Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-08-12 Thread Romain Manni-Bucau
Good news guys :) Issue was actually a side effect of OWB-1241 (so more related to qualifiers than types). In other words: use >= 2.0.5 and you are good to go :) Sorry for the noise Romain Manni-Bucau @rmannibucau | Blog | Old

Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-08-12 Thread Romain Manni-Bucau
Investigated a bit more, we already have this kind of impl (using org.apache.webbeans.container.InjectionResolver#findByParameterizedType) but it doesn't resolve generics. I'll check if I can enhance current logic. Romain Manni-Bucau @rmannibucau | Blog

Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-08-12 Thread Romain Manni-Bucau
Cause you measured it? I have no app where it is useful once the scanning configured and it just breaks multiple apps OOTB and goes against CDI do we need a way to fix it. Can be a fastMatchUntilAmbiguous or so but we need sthg ;) Le dim. 12 août 2018 17:34, Thomas Andraschko a écrit : > Then

Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-08-12 Thread Thomas Andraschko
Then -1 to set it as default :D Romain Manni-Bucau schrieb am So., 12. Aug. 2018, 15:36: > Without. Any generic injection is broken in any project and de are not spec > compliant without that flag. > > Not sure it does worth having our default for most apps but i know for very > big ones it

Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-08-12 Thread Romain Manni-Bucau
Without. Any generic injection is broken in any project and de are not spec compliant without that flag. Not sure it does worth having our default for most apps but i know for very big ones it saves some boot time. Le dim. 12 août 2018 13:13, Mark Struberg a écrit : > hmm which test is that? >

Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-08-12 Thread Mark Struberg
hmm which test is that? And does deltaspike break with or without the fastMatch flag? LieGrue, strub > Am 10.08.2018 um 14:23 schrieb Romain Manni-Bucau : > > Up, > > just realized it affects deltaspike-data which can't inject > EntityRepository String> cause of that > > Romain Manni-Bucau >

Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-08-10 Thread Romain Manni-Bucau
Up, just realized it affects deltaspike-data which can't inject EntityRepository cause of that Romain Manni-Bucau @rmannibucau | Blog | Old Blog | Github |

Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-04-22 Thread Romain Manni-Bucau
Oki, let me know if we can do it and I would be happy to do the change. Alternatively - if we can't, maybe we should introduce that flag per jar or extension potentially to enable a "mixed" app (to think if previous option doesn't become possible). Romain Manni-Bucau @rmannibucau

Re: org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-04-22 Thread Mark Struberg
I'd need to do some tests. But think it's fine. LieGrue, strub > Am 21.04.2018 um 18:48 schrieb Romain Manni-Bucau : > > Hi guys, > > do we want to switch our fastMatching flag on the 2.x? (think we cant on > 1.x) > > Rational is we would be spec compliant. I know it

org.apache.webbeans.container.InjectionResolver.fastMatching = false?

2018-04-21 Thread Romain Manni-Bucau
Hi guys, do we want to switch our fastMatching flag on the 2.x? (think we cant on 1.x) Rational is we would be spec compliant. I know it has some issue with the old way to create producers without generics (or more exactly with a wildcard) but it is not spec compliant too. I encounter that