------- Comment #2 from audriusa at bluewin dot ch 2006-01-27 16:33 -------
Created an attachment (id=10742)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=10742&action=view)
Proposed fix
This patch fixes the RMI bug 25520 (the user class is not found by the RMI
server). The user class is not found because there is no user class loader on
the stack of that thread. It can only be loaded by the loader, returned from
the Thread.currentClassLoader().
I think, there is no harm to return the thread current class loader rather than
null from the VMObjectInputStream.currentClassLoader. The system class loader
will be the parent class of this class loader anyway.
2006-01-27 Audrius Meskauskas <[EMAIL PROTECTED]>
* java/io/VMObjectInputStream.java (loaderAction.run):
If no user class loaders found on the stack, return the thread
context class loader. (currentClassLoader): Explained.
--
audriusa at bluewin dot ch changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned at gcc dot gnu |audriusa at bluewin dot ch
|dot org |
Status|UNCONFIRMED |ASSIGNED
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25520
_______________________________________________
Bug-classpath mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-classpath
