Fits generalized linear models where the parameters are subject to linear constraints. The model is specified by giving a symbolic description of the linear predictor, a description of the error distribution, and a matrix of constraints on the parameters.
