If those tests were already flaky, i.e. if the "flakiness" is not related to this migration, I think it would be better to migrate ASAP, because no-one can foresee what unexpected accident could happen in the meantime. At the same time, this would ease everyone' work. If, on the other side, these tests are flaky only on the migration (for some reason) I agree to fix them before migration; so to discover the root cause - because it could impact something else
Il giorno mar 12 dic 2023 alle ore 10:50 Francisco Javier Tirado Sarti < ftira...@redhat.com> ha scritto: > I agree the situation is good, but I need one more day to try to find out > why the flaky tests on runtimes are flaky. If we merge now with these flaky > tests, CI will be a pain (I do not like disabling these tests because they > are important). Therefore, I suggest merging this thursday/friday. > > On Tue, Dec 12, 2023 at 10:13 AM Toshiya Kobayashi < > toshiyakobaya...@gmail.com> wrote: > > > Test failures are being fixed and those repositories are in good shape (I > > think many of kogito-examples test failures will be fixed by Ricardo's > PR). > > Thank you, guys! > > > > +1 for migrating to main even if there are still some test failures > unless > > it's critical. We can fix them after the migration. > > > > Regards, > > Toshiya > > > > > > On Sat, Dec 9, 2023 at 3:55 AM ricardo zanini fernandes < > > ricardozan...@gmail.com> wrote: > > > > > I'll take a look at the `onboarding-springboot` and > > > `serverless-workflow-loanbroker-showcase`. > > > > > > I'll talk internal if someone can take the others. > > > > > > Thanks for the update! > > > > > > On Thu, Dec 7, 2023 at 12:28 AM Toshiya Kobayashi < > > > toshiyakobaya...@gmail.com> wrote: > > > > > > > Hi, > > > > > > > > > * kogito-runtimes has 1 fail in default and 0 in native > > > > > > > > kogito-runtimes test failures are only flaky tests now. I even filed > a > > PR > > > > to disable the flaky tests until they get stabilized. > > > > https://github.com/apache/incubator-kie-kogito-runtimes/pull/3321 > > > > > > > > > * kogito-examples has 10 fails in default and 5 in native > > > > > > > > Failing examples are `flexible-process-springboot`, > > > > `onboarding-springboot`, `process-decisions-rest-springboot`, > > > > `serverless-workflow-loanbroker-showcase`. > > > > > > > > I hope process / serverless-workflow guys can have a look. > > > > > > > > Thanks! > > > > Toshiya > > > > > > > > On Wed, Dec 6, 2023 at 9:18 PM Jan Šťastný <jstastn...@apache.org> > > > wrote: > > > > > > > > > Hello, > > > > > I would be against breaking main just for the sake of ease of > > > > development. > > > > > If those are test failures though, it should have minimal impact on > > > > things > > > > > like nexus deployment, the SNAPSHOTS are published also for builds > > with > > > > > failed tests. > > > > > So as long as the Jenkins builds in quarkus-3.2LTS nightlies are > not > > > red > > > > - > > > > > it should be fine. (Considering that at the moment we have ~30 > > failures > > > > in > > > > > incubator-kie-kogito-examples in the main branch.) > > > > > > > > > > Otherwise maintenance costs for the additional CI pipelines is > > minimal, > > > > > because it has only 2 environments defined (default and native > > build). > > > > > > > > > > Anyways, the intention for quarkus-3.2LTS was it would be > short-lived > > > > > branch facilitating the fixing of failed tests, the current summary > > is: > > > > > * drools is green in all nightlies > > > > > * kogito-runtimes has 1 fail in default and 0 in native > > > > > * kogito-apps has 16 fails in default and 2 in native > > > > > * kogito-examples has 10 fails in default and 5 in native > > > > > > > > > > Which is not an overly large number and hopefully gets lower during > > > this > > > > > week, so I'd suggest refresh on the topic of stability of the > branch > > at > > > > the > > > > > end of the week. > > > > > > > > > > Regards > > > > > Jan > > > > > > > > > > Dne st 6. 12. 2023 10:39 uživatel Tibor Zimányi < > tzima...@apache.org > > > > > > > > napsal: > > > > > > > > > > > Hi everyone, > > > > > > > > > > > > based on my understanding (please correct me if I am wrong), the > > > actual > > > > > > state of Quarkus 3 migration is that there are some test failures > > > that > > > > > > needs to be addressed. Currently the Quarkus 3 migrated code is > on > > > > > separate > > > > > > Quarkus 3 branches. This brings some maintenance work, e.g. the > > need > > > to > > > > > > update the branches with latest changes from main etc. As we > > > discussed > > > > > and > > > > > > decided before, the Quarkus 3 branches should be main branches at > > the > > > > > end. > > > > > > > > > > > > My proposal is to go ahead with the original plan we discussed > > > before, > > > > > even > > > > > > with test failures being present: > > > > > > - Branch current main branches for legacy purposes. Drools > branched > > > as > > > > > 8.x, > > > > > > Kogito as 1.x, Optaplanner as 8.x. > > > > > > - Merge Quarkus 3 branches to main, even with current test > > failures. > > > > > Those > > > > > > test failures could get fixed on main branch afterwards. > > > > > > - Change version on all relevant repositories to 10.0.0-SNAPSHOT. > > > > > > > > > > > > This change should free the maintenance cost of having two > separate > > > > > "main" > > > > > > branches and finish another step towards the first Apache > release, > > > > > instead > > > > > > of waiting till all tests get fixed. It will also remove the need > > to > > > > > > maintain two streams of pipelines in CI. > > > > > > > > > > > > > > > > > > What do you think please? Let's discuss and during next week we > can > > > > make > > > > > a > > > > > > final decision. > > > > > > > > > > > > Best regards, > > > > > > Tibor > > > > > > > > > > > > > > > > > > > > >
