[Rd] Undocumented S4 methods: generic coerce and siglist matrix, Massvector

John Chambers jmc at R-project.org
Thu Feb 10 21:01:26 CET 2005

Witold Eryk Wolski wrote:
> Hi,
> While checking the package I am getting all the time the following 
> complain about the function:
> Undocumented S4 methods:
>  generic coerce and siglist matrix,Massvector
> I have not defined the function nowhere in the R files.
> My gues is that this has something to do with the following definition 
> of the setAs function.
> setAs("matrix","Massvector"
>      ,def= function(from)
>      {
>        return(new("Massvector",from))
>      }
>      )

It certainly does.  From the documentation of setAs():

With this explanation as background, the function setAs does a fairly 
obvious computation: It constructs and sets a method for the function 
coerce with signature c(from, to), using the def argument to define the 
body of the method.

> The problem is not to cheat the check mechanism. I knew that the warning
> can be elminitad by by defining an alias of this function in one of the 
> Rd files.

No cheating seems needed.  The setAs() call is part of your software & 
presumably you document it somewhere.  You do have to construct the 
corresponding alias in the same .Rd file, but it doesn't amount to cheating.

It's true that a user would have to be unusually well-informed to know 
how to ask for the corresponding documentation explicitly.

> Yours
> Eryk

More information about the R-devel mailing list