THANK YOU Tama, you rock ! I've downloaded latest SNAPSHOT from here ( https://repository.apache.org/content/repositories/snapshots/org/apache/maven/apache-maven/3.9.10-SNAPSHOT/ ) and on a very simple example this works as expected. My test is really simple so I do not pretend to validate the whole thing, but at least this use-case is now correct.
I think this fix is REALLY important as I expect that some builds may be corrupted (or the output may not match the code being compiled) in some case. Thanks ! Le jeu. 14 nov. 2024 à 15:20, Tamás Cservenák <ta...@cservenak.net> a écrit : > Eh, sorry, am a bit overwhelmed.... > > It seems this is FIXED in Maven-3.9.x branch? :) > > https://github.com/apache/maven/commit/45201347c417896b57159221130333f5fa4bbfb6 > > So once 3.9.10 is out, it should be workable? Can you test the > maven-3.9.x branch? > > Thanks > T > > On Thu, Nov 14, 2024 at 3:14 PM Tamás Cservenák <ta...@cservenak.net> > wrote: > > > > Maven4 fixes this issue, yes, it is "aware" differently than Maven3 is > > (where excluded projects are fully removed from line of sight). > > > > In Maven3 is not doable sadly, at least not with some simple change. I > > did some experiments about this long time ago, see > > https://github.com/apache/maven/pull/1406 > > > > T > > > > On Thu, Nov 14, 2024 at 3:12 PM Francois Marot <francois.ma...@gmail.com> > wrote: > > > > > > Thanks for the info Tamás. > > > Do you know if: > > > 1- a workaround is planned ? > > > 2- Maven 4 will solve this problem ? > > > 3- commands like "-pl :myModule -am " also suffer from this kind of > problem > > > ? > > > > > > I'm trying to convince everyone at work to move to a large Monorepo > and my > > > argument is "learn how to use Maven and your builds will not be > longer" but > > > I may have to stop immediately before it is too late ! > > > > > > Best regards > > > > > > > > > Le jeu. 14 nov. 2024 à 14:56, Tamás Cservenák <ta...@cservenak.net> a > > > écrit : > > > > > > > Howdy, > > > > > > > > This is a problem with Maven 3 and how it "reduces" the reactor (when > > > > using -rf, -pl etc, basically when reactor < full project). > > > > This was already discussed once on this ML but let me shortly explain > > > > what happens: > > > > > > > > You basically "moved" B out of the reactor, and your "snake" looks > like > > > > this: > > > > > > > > A (in reactor) -> B (out of reactor, like any external dep) -> C (in > > > > reactor). > > > > > > > > Maven 3 cannot follow these "out then back in" links, and assumes A > > > > and C are unrelated. > > > > > > > > The 1st command simply "loses" the link between A and C, because B > got > > > > out of the reactor and built them in parallel. > > > > The 2nd command works as it is single threaded, but in fact, > > > > internally, it also does not "see" the relation. > > > > > > > > > > > > > > > > On Thu, Nov 14, 2024 at 2:33 PM Francois Marot < > francois.ma...@gmail.com> > > > > wrote: > > > > > > > > > > Hello, > > > > > > > > > > I just faced a problem and would like your input if it is really a > > > > problem > > > > > or not. > > > > > I have a multimodule project where dependencies are like: > > > > > > > > > > A --> B --> C --> D (A depends upon B which depends upon C and > so > > > > on...) > > > > > > > > > > I would have expected those 2 commands to exhibit the exact same > > > > behaviour: > > > > > > > > > > 1- mvn package -pl :A,:C -T 1C > > > > > 2- mvn package -pl :A,:C > > > > > > > > > > but the problem is that when command 1 is run, both projects are > built > > > > > simultaneously (the logs are interweaved). It should not be the > case > > > > > because A depends transitively upon C. > > > > > When command 2 is run, project C is built before project A which > is what > > > > I > > > > > expect. > > > > > > > > > > I'm using mvn 3.9.9 > > > > > > > > > > Best regards > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > *- - - - -François Marot06 50 91 96 38* > > > > > > > > --------------------------------------------------------------------- > > > > To unsubscribe, e-mail: users-unsubscr...@maven.apache.org > > > > For additional commands, e-mail: users-h...@maven.apache.org > > > > > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscr...@maven.apache.org > For additional commands, e-mail: users-h...@maven.apache.org > >
