[RFC] PAV: add missing panel timeout check

[Thomas Meyer]

# HG changeset patch
# User thomas
# Date 1327229831 -3600
# Node ID ddb05011ff2f86ccdd2d0bad1fa659941fda66f1
# Parent  7d090bfa381924100c0f8446dda2a615a75637d9
The wait loop for the panel to instantiate will result in a really *long* wait loop, e.g. when the JAR for the panel class is not found for some reason, as each loop step will wait 180s.
while waiting for the loop to finish, at least the firefox browser becomes unresponsive for all open tabs.
i'm not sure why this happens, but I guess the c++ side waits for a response it never gets and this will stall the whole browser.
maybe we should decrease the maximum wait timeout to something more appropriate?!

diff -r 7d090bfa3819 -r ddb05011ff2f plugin/icedteanp/java/sun/applet/PluginAppletViewer.java
--- a/plugin/icedteanp/java/sun/applet/PluginAppletViewer.java	Tue Dec 13 12:45:36 2011 +0100
+++ b/plugin/icedteanp/java/sun/applet/PluginAppletViewer.java	Sun Jan 22 11:57:11 2012 +0100
@@ -745,9 +745,17 @@
             long maxTimeToSleep = APPLET_TIMEOUT;
             panelLock.lock();
             try {
-                while (panel == null || !panel.isAlive())
+                while (panel == null || !panel.isAlive()) {
                     maxTimeToSleep -= waitTillTimeout(panelLock, panelLive,
                                                       maxTimeToSleep);
+                    
+                    // we already waited till timeout, give up here directly, instead of waiting 180s again in below waitForAppletInit()
+                    if(maxTimeToSleep < 0) {
+                        panelLock.unlock();
+                        streamhandler.write("instance " + identifier + " reference " + -1 + " fatalError: " + "Initialization timed out");
+                        return;
+                    }
+                }
             }
             finally {
                 panelLock.unlock();