Mailing-List: contact derby-dev-help@db.apache.org; run by ezmlm
Precedence: bulk
Reply-To: <derby-dev@db.apache.org>
Received-SPF: neutral (asf.osuosl.org: local policy)
Message-ID: <4411C6E8.5030408@sbcglobal.net>
Date: Fri, 10 Mar 2006 10:35:20 -0800
From: Mike Matrigali <mikem_app@sbcglobal.net>
Reply-To: mikem_app@sbcglobal.net
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
MIME-Version: 1.0
To: derby-dev@db.apache.org
Subject: Re: [jira] Updated: (DERBY-1067) support holdable Scrollable
 Updatable
 Resultsets
References: <1984080508.1142010760142.JavaMail.jira@ajax>
In-Reply-To: <1984080508.1142010760142.JavaMail.jira@ajax>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

I will review this patch, but probably won't get to it until monday.  If 
anyone else is reviewing, let me know so I can know when/if commit 
should happen.

Andreas Korneliussen (JIRA) wrote:
>      [ http://issues.apache.org/jira/browse/DERBY-1067?page=all ]
> 
> Andreas Korneliussen updated DERBY-1067:
> ----------------------------------------
> 
>     Attachment: DERBY-1067v2.diff
>                 DERBY-1067v2.stat
> 
> Attached is an updated patch. This patch includes changes from DERBY-690 to do 
> positioning at row location.  
> 
> The sequence number has been associated with reusable record id, and is now called "reusableRecordIdSequenceNumber"
> instead of "compressTimeStamp".
> 
> The sequence number is incremented from BaseContainer.compressContainer(..).  I noticed that a purge operation
> may cause pages with no records to be deallocated, so I have considered also incrementing the number when
> removing pages. However, some testing indicated to me that when a page is reused after being deallocated by a
> purge, the recordId is not reused within the page. It seems only truncate will cause the recordId to be reused.
> (truncate uses BaseContainer.compressContainer(..)).
Yes purge of page is not a problem.  Note these purges happen all the 
time in a db with deletes taking place, they are not specific to truncate.

> 
> Another change is that when incrementing the reusable record id sequenence number, the isDirty flag may also be set, to 
> ensure the container header gets written do disk.
> 
> Testing: 
> 
> T_AcessFactory has been extended with a test which checks that a compress
> will update the sequence number, so that the positionAtRowLocation(..) returns false.
> 
> The phaseTester has been extended to test that the version sequence number does not 
> harm a 10.1 Database.
> 
> I have currently only run these tests with this patch. 
> I will start derbyall now, and also test this patch along with DERBY-690 and DERBY-1058.
> 
> 
> 
>>support holdable Scrollable Updatable Resultsets
>>------------------------------------------------
>>
>>         Key: DERBY-1067
>>         URL: http://issues.apache.org/jira/browse/DERBY-1067
>>     Project: Derby
>>        Type: Sub-task
>>    Reporter: Andreas Korneliussen
>>    Assignee: Andreas Korneliussen
>> Attachments: DERBY-1067.diff, DERBY-1067.stat, DERBY-1067v2.diff, DERBY-1067v2.stat
>>
> 
> 
>