This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Using NetBeans 6 M10 running on JDK6u2 on Ubuntu 7.04. When comparing two memory snapshots I noticed that both snapshots have 3 char[] objects. The total number of bytes, however, is different. In one of the snapshots the total bytes is 96, in the other it is 104. I opened the Allocations Comparison window on these two snapshots and expected to see an entry for char[] but there is no entry.
Created attachment 46814 [details] profiler memory snapshot
Created attachment 46815 [details] profiler memory snapshot
All classes with similar number of allocated instances in snapshots being compared are filtered-out by mistake in snapshots diff.
Fixed for Beta1. Could you please verify also for "Record both object creation and garbage collection" memory profiling using your simple app?
>Could you please verify also for "Record both object creation and garbage collection" memory profiling using your simple app? Looks like it is working in the daily build from 2007-09-03. But now that the fix is in, I have some new questions. Perhaps I should open a new IZ. :-) I will attach two more snapshots. When I compare them I see something I don't understand: some of the entries have zero for both values. For example, javax.swing.text.Segment. In both snapshots the total bytes is 64 and the Objects Allocated is 2 - so why is there an entry in the comparisons window? Also, I don't really understand the entries for sun.java2d.pipe.Region. The objects allocated value is non-zero, but the bytes allocated value *is* zero. How can those object instances take up no space? Is the objects allocated value taken from perm gen and sun.java2d.pipe.Region has been loaded multiple times into perm gen but when these snapshots were taken there were no instances of it on the heap? In other words, all object instances have been removed by GC from the young and tenured generations but since perm gen gets GC'ed differently not all copies of the class itself got removed? Am I making any sense? :-)
Created attachment 48099 [details] sample memory snapshot
Created attachment 48100 [details] sample memory snapshot