Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>Although the DropHandler is not removed from the static AppContext map, that is not really the root cause, but merely one of the causes in the chain. (The drop handler is intended to be a singleton, and not cleared up until the AppContext class is unloaded, which in practice is never.) Use of a singleton DropHandler is by design.</p> <p>The real cause of the leak is that the DropHandler sets up an instance of TransferSupport, that is reused for each DnD operation, and during a DnD operation, provides it with a reference to the component involved in DnD. The problem is that <strong>it does not clear the reference when DnD finishes</strong>. If the DropHandler called <code>TransferSupport.setDNDVariables(null,null)</code> when the DnD exited, then the problem would go away. This is also the most logical solution, since the reference to the component is only required while DnD is in progress. Other approaches, such as clearing the AppContext map are circumventing the design rather than fixing a small oversight.</p> <p>But even if we fix this, the frame would still not be collected. Unfortunately, there appears to be another problem: When I commented out all the DnD related code, reducing to just a simple JFrame, this too was not being collected. The retaning reference was in <code>javax.swing.BufferStrategyPaintManager</code>. There is a <a href="http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6575402" rel="nofollow noreferrer">bug report</a> for this, as yet unfixed.</p> <p>So, if we fix the DnD, we hit another retention problem with repainting. Fortunately, all of these bugs hold on to only one frame (hopefully the same one!), so it is not as bad as it could be. The frame is disposed, so native resources are released, and all content can be removed, allowing that to be freed, reducing the severity of the memory leak.</p> <p>So, to finally answer your question, you are not doing anything wrong, you are just giving some of the bugs in the JDK a little air time!</p> <p><strong>UPDATE:</strong> The repaint manager bug has a quick fix - adding</p> <pre><code>-Dswing.bufferPerWindow=false </code></pre> <p>To the jvm startup options avoids the bug. With this bug quashed, it makes sense to post a fix for the DnD bug:</p> <p>To fix the DnD problem, you can add a call to this method at the end of importData().</p> <pre><code> private void cancelDnD(TransferSupport support) { /*TransferSupport.setDNDVariables(Component component, DropTargetEvent event) Call setDNDVariables(null, null) to free the component. */ try { Method m = support.getClass().getDeclaredMethod("setDNDVariables", new Class[] { Component.class, DropTargetEvent.class }); m.setAccessible(true); m.invoke(support, null, null); System.out.println("cancelledDnd"); } catch (Exception e) { } } </code></pre>
    singulars
    1. This table or related slice is empty.
    1. POMemory Leak with Swing Drag and Drop
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USmdma
    1. USmdma
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. POMemory Leak with Swing Drag and Drop
      singulars
      1. PTQuestion
    1. This table or related slice is empty.
    1. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTBountyClose
    3. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTAcceptedByOriginator
    1. COThanks for the info. I had spotted BufferStrategyPaintManager retaining some classes but was distracted with the DnD stuff. I accept that the reflection trick will work - but I don't intend to use it
      singulars
      1. PO
      1. UStom
    2. COSure, it's your choice if you use the fix or not, but now at least you know the scope of the problem and can decide how you deal with it. Like you. I'd probably not bother adding the fix, and knowing the problem can be partially mitigated, probably do nothing or at most, remove content from frames on dispose. But that is in hindsight now the extent of the problem is known. I was very curious to know why the leak was happening and what the consequences were, and it's nice to know that if we suddenly find a fix is needed, that one is available. I can sleep easy tonight! :-)
      singulars
      1. PO
      1. USmdma
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload