\item{[}{...} translated incorrectly in Markdown mode

[krlmlr]

tidyverse/tibble@0c4a581

[gaborcsardi]

@hadley I don't think this needs to be fixed before the release.

[gaborcsardi]

Actually, this seems to be fine for me:

roc <- rd_roclet()
out1 <- roc_proc_text(roc, "
    #' Title
    #'
    #' Description.
    #'
    #' \\describe{
    #' \\item{`[`}{This.}
    #' \\item{`[[`}{That.}
    #' }
    #'
    #' @md
    foo <- function() {}")[[1]]
out2 <- roc_proc_text(roc, "
    #' Title
    #'
    #' Description.
    #'
    #' \\describe{
    #' \\item{\\code{[}}{This.}
    #' \\item{\\code{[[}}{That.}
    #' }
    #'
    foo <- function() {}")[[1]]
expect_equivalent_rd(out1, out2)

Do you have a self-contained example that fails?

[krlmlr]

Adding a simple link breaks it:

roc <- rd_roclet()
out1 <- roc_proc_text(roc, "
    #' Title
    #'
    #' Description.
    #'
    #' \\describe{
    #' \\item{`[`}{This.}
    #' \\item{`[[`}{That. [subset()]. }
    #' }
    #'
    #' @md
    foo <- function() {}")[[1]]
out1

Output:

% Generated by roxygen2: do not edit by hand
% Please edit documentation in RtmpLVDp8r/file173c4115aa40
\name{foo}
\alias{foo}
\title{Title}
\usage{
foo()
}
\description{
Description.
}
\details{
\describe{ \item{\code{[}}{This.} \item{\code{[[}}
{That. [subset()]. } } [\code{}{This.} \item{}[[`}{That.
[subset()]: R:%60%7D%7BThis.%7D%0A%5Citem%7B%60[[%60%7D
%7BThat.%20[subset()
}

[gaborcsardi]

This looks like a bug indeed.

[hadley]

I see a possibly related problem with

#' @param .init If supplied, will be used as the first value to start
#'   the accumulation, rather than using `x[[1]]`. This is useful if
#'   you want to ensure that `reduce` returns the correct value when
#'   `.x` is [is_empty()].

which generates

\item{.init}{If supplied, will be used as the first value to start
the accumulation, rather than using \code{x[[1]]}. This is useful if
you want to ensure that \code{reduce} returns the correct value when
\code{.x} is [is_empty()].

[[1]: R:[1
[is_empty()]: R:is_empty()}

[hadley]

Slightly more minimal reprex:

out <- roc_proc_text(rd_roclet(), "
  #' Title
  #'
  #' `[[`. [subset()]
  #'
  #' @md
  foo <- function() {}")[[1]]

out$get_field("description")

[hadley]

Actually a minimal reprex:

add_linkrefs_to_md("`[[`. [subset()]")
#> [1] "`[[`. [subset()]\n\n[[`. [subset()]: R:[%60.%20[subset()\n"

[hadley]

The problem appears to be that the regexp is matching too much - from the first [.

I think the fix is to also disallow matching [ in the contents of the link - i.e. change [^]] to [^]\\[]. That gives me:

cat(add_linkrefs_to_md("`[[`. [subset()]"))
#> `[[`. [subset()]
#> 
#> [subset()]: R:subset()

[hadley]

That gives me this regexp:

regex("(?<=[^]]|^)\\[([^\\]\\[]+)\\](?:\\[([^\\]\\[]+)\\])?(?=[^\\[]|$)")

That doesn't break any unit tests, but I don't understand it fully, so I'll wait for approval from @gaborcsardi before changing.

[brodieG]

Similar issue using 6.0.1.9000. I only list it here for completeness for tests since these don't actually involve links, and in my use case I have the workaround of switching the order of the bullets.

#' This Fails
#'
#' @param bounds character(1L)
#'   * "[)"
#'   * "(]"

test_fun <- function(bounds) TRUE

#' This Works
#'
#' @param bounds character(1L)
#'   * "(]"
#'   * "[)"

test_fun2 <- function(bounds) TRUE

DESCRIPTION:

...
RoxygenNote: 6.0.1.9000
Roxygen: list(markdown = TRUE)

Also, this is what the failing rd looks like:

% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/test.R
\name{test_fun}
\alias{test_fun}
\title{This Fails}
\usage{
test_fun(bounds)
}
\arguments{
\item{bounds}{character(1L)
\itemize{
\item "[)"
\item "(]"
}

[)"
\itemize{
\item "(]: R:)%22%0A%20%20*%20%22(
}}
}
\description{
This Fails
}