Memory leak ChatRoom
Description
Environment
Activity
Walter Ebeling February 24, 2013 at 9:42 AM
Mass closure due to issue clean up
Walter Ebeling February 24, 2013 at 9:29 AM
Mass closing of resolved issues
Holger Bergunde February 28, 2012 at 2:20 PM
seems to be closed. reopen if there are still problems
Walter Ebeling June 20, 2011 at 6:58 PM
Please make ROAR the default and toast an option for 2.6.1.
wroot June 20, 2011 at 3:41 AM
Well. If i use Roar with SparkToaster popup option, then i don't see large chunks of memory being taken while showing popups. With a built-in toasts popups it uses much more memory and also i see that it steals focus with the built-in popups. Maybe we should already remove built-in popups and leave only the Roar.
But in general Spark is not releasing any memory when you close chat windows. Maybe that's how java apps operate, because this is the same with the JBother client i test. Maybe it will do a garbage collection and release some memory after some time. I'm also checking only with Process Explorer. Not sure which stats should i monitor (i'm looking at Working Set). Btw, if i minimize Spark roster window, then Working Set becomes very small, but if i press close button (which removes roster window and it only stays in systray), then no memory is released. But this is the same for the windows app Exodus and other ones. Working Set is declining only if you minimize the window. Maybe this is how Windows operate the memory.
I think we can disable built-in popups and close this ticket. Also, maybe we should do a check and if built-in popups are enabled in spark.properties file, then enable Roar plugin and set it to SparkToaster option.
As other users reports, after receiving lots of messages Spark can use huge amount of memory and after closing chat windows it doesn't release this memory. As i have tested this myself, i have noticed, that large memory chunks are used after every popup and it doesnt use so much memory if toasts are disabled, so probably there is some leak in toast popups. Also one should investigate garbage collection process and make Spark free uneeded memory.
Chatroom references dont get released