Historically, GR has been using Latin-1 / ISO 8859-1 encoded strings, both
internally and in its API. Starting with version 0.39.0, UTF-8 is used
internally to allow a greater set of possible characters and symbols to be
represented. The API still expects input as Latin-1, but users can set the
input encoding using the
environment variable, which currently supports
Latin-1 and UTF-8.
Some workstation types lack any support for UTF-8 or Unicode in general and convert all input strings to Latin-1 (e.g. pdf and ps) and others are restricted to a subset of characters due to missing glyphs in the fonts distributed as part of GR (e.g. cairo). We plan to address these issues in future releases.