From: "Shawn O. Pearce" <spearce@spearce.org>
To: Robin Rosenberg <robin.rosenberg@dewire.com>
Cc: git@vger.kernel.org
Subject: [JGIT PATCH 08/13] Work around Sun JVM bug "Cleared SoftReference not added to queue"
Date: Tue, 28 Apr 2009 14:12:21 -0700 [thread overview]
Message-ID: <1240953146-12878-9-git-send-email-spearce@spearce.org> (raw)
In-Reply-To: <1240953146-12878-8-git-send-email-spearce@spearce.org>
> http://bugs.sun.com/view_bug.do?bug_id=4485942
>
> Normally all soft references cleared by the garbage collector are
> added to the reference queue the reference was created with (if any).
> But if a program happens to call SoftReference.get() after the reference
> has been cleared, but before it has been added to the reference queue it
> will never be added to the reference queue.
...
> This bug can be worked around by keeping track of calls to
> SoftReference.get() that returns null. The only problem is that the
> code must also allow for the behavior to change when/if the bug is
> fixed.
The bug was opened against JDK 1.3.1, reported to still exist in 1.4,
and isn't marked fixed yet despite 1.6 having shipped.
We already have a work around for this in our code. We mark a reference
as cleared using our own boolean, as we need to work around another bug
in the ReferenceQueue implementation anyway. So, we try to enqueue the
dead reference when we find it. If its already enqueued, no harm. If
we hit the enqueue bug and enqueue it twice, no harm, our cleared flag
will skip the second removal from the queue. If the bug described in
4485942 exists, we'll enqueue the reference for the GC.
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
---
.../src/org/spearce/jgit/lib/OffsetCache.java | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/org.spearce.jgit/src/org/spearce/jgit/lib/OffsetCache.java b/org.spearce.jgit/src/org/spearce/jgit/lib/OffsetCache.java
index 12912d9..e3ed37d 100644
--- a/org.spearce.jgit/src/org/spearce/jgit/lib/OffsetCache.java
+++ b/org.spearce.jgit/src/org/spearce/jgit/lib/OffsetCache.java
@@ -232,7 +232,7 @@ private V scan(Entry<V> n, final PackFile pack, final long position) {
hit(r);
return v;
}
- n.dead = true;
+ n.kill();
break;
}
}
--
1.6.3.rc1.205.g37f8
next prev parent reply other threads:[~2009-04-28 21:15 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-04-28 21:12 [JGIT PATCH 00/13] Misc. bug fixes and cleanups Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 01/13] Fix performance problem recently introduced to DeltaPackedObjectLoader Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 02/13] Don't use ByteWindows when checking pack file headers/footers Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 03/13] Rewrite WindowCache to be easier to follow and maintain Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 04/13] Document the IllegalArgumentException thrown by WindowCache.reconfigure Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 05/13] Create the new WindowCache before clearing the old one Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 06/13] Better handle concurrent reads during a WindowCache reconfiguration Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 07/13] Clear dead OffsetCache cells when clearing a reference Shawn O. Pearce
2009-04-28 21:12 ` Shawn O. Pearce [this message]
2009-04-28 21:12 ` [JGIT PATCH 09/13] Replace inefficient new String(String) constructor to silence FindBugs Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 10/13] Replace inefficient new Long(long) " Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 11/13] Change IndexPack to use ObjectIdSubclassMap instead of ObjectIdMap Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 12/13] Use getCachedBytes in IndexPack to avoid an unnecessary copy Shawn O. Pearce
2009-04-28 21:12 ` [JGIT PATCH 13/13] Remove ObjectIdMap from the JGit library Shawn O. Pearce
2009-04-29 19:45 ` [JGIT PATCH 10/13] Replace inefficient new Long(long) constructor to silence FindBugs Robin Rosenberg
2009-04-29 20:42 ` [PATCH 12/11] Add a test for LongMap, IndexPack's helper class Shawn O. Pearce
2009-04-29 20:10 ` [JGIT PATCH 09/13] Replace inefficient new String(String) constructor to silence FindBugs Robin Rosenberg
2009-04-29 20:22 ` Shawn O. Pearce
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1240953146-12878-9-git-send-email-spearce@spearce.org \
--to=spearce@spearce.org \
--cc=git@vger.kernel.org \
--cc=robin.rosenberg@dewire.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).