Yes, it is a bug, thanks.
It needs a test for each > 0 if length.out > 0 is set. Added for 2.5.0.
On Fri, 30 Mar 2007, Hiroyuki Kawakatsu wrote:
>> rep(2,0)
> numeric(0)
>> rep(2,0,2)
> [1] 2 2
>> rep(2,0,2,0)
> Floating exception (core dumped)
>
> #yes, i should not be doing this but i mistakenly did
> #par(oma=rep(2,0,2,0)) when i meant c(2,0,2,0)
> #
> #both on
> #2.4.1 Patched (2007-03-21 r40875); x86_64-unknown-freebsd6.2
> #and
> #2.5.0 alpha (2007-03-30 r40957); i386-pc-mingw32
>
>
--
Brian D. Ripley, ripley at stats.ox.ac.uk
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595