db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Knut Anders Hatlen <Knut.Hat...@Sun.COM>
Subject Re: (DERBY-822 followup) Server prefetch when QRYROWSET is 0 (zero)?
Date Fri, 30 Jun 2006 18:49:08 GMT
Army <qozinx@gmail.com> writes:

> Knut Anders Hatlen wrote:
>> Bryan Pendleton <bpendleton@amberpoint.com> writes:
>
> Thanks to both Knut Anders and Bryan for the excellent and timely
> replies to my question.
>
>> When I look at writeQRYDTA and writeFDODTA in DRDAConnThread, it seems
>> to me that we actually break the DRDA spec in two ways:
>>
>>   1) QRYROWSET=0 has the same effect as QRYROWSET=1
>>   2) For forward-only/read-only result sets, QRYROWSET will be ignored
>>      (both on OPNQRY and CNTQRY) - haven't verified this since the
>>      client doesn't send QRYROWSET in this case.
>>
>> 1 is causing the reported problem with the ODBC driver. I don't know
>> if 2 is a problem. The client driver never sends an explicit QRYROWSET
>> when the result set is forward-only, but maybe the ODBC driver does.
>
> Thank you for tracking these down.  I filed #1 as a bug (DERBY-1468)
> and #2 as a "task" for further investigation (DERBY-1469).

Thanks!

> So the next questions are 1) should DERBY-1468 be marked as a
> "Regression" since existing apps that use ODBC against 10.1 will see
> incorrect behavior if they upgrade to 10.2?

I think marking it as regression is appropriate. The 822 patch broke
some of the implicit assumptions in the code.

> And 2) Ummm...anyone out
> there interested in fixing DERBY-1468 for the 10.2 deadline? :)

I guess it's my duty since I'm the source of the regression... Will
have a look at it.

-- 
Knut Anders

Mime
View raw message