Please allow uncommented code output to R script with chunk options eval=FALSE purl=TRUE

[alchemit]

It seems to be knitr::purl feature to comment out the code chunks with options eval=FALSE, purl=TRUE in the extracted R script, and I could not find a workaround (option coment="" doesn't help). Let me show some use cases where this feature is a show stopper:

• The source document explains how to "play with data" on a specific example to achieve the desired result, but the extracted script should be usable with any data set. The script may thus ask the user about input/output file names or parameters relevant to its purpose. However, evaluation of something like readLines("stdin", n = 1) hangs the knitting process, at least in Rstudio.
• The source document explains how to use the generated script, for example describing its command line options. Of course there are no CLI options when knitting so eval=FALSE makes perfect sense then, but the code should be working in the extracted script.

[cderv]

Thanks for your feedback.

Can you share an example document showing why you need eval = FALSE in the .Rmd, while wanting to have code uncommented in script ?

eval = FALSE means to no run the code, and purl() is meant to give a R script version of what is executed in the Rmd doc. So why it comments unevaluated code.

I understand your use case, but probably not enough to see what could be done. Having a real example that shows how you currently do things may help find solution playing with existing options.

Thank you