[Rd] Possible bug in accessing methods documentation? (PR#9291)

[Duncan Murdoch]

On 10/11/2006 4:26 PM, murdoch at stats.uwo.ca wrote:
> On 10/11/2006 2:48 PM, Seth Falcon wrote:
>> Hi,
>>
>> Reading help("Documentation"), I'm led to believe that a help call
>> like:
>>
>>     ?myFun(x, sqrt(wt))
>>
>> Will search for help on the appropriate method in the case that myFun
>> is generic.  This isn't working for me.  Here is an example using the
>> Biobase package:
>>
>> ## If Biobase is not installed
>> source("http://bioconductor.org/biocLite.R")
>> biocLite("Biobase")
>>
>>> library("Biobase")
>>> data(sample.ExpressionSet)
>>> class(sample.ExpressionSet)
>> [1] "ExpressionSet"
>> attr(,"package")
>> [1] "Biobase"
>>> z <- featureNames(sample.ExpressionSet)
>>> z[1:2]
>> [1] "AFFX-MurIL2_at"  "AFFX-MurIL10_at"
>>> ?featureNames(sample.ExpressionSet)
>> Warning message:
>> no method defined for function 'featureNames' and signature 'object = "missing"' in: .helpForCall(e1Expr, parent.frame()) 
>> Error in .helpForCall(e1Expr, parent.frame()) : 
>>         no documentation for function 'featureNames' and signature 'object = "missing"'
>>
>>
>> Am I expecting the wrong thing or is this broken?  
> 
> I think it's broken.  The line that looks broken is in .helpForCall, 
> where it has
> 
> methods::"elNamed<-"(sigClasses, arg, class(argVal))
> 
> Looking at the elNamed<- function, I think this is supposed to be 
> equivalent to
> 
> sigClasses[[arg]] <- class(argVal)
> 
> but it is not making any assignment.  I don't understand the point of 
> doing the assignment that way, or why it's not working, so I'll leave 
> this for someone else.

Actually, why it's not working is obvious:  it's not assigning the 
result back into sigClasses.  But I still don't understand what it is 
doing that is different from the version below.

Duncan Murdoch