Hello,

We just upgraded from Koha 20.11 to 22.11, and we are facing an issue for the OAI-PMH service :

Test case (Koha Version 22.11.04) :

- Select a document having a title containing accents (eg : "Préparation à l'oral d'espagnol")
- Everything is correctly displayed in Koha
- Query the OAIPMH service for this record (cgi-bin/koha/oai.pl?verb=GetRecord&metadataPrefix=oai_dc&identifier=<oai_prefix>:<biblionumber>)
- The fields with accents are badly displayed. (For this example I get "Prp̌aration  l̉'oral d'espagnol"

======

My assumption is that the metadata gets encoded twice, and the accents are then badly encoded in the result.

I spotted exactly the line in the code where the data is badly encoded, but I can't find why.

The major difference between 20.11 et 22.11 is in Koha/OAI/Server/Repository.pm, line 181 : 
Before, we used the C4::Biblio::GetMarcBiblio function (now deleted), and now it's $biblio->metadata->record

For info, $biblio->metadata->metadata is correctly encoded

Thanks