Paul Eggert wrote: > I guess that the bug lies in the printing of a long double value that > is not exactly representable as a double.
This hardly explains it. Jim observes a call scale10_round_decimal_long_double (x, n=1) where x is some valid value. The code of scale10_round_decimal_long_double does not use the 'double' type; all it uses are 'long double', 'int', and 'unsigned int'. Bruno
