I see that getdate.y normally zeroes out the nanoseconds field (e.g.
$$.tv_nsec = 0).

Is it likely that changing the parser to accept floating-point seconds
fields would change the meaning of time specifiers that previously
were interpreted differently?   It doesn't immediately seem so to me,
but I'm worried that a change of interpretation here could cause
behaviour changes that are somewhat insidious.

James.


Reply via email to