pfr(): multiple fpc-terms don't work

[fabian-s]

pfr(pasat ~ fpc(cca), data=DTI[complete.cases(DTI),])
pfr(pasat ~ fpc(rcst), data=DTI[complete.cases(DTI),])
# so far so good

pfr(pasat ~ fpc(rcst) + fpc(cca), data=DTI[complete.cases(DTI),])
# Error in xt$X %*% sm$X : non-conformable arguments
## xt$X is 229 x 55 (55 is the gridlength for rcst)
## sm$X is 93 x 40 (93 is the gridlength for cca)

pfr(pasat ~ fpc(cca) + fpc(rcst), data=DTI[complete.cases(DTI),])
# Error in xt$X %*% sm$X : non-conformable arguments
## xt$X is 229 x 93 (gridlength for cca)
## sm$X is 55 x 40 (55 is the gridlength for rcst)

seems like the order of the terms matters here and that either xt$X or sm$X use the wrong functional covariate....

(tested with refundDevel @ ce08fcb)

@jgellar :
I suspect this

if ("xt" %in% names(x$call)) {
        xtvars <- all.vars(x$call$xt)
        if (length(xtvars)) {
          sapply(xtvars, function(xtvar) {
            xtvarval <- eval(as.name(xtvar), envir = evalenv, enclos = frmlenv)
            #assign into parent of newfrmlenv because these are not
            #necessarily covariates so list2df(newfrmlenv) below would fail
            assign(x = xtvar, value = xtvarval, envir = parent.env(newfrmlenv))
            invisible(NULL)
          })
        }
      }

may be the section of pfr where this goes wrong, but I'm not sure. Hope it's ok to assign this to you...