Sunbug 6934356: Vector.writeObject() synchronization risks serialization deadlock

Dr Andrew John Hughes gnu_andrew at
Fri Dec 17 01:36:29 UTC 2010

On 16 December 2010 16:39, Neil Richards <neil.richards at> wrote:
> On 13 December 2010 20:04, Neil Richards <neil.richards at> wrote:
>> On 13 December 2010 18:46, Alan Bateman <Alan.Bateman at> wrote:
>>> I haven't looked at your patch in detail yet but I suspect there may be
>>> similar issues in other places (like j.u.Hashtable) once you are done with
>>> Vector.
>> You are correct, there is a similar issue in Hashtable.
>> I'm currently in the process of working up another fix and testcase
>> for that, which I hope to submit separately shortly.
> I now notice that there isn't a bug in Java bug database exactly for
> the problem in Hashtable - though it's obviously related to 6934356,
> and to 4741471 (which removed writeObject() synchronization from the
> unsynchronized collection classes).
> (As in Vector, the problem in Hashtable has existed since 1.4).
> In this case, can I raise an entry in the OpenJDK bug database for the
> problem in Hashtable, or do I need to get it raised in the Java bug
> database?

Commits to OpenJDK still require a ID.  There was a plan nearly
two years ago to actually make the OpenJDK bugzilla more useful
(see but I've not heard of any
progress on this.

While we're still stuck with the system, It's easiest to
just ask one of the
Oracle developers to allocate a bug ID.  They seem to be able to do so
fairly instantaneously
whereas it takes a lot longer externally. tells you how to
provide an appropriate
Mercurial changeset for commit.  Commit comments are in a rigid format
controlled by a proprietary jcheck
script running on the OpenJDK servers and this will reject commit
matches that don't match the format or have a
duplicate bug IDs as well as patches which don't meet the whitespace
formatting requirements (and probably numerous
other checks that we don't know about).

> Any guidance gratefully received,
> Thanks,
> Neil
> PS: What is the standard way of referring to a bug from one system or
> the other? Or are the bug numbers from each system sufficiently far
> apart to avoid ambiguity that way?
> --
> Unless stated above:
> Work email: neil_richards at
> IBM United Kingdom Limited - Registered in England and Wales with number 741598.
> Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU

Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (

Support Free Java!
Contribute to GNU Classpath and the OpenJDK

PGP Key: 94EFD9D8 (
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8

More information about the core-libs-dev mailing list