Should something about this be appropriate in "R Packages: Learn how to
create a package, the fundamental unit of shareable, reusable, and
reproducible R code" by Hadley Wickham and Jennifer Bryan -- in their
current chapter 17. Vignettes?
https://r-pkgs.org/vignettes.html
I recently studied that book and created two packages using my
understanding of their recommendations, including using Roxygen2. I
remember wondering about things like what Duncan recommends.
Spencer Graves
On 12/14/25 15:13, Duncan Murdoch wrote:
I'm writing a procedure here that I always forget, and have to
reconstruct each time. Maybe I'll remember I put it here, and maybe
someone else will save some time by using it:
A useful check on vignette reproducibility is to have a
vignette.Rout.save file in the vignettes directory. This is a file of
output from running the R code in the vignette, and R CMD check compares
the output it sees to your saved file.
A problem I always have is remembering the way to produce this file from
the vignette. The easiest way I know is the following:
- create a file vignette.Rout.save in the appropriate place,
containing junk.
- Then build the package and run R CMD check on it.
The check will complain about the mismatch between the junk in the save
file and the actual output produced from your vignette, and will save a
file with a name like "vignette.Rnw.log" or "vignette.Rmd.log" in the
package.Rcheck directory. That's the file that should become
vignette.Rout.save.
I assume there's a good reason why the file isn't called "vignette.Rout"
or similar, but using "log" sure makes it hard to remember.
Duncan Murdoch
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
