Nordlöw: > I was specifically interested in something that exercises > (random samples) potentially _all_ code points for string, > wstring and dstring (all code units that is). That's harder. Generating all uints and then testing if it's a Unicode dchar seems possible. Bye, bearophile