Hi, from FindBugs point of view, the biggest issue was the removal of "intermediate" org.apache.bcel.classfile.StackMapTable / org.apache.bcel.classfile.StackMapTableEntry classes. The removal of Serializable is not important for us at all, so I don't care.
We had a discussion here on the list regarding StackMapTable* and as far as I understood, there is no plan to restore those two types. Therefore I've adopted FB code now (not on master, but soon) and included replacements for the missing classes to the binary distribution, so that the downstream plugins provider will at least do not crash due the missing classes. However, the behavior is changed and without source code adaptation code relied on those two classes will not work. I'va also tried (smoke tests) the biggest known (to me) plugin "fb-contrib" from Dave and found only one BC issue left - https://github.com/mebigfatguy/fb-contrib/issues/116. This was also caused by some refactorings in not released code (ParameterAnnotationEntry). So from my limited FindBugs point of view we can release BCEL6 as it is today on trunk. BTW, if you will, you can compare current BCEL 6 jar from trunk with the latest "FB BCEL patch" jar here: https://github.com/findbugsproject/findbugs/blob/master/findbugs/lib/bcel-6.0-SNAPSHOT.jar. I think Clirr should be able to do a diff here. Don't ask me from which revision this patched jar was created, I have no clue. This is the BCEL API state we "exported" to downstream since 3 years :-( Kind regards, Andrey Loskutov http://google.com/+AndreyLoskutov > Gesendet: Sonntag, 12. Juni 2016 um 14:12 Uhr > Von: sebb <seb...@gmail.com> > An: "Commons Developers List" <dev@commons.apache.org> > Betreff: Re: [BCEL] Latest Clirr report > > On 12 June 2016 at 12:43, Benedikt Ritter <brit...@apache.org> wrote: > > sebb <seb...@gmail.com> schrieb am So., 12. Juni 2016 um 13:38 Uhr: > > > >> On 12 June 2016 at 12:00, Benedikt Ritter <brit...@apache.org> wrote: > >> > Hi, > >> > > >> > I've uploaded the verbose Clirr report to my space at home.apache.org > >> [1]. > >> > Do we have to fix all of this in order to release 6.0? > >> > >> No, most of the changes don't break strict BC. > >> > >> The main changes to the interfaces don't break BC and are highly > >> unlikely to affect source because there are abstract implementation > >> classes. > >> > >> See the changes.xml file and release notes for some details. > >> > > > > Thank you, I will have a look. Can we compile a list of things we need to > > fix for 6.0? > > Unfortunately the lack of updates has caused downstream users to rely > on non-released code. > I think we should try and ensure that we don't make things worse. > > Therefore I don't think we can compile a full list without assistance > from FindBugs and other consumers of the code. > > I suspect this will be an iterative process to try and find the best > compromise that suits these users as well as users who have only used > formal releases. > > > Benedikt > > > > > >> > >> > Benedikt > >> > > >> > [1] > >> > > >> http://home.apache.org/~britter/commons/bcel/6.0-SNAPSHOT/clirr-report.html > >> > > >> > sebb <seb...@gmail.com> schrieb am Mi., 8. Juni 2016 um 11:54 Uhr: > >> > > >> >> On 8 June 2016 at 10:46, Benedikt Ritter <brit...@apache.org> wrote: > >> >> > Hello sebb, > >> >> > > >> >> > sebb <seb...@gmail.com> schrieb am Mi., 8. Juni 2016 um 11:21 Uhr: > >> >> > > >> >> >> Was that the full report, or the 'quietened' one? > >> >> >> There's a Clirr suppression file which removes some diffs that are > >> >> >> considered non-breaking (as per the changes description) > >> >> >> (otherwise the report is rather difficult to read) > >> >> >> > >> >> >> To show all the errors, use -P!quieten-clirr or > >> -Dclirr.allDifferences > >> >> >> > >> >> > > >> >> > The report is the result of running mvn clean site. So if > >> quiten-clirr is > >> >> > activated by default it's the 'quitened' one (sorry I currently don't > >> >> have > >> >> > time to look into the pom). > >> >> > >> >> Yes, it is the default. > >> >> > >> >> > I can upload the "real" report to home.apache.org tonight when I'm at > >> >> home. > >> >> > >> >> OK, thanks. > >> >> > >> >> > Benedikt > >> >> > > >> >> > > >> >> >> > >> >> >> On 8 June 2016 at 09:35, Benedikt Ritter <brit...@apache.org> wrote: > >> >> >> > Hi all, > >> >> >> > > >> >> >> > I've published the BCEL website after reverting the package rename. > >> >> Here > >> >> >> > [1] is the Clirr report of the current code base. > >> >> >> > > >> >> >> > Benedikt > >> >> >> > > >> >> >> > [1] > >> http://commons.apache.org/proper/commons-bcel/clirr-report.html > >> >> >> > >> >> >> --------------------------------------------------------------------- > >> >> >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > >> >> >> For additional commands, e-mail: dev-h...@commons.apache.org > >> >> >> > >> >> >> > >> >> > >> >> --------------------------------------------------------------------- > >> >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > >> >> For additional commands, e-mail: dev-h...@commons.apache.org > >> >> > >> >> > >> > >> --------------------------------------------------------------------- > >> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > >> For additional commands, e-mail: dev-h...@commons.apache.org > >> > >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
