Parameter names aren't validated

[Kodiologist]

With the latest XGBoost (9c1103e) on R, no error or warning seems to be produced for unrecognized parameter names, despite #5157.

> library(xgboost)
> m = xgboost(params = list(gama = 2), nrounds = 10, data = as.matrix(mtcars[, c("disp", "hp", "drat")]), label = mtcars$mpg)

[trivialfis]

I'm enabling parameter validation interface by interface, only went through 2 Python interfaces (normal train and scikit-learn). You can try to set validate_parameters to TRUE on R to see if it helps.

Will test through R package for validation as the next step, just being careful so that users won't silent XGBoost due to false alarms.

[Kodiologist]

validate_parameters = TRUE indeed prints Parameters: { gama } might not be used.