Hi, Petr thanks for testing and pointing to Quarkus. It is a very good place to check.
I also checked with Quarkus, tried to violate some of the rules and was reported as errors. Root cause for such improvement is that we replaced maven-dependency-tree with a direct resolver API. I expected some improvement but the test exceeded my expectations :-) in positive way pon., 23 sty 2023 o 14:32 Petr Široký <petr.sir...@proton.me.invalid> napisał(a): > Hello, > > great stuff! Looking forward to that release. > > I gave the latest 3.2.0-SNAPSHOT (26697b16) a quick try with Quarkus repo ( > https://github.com/quarkusio/quarkus), since the project is pretty big > (more than 1k sub-modules) and it also uses enforcer quite extensively > (multiple executions). The current enforcer version is 3.0.0-M3, since > there were multiple bugs in 3.1.0 which I think prevented the upgrade. > > Here are my findings: > - 'mvn validate' still succeeds, there are no enforcer errors. This is I > think good. I obviously did not check if every rule of every module is > being executed, but I did a quick comparison in few modules and everything > _seems_ to be in order. > > - what seems to be very positive is the improvement in execution time. I > compared a sequential and parallel runs, and also mvnd run (using JDK 17 + > Maven 3.8.7 / Maven Deamon 0.8.2): > > 1) Sequential run: > $ mvn validate -Dversion.enforcer.plugin=3.0.0-M3 > === Took ~140s > > $ mvn validate -Dversion.enforcer.plugin=3.2.0-SNAPSHOT > === Took ~35s > > > 2) Parallel run (with 8 threads): > $ mvn validate -T8 -Dversion.enforcer.plugin=3.0.0-M3 > === Took ~30s > > $ mvn validate -T8 -Dversion.enforcer.plugin=3.2.0-SNAPSHOT > === Took ~10s > > > 3) Maven deamon run (with 8 threads and multiple executions to let the > daemon "warm-up"): > $ mvnd validate -T8 -Dversion.enforcer.plugin=3.0.0-M3 > === Took ~80s > $ mvnd validate -T8 -Dversion.enforcer.plugin=3.2.0-SNAPSHOT > === Took ~8s > > Note: I don't quite understand why the Maven daemon run with version > 3.0.0-M3 is so much slower comparing to the non-daemon run (80s vs 35s). I > would have to investigate further, so let's ignore those numbers for now I > guess. > > Are we expecting this kind of improvement? Or could this be somehow caused > by e.g. mistakenly not executing some of the rules? I would need to again > dig deeper, but before that it would be good to know if this is maybe > something we are expecting. > > Thanks, > Petr > > > ------- Original Message ------- > On Friday, January 20th, 2023 at 19:09, Slawomir Jaranowski < > s.jaranow...@gmail.com> wrote: > > > > > > > > Hi, > > > > I'm going to release the next version of Maven Enforcer plugin and build > in > > rules. > > Probably in next week - about 24-26 of January > > > > Prepared release note: > > > https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12317520&version=12351838 > > > > The most important: > > - New Enforcer API - https://issues.apache.org/jira/browse/MENFORCER-455 > > - Get rid of maven-dependency-tree dependency - > > https://issues.apache.org/jira/browse/MENFORCER-435 > > - Verify working with Maven 4 - > > https://issues.apache.org/jira/browse/MENFORCER-447 > > > > I will appreciate any tests on your projects. > > If something else should be added please let me know. > > > > -- > > Sławomir Jaranowski > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org > For additional commands, e-mail: dev-h...@maven.apache.org > > -- Sławomir Jaranowski
