At 10:16 PM 12/16/2011 -0500, Chris Nighswonger wrote:
On Fri, Dec 16, 2011 at 10:12 PM, Chris Nighswonger <cnighswonger@foundations.edu> wrote:
On Fri, Dec 16, 2011 at 7:09 PM, Paul <paul.a@aandc.org> wrote:
Lines 3006-3008 of biblio.pm called from additem.pl:
       # get title of the record (to store the 10 first letters with the index)
       my ( $titletag, $titlesubfield ) = GetMarcFromKohaField( 'biblio.title', '' );    # FIXME: should be GetFrameworkCode($biblionumber) ??
       $title = lc( $record->subfield( $titletag, $titlesubfield ) );

Before I start looking elsewhere, can someone kindly help me with the use of FIXME comments?
[snip]
Based on the commit message it looks like Joe was thinking that the second argument passed to GetMarcFromKohaField might should be populated with the results of GetFrameworkCode($biblionumber). In any case, it is still more likely to be corrupted data than the empty framework parameter passed to GetMarcFromKohaField that is causing your error.

Just back from a few days holidays, I cannot find "corrupted data" (this was an sql dump of our 3.2 data file), EXCEPT that we appear to have completely lost "authorities."  Could this be the "empty framework parameter" you mention?

While I was away, our most experienced cataloguer entered 25 [biblios + items] at my request for debugging purposes. The biblios went fine, the items all produced the biblio.pm line 3009 error, but both OPAC and staff search find the items perfectly.  However our "authorities" (personal names) which has some 15,000 entries is unsearchable (only the latest 25 appear.)

Perhaps I'm way off-track, but this may have something to do with re-indexing zebra.  I thought I had set up a cron job at 1 min intervals (as on the old server) but using both users (root and Koha) $ crontab -l gives nothing - so I am trying to set it up again, but find at <http://wiki.koha-community.org/wiki/Koha_on_Ubuntu> that

"Do Either Option#1 [OPTION 1 is Highly RECOMMENDED for Koha version 3.06.02] or Option #2 DO NOT DO BOTH.
Option #1: Use the Zebra Queue Daemon to manage indexing."

I had always understood that the "daemon" was deprecated in favour of a cron job (has this changed for 3.6?) and am now afraid that the "daemon" may have been working in the background damaging {something} as our searches (staff and OPAC) find titles, authors etc -- just not "authorities."

Fearing I'm getting into a potential mess here, could anyone please come up with suggestions?

Thanks in advance and happy new year,
Paul