On 23 March 2012 07:31, Benedikt Ritter <benerit...@googlemail.com> wrote: > Am 22. März 2012 17:28 schrieb <s...@apache.org>: >> Author: sebb >> Date: Thu Mar 22 16:28:47 2012 >> New Revision: 1303878 >> >> URL: http://svn.apache.org/viewvc?rev=1303878&view=rev >> Log: >> Allow testing of dynamically loaded CSVLexers >> > > I'm not sure if I understand where you are going with this. Don't you > think it's a bit over engineered to put reflection into the > performance test?
No, see below. > Are you doing that, because you want to test new > CSVLexer implementations using the CSVLexer1? To test new subclasses of the abstract Lexer. > The performance test starts to get complex. It actually gets more complex to keep adding/subtracting new tests. I originally did that at first locally, and it was a pain. > Maybe you should explain to the ML, what you are planning? This makes it much easier to test new implementations. Just create the new Lexer implementation, and then run the test. No need to update the test harness. And when the test is completed, the implementation can be deleted without having to fix the performance test again. Note: I don't plan to extend the reflective invocation further. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
