Thanks for input. It might be best to wait a bit to review this, as I'm making good progress and where we end up will look a bit different. Also I'm evaluating use of the newest version findbugs-3.0.0, just released.
I'll post a follow on message when things settle into place. On Wed, Aug 6, 2014 at 10:19 AM, Chris Bowditch <bowditch_ch...@hotmail.com> wrote: > Hi All, > > Yes it is sensible to add FindBugs into the default target, but it should > be done after the committers vote to accept the new FindBugs configuration. > At the moment Findbugs is not yet accepted by everyone due to the > sub-optimal configuration. > > Glenn; thanks for doing the initial analysis of the exclusions. I'm happy > with the approach you propose to improve the Findbugs configuration. it > would be good if the other committers could review and give feedback also. > > Thanks, > > Chris > > On 04/08/2014 23:00, Glenn Adams wrote: > >> >> >> On Mon, Aug 4, 2014 at 3:47 PM, Luis Bernardo <lmpmberna...@gmail.com >> <mailto:lmpmberna...@gmail.com>> wrote: >> >> >> I suggest also to make running findbugs part of the default ant >> task so that we don't forget to do it! >> >> >> Good suggestion. >> >> >> >> On 8/4/14, 5:43 PM, Glenn Adams wrote: >> >>> I have performed a brief study of the current findbugs exclusions >>> file. At present there are 1094 exclusions covering 101 distinct >>> exclusions types. Of these types, half (~50) have 3 or fewer >>> exclusions. >>> >>> I plan to start cleaning up these exclusions by first fixing all >>> types having 10 or fewer exclusions. This will leave >>> approximately 25 types of exclusions to handle in a second fix pass. >>> >>> Eventually, I will identify in the exclusions file all exclusion >>> types that should not be permitted to be added, i.e., that must >>> be fixed. I expect that some exclusion types will remain, such as: >>> >>> * BC_UNCONFIRMED_CAST >>> * EI_EXPOSE_REP >>> * EI_EXPOSE_REP2 >>> * PZLA_PREFER_ZERO_LENGTH_ARRAYS >>> * UWF_FIELD_NOT_INITIALIZED_IN_CONSTRUCTOR >>> >>> >>> Those that remain (and added in the future) should be subject to >>> an explicit design choice, and not merely to silence a warning. >>> >>> Once I sort this out, I will ask for a vote on enabling findbugs >>> in nightly builds. >>> >> >> >> >