maxit argument in additivePenal causes R to crash

[ykarayankov]

The issue occurs when running the following code:

# transform data
  data(dataAdditive)
  test_data <- dataAdditive[200:800,]
  test_data$t2 <- test_data$t2 * 0.345
  
additivePenal(survival::Surv(t1, t2, event) ~ cluster(group) +
                                 var2 + var1 + slope(var1), correlation=TRUE, 
                               data = test_data,
                               n.knots = 8, kappa = 1,
                               print.times = FALSE, maxit = 10)

It seems to be related to the number of observations in the dataset and a minimum number for maxit before it causes R to crash. For example when using test_data <- dataAdditive[200:800,] and maxit = 15 this gives an expected non-convergence warning. However, when maxit = 10 using this data the session will crash. If instead we use test_data <- dataAdditive[400:800,] then maxit can be reduced to 4 before it crashes.