lm varImp does not return all features

[omarwagih]

varImp seems to return one less variable when tuneGrid = expand.grid(intercept = FALSE) is set.
Is this a bug?

To reproduce:

# Train on iris data
mod = train(Petal.Width~., data=iris[,1:4], method='lm', 
                tuneGrid  = expand.grid(intercept = FALSE))

# 3 features + label
head(x$trainingData)

# only 2 features returned
varImp(x)$imp

[hadjipantelis]

Yes, it is/was a bug. Thank you for mentioning it. A PR fixing it has been made.

[topepo]

Closed via #858