New IcedTea6 Repository

[Joseph D. Darcy]

Mark Wielaard wrote:
> On Mon, 2008-02-18 at 12:33 +0100, Mark Wielaard wrote:
>> On Sun, 2008-02-17 at 14:24 -0800, Joseph D. Darcy wrote:
>>> If you send me Sun bug ids (here or over on jdk6-dev at openjdk.java.net), 
>>> I can lookup if they are expected to be fixed or not in the OpenJDK 6 
>>> code.  The OpenJDK 6 code branched from JDK 7 from around build 20 to 22 
>>> so most fixes should be present in OpenJDK 6 too.  We're also interested 
>>> in having external fixes migrate back.
>> We try to always add a Sun Bug ID URL to any icedtea bugzilla entry. But
>> this is sometimes a little hard since it can take days (sometimes even
>> weeks) for sun bug ids to show up publicly. See the URL field of the
>> bugzilla entry. (I just added a couple.)
>>
>> After a quick look the following bugzillas/bugids seem relevant:
> 
> And just remembered that Keith has been trying to push through some
> fixes from icedtea to openjdk:

Hello.

> http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=60
> http://bugs.sun.com/view_bug.do?bug_id=6593946
> "ByteBuffer.compact() does not discard mark"
> patches/icedtea-bytebuffer-compact.patch

There was a little discussion about the appropriateness of the fix after 
Keith proposed it:
http://mail.openjdk.java.net/pipermail/core-libs-dev/2007-October/000077.html

Once OpenJDK 6 settles down a bit, I plan to write some guidelines for 
discussion of what kinds of bug fixes are appropriate for OpenJDK 6. 
For the particular issue at hand, the change could cause some existing 
programs to throw an exception where they did not before.  If we find 
long-standing behavior which contradicts the spec, at times we prefer to 
  preserve behavioral compatibility and update the specification to 
match the behavior.  In other cases, we do the reverse.  In a feature 
release like JDK 7, we generally feel freer to change behaviors like 
these than in an update release; the right balance in this regard for 
OpenJDK 6 may be between feature and update release.

> http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=70
> http://bugs.sun.com/view_bug.do?bug_id=6593486
> "IcedTea should use system tzdata instead of built-in db"
> icedtea-use-system-tzdata.patch

Yes, I want to consider incorporating that in OpenJDK 6 too.

Regards,

-Joe