Rmd: Markdown fenced code blocks with language identifier cause Error: Malformed file!

[salim-b]

Problem

Consider the following reprex:

example_file <- tempfile()

readr::write_file(path = example_file,
                  x = '# Rmd example file

## R code chunk

The following is linted fine:

```{r}
print ("hi" )
```

## Plain code block

The following fenced code block _without_ a language identifier is not linted (also fine!):

```
print ("hi" )
```

## Code block with language identifier

The language identifier of the following code block triggers a \'malformed file\' error (bad!):

```r
print ("hi" )
```
')

lintr::lint(filename = example_file)
#> Error: Malformed file!

^{Created on 2020-07-03 by the reprex package (v0.3.0)}

Markdown code blocks with a language identifier are valid R Markdown syntax and should simply be excluded from linting.

Possibly related

If I understand it correctly, R Markdown files should be supported by lintr as of 50ab86d fixing #32.

In #252, a similar issue was raised which has been solved by fdd1950, assumedly without properly handling language identifiers. I'm not familiar with lintr's internals and didn't dive into the code, but I guess the R code extraction logic in https://github.com/jimhester/lintr/blob/master/R/extract.R is erroneous. Maybe @russHyde or @jimhester could have a look at this?

Maybe also related: #322

[russHyde]

could you confirm this error using the current master branch of lintr, we have a test case that covers the case:

```r
abc = 123
```

[salim-b]

could you confirm this error using the current master branch of lintr

Just tested the current master branch: The issue is gone, no more error. 👍 I should have tested this before opening this issue 🙈 Thanks for the hint!