[BioC] Note in check report was: [Bioc-devel] Recent changes in R

Seth Falcon sfalcon at fhcrc.org
Tue May 15 02:05:34 CEST 2007


"Holger Schwender" <holger.schw at gmx.de> writes:
> Hi Seth,
>
> while following your suggestion and taking a look on the check
> reports of siggenes 1.11.0 (for BioC 2.1) I found the following
> note:
>
> checking R code for possible problems ... NOTE
> filterALL: no visible binding for global variable 'ALL'
> filterALL: no visible binding for global variable 'ALL'
> .
> .
> .
> make.tablecode: no visible global function definition for 'lookUp'
> make.tablecode: no visible global function definition for 'getTDRows'
>

Thanks for mentioning that.  This is yet another new R feature: R CMD
check now runs some code checks from the codetools package.  The
results are currently reported using NOTE.

> I think that has something to do with namespace, since all the
> functions mentioned here (e.g., lookUp) are either from the annotate
> or the the genefilter package which have both namespaces and are in
> the suggests field of siggenes. Can I ignore this, since it is
> "only" a note and not a warning or error? Or should I do something,
> e.g. adding

The messages about "no visible global function definition" are worth
investigating.  This can be an indication of a typo in a function
name, a missing function definition, or a missing Depends listing.

It can also be a false positive.  One case where you will see false
positives is if you Suggest a package and call require() to provide
optional functionality.  If you see these messages for function that
you know will be made available by a call to require(), then you can
ignore those messages.

For siggenes, I think you have some of both going on.  I think the
message about variable ALL could be a real problem since you do not
Depends or Suggests the ALL package.

As you noted, lookUp is from annotate and you have annotate in
Suggests and call require -- but you do not check the return value and
that is really essential.  I don't recognize getTDRows off the top of
my head, but you should investigate similarly.

Best,

+ seth

-- 
Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center
http://bioconductor.org



More information about the Bioconductor mailing list