Hi, Maybe it was the concerns that inherited thread local will leak and likely keep the issue you can hit maven project methods without coming from an initialized thread and for a plain threadlocal the fact it is never resetted (so leaks too even if not crazy in terms of mem and this one is solvable compared to an inherited flavor which is uncontrolled), and has the same issue that depending your thread path you can miss the value using maven project (in a scripting based plugin or lazily init context for ex). A maven project artifacts should be thread safe generally speaking but also shared - it is why it can go through threads actually - so think thread safety shouldbe better handled. The mitigation to accept such a risky impl was to not enable it in concurrency=1 to keep the old behavior but was not done in 3.8.x AFAIK so maybe it would need more love before another iso release?
Hope it was what you had in mind. Romain Manni-Bucau @rmannibucau <https://twitter.com/rmannibucau> | Blog <https://rmannibucau.metawerx.net/> | Old Blog <http://rmannibucau.wordpress.com> | Github <https://github.com/rmannibucau> | LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book <https://www.packtpub.com/application-development/java-ee-8-high-performance> Le ven. 17 sept. 2021 à 00:41, Falko Modler <f.mod...@gmx.net> a écrit : > Hi Michael, > > thanks for moving this forward! > > Great to see that you are considering to include MNG-7235, as both PRs > were created by a fellow Quarkus team member after I advertised to > report this issue (that came up during a recent relocation in Quarkus). > > Regarding MNG-7212: > IIRC, I saw someone else reporting some issue in relation to MNG-7212, > but I cannot find that conversation anymore. I'm probably confusing it > with MNG-7220 (the regression I fixed myself), that was reported at > least twice. > Anyway, in case Oliver is not available, I can help with resolving the > merge conflict and adjusting the unit test to the inheritance semantics. > Just let me know. (I suppose I'd need to create a new PR then?) > > Cheers, > > Falko > > Am 16.09.2021 um 22:50 schrieb Michael Osipov: > > Folks, > > > > all important tickets for Maven 3.8.2 have been addressed in > > maven-3.8.x branch. > > > > There are only three issues left: > > * MNG-7235: In the works and requires review > > * MNG-7253: Should be a no-brainer to merge > > * MNG-7212: Very hard for me to tell for two reasons: (1) I don't want > > it to intermingle with Falko Modler's subsequent patch, (2) I'd like > > to see a real failing usecase where this is required. Ultimately, I am > > really inclined to back this out for 3.8.4 if necessary. > > > > Here you will find a precompiled version based on > > 8c66edc0417f37540b3a6ef4b266ff88bdbd8327: > > http://home.apache.org/~michaelo/maven/3.8.x/ > > > > Please test and let me know whether you still see errors. If I don't > > read any objections I will start the new release somewhere around > > 2021-W39. > > > > Michael > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org > > For additional commands, e-mail: dev-h...@maven.apache.org > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org > For additional commands, e-mail: dev-h...@maven.apache.org > >
