[jira] Commented: (VFS-50) Threading issue with AbstractFileSystem.fireEvent

["Chris DiGiano (JIRA)" <ji...@apache.org>]

    [ http://issues.apache.org/jira/browse/VFS-50?page=comments#action_12412770 ] 

Chris DiGiano commented on VFS-50:
----------------------------------

Here's a patch that makes a copy of the listener list to avoid race conditions:

Index: src/java/org/apache/commons/vfs/provider/AbstractFileSystem.java
===================================================================
--- src/java/org/apache/commons/vfs/provider/AbstractFileSystem.java    (revision 408551)
+++ src/java/org/apache/commons/vfs/provider/AbstractFileSystem.java    (working copy)
@@ -484,9 +484,12 @@
         synchronized (listenerMap)
         {
             final FileObject file = event.getFile();
-            final ArrayList listeners = (ArrayList) listenerMap.get(file.getName());
-            if (listeners != null)
+//            final ArrayList listeners = (ArrayList) listenerMap.get(file.getName());
+//            if (listeners != null)
+            if (listenerMap.containsKey(file.getName()))
             {
+                // Make a copy of the listeners to avoid race conditions. (Chris DiGiano)
+                final ArrayList listeners = new ArrayList((ArrayList) listenerMap.get(file.getName()));
                 final int count = listeners.size();
                 for (int i = 0; i < count; i++)
                 {


> Threading issue with AbstractFileSystem.fireEvent
> -------------------------------------------------
>
>          Key: VFS-50
>          URL: http://issues.apache.org/jira/browse/VFS-50
>      Project: Commons VFS
>         Type: Bug

>     Reporter: Chris DiGiano
>     Priority: Minor

>
> I make heavy use of DelegateFileObject in my code and recently ran into a threading issue when a change in one file was trying to automatically update dependent DelegateFileObjects. While events where still being propagated, one DelegateFileObject removed itself as an event listener. This caused an index-of-range problem in AbstractFileSystem.fireEvent because the listener list had shrunk, but the "count" had already been fixed.
> It appears that the current implementation of fireEvent is not entirely thread safe, if event handlers can still change the listener list in place. I will attach a patch that works for me.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira


---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org