Using letters that are not in the standard english alphabet is causing issues when running workspaces.
To take one exemple, passing a user parameter to a child workspace using WorkspaceRunner, will fail if the parameter contains ie. norwegian letters. “å” becomes “ Ă¥” (unicode \u00E5 ). If the parameter is a file path, child workspace will fail finding the folder/file.
A workaround is to manually convert unsupported unicode characters with StringPairReplacer.
This is suboptimal for 3 main reasons:
- Most often we need to handle these characters in inputs and outputs.
- Non english letters are meaningful, they convey sense for users. Scaling down to an english subset can sometimes create confusion between two similar words that are differentiated with these extra letters.
- Creating datastructure (and being able to handle these characters in our FME flows) in native language is adding value to our customers.