R CMD check tells me 'no visible binding for globalvariable ', what does it mean?

On Apr 16, 2010, at 4:38 AM, Henrik Bengtsson wrote:

I think what people are also thinking about is that the policy for
publishing a package on CRAN is that it have to pass R CMD check with
no errors, warnings *or* notes.

Can you cite your reference, please? I see only (R-ext 1.5 Submitting a package to CRAN):

"Please ensure that you can run through the complete procedure with only warnings that you understand and have reasons not to eliminate. In principle, packages must pass R CMD check without warnings to be admitted to the main CRAN package area. If there are warnings you cannot eliminate (for example because you believe them to be spurious) send an explanatory note with your submission."

It talks explicitly about warnings, notes are not mentioned at all... That said, you should examine all notes and make sure they are not indications of problems.

In my experience, if a package is new or previously checked without notes, the CRAN maintainers will likely ask you to look at them to make sure they aren't problems, but there isn't any difficulty in getting a package on CRAN if it has notes.  A whole lot of packages on CRAN have notes even when checked on r-release.

CMD check notes are the R equivalent of old-time lint warnings in C, and as the First Commandment says:
 	 Thou shalt run lint frequently and study its pronouncements with care, for verily its perception and judgement oft exceed thine.
and the prophet (Henry Spencer) expands on this:
  ``Study'' doth not mean mindless zeal to eradicate every byte of lint output-if for no other reason, because thou just canst not shut it up about some things-but that thou should know the cause of its unhappiness and understand what worrisome sign it tries to speak of.


          -thomas

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle