Rick Hillegas <[EMAIL PROTECTED]> writes:

> Hi Dyre,
>
> These two issues were marked as blockers by the people who logged
> them. Is that anything more than a statement of how much pain these
> issues cause those people? 

Probably not. But I wanted to get an impression of which open bugs could
be candidates for 10.4, and the high-priority issues seemed like a good
place to start. 

> If I were to categorize issues according to
> the severity of their impact, it would look something like this (in
> declining order of significance):
>
> 1) Engine crashes
> 2) Data corruption
> 3) Wrong results
> 4) Failures (e.g., a query just aborts)
> 5) Misleading messages, typos, ugly formatting, ...

Agreed. Except that I think I would move 1) below corruption and wrong
results, but that's just my opinion...

> These two issues seem to fall into category (4). 

> If I were release
> manager, I would downgrade these to the level of other NPEs and query
> aborts--beneath the significance of issues in category (3).

Agreed, and this is more or less what I was planning to do. In the case
of DERBY-3260 it seemed so close to being fixed that I thought it might
be easier just to get it checked in. But if it, for some reason, cannot
be fixed, I agree that it should be downgraded.

Wrt. DERBY-3303 I think that it is actually in category 3) (as I think
Army has said, as well).  

> You, of course, are the release manager. It's up to you to publish
> your criteria for ranking issues. It may also be helpful to ask the
> user community to vote on which issues are most important to them.

Well, I welcome any input on how to categorize bugs :)

>
> [EMAIL PROTECTED] wrote:
>> Currently we have two JIRA issues classified as "Blockers":
>>
>> https://issues.apache.org/jira/browse/DERBY-3260
>>
>> and
>>
>> https://issues.apache.org/jira/browse/DERBY-3303
>>
>>
>> DERBY-3260 has a patch which has been reviewed favorably by Knut, but I
>> feel that the following quote from his comment deserves some attention:
>>
>> "It would be interesting, though, if someone could shed some light on
>> why the synchronization was commented out in the first place. This piece
>> of code hasn't been changed since the code was donated to Apache, so the
>> commit log doesn't tell us anything about it."
>>
>>
>> It seems like both DERBY-3303 and its "Critical" cousin DERBY-3231
>> are caused by rev 516454 (fix for DERBY-681), (a fix was checked in for
>> DERBY-3231, but it remains open). Is there any chance of these being
>> fixed in time for 10.4?
>>
>>   
>

-- 
dt

Reply via email to