thing subprojects that rely on
Tom's PMD reports could do is adjust their Gump descriptors (or
just their build files) so that Gump will run PMD using their custom PMD
configuration and generate subproject-specific reports on a daily basis.
But I guess that would require Gump to publish the reports.  Short
of developing a comprehensive opt-in cross-subproject process plan, I
think Gump's a good place to focus this sort of stuff so ad hoc process
additions that require asynchronous recurrent generation of artifacts are
localized in a single place.


To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to