[Bioc-devel] no visible binding warning messages

[Seth Falcon]

Hi Keith,

Seems like there are two issues here: namespaces and the 'NOTE'
messages from codetools.  Let's handle the 'NOTE' items first.

One of the new features of R 2.6.0 is that R CMD check runs some of the code
analysis tools provided by the codetools package.  These checks are
not perfect, but in my experience they very often point you to real
bugs in your code.

So when you see a NOTE like:
> affyPlotMA : plotFunction: no visible binding for global variable
>    .affylmGUIglobals

This is telling you that in plotFunction, the codetools analysis was
not able to be certain that .affylmGUIglobals was previously defined.
Just reading over the code quickly I can't tell if this is a false
positive or not.  You have

    Try(assign(".affylmGUIglobals",affylmGUIglobals,.GlobalEnv))

In the PlotOptions function, but when does that get called?  For
storing a set of global config variables for your pacakge, you might
consider creating an environment object at package level and putting
things there:

   .affylmGUIglobals <- new.env(parent=emptyenv(), hash=TRUE)

This might be cleaner and will avoid the NOTE.


Keith Satterley <keith at wehi.EDU.AU> writes:
> I tried a NAMESPACE file consisting of
> ###
> #  All variables exported other than those starting with "."
> exportPattern("^[^\\.]")
> import("limma")
> import("tcltk")
> ###
>
> and changed my .First.lib function to .onLoad, but I ran into problems with the 
> .onLoad function and it didn't help with the error messages. I have some way to 
> go yet before I understand how I should apply a NAMESPACE and a
> .onLoad function.

Adding a NAMESPACE is a really good thing to do.  If you provide us
with some further details about what didn't work, I bet we can get you
going.

> One other point, each variable name in the error message in my DOS Cmd window 
> has a hex character immediately before it and just after it. The character 
> before each variable name is E6 and the character after each name is C6 - this 
> seems a bit odd, but not a real problem as the messages are clear!

I suspect this is due to recent changes in R's sQuote function which
is now trying to display fancy single quotes.

+ seth

-- 
Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center
BioC: http://bioconductor.org/
Blog: http://userprimary.net/user/