[Rd] Running examples with R5 constructor from package failing

Duncan Murdoch murdoch.duncan at gmail.com
Sat Feb 1 23:03:28 CET 2014


On 14-02-01 4:21 PM, Imanuel Costigan wrote:> You referred to another 
package that exports an R5 class. Which one?

I was looking at lme4, which doesn't have any reference classes, but 
should be relevant since the setClass function behaves a lot like 
setRefClass.

The ascii package does use setRefClass.  I don't think it uses 
exportClass though...

Duncan Murdoch

 >
 > Thanks for your help.
 >
 > On 02 Feb, 2014,at 12:57 AM, Duncan Murdoch <murdoch.duncan at gmail.com>
 > wrote:
 >
 >> On 14-01-31 10:14 PM, Imanuel Costigan wrote:> I am having an issue with
 >> reference classes (R5) in a package I am developing. They have both been
 >> illustrated in the `roxygentest` package I've got hosted on Github [1].
 >> >
 >> > When I run `R CMD check .` on the source directory of my package I am
 >> told that:
 >> >> Error: could not find function "Blob"
 >> >
 >> > However, when I install and load the package, I can successfully run
 >> the command in the example.
 >> >
 >> > Does someone know why this is happening?
 >> >
 >> > The relevant definitions and the example (in Rd format):
 >> >
 >> > R5 definition:
 >> > ```
 >> > draw.Blob <- function (type = 'small')
 >> > {
 >> > 'Something about blobs'
 >> > type <<- 'small'
 >> > }
 >> >
 >> > Blob <- setRefClass (
 >> > Class = "Blob",
 >> > fields = list(type = 'character'),
 >> > methods = list(draw = draw.Blob)
 >> > )
 >> > ```
 >> >
 >> > Example:
 >> > ```
 >> > \examples{
 >> > Blob(type = 'small')
 >> > }
 >> > ```
 >>
 >> The symptoms make it look as though the *function* Blob is not exported.
 >> I see in your source that you do export the *class* Blob, but the docs
 >> don't make it clear to me that this also implies export of the generator
 >> function.
 >>
 >> So I'd add
 >>
 >> export(Blob)
 >>
 >> to your NAMESPACE. I just checked another package that exports a class,
 >> and it separately exported the generator function, so I guess this is
 >> known.
 >>
 >> Duncan Murdoch
 >>
 >>
 >>



More information about the R-devel mailing list