https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=29268 --- Comment #7 from Martin Renvoize (ashimema) <martin.renvoize@openfifth.co.uk> --- Created attachment 211528 --> https://bugs.koha-community.org/bugzilla3/attachment.cgi?id=211528&action=edit Bug 29268: Correct cni handling for get_marc_host get_marc_host searched the Control-number index for the complete $w, for example "(OSt)123", which never matches a 001, then for the number alone. It never used the control number identifier (003), so two hosts with the same 001 and different 003 could not be told apart, and the $w of other organizations were ignored. Only numeric control numbers were accepted. Add get_host_queries, which builds the searches from every 773$w, most specific first: 1. (OrgCode)number from our own MARCOrgCode: 001 and 003 2. (OrgCode)number of any other organization: 001 and 003 3. number without OrgCode: 001 4. the number of 1. and 2. alone, for hosts that have no 003 get_marc_host uses the first one that finds a single record, as before, and still falls back to 773$0 and 773$9 when none does. The control number may now contain letters, and repeated 773 and $w are all used. To test: 1. prove t/db_dependent/Koha/Biblio/host_record.t 2. Set MARCOrgCode to OSt and add the hosts: 001 123 + 003 OSt, 001 123 + 003 XYZ, and 001 ocm555 without 003 3. Add children with 773$w "(OSt)123", "(XYZ)123", "123" and "(OSt)ocm555" 4. In the staff client add the children to a basket and look at the host shown for each: "(OSt)123" finds the OSt host, "(XYZ)123" the XYZ host, "(OSt)ocm555" the host without 003, and "123" none as it is ambiguous -- You are receiving this mail because: You are the assignee for the bug. You are watching all bug changes.