a new way to crash R? (PR#8981)

This works for me on both Linux and Windows.

Please check your memory usage: it does need about 900Kb of VM, and as you 
have less RAM than that installed you need to set --max-mem-size=1G or 
some such.

(It is quite likely that this is a Windows memory allocation failure: that 
has been reported before but not tracked down.)

On Wed, 14 Jun 2006, jritter at umn.edu wrote:

This is a multi-part message in MIME format.
--------------010500030900050501010804
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Dear R Team,

First, thank you for incredibly useful software!

Now the bad news:  The attached script (or the original version with
real data) will reliably crash R on my machine.  I am using:

R version: either 2.2.1 or 2.3.1
Windows 2000 Professional, Service Pack 4
512 MB of RAM

On my machine the script will crash R on line 42 [  probits21 <-
lapply(...  ].

In both this script and the "real" one, memory.size() indicates that R
is about 512 MB just before executing the command that crashes it.  In
other contexts, R has been able to allocate virtual memory on this
computer without any trouble.

Thanks again for all your work on R.

Regards,

Joe Ritter

--------------------------------------------------------------------------
Joseph A. Ritter
Associate Professor
Humphrey Institute of Public Affairs
University of Minnesota
Minneapolis, MN  55455
U.S.A.
612-625-0442

--------------010500030900050501010804
Content-Type: text/plain;
name="crash.r"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="crash.r"

x1 <- rnorm(10000)
x2 <- rnorm(10000)
x3 <- rnorm(10000)
x4 <- rnorm(10000)
x5 <- rnorm(10000)
x6 <- rnorm(10000)
x7 <- rnorm(10000)
x8 <- rnorm(10000)
y <- round(runif(10000))

specs <- list()
specs$A <- formula(y~x1+x2+x3+x4+x5+x6+x7+x8)
specs$B <- specs$A
specs$C <- specs$A
specs$D <- specs$A
specs$E <- specs$A
specs$F <- specs$A
specs$H <- specs$A
specs$I <- specs$A
specs$J <- specs$A

probits1 <- lapply(specs, glm, family=binomial(link="probit"))
probits2 <- lapply(specs, glm, family=binomial(link="probit"))
probits3 <- lapply(specs, glm, family=binomial(link="probit"))
probits4 <- lapply(specs, glm, family=binomial(link="probit"))
probits5 <- lapply(specs, glm, family=binomial(link="probit"))
probits6 <- lapply(specs, glm, family=binomial(link="probit"))
probits7 <- lapply(specs, glm, family=binomial(link="probit"))
probits8 <- lapply(specs, glm, family=binomial(link="probit"))
probits9 <- lapply(specs, glm, family=binomial(link="probit"))
probits10 <- lapply(specs, glm, family=binomial(link="probit"))
probits11 <- lapply(specs, glm, family=binomial(link="probit"))
probits12 <- lapply(specs, glm, family=binomial(link="probit"))
probits13 <- lapply(specs, glm, family=binomial(link="probit"))
probits14 <- lapply(specs, glm, family=binomial(link="probit"))
probits15 <- lapply(specs, glm, family=binomial(link="probit"))
probits16 <- lapply(specs, glm, family=binomial(link="probit"))
probits17 <- lapply(specs, glm, family=binomial(link="probit"))
probits18 <- lapply(specs, glm, family=binomial(link="probit"))
probits19 <- lapply(specs, glm, family=binomial(link="probit"))
probits20 <- lapply(specs, glm, family=binomial(link="probit"))
probits21 <- lapply(specs, glm, family=binomial(link="probit"))
probits22 <- lapply(specs, glm, family=binomial(link="probit"))
probits23 <- lapply(specs, glm, family=binomial(link="probit"))
probits24 <- lapply(specs, glm, family=binomial(link="probit"))
probits25 <- lapply(specs, glm, family=binomial(link="probit"))
probits26 <- lapply(specs, glm, family=binomial(link="probit"))
probits27 <- lapply(specs, glm, family=binomial(link="probit"))
probits28 <- lapply(specs, glm, family=binomial(link="probit"))
probits29 <- lapply(specs, glm, family=binomial(link="probit"))
probits30 <- lapply(specs, glm, family=binomial(link="probit"))
probits31 <- lapply(specs, glm, family=binomial(link="probit"))

--------------010500030900050501010804--

On 6/26/2006 7:31 AM, ripley at stats.ox.ac.uk wrote:

This works for me on both Linux and Windows.

Please check your memory usage: it does need about 900Kb of VM, and as you 
have less RAM than that installed you need to set --max-mem-size=1G or 
some such.

(It is quite likely that this is a Windows memory allocation failure: that 
has been reported before but not tracked down.)

I can make it reproducible now.  It happens relatively quickly when I 
set max-mem-size=100M.  In R-devel, there's a call to malloc at line 
1952 of memory.c, and as R is running out of memory, that returns a -1 
instead of a zero.  This causes a seg fault a few lines later.

The malloc code is quite complicated, so I can't see exactly why we're 
getting the -1.