| Using CopyFilesAndFoldersOperation from a WorkspaceJob [message #334891] |
Sat, 07 March 2009 20:47 |
 eric
Messages: 1
Registered: July 2009 |
Junior Member |
|
|
I am having problems using CopyFilesAndFoldersOperation from a
WorkspaceJob. In particular, I am trying to use the methods
copyResourcesInCurrentThread() in a backgroundable job, but I am running
into exceptions resulting from invalid thread access as a result of the
call to IProgressService.run() on line 559. For reference, I am using the
3.4.0 version of the org.eclipse.ui.ide plugin.
It seems to me that when called from copyResourcesInCurrentThread() the
internal implementation of copyResources() (declared on line 532) should
not make its call to copyResources() within a runnable. Apart from the
current problem I am encountering with invalid thread access, I do not
think it is consistent with the javadoc of copyResourcesInCurrentThread()
to make the copyResources() call in this way.
I believe that the implementation of copyFileStores() (on line 776) is
more correct since it uses the "fork" argument to determine whether or not
to perform the actual copy in a runnable. It seems that copyResources()
should be implemented similarly; i.e. lines 551-564 should be replaced
with the following:
if (fork) {
IRunnableWithProgress op = new IRunnableWithProgress() {
public void run(IProgressMonitor monitor) {
copyResources(resources, destinationPath, copiedResources,
monitor);
}
};
try {
PlatformUI.getWorkbench().getProgressService().run(fork, true, op);
} catch (InterruptedException e) {
return copiedResources[0];
} catch (InvocationTargetException e) {
display(e);
}
} else {
copyResources(resources, destinationPath, copiedResources, monitor);
}
That implementation fixes the invalid thread access and would seem to be
the behavior which the javadoc specifies.
Does this make sense? If so, I will file a bug on the existing
implementation.
Thanks,
Eric
|
|
|
Powered by
FUDForum. Page generated in 0.02655 seconds
] 
Search
Help
Register
Login
Home
