Disclaimer: I just have my tablet here, I cannot actually do anything in Git or an IDE now.
Anyway, a little hint: I counted about 38 or so commits in between 3.1 and 3.2, i.e. with 'git bisect' (one of my favourite bug hunting tools) you should need max. 6 iterations in order to identify the exact commit which caused your current problem. Regards -- Alexander Kriegisch http://scrum-master.de > Am 21.01.2016 um 17:30 schrieb Tibor Digana <[email protected]>: > > This issue must be really annoying to users, I agree. > The way would be to find the exact commit in GitHub between Versions 3.1 > and 3.2 which caused this issue. > This means compiling the maven-compiler-plugin in a loop and testing the > project attached over and over again. > >> On Thu, Jan 21, 2016 at 3:44 PM, Thomas Broyer <[email protected]> wrote: >> >>> On Wed, Jan 20, 2016 at 8:43 PM Andreas Gudian <[email protected]> wrote: >>> >>> ** New Feature >>> * [MCOMPILER-203] - Allow compiler-plugin to specify annotation >>> processor dependencies >>> >> >> Great! But… >> >> I don't get how you can add a feature about annotation processors and not >> at the same time fix one of the most annoying issue when using annotation >> processors that generate Java sources: >> https://issues.apache.org/jira/browse/MCOMPILER-235 >> >> Is everyone always doing clean builds? or everyone sets >> useIncrementalCompilation to false? >> (or maybe everyone actually still uses 3.1; along with the >> build-helper-maven-plugin if they need the generated sources added as >> source roots in downstream plugins) >> >> /me frustrated >> > > > > -- > Cheers > Tibor
smime.p7s
Description: S/MIME cryptographic signature
