Re: [Rd] incorrect usage of glmer crashes R (PR#12375)

From: <maechler_at_stat.math.ethz.ch>
Date: Fri, 08 Aug 2008 20:15:14 +0200 (CEST)


This is about package 'lme4' and strictly speaking would not belong to R-bugs.

But you have reported a bug indeed; thank you.

>>>>> "s" == susscorfa <susscorfa_at_zonnet.nl>
>>>>> on Thu, 7 Aug 2008 12:20:41 +0200 (CEST) writes:

    s> Full_Name: susscorfa Version: 2.7.1 OS: ubuntu Submission     s> from: (NULL) (129.125.177.31)

    s> Incorrect implementation of the grouping variable in the     s> function glmer crashes R

{the correct term is "segmentation fault" or "segfault"}

    s> a small example:

    s> require(lme4); 
    s> a<-data.frame(b=rpois(1000,10), c=gl(20,50), d=rnorm(1000,3), e=rnorm(1000,5),
    s> f=rnorm(1000,2)+5);  
    s> glmer(b~d+f|c+(e), family=poisson, data=a)

Yes, indeed this segfaults; similarly for lmer().

The development version of the 'lme4' package (on R-forge) no longer does so since a few minutes ago.

    s> It crashes R on debian linux (2 independant systems) as well as windows

    s> The correct function usage is:
    s> glmer(b~d+f+(e|c), family=poisson, data=a)

    s> I suppose it should just eveluate as wrong usage of a     s> function but should not crash R

Yes, it now gives an error message about the incorrect specification.

Martin Maechler, ETH Zurich



R-devel_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel Received on Fri 08 Aug 2008 - 18:21:35 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Sat 09 Aug 2008 - 06:36:15 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-devel. Please read the posting guide before posting to the list.

list of date sections of archive