Thanks much for your response, Kris. I have been able to make things work for my purposes by applying the patch
suggestedby Daniel Migowski in the thread referenced below. I tried to implement your suggestion as well, but know
onlyenough to be dangerous. Returning Integer.MAX_VALUE from the Oid.TEXT case in the getPrecision method of
TypeInfoCache.javahad no discernible effect - I must have been coding in the wrong place. At any rate, I will look
forwardto that fix in the next release.
Thanks again,
--Ken
>>> On 12/2/2008 at 4:40 PM, in message
<Pine.BSO.4.64.0812021736080.1073@leary.csoft.net>, Kris Jurka
<books@ejurka.com> wrote:
>
> On Tue, 2 Dec 2008, Ken Boss wrote:
>
>> ... that postgres text datatypes appear to Crystal as String fields of
>> length -1, which confuses the software and prohibits their use. This is
>> particularly problematic given postgresql's proclivity for casting the
>> results of concatenations, unions, etc. to text on-the-fly.
>>
>> I see a long thread related to this topic from about a year ago here:
>> http://archives.postgresql.org/pgsql-jdbc/2007-12/msg00145.php , but it
>> seems to have run its course without coming to any specific conclusions.
>
> The conclusion that Oliver and I seemed to come to was that the type
> returned should not be changed, but instead the precision should be
> changed from -1 to Integer.MAX_VALUE. This change hasn't been made yet,
> but that will happen before the 8.4 release. This change won't be
> backpatched to earlier driver releases for backwards compatibility
> reasons.
>
> Kris Jurka