Skip to content
Prev 32039 / 398506 Next

mcl models, percentages

On Wed, 14 May 2003, Thomas Lumley wrote:

            
I tried to change 'eps' to 1.e-8, found yet another way of crashing  R:
------------------------------------------------------------------------
control = list(eps = 1.e-8))

Program received signal SIGSEGV, Segmentation fault.
0x405526a0 in agfit3 (maxiter=0x0, nusedx=0x1, nvarx=0x1, start=0x8f267b8, 
    stop=0x8f26770, event=0x8f25f98, covar2=0x8f26728, offset=0x8f266e0, 
    weights=0x8f26698, nstrat=0x8ec4ab0, strata=0x8ec4a90, 
sort1=0x8f25f60, 
    sort2=0x8f25f28, means=0x8e7ed30, Rf_beta=0x8e7ed08, u=0x8e7ece0, 
    imat2=0x8e7ecb8, loglik=0x8f25ef0, flag=0x8ec4a70, work=0x8f249a0, 
    eps=0x8e7ec90, tol_chol=0x0, sctest=0x8e7ec68) at agfit3.c:263
263     agfit3.c: No such file or directory.
        in agfit3.c
-------------------------------------------------------------------------
It's of course my fault; 'control' should be a call to 'coxph.control'.
But should  R  crash...?

G?ran
[...]

---
 G?ran Brostr?m                    tel: +46 90 786 5223
 Department of Statistics          fax: +46 90 786 6614
 Ume? University                   http://www.stat.umu.se/egna/gb/
 SE-90187 Ume?, Sweden             e-mail: gb at stat.umu.se