[R] "R CMD check" does not find a mistake

[Martin Maechler]

>>>>> "CG" == Christophe Genolini <cgenolin at u-paris10.fr>
>>>>>     on Tue, 09 Jun 2009 16:17:15 +0200 writes:

    CG> Hi the list, I build a package. They was a mistake in
    CG> it, but R CMD check did not find it. Is that normal ?

    CG> Here is what Kurt gets (which is right, I did this
    CG> mistake):

    CG> --- 8< ----------------
    CG> * checking for code/documentation mismatches ... WARNING
    CG> S4 class codoc mismatches from documentation object 'LongData-class':
    CG> Slots for class 'LongData'
    CG> Code: id other time traj varName
    CG> Docs: id time traj varName
    CG> --- 8< ----------------

Others have already hinted at the solution:

The problem only shows in  "R-devel" (2.10.x).

Note that I had sent an explicit message to this list
*exactly* in order to advise all package authors :

 Subject: [Rd] R-devel:codocClasses() now finds more --> R CMD check warnings
 Date: Wed, 3 Jun 2009 11:45:45 +0200

i.e. only 6 days before you saw the problem....
    ....
    yes, I am sometimes too optimistic, assuming that people
    actually read what I write ....

Regards anyway!
Martin

--
Martin Maechler, ETH Zurich