Curt Arnold (JIRA) wrote:
> https://issues.apache.org/jira/browse/LOGCXX-161
>
> Curt Arnold commented on LOGCXX-161:
> ------------------------------------
>
> Thanks for clarification. I wasn't particularly comfortable with the
> patch since it was possible that the Action might have a longer life
> that the Pool passed on the rollover method. I took a different tack
> on the problem and hope I have something that will address your
> concerns. I would appreciate your testing and feedback.
>
Thanks for your solution. I have done only a small amount of testing so
far, and can confirm that Actions don't increase the working set anymore.
During my test with perhaps 100 rollovers (each generating 10 Actions) I
have encountered only one increase in the working set. As I have built
APR with the VC debug heap support, I know that the increase was caused
by a pool allocation:
...\memory\unix\apr_pools.c(330) : {863474} normal block at 0x00E8DD68,
16384 bytes long.
There is no evidence yet that this is a permanent increase of the
working set; I had no time to delve deeper into this (I imagine I'd need
a helper library that does log callstacks to find out where log4cxx
allocates this pool block. Do yo know of such a library?).
Regards, Peter
--
_ _ Peter Steiner <[EMAIL PROTECTED]>
/ /_/ / Hug-Witschi AG <http://www.hugwi.ch/>
/ _ / Electronic Engineering
/_/ /_/ _ _ Auriedstrasse 10
/ / / / / / CH-3178 Boesingen
/ /_/ /_/ / Tel +41 31 740 44 44
/_ _ _ _ _/ Fax +41 31 740 44 45
