Hi Walter,

Attached example generates mentioned internal error. I suppose the problem
concerns implicit double->float
conversion. Earlier (8.28), compiler worked OK.

Nic Tiger.