mark oddly interprets input as file where documentation suggests otherwise?

[LukasWallrich]

I frequently use markdown::mark(), yet now ran into an issue where "*F*(5, 7165) = 358.25, *p* < .001" was interpreted as a filename. Having never looked at the documentation before, it took me a while to figure out where that error came from.

According to the docs a single character string input will be treated as a file only when the file exists or it has a file extension - neither seems to be the case here?

markdown::mark("*F*(5, 7165) = 358.25, *p* < .001")
#> Warning in file(con, "r"): cannot open file '*F*(5, 7165) = 358.25, *p* <
#> .001': No such file or directory
#> Error in file(con, "r"): cannot open the connection

^{Created on 2023-04-28 with reprex v2.0.2}

[yihui]

In this case, .001 was treated as a "file extension":

markdown/man/mark.Rd

Lines 24 to 30 in 0f64983

	\item{file}{Path to an input file. If not provided, it is presumed that the
	\code{text} argument will be used instead. This argument can also take a
	character vector of Markdown text directly. To avoid ambiguity in the
	latter case, a single character string input will be treated as a file only
	when the file exists or it has a file extension. If a string happens to
	have a \dQuote{file extension} and should be treated as Markdown text
	instead, wrap it in \code{I()}.}

To ensure the input is text, use either I() or the text argument:

markdown::mark(I("*F*(5, 7165) = 358.25, *p* < .001"))
markdown::mark(text = "*F*(5, 7165) = 358.25, *p* < .001")

I think the file extension heuristics in mark() should be improved and I'll see what I can do. Thanks for the report!

[yihui]

Should be fixed now. Thanks again!

[LukasWallrich]

Thank you!