Created attachment 10242 [details] Screenshots depicting both functioning and non-functioning unicode searches I have in my list of Z39.50 servers, a couple of servers which return the unprintable code {FFFD} when it is supposed to put a ÿ character. Other servers with the same search work fine. Trying the same search in MarcEdit's Z39.50 client, it comes out with the correct character. The servers in question are (and these are exactly the settings I have in my Koha installation): Australian National University host: library.anu.edu.au port: 210 databse: innopac syntax: USMARC encoding: MARC-8 and University of Melbourne host: cat.lib.unimelb.edu.au port: 210 databse: innopac syntax: USMARC encoding: MARC-8 Steps to reproduceː With the above servers added to Z39.50 Client Targets list: 1) Search for 'Francis Alys' in Title 2) Look down the list for title 'Francis Alÿs' by Author 'Medina, Cuauhtm̌oc' 3) See that there is a � character where there should be a ÿ Actual resultː Result from Australia National Uni - Koha: Francis Al�s / Medina, Cuauht�moc Expected Resultː Francis Alÿs / Medina, Cuauhtm̌oc Result from Aust National Uni - MarcEdit: Francis Alÿs / Medina, Cuauhtm̌oc Additional information: Attached are screenshots of 2 cases where the characters don't show properly (University of Melbourne and Australian National University), a shot of the same search with La Trobe University, which appears correctly, and a shot of the Australian National University search done (successfully) on MarcEdit's Z39.50 Client. Here is a link to the catalogued item on Australian National Uni's own library system: http://library.anu.edu.au/record=b2298563
What happens if you tell it to get utf8 instead of marc8?
Exactly the same result with both marc-8 and utf8.
This problems seems to be resolved by changing z3950 parameters. I tried the first target with UTF-8 and received the expected response with Francis Alÿs. Please also have a look at a similar problem: bug 8610. Closing this report now.