Tim Bunce wrote:
> On Fri, May 23, 2003 at 05:40:43PM -0500, Stephen Clouse wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On Fri, May 23, 2003 at 11:28:12PM +0100, Tim Bunce wrote:
>>> On Fri, May 23, 2003 at 05:25:56PM -0400, HARMON,GARY
>>> (HP-Roseville,ex1) wrote:
>>>> DBD-Oracle 1.13 build, I get one test failing with the "expected
>>>> 'trailing' but got 'trailing ' for VARCHAR2" error.
>>>>
>>>> Tim, at
>>>> http://www.mail-archive.com/[EMAIL PROTECTED]/msg16938.html I see
>>>> you've recently requested some information on this error.  Hope
>>>> this helps.
>>>
>>> Oracle 9.2 does it, Oracle <9.2 doesn't. Sure seems like an Oracle
>>> bug/feature (unless it's an obscure side effect of the way I'm using
>>> OCI calls, but that seems unlikely).
>>>
>>> I think someone has opened a TAR with Oracle on this issue (I hope
>>> so).
>>> But I've no news yet.
>>
>> Ugh.  I meant to give an update on this but it slipped my mind.
>>
>> I had a TAR open with Oracle but I was totally unable to recreate
>> the problem at the OCI level.  My only guess is that the problem
>> lies somewhere in OCIBindDynamic, but I've never used that call
>> before, so I was in totally uncharted territory -- and considering
>> my current workload, I had to give up.
>>
>> I can reopen the TAR if I can get a real OCI test case.  If you or
>> someone else can throw something together, I'll pass it along and
>> work on the follow-up with Oracle.

FYI, the test still fails on the newly released 9.2.0.4 patchset. There's
fixed bugs in there that looked like they could have been related, but
apparently aren't, e.g.:

9202  2314238  VARCHAR2 Bind variables loose any TRAILING SPACES when used
over a DATABASE LINK or bound dynamically

9204  2886004  Reverse default behavior for blank trimming of Control
characters

-- 
Andy Hassall ([EMAIL PROTECTED]) icq(5747695) (http://www.andyh.co.uk)
Space: disk usage analysis tool (http://www.andyhsoftware.co.uk/space)

Reply via email to