Sian January wrote: > Hi Stefano, > > That's a great tool and it looks like it's going to be really helpful for > finding bugs. I have been looking through some of the findbugs reports for > luni and have already found 2 or 3 fairly serious issues. As Tim mentioned > it would be good to be able to mark issues that have been investigated as > there are a fair number of false positives. Also is it possible to change > some of the rules? For example I think "Method invokes inefficient Number > constructor; use static valueOf instead" should be allowed within the class > itself (E.g. it should be ok for other methods in the Byte class to call > new > Byte()). What do you think?
Absolutely. Increasing the signal/noise ratio of all those report is a very high priority of mine: people tend to ignore noisy information and the signal gets lost. I'll send an email with instructions on what to do when I set it up. > > Thanks, > > Sian > > > On 23/11/06, Stefano Mazzocchi <[EMAIL PROTECTED]> wrote: >> >> Find new reports on melody: >> >> 1) findbugs reports 2895 potential bugs of 207 of which serious >> >> 2) cpd now reports against both classlib/java and drlvm/c (I will do >> classlib/c and drlvm/java later) >> >> 3) removed locales and messages from cpd on classlib/java >> >> 4) I've cleaned up the code for the cpd resport which now it's cleaner. >> >> 5) I've written a script that checks for "not implemented" methods that >> are not reported as such and that might shield or fool JAPI results >> >> comments, suggestions and criticism welcome. >> >> -- >> Stefano. >> >> > > -- Stefano.
