extractProb Incorrect dimensions when X Has One Column

[ray-p144]

When using the extractProb function on a model that was trained with only one column in X, the following error is thrown:

Error in newdata[, object$method$center, drop = FALSE] : 
  incorrect number of dimensions

The problem seems to be caused by this line when the data.frame is turned into a vector.

Minimum reproducible example:

library(caret)
TrainData <- data.frame(x = runif(100))
TrainClasses <- sample(c("Y", "N"), 100, TRUE)
fit <- train(TrainData, TrainClasses,
             method = "glm",
             preProcess = c("center", "scale"),
             trControl = trainControl(method = "cv", classProbs = TRUE))
extractProb(list(fit))

I would suggest adding the argument drop = FALSE like the extractPrediction function has in this line

[topepo]

Yep! That bug has been around for _more+ than

> difftime(now(), ymd_hm("2007-10-05 12:21"), units = "weeks")
Time difference of 542.9203 weeks