Miller handles strings with any characters other than 0x00 or 0xff, using explicit UTF-8-friendly string-length computations. (I have no plans to support UTF-16 or ISO-8859-1.)

By and large, Miller treats strings as sequences of non-null bytes without need to interpret them semantically. Intentional support for internationalization includes:

Meanwhile, regular expressions and the DSL functions sub and gsub function correctly, albeit without explicit intentional support.

Please file an issue at if you encounter bugs related to internationalization (or anything else for that matter).