More informative error msgs: initial variance misspecification

[famuvie]

MRE: Take the following correctly-specified model and successively switch each value of initial variance by -1.

remlf90(
  fixed   = phe_X ~ 1,
  genetic = list(
    model = 'add_animal',
    pedigree = globulus[, 1:3],
    id = 'self',
    var.ini = 1),
  generic = list(
    bl = list(
      model.matrix(~ 0 + bl, globulus),
      diag(nlevels(globulus$bl)),
      var.ini = 1
    )
  ),
  data    = globulus,
  var.ini = list(resid = 1)
)

You then get the following error messages:

• For the genetic component:

  Error in validate_variance(var.ini, dimension = rep(dim * ncol(as.matrix(response)),  : 
    x must be a SPD matrix in the genetic component.

  The line about validate_variance(...) is not meaningful for the user. The relevant part is the second line. However, x is a meaningless internal variable name.

  • Use stop(..., call. = FALSE) to remove the call in the first line
  • Use substitute(x) to recover the original argument name var.ini

• For the generic component:

  Error in check_generic(generic, response = responsem) : 
    x must be a SPD matrix in the generic component. in generic component bl

  Same as before, plus the fact the the phrase is not well constructed as it identifies the source of the problem in two different ways.

• For the resid component:

  Error in FUN(X[[i]], ...) : x must be a SPD matrix in the 

  Same as the first case, plus the fact the the phrase is not well constructed as it does not identify the source of the problem.