Replying to
@vitaut The workaround is to cast all floats and doubles to int since the implementer hasn't (yet) had a chance to make the int to char* conversion run in negative time.
Software developer since the mid-80s, looking forward to retirement so I can work on one or more of the BSDs.