completely agree. You might have seen my other post. The most pressing consequence of the Classloader stuff is that runtime enhancement is basically broken atm. At least if you have an EE or Servlet container. Maybe we can work around this by wrapping the Instrumentation somehow and do all the PCEnhancer stuff in a new temporary ClassLoader instead of 'polluting' the SystemClassLoader.
But I will focus on getting 2.2.0 out of the door for now ;) LieGrue, strub ----- Original Message ----- > From: Pinaki Poddar <[email protected]> > To: [email protected] > Cc: > Sent: Monday, January 9, 2012 6:46 PM > Subject: Re: [jira] [Commented] (OPENJPA-2057) Rethinking ClassLoading > architecture > > it broke *lots* of internal tests and we didn't have the *time to > investigate* all of them. > > 1. number of failed tests is not that critical -- because often few causes > result into many failures. > > 2. There will never be a *time to investigate* unless we think the issue to > be important. I believe that the work at the stage of commit took a month. > And I had no illusion that it will run everything at its first commit. Nor I > have the requisite machinery to investigate *lots of internal tests* and, > unfortunately, those who have did not have the *time to investigate*. > Unless the attitude to this kind of structural modification change, and we > proactively and jointly collaborate to address such issues, I see little > possibility to resolve such complex undertakings. > > > ----- > Pinaki Poddar > Chair, Apache OpenJPA Project > -- > View this message in context: > http://openjpa.208410.n2.nabble.com/jira-Created-OPENJPA-2057-Rethinking-ClassLoading-architecture-tp6881585p7168780.html > Sent from the OpenJPA Developers mailing list archive at Nabble.com. >
