inheritDotParams does not show dot parameters

[nuno-agostinho]

For instance, in the following code:

#' Pre-styled button
#' 
#' @param id Character: button identifier
#' @param label Character: label
#' @param class Character: class
#' @inheritDotParams actionButton -inputId -label
#' 
#' @importFrom shiny actionButton
#' @return HTML for a pre-styled button
processButton <- function(id, label, ..., class="btn-primary") {
    button <- actionButton(id, label=label, class=class, type="button", ...)
    return(button)
}

The dot parameters should include the dot argument from actionButton, as the function is simply a wrapper to the other function and the dot argument is desired.

[hadley]

What is the dot argument?

[nuno-agostinho]

What I meant by dot argument is the ... from actionButton. As you can see in that documentation page, it describes the ... as "Named attributes to be applied to the button or link."

I think that in cases like the one I show above, it could make sense to inherit the description for the ... argument.

[hadley]

I really don't understand what you want here. Perhaps you could include the roxygen comments for actionButton and describe what you want to be inherited.

[privefl]

Say you want to document this function:

#' Title
#'
#' @inheritDotParams glue::glue
#'
#' @return
#' @export
#'
#' @examples
system_glue <- function(...) {
  system(glue::glue(...))
}

The "dots argument" (of glue::glue()) doesn't appear in the documentation of system_glue():

[pnacht]

I've recently bumped into this issue myself.

Let's say I have two functions:

• foo is a function which accepts dots:

#' @param a a letter
#' @param b another letter
#' @param ... some dots
foo <- function(a, b, ...) {
  # do stuff
}

• bar is a function which calls foo:

#' @inheritParams foo
bar <- function(...) {
  foo(...)
}

In this case, I'd expect the documentation of bar's dots argument to be "Arguments passed on to foo" followed by three items: a, b, and .... However, I only get a and b, losing foo's definition of its own ... argument:

[pearsonca]

Yep, just ran into this myself (see the bot note)

There's definitely an annoying recursion issue here (i.e., if the deferred to documentation itself inheritDots off to another function). I'd say the reasonable solution is only go one level (rather than trying to recursively pull them all in) - seems plausible to pull the top level note (e.g. in terms of that scales PR, if something referred to label_percent to inheritDotParams, it would get the concretely defined label_percent params, then the headline for ...s ("Arguments passed on to label_number"), but not the enumeration within that ...s arg.

[d-morrison]

I've been running into this issue as well; I agree with @pearsonca's suggestion. @hadley, would it be reasonable to reopen this issue based on the further discussion?

[hadley]

@d-morrison after an issue has been closed for 2+ years, I'd recommend that you open a new issue with a clear description of the problem.