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 201307262300) VM: Java HotSpot(TM) 64-Bit Server VM, 23.25-b01, Java(TM) SE Runtime Environment, 1.7.0_25-b16 OS: Windows 7 User Comments: GUEST: invoking code completion in php Trait GUEST: autocomplite MackSix: Running Source>>Inspect--Windows 8 profile conformance. GUEST: I were typing this code: $seats[$i] = array($i + 1); When below line were ready suggestions popup appeared with slowdown $seats[$i] = array( Maximum slowness yet reported was 6850 ms, average is 4336
Created attachment 137914 [details] nps snapshot
The latest report (683492) contains info about a slowness caused by class o.n.m.analysis.RunAnalysis. (Older reports are related to PHP code completion. They seem to be already fixed, as they come from NetBeans 7.3 beta 2 or older.) Reassigning to editor/Hints & Annotations. Please evaluate, thank you.
(I investigated only the attached report, as that was the reason this was reassigned into the editor.) Sorry, but all I see in editor.hints is a call to Dialog.setVisible, which takes 3980ms wall-clock time. As the Swing client are required to call Swing methods from AWT Event Dispatch Thread, I cannot replan this call. So, its this call that will need to be made faster. Ultimately, all the time (according the profiler snapshot) is spent in org.netbeans.core.windows.services.NbPresenter.run, which takes 3980ms of the wall-clock time.
There is analysis.ui.NextError.actionPerformed() which waits for AsynchChildren.getNodes() to finish in EDT, that doesn't seem right. Please evaluate (or reassign to analysis.ui owner), thanks.
Standa, could you please help me understand where NextError.actionPerformed() is called in the profiler snapshot #683492 (which is the one refered by both Jarda and me)? I see 3980ms wall-clock time are spent in AWT thread there, out of which 3980ms are spent in org.netbeans.core.windows.services.NbPresenter.run. I can see some nodes work being done in the snapshot, but that appears to run in a request processor, which should be fine, I hope. Or were you by any chance refering to another snapshot, accidentally forgetting to mention that? If so, I guess #683492 should be filled separately (I can do that, no problem). Anyway, I am going to fix the problem in NextError.
Created attachment 138761 [details] screenshot I hope I'm not reading the snapshot wrong but that's what I get when I open that snapshot in NetBeans and expand AWT-EventQueue-0 node.
(In reply to Stanislav Aubrecht from comment #6) > Created attachment 138761 [details] > screenshot > > I hope I'm not reading the snapshot wrong but that's what I get when I open > that snapshot in NetBeans and expand AWT-EventQueue-0 node. Hm, sorry - when I had it opened before, it appeared there is nothing under NbPresenter.run - but after re-opening the snapshot, I see there is. My apologies for this, and thanks for pointing this out. Fixed: http://hg.netbeans.org/jet-main/rev/646e7751e47d
Integrated into 'main-silver', will be available in build *201308162300* on http://bits.netbeans.org/dev/nightly/ (upload may still be in progress) Changeset: http://hg.netbeans.org/main-silver/rev/646e7751e47d User: Jan Lahoda <jlahoda@netbeans.org> Log: #214744𹁘: computing next/previous nodes asynchronously