[egit-dev] Hudson CI bot reconfigured
As the egit ui tests are still not really reproducable and fail the build a bit too often
due to spurious problems I reconfigured the Hudson CI job which is validating new
patchsets pushed to Gerrit for code review in the following way:
- build fails if compilation or egit core tests fail, in this case the
Hudson CI bot votes "Verified -1"
- if egit ui tests fail the build does not fail but is marked unstable
in this case the Hudson CI bot votes "Verified 0" and "Code Review -1"
this allows to overrule this vote, only do so if you ran the ui tests successfully
in your local environment
This is surely not a perfect setup but prevents too many negative votes caused
by shaky ui tests. I think we should use this configuration until we reached more
stable ui tests then we should go back to the more strict configuration.
The continuous integration build running on master was not reconfigured so it will
still fail if ui tests don't pass.
Any thoughts on this ?