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.
Build: NetBeans IDE Dev (Build 090826) VM: Java HotSpot(TM) Client VM, 11.3-b02, Java(TM) SE Runtime Environment, 1.6.0_13-b03 OS: Linux, 2.6.27-7-generic, i386 User Comments: vv159170: started IDE previously closed with 6 projects Maximum slowness yet reported was 51346 ms, average is 51346
Created attachment 86704 [details] nps snapshot
IMO most of the time is spent in File operation. Please re-evaluate. Thanks.
What exactly you want us to re-evaluate? AWT is blocked by ActiveConfigAction for 51s. It blocks and waits for some background activity. It shall not block for such a long time. Thinking about it, 51s is a lot more than UI guidelines allow. Making P2: http://performance.netbeans.org/responsiveness/whatisresponsiveness.html
Thanks for valuable advice. Setting priority back to P3 since 42s of those 51s is spent in File.lastModified(). Reassigning to apisupport for further evaluation.
Usual issue of being too slow to load module list. *** This issue has been marked as a duplicate of 169040 ***
Although it is blocked by loading module universe, this won't get resolved with fix of issue #169040. Here's the probable setup according to the snapshot: 1) 2 projects are being open during startup: one project with some configurations (e.g. J2SE) and an apisupport project 2) user does something (not sure what) which triggers ActiveConfigAction.refreshView() and PM.findProject in EQ, which probably gets blocked by loading NBM project. Strange thing is that both threads need only read lock on PM.mutex() and I don't see anything else that would block them. Possible solution would IMHO be to refresh view in background or somehow break the notification chain in AWT thread. Can't really be solved by any speedup of module list loading as it is blocked for 42 s in native lastModified() method.
I don't want to juggle with threads when refreshing ActiveConfigAction just because of this one single report that is moreover suspicious because of 42 sec. spent in simple native IO call. Closing the issue until there are more similar reports of this performance problem.
*** This bug has been marked as a duplicate of bug 175348 ***