db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rick Hillegas <rick.hille...@oracle.com>
Subject Re: First Draft 10.8.2 release notes
Date Mon, 12 Sep 2011 13:00:41 GMT
On 9/9/11 2:40 PM, Myrna van Lunteren wrote:
> On Fri, Sep 9, 2011 at 8:53 AM, Dag H. Wanvik<dag.wanvik@oracle.com>  wrote:
>> Hi,
>>
>> thanks for assembling this, Myrna.
>>
>> Myrna van Lunteren<m.v.lunteren@gmail.com>  writes:
>>
>>>   DERBY-5331 Incorrect use of CharsetEncoder in DDMWriter
>>>   DERBY-5354 Remove unnecessary dita.regex build target
>>>   DERBY-5318 Use assertDirectoryDeleted in ReplicationRun and remove dead code
>>>   DERBY-5314 Enable i18n tests in non-English locales
>>>   DERBY-5291 testDerby4137_TransactionTimeoutSpecifiedNotExceeded
>>>   DERBY-5262 Running JUnit tests with Java 1.4.2 fails if the package private
>>>   DERBY-5253 Engine code references
>>>   DERBY-5252 make GrantRevokeTest pass in non-English locale
>>>   DERBY-5251 make ErrorCodeTest pass in non-English locale
>>>   DERBY-5247 Warnings regarding XPath displayed when generating JavaDoc
>>>   DERBY-5211 Make SysinfoCPCheckTest pass in non-English locale
>>>   DERBY-2625 SEVERE error involving column-width property
>>>   DERBY-2623 SEVERE error involving column-number property
>>>   DERBY-3337 convert jdbcapi/derbyStress.java to JUnit
>>>   DERBY-5084 convert ijConnName.sql to a ScriptTest junit test
>>>   DERBY-1903 Convert largedata/LobLimits.java to junit
>> These are not user visible changes and could be left out, I think.
>>
>> Rest looks good to me!
>>
>> Thanks,
>> Dag
>>
> Thanks for the review (Rick too)!
>
> I did a little experiment, to see if I could tweak my filter, and I
> succeeded in getting most - but not all - issues you mentioned by
> removing the 'Test' component. I didn't think I should manually adjust
> the release notes.
That sounds like the right approach. Hand-editing the release notes 
would be a brittle solution.
> However, I then checked on what Rick has done for e.g. 10.8.1 and test
> issues, which end-users wouldn't see, were included in the release
> notes for that release also, look e.g. for DERBY-5174 in those.
>
> I remember - maybe incorrectly - that Apache instructs us to include a
> record of *all* changes that went into a release, I think that's why
> we're including test and build changes (as well as changes to source
> that hasn't been in any official release). I think this is related to
> the fact that our apache release also includes a source distribution.
I believe that a similar discussion came up in the past about whether we 
should exclude documentation changes from our release notes. I imagine 
there are people out there who are only interested in behavioral changes 
to the product. The extra detail in our release notes won't be useful to 
them.

I don't get the sense that we're swamping signal with noise. However, 
that's just my opinion.
> So I think those bugs will stay in the release notes...
>
> Rick, as the most recent release manager, if I'm going off track with
> my reasoning, let me know...
If people feel strongly that the release notes are too verbose, then we 
should discuss how to flag noise issues in JIRA. That way we can 
programmatically exclude them from the filters. It's late in the day to 
do this for 10.8.2 but we could consider this change for 10.9.

Thanks,
-Rick
> Myrna
>


Mime
View raw message