R devel - Dec 2003 - Maybe some bug on sapply and cbind (PR#5628)

Full_Name: Drouilhet R?my
Version: 1.8.1
OS: Linux
Submission from: (NULL) (195.221.43.136)


Hi!!!

Is there some mistake?
This is some part of one of my R programs. I created this function to show you
the trouble!!!

bugTest<-function() {
  paramnames <- c("mean","sd")
  param <- list(expression(1),expression(2))
  paramtmp <- list(mean=0,sd=1)
  if(length(param)) {
      sapply(1:length(param),function(i)
paramtmp[[paramnames[i]]]<<-param[[i]])
  }
  paramtmp
}

If you run :
> bugTest()
$mean
expression(1)

$sd
expression(2)

Everything seems Ok!!! But, now, continue by running :
> cbind(1,rep(2,1))
Segmentation fault

In fact, every cbind, rbind applying vectors seem to fail.

Now, change in the bugTest the sapply by some for instance:

bugTest2<-function() {
  paramnames <- c("mean","sd")
  param <- list(expression(1),expression(2))
  paramtmp <- list(mean=0,sd=1)
  if(length(param)) {
      for( i in 1:length(param)) paramtmp[[paramnames[i]]]<-param[[i]]
  }
  paramtmp
}
>  bugTest2()
$mean
expression(1)

$sd
expression(2)
>  cbind(1,rep(2,1))
     [,1] [,2]
[1,]    1    2

NOW, EVERYTHING IS OK!!!

I hope this could be helpful for us!!

Bye

I can confirm this bug on Solaris 2.8 running R-1.8.1.  Here's a simplified
version:

y <- list()
sapply(1, function(i) y[[i]] <- expression(1))
cbind(1, rep(2,1))

  [ Process R segmentation Fault at Tue Dec  9 09:20:33 2003 ]

Curiously, replacing the third line with
  cbind(1, 2)
does not casue a crash, even though rep(2,1) == 2 !

> R.version
platform sparc-sun-solaris2.8
arch     sparc               
os       solaris2.8          
system   sparc, solaris2.8   
status                       
major    1                   
minor    8.1                 
year     2003                
month    11                  
day      21                  
language R                   
-- 
                              -- David Brahm (brahm@alum.mit.edu)