Frederic Demians wrote: > What impact has this patch on performances? > > There isn't any issue with MARC21 biblio records. For me the issue > comes from MARC::Record which doesn't handle properly UNIMARC records > encoding. More precisely, in MARC::File::USMARC::decode function which > is used by MARC::Record even for UNIMARC records. If you force raw > ISO2709 UNIMARC record to UTF8 as MARC21 understand it, it solves the > problem: > > substr( $rawmarc, 9, 1 ) = 'a'; > my $record = MARC::File::USMARC::decode( $rawmarc ); Problem is that the label, in my opinion, should not be changed since it is a user input. My understanding is that this is a MARC::Record (MARC::File::USMARC) issue, I sent my script on perl4lib to detail the problem and explain what I have to do to fix that. But had no feedback. Your solution could be a quick fix for display purposes. But would not be right if you have to save the record since your label would not be UNIMARC but sort of UNIMARC21.
Anyhow the first problem addressed by the patch sent should not be forgotten though. Since it is really a problem for UNIMARC XSLT usage (see XSLT.pm modification). _______________________________________________ Koha-patches mailing list [email protected] http://lists.koha.org/mailman/listinfo/koha-patches
