public static List extractSelectedObjects(TreeView tv) { List retList = new ArrayList(); for (Path p : tv.getSelectedPaths()) { Object currBranch = tv.getTreeData(); for (int i : p) { currBranch = ((List) currBranch).get(i); } retList.add(currBranch); } return retList; }
Monday, August 16, 2010
Extracting selected pivot tree items
The pivot TreeView only returns selected paths. A path is a list of list indexes which refer to the underlying data model of a tree. Unfortunately, the treeview doesn't have a convenience method for extracting the items referenced by those paths. Here's that missing method:
Thursday, August 12, 2010
Preventing grizzly timeouts
In grizzly, the default timeout for an http process is 5 minutes. This applies to the total activity time. In most cases, this is fine, however for long uploads you may need to extend this. Oh say for uploading 50g files. There is a switch for disabling timeouts, however this is global. What we need is a timeout for connections that are still active. Inactive connections should close 5m after last activity.
Grizzly logs the following:
Aug 12, 2010 3:17:25 PM com.sun.grizzly.http.KeepAliveThreadAttachment timedOut WARNING: GRIZZLY0023: Interrupting idle Thread: Grizzly-8080(2). Aug 12, 2010 3:17:27 PM com.sun.grizzly.http.KeepAliveThreadAttachment timedOut WARNING: GRIZZLY0023: Interrupting idle Thread: Grizzly-8080(3).The somewhat messy solution is the pick apart the thread that your request is running in and toggle timeouts as needed.
while ((read = in.read(byteArray)) != -1) { if ((System.currentTimeMillis() - lastXmit) > 5000) { HttpWorkerThread thread = (HttpWorkerThread)Thread.currentThread(); ((SelectionKeyAttachment) thread.getProcessorTask().getSelectionKey().attachment()).setTimeout(lastXmit); lastXmit = System.currentTimeMillis(); } buffer.clear(); buffer.limit(read); openFile.write(buffer); }
Subscribe to:
Posts (Atom)