[Rd] [PATCH] suggestions for R-lang manual

[Martin Maechler]

>>>>> Scott Kostyshak <skostysh at princeton.edu>
>>>>>     on Thu, 27 Feb 2014 16:43:02 -0500 writes:

    > On Thu, Nov 21, 2013 at 1:17 AM, Scott Kostyshak <skostysh at princeton.edu> wrote:
    >> Attached is a patch with suggestions for the R-lang manual at r64277.
    >> 
    >> Below are a few comments (some are implemented in the patch):
    >> 
    >> In the section "Objects", there is a table introduced by "The
    >> following table describes the possible values returned by typeof". One
    >> of the results is "any". Can "any" be returned by "typeof()" ?

ANYSXP  is a valid internal type on the C level, and
src/main/util.c  will make  typeof(ob) return "any"
if you can get your hands at an R level object of that type.
I'd guess you can only get it currently by using .Call() and 
using your own C code, .. but at least that way it must be possible.

    >> Regarding the "Recycling rules" section,
    >> 
    >> -One exception is that when adding vectors to matrices, a warning is not
    >> -given if the lengths are incompatible.
    >> - at c Is that a bug?
    >> -
    >> 
    >> was this a bug that was fixed? 

I did not investigate in details, but yes, I vaguely remember we
had fixed that.  So indeed, it's fine you omitted the para in
your patch.

    >> I see the following behavior:
    >> 
    >>> myvec <- 1:3
    >>> mymat <- matrix(1:12, ncol=2)
    >>> myvec <- 1:5
    >>> myvec + mymat
    >> [,1] [,2]
    >> [1,]    2    9
    >> [2,]    4   11
    >> [3,]    6   13
    >> [4,]    8   15
    >> [5,]   10   12
    >> [6,]    7   14
    >> Warning message:
    >> In myvec + mymat :
    >> longer object length is not a multiple of shorter object length
    >>> 
    >> 
    >> Regarding
    >> 
    >> -The arguments in the call to the generic are rematched with the
    >> -arguments for the method using the standard argument matching mechanism.
    >> -The first argument, i.e.@: the object, will have been evaluated.
    >> -
    >> 
    >> this information is duplicated. See a few paragraphs up "When the
    >> method is invoked it is called..."

    >> Scott

Thank you, Scott.
Indeed, I've finally carefully looked at the patch, and applied
it - for R-devel, to become R 3.1.0 in April.

Martin


    >> --
    >> Scott Kostyshak
    >> Economics PhD Candidate
    >> Princeton University

    > The patch still applies cleanly (one offset) on r65090.

    > Best,
    > Scott