So you also think we should make it a default behavior to NOT scan a JAR with extension and without beans.xml?
Mark Struberg <strub...@yahoo.de.invalid> schrieb am Do., 23. Sept. 2021, 21:26: > I personally did also wonder about this rather weird rule in the spec some > years ago. > Back then we concluded that it is rather an unintended behaviour. I'm > surprised that it's now used intentionally. If you have an Extension in a > jar and only want to use programmatic beans, then why the hack does one add > e.g. an @ApplicationScoped annotation? This really sounds weird. So I'd > rather simply use the existing jar-exclude in the scanning. > > > LieGrue, > strub > > > > Am 22.09.2021 um 11:19 schrieb Romain Manni-Bucau <rmannibu...@gmail.com > >: > > > > Le mer. 22 sept. 2021 à 11:17, Thomas Andraschko < > > andraschko.tho...@gmail.com> a écrit : > > > >> I really prefer to make the spec the default behavior, nobody expects a > JAR > >> to be scanned without having a beans.xml but a extension > >> normally if you would like to use CDI, but avoid scanning, you add a CDI > >> extension but no beans.xml, thats exactly like MF and Mojarra is > >> implemented > >> > > > > Just to correct that, several people expect a jar with an extension to be > > scanned in annotated mode without a beans.xml when it was full OWB from > the > > start. > > This is why switching will silently break users. > > So maybe a toggle with the same default than today + warning is saner. > > > > > >> > >> Lets wait for other opinions :) > >> > >> > >> Am Mi., 22. Sept. 2021 um 11:08 Uhr schrieb Romain Manni-Bucau < > >> rmannibu...@gmail.com>: > >> > >>> Think we should prefer our users over the spec for that so I'd be to > >>> support it with a toggle to enable it > >>> (org.apache.webbeans.skipJarWithExtensionScanning=false by default). > >>> If needed we can add a bdascannerservice with it at true > >> (WebSpecScanner). > >>> And in 2.1 we can change it > >>> > >>> 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 mer. 22 sept. 2021 à 11:06, Thomas Andraschko < > >>> andraschko.tho...@gmail.com> a écrit : > >>> > >>>> Hi, > >>>> > >>>> I discussed with Romain about OWB-1298. > >>>> > >>>> Mojarra has NO beans.xml but a CDI Extension, which currently breaks > >> OWB > >>>> as: > >>>> - WebsocketUserManager is scanned > >>>> - WebsocketUserManager is added in the extension > >>>> > >>>> in the specs (12.1) its defined, that: > >>>> > >>>> An archive which: > >>>> > >>>> - > >>>> > >>>> contains a beans.xml file with the bean-discovery-mode of none, or, > >>>> - > >>>> > >>>> contains an extension and no beans.xml file > >>>> > >>>> is NOT a bean archive. > >>>> > >>>> > >>>> Which means that OWB should NOT scan Mojarra and therefore users can > >> use > >>>> Mojarra with OWB, without adding it to the scan exclusions or setting > >>> other > >>>> properties. > >>>> > >>>> If MyFaces wouldnt be in the default exclusions, it would also break > MF > >>> and > >>>> maybe other libs. > >>>> > >>>> We should fix it and make it the default behavior. > >>>> > >>>> It should even improve startup as we skip more archives per default. > >>>> > >>>> Romain mentioned that there could be cases, where it could break apps. > >> So > >>>> we should maybe introduce a new config property to enable the old > >>> behavior > >>>> again, which also scans archives with extensions but without beans.xml > >>>> > >>>> this is also related to org.apache.webbeans.scanBeansXmlOnly > >>>> > >>>> WDYT? > >>>> > >>>> > >>>> Best regards, > >>>> > >>>> Thomas > >>>> > >>> > >> > >
