Re: GiST indexing question

Dear Yeb,

Thanks for the answer.

On Sat, Dec 11, 2010 at 3:10 PM, Yeb Havinga <yebhavinga(at)gmail(dot)com> wrote:
> On 2010-12-11 06:09, Greg Landrum wrote:
>> I guess I must be doing something stupid, but neither the docs nor the
>> internet have been particularly helpful in letting me know what.
>
> Yeah, a hint might help for that error message. But since it doesn't, a
> quick way for problems like this is to increase log_error_verbosity to
> verbose, trigger the error again. Then the sourcefile and linenumber are
> logged as well. If it is then still unclear what's causing the error, you
> can use that information in gdb to set a breakpoint on that source location.
>>
>> In case it's helpful, the full extension code (without the
>> non-functioning attempts at adding == to the index) is here:
>> http://rdkit.svn.sourceforge.net/viewvc/rdkit/trunk/Code/PgSQL/rdkit/
>
> If you take a look at
> http://rdkit.svn.sourceforge.net/viewvc/rdkit/trunk/Code/PgSQL/rdkit/rdkit_gist.c?revision=1481&view=markup
> and then line 651: that's the error being thrown. You need to extend
> gmol_consistent to handle the strategy 6 number.

Doh! That was it. No big wonder I couldn't find anything helpful on
the 'net. Now I'm really glad I included the link to the code.

Thanks again for both the answer and for providing a description of
how to track down future problems.

Best Regards,
-greg