[R] Hmisc's aregImpute segfaults R-1.7.0 under linux

[Kieran Healy]

Hello -

When trying to use Hmisc library's aregImpute function on R 1.7.0, I 
got the following error -- shown here using the example code from the 
help page --- under both Linux and Mac OS X 10.2.5:

set.seed(3)
x1 <- factor(sample(c('a','b','c'),1000,T))
x2 <- (x1=='b') + 3*(x1=='c') + rnorm(1000,0,2)
x3 <- rnorm(1000)
y  <- x2 + 1*(x1=='c') + .2*x3 + rnorm(1000,0,2)
orig.x1 <- x1[1:250]
orig.x2 <- x2[251:350]
x1[1:250] <- NA
x2[251:350] <- NA

# Use 100 imputations to better check against individual true values
f <- aregImpute(~y + x1 + x2 + x3, n.impute=100)
Iteration:1 Error in whichClosest(pti[j], pti[nai]) : Incorrect number 
of arguments (7), expecting 8 for bincode

I upgraded to the most recent version of Hmisc (verson 1.5-3) on both 
platforms. This fixed the problem on OS X and aregImpute runs fine 
there. But with Hmisc 1.5-3, aregImpute now causes R to segfault on my 
linux box, viz:

f <- aregImpute(~y + x1 + x2 + x3, n.impute=100)
Iteration:1
Process R segmentation fault at Tue Apr 22 14:29:18 2003

I'm running an updated version of SuSE 7.1, if that's relevant. Is this 
a known issue?

Kieran

--
Kieran Healy, http://www.u.arizona.edu/~kjhealy
Asst Professor, Sociology Dept, University of Arizona.