Project Proposal: JDK 7 Update

Volker Simonis volker.simonis at gmail.com
Sat Jun 18 16:22:31 UTC 2011


On Sat, Jun 18, 2011 at 7:44 AM, David Holmes <David.Holmes at oracle.com> wrote:
> Volker Simonis said the following on 06/18/11 00:18:
>>
>> Sorry, I apologize if my mail was to rude or questioned the integrity
>> of anybody on this list. I definitely didn't wanted to hurt anybody.
>>
>> I'm just not very satisfied with how the things are going with the HSX
>> repositories compared to the HS versions in the official Oracle JDK
>> and wanted to prevent this situation for the JDK 7 update releases. If
>> my fears are without cause - that's just great!
>
> <snip>
>>
>> Just have look at he HSX (HotSpot express) repositories on
>> hg.openjdk.java.net/ and compare the changesets there with the
>> changesets which went into the HSX version in the official Oracle
>> JDK6uXX versions
>> (http://www.oracle.com/technetwork/java/javase/releasenotes-136954.html).
>> The difference that you'll notice is exactly the point of my mail.
>
> I'm afraid I'm not getting your point at all, what exactly is your concern
> with the HSX release model as it relates to the JDK6 update train? What is
> it that you see, or don't see, in the release notes that concerns you?
>

There are changes in the official 6uXX releases which don't get
backported to the corresponding OpenJDK hsx repositories. These are
mainly security fixes but I also recall another example which was
related to compressed strings (see
http://mail.openjdk.java.net/pipermail/hotspot-dev/2011-March/003907.html).

As you probaly now, the official JDK6 update releases contain HSX
version with incremented sub-numbers (HS20.1bxx, HS20.2bxx). They are
all different from the HSX repository in the OpenJDK which usually
remain at a certain build number and dont increment the sub-number. I
would definitely prefer to have one "golden" HSX repositotry in the
OpenJDK which contains ALL the changes of the corresponding HSX
versions from the closed JDK (maybe with a certain delay for security
fixes which would be no problem).


> Regards,
> David Holmes
>



More information about the discuss mailing list