I have a document with a MSWord file that contains non-latin (e.g. icelandic/cyrillic) letters in its name.
MSWord shows it correctly in the open dialog. Altough wen I try to open the file, an error occurs.
The problem is that a name of a file to open comes to Nuxeo (WSSFilter) spoiled. Seems that it is being converted from UTF to ISO-8559-1, so each non-latin letter is represented as two symbols. So, Nuxeo cannot recognize the file.
I've made the simplest possible fix: introduced an additional comparison into loop in SimpleNuxeoBackend.resolveLocation method.
This loop iterates through documents and tries to match a document's file name with a given one.
My check compares the file name converted to ISO-8559-1 as well.
This approach works well, but probably you can find better fix of this problem.