Re: getProcedureColumns

On Feb 14, 2007, at 1:37 AM, Kris Jurka wrote:

>
>
> On Tue, 6 Feb 2007, Jeffrey Cox wrote:
>
>> Finally got these patches finished up. Sorry for the delay ( had a
>> machine learning program fighting me all weekend). Please let me
>> know if this is not what you need, or more it turns out there is
>> something not working. All seems to be working, and am actually
>> using the resulting jar in a external app. Though the tests get
>> the job done, they are a bit 'brute force'. i figure they can be
>> re-factored later. probably goes for the other code as well I
>> suppose.
>>
>
> This doesn't work on an 8.0 server so they're unlikely to work on
> 7.4 or 7.3 servers either all of which the driver must support.
> Right now I don't have all of these handy to test with, so if you
> don't either we'll have to wait for me to return to my normal dev
> machine.

I can make them handy, but when you say they don't work do you mean
it crashes, or what exactly. I ask only because the DBMS seems to
have varying support for the method overall. below 7.3 not being able
to support the call at all, below 8.0 not supporting argument names
and below 8.1 not supporting argument modes. when I say not
supporting, I mean there is not column for the data in 'pg_proc'. I
didn't actually try this against those over servers, but looked it up
in the respective documentation of each server version. I suppose
that information might reside in another table, I will have to
install the other versions and find out.

>
> Some other notes:
>
> 1) It's better to send all your changes as one patch instead of one
> per file. It makes it easier to apply.

good to know, would you like me to resend them?

>
> 2) You've sent a patch for AbstractJdbc2Array that only adds
> something that's commented out.

My bad... I had made a number of changes in that file originally,
then after some discussion, changed direction. I must have missed the
comments.

>
> 3) When doing version comparisons you want
> connection.haveMinimumServerVersion, not MinimumCompatibleVersion.
> The Compatible idea is supposed to indicate what version of the
> driver the current driver should imitate (although it's rarely used).

Also good to know, changed and the patch is attached.

Jeff