bioinformatics-ua/dicoogle

Web services output ISO-8859-1 text in some endpoints

Enet4 opened this issue · 1 comments

Enet4 commented

A significant portion of web services in Dicoogle return their text response in a Latin1 encoding, which is not always compatible with the content to be transferred. This happens in those implemented via servlets, such as /search and /searchDIM.

This is assumed to be the default behavior from Jetty, although UTF-8 is a more reasonable default nowadays.

Enet4 commented

In accordance to the information presented here, Jetty includes a default set of encodings per MIME type in org/eclipse/jetty/http/encoding.properties. This file can be seen inside the Dicoogle jar with the default properties, which are far from ideal: it is set up to output text in ISO-8859-1 for almost all MIME types.

I could not yet find a way to override this file in the building process.