Hi Matt, Can you provide a patch to add an option to restore the behavior from we had in 2.4.2?
Gary On Sun, Oct 29, 2017 at 2:32 PM, Matt Sicker <boa...@gmail.com> wrote: > It's been a while since I looked at that, but from what I can tell, if the > method name needs to be tracked as well as the class name, then the > SecurityManager solution isn't practical at all. Pre Java 9 would need to > use a Throwable no matter what, while Java 9+ can use StackWalker. > > On 29 October 2017 at 10:51, Gary Gregory <garydgreg...@gmail.com> wrote: > > > On Oct 29, 2017 02:21, "Pascal Schumacher" <pascalschumac...@gmx.net> > > wrote: > > > > Am 29.10.2017 um 09:09 schrieb Mark Thomas: > > > > > On 29 October 2017 03:54:40 GMT+00:00, "Bruno P. Kinoshita" < > > > brunodepau...@yahoo.com.br.INVALID> wrote: > > > > > >> Hi Gary, > > >> > > >> Started the tests in Maven command line, found which tests failed. > > >> Executed the tests in Eclipse, found which class was related to the > > >> failure. Then did a diff between both tags. > > >> > > >> git diff POOL_2_4_2 POOL_2.4.3-RC1 -- > > >> ./src/main/java/org/apache/commons/pool2/impl/ > DefaultPooledObject.java > > >> > > >> The short diff indicates that the failures started possibly due to the > > >> replacement of Exceptions in the DefaultPooledObject by a CallStack. > > >> Looks like the CallStack does not keep track of all the methods called > > >> (due to a security context manager from what I could tell?). Only > > >> classes. > > >> > > >> The following pull request updates pool to 2.4.3, and changes the unit > > >> tests to, instead of looking for method names in the log writer > output, > > >> to look for the test class name. > > >> > > >> https://github.com/apache/commons-dbcp/pull/8 > > >> > > >> > > >> Feel free to review and merge if you agree it's a good solution. > > >> > > > I'm not sure it is a good solution. I need to look at what is going on > in > > > more detail but if the method information has been lost then that will > > make > > > tracking down the root cause of a pool leak much, much harder. > > > > > > > I guess this was caused by: > > > > [POOL-320]: Use more efficient stack walking mechanisms for usage > tracking > > > > https://github.com/apache/commons-pool/commit/3994baf0f3ce59 > > b73bd36e869320275d757d1884 > > > > > > Matt, > > > > Any thoughts on the missing method names? > > > > Gary > > > > > > -- > Matt Sicker <boa...@gmail.com> >
