[Koha-zebra] indexing with File Record IDs, or General Record IDs ?

[Paul POULAIN]

Hello,

i'm playing with zebra & IDs.
The doc says "general record IDs are highly slower than File Record IDs".
But I suspect that with Zoom/Perl, we won't have the problem encountered 
with "parsing the whole directory to find what to do when running zebraidx"

So I think we should use General Record IDs. Am I right here ? If I am, 
then I need help to know how to setup zebra.cfg to explain that MARC 
field 090$a is the field to get ID from ?
I don't understand what to write with the doc...
--
Paul POULAIN
Consultant indépendant en logiciels libres
responsable francophone de koha (SIGB libre http://www.koha-fr.org)