From b09964e97f65f7ca6e7dfaf33e34f2bbefe4e02b Mon Sep 17 00:00:00 2001 From: Stephan Bergmann Date: Thu, 11 Dec 2014 13:27:26 +0100 Subject: update documentation Change-Id: Icd966a850b7c5e276e8b1d74566a4ea02e5b4a5c --- jurt/com/sun/star/lib/util/AsynchronousFinalizer.java | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'jurt') diff --git a/jurt/com/sun/star/lib/util/AsynchronousFinalizer.java b/jurt/com/sun/star/lib/util/AsynchronousFinalizer.java index 71743b25241b..908a53e8bc0c 100644 --- a/jurt/com/sun/star/lib/util/AsynchronousFinalizer.java +++ b/jurt/com/sun/star/lib/util/AsynchronousFinalizer.java @@ -24,15 +24,17 @@ import java.util.LinkedList; * Helper class to asynchronously execute finalize methods. * *

Current JVMs seem not to be robust against long-running finalize methods, - * in that such long-running finalize methods may lead to OutOfMemoryErrors. This - * class mitigates the problem by asynchronously shifting the bodies of + * in that such long-running finalize methods may lead to OutOfMemoryErrors. + * This class mitigates the problem by asynchronously shifting the bodies of * potentially long-running finalize methods into an extra thread. Classes that * make use of this in their finalize methods are the proxies used in the * intra-process JNI UNO bridge and the inter-process Java URP UNO bridge (where * in both cases finalizers lead to synchronous UNO release calls).

* - *

If JVMs are getting more mature and should no longer have problems with - * long-running finalize methods, this class could be removed again.

+ *

Irrespective whether JVMs are getting more mature and should no longer + * have problems with long-running finalize methods, at least the JNI UNO bridge + * needs some way to stop finalization of proxies (to C++ objects) well before + * process exit, as provided by drain().

*/ public final class AsynchronousFinalizer { public AsynchronousFinalizer() { -- cgit