[R] Crash on Windows with package wle; was: Re: [R] question about simulation

[Uwe Ligges]

Lily wrote:

> Thanks for your response.I am working on R-1.7.1 under
> windows. Here is part of my code for simulation:
> 
> x1<-...
> y1<-...
> beta<-c(1,3,5,7,9))
> mn<-0
> for ( i in 1:3) {
> 
> e1<-rnorm(40,mean=0,sd=1)
> y1<-x1%*%beta+e1
> result<-mle.cv(y1~x1)
> result2<-result[1, ]
> mn<-rbind(mn, result2)
> 
> }
> mn

A *reproducible* example, please!
That means it runs without modifications and reproduces the error in a 
new R session - it's not the idea that each helper has to generate 
appropriate data him/herself.


You could have told you are using
  Package: wle
Is your version of wle recent, i.e.
  Version: 0.7-5 ?

Uwe Ligges


BTW: I may point you to R's easy to use debugging tools. It's worth 
learning to debug your functions for other possibly forthcoming tasks as 
well.


> Thanks!! 
> 
> 
> --- Uwe Ligges <ligges at statistik.uni-dortmund.de>
> wrote:
> 
>>Lily wrote:
>>
>>
>>>I am running a 1000 simulations, it works for 2
>>>simulations. However, I get the following error
>>>message whenever I run it more than 3 times:
>>>
>>>"The instruction at '0*11044080' referenced memory
>>
>>at
>>
>>>"o*3ff00000". The memory could not be "written".
>>>and, I can also get something like "exception:
>>
>>access
>>
>>>violation (0*c0000005). Address:0*11044080".
>>
>>You found a bug.
>>
>>I guess you are on Windows (which?) and you are
>>talking about R-1.7.1 
>>(please tell us these details).
>>
>>Can you provide a minimal example that reproduces
>>the error with 
>>R-1.7.1, please?
>>
>>Uwe Ligges
>>
> 
> 
> 
> __________________________________
> Do you Yahoo!?

> http://sitebuilder.yahoo.com