JDK 7 Updates: Policy Changes

[Andrew Haley]

On 23/07/15 21:27, Andrew Hughes wrote:
> 
>> So, the situation with OpenJDK 7 would be as with other projects, and
>> people don't need to be, say, JDK9 committers to modify bug reports.
>> And we can grant Author status to suitably-qualified contributors to
>> OpenJDK 7.
> 
> Can we?

Yes.

> As I say, it's not clear how this is done. I also don't believe
> someone should first need to be "a suitably-qualified contributor"
> to file a bug report.

Neither do I, but that's something to be fixed as part of of the
OpenJDK project itself rather than trying to do an end-around with
another bug database.

> People testing and filing issues are just as valuable as those
> fixing them.
>
> Other development projects, yes. I think the potential bug reporters for
> a package which is installed on many user's desktops (java-1.7.0-openjdk
> and the like in other distros) is different from that of something
> like Shenandoah, which are still experimental and where they'd probably
> have had to build it themselves first.
> 
>> It has the downside that end-users can't create bugs, but that's no
>> different from the other OpenJDK projects.  Sure, it'd be nice to get
>> that fixed for all projects.
> 
> It's different from the IcedTea bugzilla and our existing OpenJDK 6 JIRA,
> which are closer equivalents of OpenJDK 7. 

Indeed it is.

> It's also different from what OpenJDK 7 had before, where bugs could
> enter the database via Oracle's reporting system.

That's an interesting point, although from what I remember it has not
always been successful, and there's always been that "black hole"
feeling after a bug was reported and before it became visible.

> I don't think it's something that will ever be "fixed". It's done by
> design so that bugs for Oracle's binaries have to go through their
> triaging system first.
> 
> I'm hearing lots of downsides, but I have yet to hear any benefits
> of using the OpenJDK bug system over the existing system we have
> for OpenJDK 6, which has worked well for the last couple of years.

Using a separate bug database is not a real solution, though.  It's
not at all clear to me that we benefit substantially enough from going
outside the OpenJDK bug database but we lose continuity, the ability
to search and link bugs, and so on.  I'm not convinced that the
ability of non-Authors to create and edit bug reports is worth doing
something so different from the rest of OpenJDK.

The semi-detached status that the OpenJDK community outside Oracle has
is not doing anyone any good.  It's a hangover from the days when we
really did need to use our own infrastructure, not something to be
perpetuated.  By working outside OpenJDK we also lose some of the
incentive to get OpenJDK itself fixed.

Andrew.