there is a bug in openjdk7.

[Dalibor Topic]

On 8/6/13 12:16 AM, Andrew Hughes wrote:
> The problem is this change:
> 
> changeset:   6767:6bf5fa233df0
> user:        mchung
> date:        Tue Jul 02 13:23:35 2013 -0700
> summary:     8014925: Disable sun.reflect.Reflection.getCallerClass(int) with a temporary switch to re-enable it
> 
> as mentioned in [1] is missing from OpenJDK 7 u25, though clearly not Oracle u25. 

"clearly not" ;)


As the summary says, this is a fix for the issue 8014925. A quick look at http://bugs.sun.com/view_bug.do?bug_id=8014925 shows

"Fixed Versions:
	
7u40 (b33)"

If this was fixed in a different release, the "Fixed Versions" value would be different. 



For anyone following along the development of 7 updates, it's pretty trivial to look at the archives, and see that 

a) the fix was pushed on July 2nd: http://mail.openjdk.java.net/pipermail/jdk7u-dev/2013-July/006989.html - just like the changeset metadata says

b) the discussion about the change started on June 18th: http://mail.openjdk.java.net/pipermail/jdk7u-dev/2013-June/006791.html

c) that mail starting the discussion explicitly says:

"[2] 8016814: sun.reflect.Reflection.getCallerClass returns the wrong 
stack frame"

"This also fixes a regression introduced in 7u25 in which 
getCallerClass(int) is now a Java method that adds an additional frame 
that wasn't taken into account."

d) and for bonus points, note that Oracle JDK 7u25 was released on June 18th: http://www.oracle.com/technetwork/topics/security/javacpujun2013-1899847.html



So, barring a ripple in the time-space continuum, the "clearly not" charge is entertainingly wrong.

> The 2.3 JDK
> tree is synchronised with u25, but versions appear to have been added with no public changes:
> 
> icedtea7-2.3/jdk $ hg in ../../jdk7/jdk -r jdk7u25-b32
> 
> comparing with ../../jdk7/jdk
> searching for changes
> changeset:   6408:173b399abd8a
> tag:         jdk7u25-b16
> user:        katleman
> date:        Tue Jun 11 11:45:28 2013 -0700
> summary:     Added tag jdk7u25-b30 for changeset 0d68732acbd4
> 
> changeset:   6409:55f01444cf58
> tag:         jdk7u25-b32
> user:        katleman
> date:        Wed Jun 12 20:18:31 2013 -0700
> summary:     Added tag jdk7u25-b16 for changeset 173b399abd8a
> 
> $ hg log -R ~/projects/openjdk/upstream/icedtea7-2.3/jdk -k 8016814
> $

Yes, as discussed last year on this list, 

"Build # greater than b30 is to address Oracle customer issues and those 
builds are meant only for our customers."

http://mail.openjdk.java.net/pipermail/jdk7u-dev/2012-May/003197.html
 
> I can pull 8014925 into the 2.3.x tree and do a 2.3.13 release, but there needs to
> be much better project management of 7u, as this is getting beyond a joke.

That by itself would be a rather bad idea, I think, for a number of reasons:

a) Given that 8014925 disables sun.reflect.Reflection.getCallerClass(int), while 7u25 doesn't, you'd likely encounter some interesting bug reports
b) To better understand what kind of bug reports you'd get, I'd suggest reading up on the threads around 8014925 on mlvm-dev and core-libs-dev
c) In particular, make sure you understand 8021946

Anyhow, from reading https://bugzilla.redhat.com/show_bug.cgi?id=976693
it sounds like Red Hat has already solved this issue as of 1.7.0.25-2.4.1.0.el6 for 
RHEL a while ago, so I am curious why the change has not been picked up in 
Icedtea yet, allowing Ubuntu users to benefit from it as well ...

cheers,
dalibor topic
-- 
Oracle <http://www.oracle.com>
Dalibor Topic | Principal Product Manager
Phone: +494089091214 <tel:+494089091214> | Mobile: +491737185961 <tel:+491737185961>
Oracle Java Platform Group

ORACLE Deutschland B.V. & Co. KG | Kühnehöfe 5 | 22761 Hamburg

ORACLE Deutschland B.V. & Co. KG
Hauptverwaltung: Riesstr. 25, D-80992 München
Registergericht: Amtsgericht München, HRA 95603
Geschäftsführer: Jürgen Kunz

Komplementärin: ORACLE Deutschland Verwaltung B.V.
Hertogswetering 163/167, 3543 AS Utrecht, Niederlande
Handelsregister der Handelskammer Midden-Niederlande, Nr. 30143697
Geschäftsführer: Alexander van der Ven, Astrid Kepper, Val Maher

Green Oracle <http://www.oracle.com/commitment> Oracle is committed to developing practices and products that help protect the environment