hg: jdk6/jdk6/corba: 6746424: Remove build dependency on findbugs and FINDBUGS_HOME

[Andrew John Hughes]

2009/5/7 Kelly O'Hair <Kelly.Ohair at sun.com>:
>
> Andrew John Hughes wrote:
>>
>> 2009/5/6  <kelly.ohair at sun.com>:
>>>
>>> Changeset: 452dfe027c41
>>> Author:    ohair
>>> Date:      2009-05-05 16:58 -0700
>>> URL:       http://hg.openjdk.java.net/jdk6/jdk6/corba/rev/452dfe027c41
>>>
>>> 6746424: Remove build dependency on findbugs and FINDBUGS_HOME
>>> Reviewed-by: robilad
>>>
>>> ! make/common/shared/Defs-utils.gmk
>>>
>>>
>>
>> Why the removal of the lines regarding Ant? This doesn't seem to be
>> mentioned in this bug.
>
> The corba builds never used ant, this was just carry over from these files
> being copied from the jdk repository.
> Ultimately these makefiles in corba will go away, or the whole thing
> will change. So these lines being deleted in corba is just clean up.
>

Seeing the CORBA build cleanup would be very nice :)
At least we got rid of the whole Scheme dependency nonsense.

However, if these makefiles going away means replacing them with an
Ant build, then that sounds even worse...

>>
>> Note that last time I looked, you could just run FindBugs directly
>> from its website via jaxws and that worked fine on GNU Classpath using
>> IcedTea's jaxws implementation.
>
> The removal of findbugs from the makefile and the build dependencies
> is just a build simplification issue.
> Initially I had hoped that running findbugs during the build would be
> just part of the build, but it just isn't practical. It takes too long
> to run on every build.
>
> Don't get me wrong, I love findbugs, and want to use it regularly, it
> just doesn't make sense to burden the build system with it.
> I need a better way to get it run regularly, maybe with a separate
> continuous job run on a regular basis, separate from the building.
>

I agree; the build is already long enough.  It's even longer if you
also run the JTReg and Mauve tests as I believe the distros do.  It's
really the wrong place for FindBugs IMO anyway; it works best as a
motivation to clean up a particular bit of code, spotting bugs you
wouldn't otherwise notice on long forgotten bits of the tree.  Having
it nag every time you build for a completely unrelated reason is
enough to probably drive people insane ;) Plus, there's probably
people who build OpenJDK with no intention of going into the code to
fix such bugs (or who even have a clue what they mean for that
matter).

On the contrary, having it run regularly as you suggest AND posting
the results publicly would be a great motivation for hackers to get
involved and fix some low-hanging fruit.

> -kto
>
>



-- 
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8