Skip to content

Bug in pexp (PR#1590)

1 message · Martin Maechler

#
dtrenkler> I wonder if something like this has been reported
    dtrenkler> before:

    >> pexp(85:86,0.438)
    dtrenkler> [1] 1 -Inf

yes, it has been reported (May 11, R-help) for pweibull() and
was tracked to a bug in the new R internal expm1() function
which only shows when no system-internal expm1() function is
present. Hence the problem has only been seen on Windows.

It has been fixed in "R-patched" for quite a while, and we
recommend you get "R-patched" pre-compiled for Windows
by Prof Ripley

	  http://www.stats.ox.ac.uk/pub/bdr/RWin/Rpre/SetupRpatched.exe

Thank you for reporting it anyway (it was not obvious if you haven't
followed R-help closely).