[R-pkg-devel] R CMD build: "Error in if (any(update))"

| I've got an R package that I periodically create in what I assume is
| an old-fashioned way:
| 
| Within R: use "package.skeleton"
| Outside of R: update e.g. .rd files as appropriate
| Command window: run the following commands
| R CMD build packageName
| R CMD check packageName

Close. This should check the .tar.gz produced by the preceding command:

  R CMD check packageName_a.b.c.tar.gz     # where a.b.c is current version

| R CMD INSTALL --build packageName

Not sure you need --build; it creates a binary for redistribution but only 
for system that match your system closely.  I would recommend

  R CMD INSTALL packageName_a.b.c.tar.gz

instead.

| I am now getting an error on the R CMD build stage. Initially I had
| wanted to remove a package dependency (by not passing some functions
| to package.skeleton and editing my NAMESPACE and DESCRIPTION file
| appropriately), but on getting the error I tried to revert everything
| back to my last successful build (Feb 2019), however, the error
| persisted:
| R CMD build packageName
| * checking for file 'packageName/DESCRIPTION' ... OK
| * preparing 'packageName' :
| * checking DESCRIPTION meta-information ... OK
| * checking for LF line-endings in source and make files
| * checking for empty or unneeded directories
| * looking to see if a 'data/datalist' file should be added
| Error in if (any(update)) { : missing value where TRUE/FALSE needed
| Execution halted
| 
| I am guessing that the error could be due to any of the last three
| lines that don't have an "OK", but I am completely stuck as to how to
| narrow it down further, and Dr. Google has already failed me.

I would think it has to do with your package sources, and a possible change
in R 3.6.0.  Checks get better over time so it may set off a flag for
something that passed before.

But we can't really tell without access to your code.

Hth, Dirk

http://dirk.eddelbuettel.com | @eddelbuettel | edd at debian.org

You might be able to find where the error occurred by putting the following
lines in ~/.Rprofile

cat("~/.Rprofile: setting alternate error handler\n")
options(
    error=quote({
        dump.frames()
        writeLines(c("Stack trace", paste0("  ",names(last.dump))))}))

Bill Dunlap
TIBCO Software
wdunlap tibco.com

Hi,

(Re-posting here as per advice from r-help)

I've got an R package that I periodically create in what I assume is
an old-fashioned way:

Within R: use "package.skeleton"
Outside of R: update e.g. .rd files as appropriate
Command window: run the following commands
R CMD build packageName
R CMD check packageName
R CMD INSTALL --build packageName

I am now getting an error on the R CMD build stage. Initially I had
wanted to remove a package dependency (by not passing some functions
to package.skeleton and editing my NAMESPACE and DESCRIPTION file
appropriately), but on getting the error I tried to revert everything
back to my last successful build (Feb 2019), however, the error
persisted:
R CMD build packageName
* checking for file 'packageName/DESCRIPTION' ... OK
* preparing 'packageName' :
* checking DESCRIPTION meta-information ... OK
* checking for LF line-endings in source and make files
* checking for empty or unneeded directories
* looking to see if a 'data/datalist' file should be added
Error in if (any(update)) { : missing value where TRUE/FALSE needed
Execution halted

I am guessing that the error could be due to any of the last three
lines that don't have an "OK", but I am completely stuck as to how to
narrow it down further, and Dr. Google has already failed me.

Using R version 3.6.0, R tools version 3.5.0.4 (I don't recall if had
different versions previous time I built this package in Feb 2019).
Attempt to reproduce error with dummy package failed (i.e. no issues
with R CMD build).

Cheers,
Rob.

______________________________________________
R-package-devel at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel

Thanks for the replies. I am still stuck, although agree it is most
likely due to change to R version 3.6.0. I did however manage to
create a dummy package that now does re-create the same error, this
required the inclusion of a data object. The R CMD build error occurs
regardless of whether there is a sysdata.rda in the R subdirectory,
and whether there is a datalist in the data subdirectory. Apologies
for the lengthy "code" that follows:

In R:
myCol <- rgb(0.9,0.5,0.5)
myPlot <- function(x){
   plot(x, col = myCol)
   return(NULL)
}
package.skeleton(name = "myPackage", list = c("myPlot", "myCol"))

"Optionally" in R:
save(list = c("myCol"), file = "sysdata.rda")

my DESCRIPTION file:
Package: myPackage
Type: Package
Title: Test Package
Version: 1.0
Date: 2019-06-11
Author: Rob Foxall <rob.foxall at test.com.au>
Maintainer: Rob Foxall <rob.foxall at test.com.au>
Description: To test making a package
License: GPL-2
LazyLoad: yes

my NAMESPACE file:
exportPattern("^[[:alpha:]]+")

contents of myPackage directory:
directories "data", "man", "R", files "DESCRIPTION", "NAMESPACE",
"Read-and-delete-me"
contents of data subdirectory:
myCol.rda, (optionally) datalist
contents of man subdirectory:
myCol.Rd, myPackage-package.Rd, myPlot.Rd
contents of R subdirectory:
myPlot.R, ("optionally") sysdata.rda

The ".Rd" files:
myCol.Rd:
\name{myCol}
%
\alias{myCol}
%
\docType{data}
%
\title{A colour}
%
\description{It's a colour}
%
\usage{data(myCol)}
%
\format{
It is a colour, defined by:
rgb(red=0.9,green=0.5,blue=0.5)
}
%
\details{
Other, standard colours are available via sensible names. See \code{colours}. }
%
\source{See Me}
%
\references{Albers, J. (1975) \emph{Interaction of Colour}. Yale
University Press}
%
\examples{
myPlot(3)
}
%
\keyword{sysdata}

myPackage-package.Rd:
\name{myPackage-package}
%
\alias{myPackage-package}
%
\alias{myPackage}
%
\docType{package}
%
\title{Test package}
%
\description{Testing making a package}
%
\details{
\tabular{ll}{
Package: \tab myPackage\cr
Type: \tab Package\cr
Version: \tab 1.0\cr
Date: \tab 2019-06-11\cr
License: \tab GPL-2\cr
LazyLoad: \tab yes\cr
}
%
Test package, dummy plot with dummy colour
}
%
\author{
Rob Foxall <rob.foxall at test.com>
Maintainer:  <rob.foxall at test.com>
}
%
\references{N.A.J. Hastings, Brian Peacock, Merran Evans (2000)
\emph{Statistical Distributions} John Wiley & Sons Inc}
%
\keyword{package}
%
\seealso{\code{\link{plot}}}
%
\examples{
myplot(3)
}

myPlot.Rd
\name{myPlot}
%
\alias{myPlot}
%- Also NEED an '\alias' for EACH other topic documented here.
\title{Do a plot of a number}
%
\description{Plots a number with a set colour}
%
\usage{myPlot(x)}
%
\arguments{
  \item{x}{A number}
}
%
\details{Plots a number with a colour.}
%
\value{Returns null}
%
\author{Rob Foxall}
%
\note{There is no note}
%
\seealso{\code{\link{plot}}}
%
\examples{myPlot(3)}
%
\keyword{KPI}

You might be able to find where the error occurred by putting the following lines in ~/.Rprofile

cat("~/.Rprofile: setting alternate error handler\n")
options(
    error=quote({
        dump.frames()
        writeLines(c("Stack trace", paste0("  ",names(last.dump))))}))

Bill Dunlap
TIBCO Software
wdunlap tibco.com


On Mon, Jun 10, 2019 at 6:38 AM Rob Foxall <rfoxall27 at gmail.com> wrote:

Hi,

(Re-posting here as per advice from r-help)

I've got an R package that I periodically create in what I assume is
an old-fashioned way:

Within R: use "package.skeleton"
Outside of R: update e.g. .rd files as appropriate
Command window: run the following commands
R CMD build packageName
R CMD check packageName
R CMD INSTALL --build packageName

I am now getting an error on the R CMD build stage. Initially I had
wanted to remove a package dependency (by not passing some functions
to package.skeleton and editing my NAMESPACE and DESCRIPTION file
appropriately), but on getting the error I tried to revert everything
back to my last successful build (Feb 2019), however, the error
persisted:
R CMD build packageName
* checking for file 'packageName/DESCRIPTION' ... OK
* preparing 'packageName' :
* checking DESCRIPTION meta-information ... OK
* checking for LF line-endings in source and make files
* checking for empty or unneeded directories
* looking to see if a 'data/datalist' file should be added
Error in if (any(update)) { : missing value where TRUE/FALSE needed
Execution halted

I am guessing that the error could be due to any of the last three
lines that don't have an "OK", but I am completely stuck as to how to
narrow it down further, and Dr. Google has already failed me.

Using R version 3.6.0, R tools version 3.5.0.4 (I don't recall if had
different versions previous time I built this package in Feb 2019).
Attempt to reproduce error with dummy package failed (i.e. no issues
with R CMD build).

Cheers,
Rob.

______________________________________________
R-package-devel at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel