[Rd] Strange behaviour of the ':' operator

[Hans W Borchers]

Is the following really intended behaviour of the ':' operator,

    > s <- pi - 3.0 + 1e-07
    > x <- s:pi
    > x
    [1] 0.1415928 1.1415928 2.1415928 3.1415928

though the last entry in the range vector is greater than pi?

    > x[4] > pi; x[4] - pi
    [1] TRUE
    [1] 1e-07

and the same, of course, for the seq() function.
I would understand this behaviour for 1e-14 or so, but it seems
unexpected for such a distinct difference as 1e-07.