IMHO this is not a bug of Wildfly/Weld! Reading the CDI-4.0 spec Wildfly behaves exactly as it is described.
Am Do., 22. Feb. 2024 um 11:24 Uhr schrieb Romain Manni-Bucau < rmannibu...@gmail.com>: > Maybe I get it wrong but if it is a bug in wildfly we should just get it > fixed then integrate the new version IMHO, not sure the point to change > everything on our side - which is valid from a spec standpoint - to hide > other bugs (we test the most used case IIRC). > > 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 jeu. 22 févr. 2024 à 10:48, Thomas Frühbeck <t.fruehb...@gmail.com> a > écrit : > > > My proposed change is restricted to arquillian tests only! > > I verified, that at least Wildfly-31 defnitely removes all "default" > beans > > on "<trim/>" from beans. > > By setting "<trim/>" we would have to rewrite _all_ our relevant test > > classes. > > The current tests relied on discovery-mode="all", what makes sense as > > arquillian itself is already heavily restricted. > > A use of trim means a redesign that IMHO does not really add to quality > > assurance on this level. > > Thomas > > > > Am Do., 22. Feb. 2024 um 08:42 Uhr schrieb Romain Manni-Bucau < > > rmannibu...@gmail.com>: > > > > > Hmm, it stays the default but not the beans.xml we recommend nor > provide > > so > > > not sure, trim should stay what is recommended in apps IMHO, annotated > > is a > > > broken mode by design so not sure it is good to use it in real CDI app. > > > > > > 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 jeu. 22 févr. 2024 à 08:39, Thomas Andraschko < > > > andraschko.tho...@gmail.com> a écrit : > > > > > > > +1 > > > > > > > > Thomas Frühbeck <t.fruehb...@gmail.com> schrieb am Do., 22. Feb. > 2024, > > > > 07:42: > > > > > > > > > The current CDI 4.0 spec has moved to default bean-discovery-mode > > > > > "annotated". > > > > > > > > > > All Arquillian tests - I have seen yet - are using Asset.Empty as > > > > > "beans.xml" which causes all tests to fail - at least on > > > Wildfly-31.0.0. > > > > > This seems to be in accordance to spec. > > > > > > > > > > I propose to replace all - yet unspecific usages of Asset.Empty > with > > > the > > > > > following default: > > > > > ArchiveUtils: > > > > > public static final Asset beansXmlAll = new StringAsset("<beans > > > > > bean-discovery-mode=\"all\"/>") > > > > > > > > > > Using this default beans.xml all Arquillian tests in > > "wildfly-managed" > > > > are > > > > > OK - at least in Deltaspike-Core. > > > > > > > > > > Please note, that usage of "<trim/>" is here unwanted, as it makes > > all > > > > > default beans unavailable. > > > > > All test I have seen yet seem to rely heavily on the previous > default > > > > > discovery-mode="all"! > > > > > > > > > > Please comment. > > > > > Thomas > > > > > > > > > > > > > > >
