From mailing list: Some Z39.50 servers available around here seem to use different encodings for search queries and marc entries. For example, catalogo.bnportugal.pt which accepts latin1 for search queries but uses ISO-5426 for database entries. In yaz-client, I am able to interact with this server by explicitly setting these character encodings: marccharset iso5426 querycharset latin1 open catalogo.bnportugal.pt But in Tellico there is only one "character set" option, so either non-ASCII searches don't work or non-ASCII content in the results is garbled. Looking at the code (If I am looking in the right place - z3950connection.cpp?), it seems it's not possible to set separate encodings for incoming and outgoing data (m_sourceCharSet is used for both?). https://mail.kde.org/pipermail/tellico-users/2020-April/001502.html
Git commit 49156ff43a682345da31e97881cfb66bb0d51f7c by Robby Stephenson. Committed on 05/04/2020 at 18:28. Pushed by rstephenson into branch 'master'. Separate query character set encoding from response for z39.50 server By default, the response is assumed to be the same as the query if it is empty. FIXED-IN: 3.3 M +4 -0 ChangeLog M +32 -26 src/fetch/z3950connection.cpp M +6 -5 src/fetch/z3950connection.h M +51 -19 src/fetch/z3950fetcher.cpp M +6 -3 src/fetch/z3950fetcher.h M +25 -2 src/tests/z3950fetchertest.cpp M +1 -0 src/tests/z3950fetchertest.h https://commits.kde.org/tellico/49156ff43a682345da31e97881cfb66bb0d51f7c