Potential bug with stepGAICAll.A

[ChristK]

Hi there

Many thanks for all your hard work with this package.

I think there is a potential bug with the stepGAICAll.A() function when it is used with 1-parameter distributions. In particular this is the only function from the stepGAIC family of functions that lacks a trace argument. But then in the function definition there are the following lines

if (npar == 1) {
        if (trace) {
            cat(
                "---------------------------------------------------",
                "\n"
            )
        }
... # more code follows here
}

Which of course then triggers an error when the distribution has 1 parameter Error in if (trace) { : argument is not interpretable as logical.

MRE

library(gamlss)

set.seed(42)
n <- 1e3
y <- rbinom(n, 1, 0.1)
a <- rnorm(n)
b <- rnorm(n)

mod_min <- gamlss(y ~ 1, family = BI())
stepGAICAll.A(mod_min, scope = list(lower = ~1, upper = ~ a + b))
# Error in if (trace) { : argument is not interpretable as logical

A potential fix would be to redefine the first codeblock like below (i.e. remove the if (trace) part)

if (npar == 1) {
            cat(
                "---------------------------------------------------",
                "\n"
            )
... # more code follows here
}

Or you could properly add a trace argument like with the other stepGAIC functions.

cheers
Chris